Hypoxia induces selective neutrophil degranulation to promote endothelial damage in COPD Katharine Mary Lodge This dissertation is submitted for the degree of Doctor of Philosophy Newnham College September 2018 Declaration This dissertation was composed on the basis of work carried out under the supervision of Dr Wei Li in the Division of Respiratory Medicine, Department of Medicine, University of Cambridge and Professor Alison Condliffe in the Department of Infection, Immunity and Cardiovascular Disease, University of Sheffield. It is the result of my own work and includes nothing which is the outcome of work done in collaboration except where specified in the text. It is not substantially the same as any that I have submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution. I further state that no substantial part of my dissertation has already been submitted, or, is being concurrently submitted for a degree or diploma or other qualification at the University of Cambridge or any other University or similar institution. This dissertation (excluding figures, tables, appendices and bibliography) does not exceed the word limit prescribed by the Clinical Medicine and Clinical Veterinary Medicine Degree Committee. Katharine Mary Lodge September 2018, Cambridge Abstract Hypoxia induces selective neutrophil degranulation to promote endothelial damage in COPD Katharine Lodge Neutrophils are the key effector cells of innate immunity, rapidly recruited to defend the host against invading pathogens. They deploy pre-formed proteases packaged within cytoplasmic granules, generate reactive oxygen species and release neutrophil extracellular traps (NETs). However, these anti-microbial strategies can also cause substantial bystander tissue injury. Infected and inflamed tissues are profoundly hypoxic, and this oxygen depletion both compromises neutrophil-mediated bacterial killing and enhances the secretion of destructive granule contents. Persistent neutrophilic infiltration of hypoxic tissues characterises diseases such as chronic obstructive pulmonary disease (COPD), which is also associated with endothelial dysfunction. Proteases such as neutrophil elastase (NE) have been implicated in COPD pathogenesis but the precise mechanisms of neutrophil-mediated tissue damage are unknown. In this thesis, I show that hypoxia synergises with neutrophil priming agents to enhance the release of NE in a PI3Kγ-dependent manner, with hypoxic supernatants inducing substantial endothelial cell dysfunction and death. To further identify potential histotoxic mediators, I used a proteomics approach to provide a full (and entirely novel) characterisation of the hypoxic versus normoxic neutrophil secretome, revealing upregulation of a number of protein candidates (both granule-associated and, surprisingly, a subset of cytoplasmic proteins) following hypoxic incubation and stimulation. As this unexpected pattern of protein release did not segregate with neutrophil granule contents, alternative secretion mechanisms were examined; however, hypoxia did not enhance the release of NETs or neutrophil-derived microvesicles. Importantly, neutrophils isolated from exacerbating COPD patients, compared with those from healthy controls, exhibited even further augmented release of selected cytotoxic granule (NE, resistin and neutrophil gelatinase-associated lipocalin) and cytoplasmic (cyclophilin A) proteins under hypoxia. Furthermore, a plasma signature of increased protease (NE and proteinase 3) activity was identified in COPD versus healthy control plasma. In conclusion, hypoxia engenders a destructive neutrophil phenotype, with enhanced release of histotoxic proteins and increased capacity to cause endothelial injury. This may contribute to local and distant tissue damage in the clinical setting of COPD, with further relevance to a wide range of chronic inflammatory diseases, underpinned by endothelial dysfunction. Acknowledgements First and foremost I would like to thank my supervisors, Alison Condliffe and Wei Li, for their unwavering support, guidance and encouragement during my PhD studies. When Alison moved to Sheffield to take up a well-deserved Professorship, I knew that she would continue to supervise me with the same care and dedication that she had provided right from the start of my PhD. I could not have hoped for more intelligent, capable, conscientious and approachable supervisors. I would especially like to thank my sponsor, Edwin Chilvers, whose door was always open for advice and I very much appreciate many insightful scientific discussions. The lab environment cultivated by Edwin ensured both intellectual stimulation and fun social occasions - I had a wonderful time! Throughout my PhD studies I have received help from numerous colleagues in Cambridge and Sheffield. I would particularly like to thank Alexi Crosby, Paul Upton, Anne-Katrien Stark and Sabine Suire for their donation of mouse legs, and Elisabet Ferrer and Merete Long for their substantial assistance with microvesicles, even in the face of adversity! I would also like to thank Arlette Vassallo for her help with the ROS assay, which worked perfectly every time! My office and lab colleagues in the Morrell, Chilvers and Summers groups, notably Arlette Vassallo, Alex Wood, Kim Hoenderdos, Ben Dunmore, Alexi Crosby, Elisabet Ferrer and Eleo Fox, made the lab a joy to work in. I especially value the firm friendship of Arlette and Alex, who kept me plied with tea and wine, and kept me sane – or thereabouts! I would like to express my gratitude to the Wellcome Trust and the British Medical Association Foundation for Medical Research for funding me and for all the blood donors who made this work possible. Finally, I’d like to give a very special thanks to my mum, and my partner, Ruaridh Buchanan, who now know nearly as much about neutrophils as I do! You were always there for me with unswerving support, and helped me to relax, recharge and keep going in the face of frustration. I could not have done my PhD without you. Declaration ………………………………………………………………………………..……..….. 1 Abstract …………………………………………………………………………………..……..…… 2 Acknowledgements ……………………………………………………………………..……..…… 3 List of Figures ……………………………………………………………………………..….….…. 8 List of Tables ……………………………………………………………………………..….….….. 9 List of Abbreviations ………………………………………………………………………………. 10 1 Introduction .................................................................................................................. 13 1.1 The innate immune system ................................................................................... 13 1.2 Neutrophils ........................................................................................................... 13 1.2.1 Effector functions of neutrophils ........................................................................ 14 1.2.1.1 Priming and activation ................................................................................ 14 1.2.1.2 Recruitment ................................................................................................ 15 1.2.1.3 Phagocytosis and Reactive Oxygen Species .............................................. 16 1.2.1.4 Granule formation and composition ............................................................ 17 1.2.1.5 Degranulation ............................................................................................. 18 1.2.1.6 Degranulation: receptors and intracellular signalling pathways ................... 20 1.2.1.7 Degranulation: vesicle trafficking and fusion machinery.............................. 22 1.2.1.8 Degranulation: differential exocytosis ......................................................... 22 1.2.1.9 Neutrophil Extracellular Traps .................................................................... 25 1.2.1.10 Neutrophil-derived microvesicles ................................................................ 26 1.2.1.11 Apoptosis and clearance ............................................................................ 27 1.2.1.12 Role of neutrophils in disease .................................................................... 27 1.2.2 The impact of hypoxia on neutrophils ................................................................ 29 1.2.2.1 Relevance of hypoxia to neutrophils ........................................................... 29 1.2.2.2 Sensing of hypoxia by neutrophils .............................................................. 30 1.2.2.3 Effect of hypoxia on neutrophil recruitment ................................................. 31 1.2.2.4 Effect of hypoxia on phagocytosis and ROS generation ............................. 32 1.2.2.5 Effect of hypoxia on degranulation ............................................................. 33 1.2.2.6 Effect of hypoxia on NET production .......................................................... 34 1.2.2.7 Effect of hypoxia on neutrophil-derived microvesicle generation ................. 35 1.2.2.8 Effect of hypoxia on neutrophil apoptosis ................................................... 35 1.2.2.9 Overall effect of hypoxia on neutrophil functions ........................................ 36 1.2.3 The role of neutrophils in COPD ........................................................................ 38 1.3 Endothelial cells .................................................................................................... 40 1.3.1 The role of neutrophils in endothelial dysfunction and cardiovascular disease .. 41 1.3.1.1 Direct neutrophil-endothelial interactions .................................................... 41 1.3.1.2 Interaction of endothelium with secreted neutrophil products ..................... 42 1.3.1.3 Interaction of endothelium with neutrophil-derived microvesicles ................ 44 1.3.2 The role of neutrophils in endothelial dysfunction and cardiovascular disease in COPD ......................................................................................................................... 45 1.4 Summary .............................................................................................................. 46 1.5 Hypothesis and aims ............................................................................................. 47 1.5.1 Hypothesis ........................................................................................................ 47 1.5.2 Specific aims ..................................................................................................... 47 2 Materials and methods ................................................................................................. 49 2.1 Materials ............................................................................................................... 49 2.2 Neutrophils ........................................................................................................... 53 2.2.1 Neutrophil preparation ....................................................................................... 53 2.2.1.1 Neutrophil isolation from human whole blood ............................................. 53 2.2.1.2 Neutrophil isolation from murine bone marrow ........................................... 55 2.2.1.3 Generation of neutrophil supernatants ........................................................ 55 2.2.1.4 Optimisation of neutrophil cell lysate generation ......................................... 56 2.2.1.5 Generation of neutrophil cell lysates ........................................................... 61 2.2.2 Working under hypoxia ...................................................................................... 61 2.2.3 Neutrophil functional assays .............................................................................. 62 2.2.3.1 Assessment of shape change .................................................................... 62 2.2.3.2 Assessment of apoptosis by morphology ................................................... 64 2.2.3.3 Assessment of apoptosis by flow cytometry ............................................... 65 2.2.3.4 Assessment of neutrophil reactive oxygen species production ................... 65 2.2.3.5 Assessment of neutrophil extracellular trap production ............................... 66 2.2.4 Neutrophil degranulation assays ....................................................................... 67 2.2.4.1 Assessment of neutrophil elastase release................................................. 67 2.2.4.2 Optimisation of the NE activity assay for murine neutrophils ...................... 67 2.2.4.3 Assessment of human NE release with PI3Kinase inhibition ...................... 69 2.2.4.4 Assessment of NE release from murine neutrophils with PI3Kinase mutations .................................................................................................................. 69 2.2.4.5 Assessment of MPO release ...................................................................... 70 2.2.5 Protein detection methods ................................................................................. 70 2.2.5.1 Sample preparation for gel electrophoresis ................................................ 70 2.2.5.2 SDS-PAGE................................................................................................. 71 2.2.5.3 Western blotting ......................................................................................... 72 2.2.5.4 Silver staining ............................................................................................. 73 2.3 Endothelial Cells ................................................................................................... 73 2.3.1 Endothelial cell culture ....................................................................................... 73 2.3.2 Endothelial cell functional assays ...................................................................... 74 2.3.2.1 Assessment of endothelial ICAM-1 expression by flow cytometry .............. 74 2.3.3 Endothelial cell viability assays .......................................................................... 75 2.3.3.1 Assessment of cell detachment by immunofluorescence ............................ 75 2.3.3.2 Assessment of cell viability by MTT assay .................................................. 75 2.3.3.3 Assessment of apoptosis by flow cytometry ............................................... 76 2.4 Proteomics ............................................................................................................ 79 2.4.1 Generation of neutrophil supernatants for proteomics ....................................... 79 2.4.2 Two dimensional difference gel electrophoresis ................................................ 79 2.4.3 10-plex tandem mass tag-labelled mass spectrometry ...................................... 82 2.4.4 Quantification of supernatant resistin, NGAL, cyclophilin A, S100A9, S100A8 and S100A8/A9 by ELISA ................................................................................................... 84 2.5 Microvesicles ........................................................................................................ 84 2.5.1 Microvesicle isolation from Histopaque®-1077-prepared neutrophils................. 84 2.5.2 Quantification of neutrophil-derived microvesicles by flow cytometry ................. 86 2.5.3 Preparation of microvesicle lysates for western blotting ..................................... 86 2.6 COPD study .......................................................................................................... 87 2.6.1 COPD patient recruitment ................................................................................. 87 2.6.2 Obtaining plasma and serum from whole blood ................................................. 88 2.6.3 Quantification of Aα-Val360 and Aα-Val541 ........................................................... 88 2.6.4 Quantification of plasma microvesicles by flow cytometry .................................. 89 2.7 Statistics ............................................................................................................... 89 3 Effect of hypoxia on NE release and neutrophil-mediated endothelial dysfunction ....... 92 3.1 Introduction ........................................................................................................... 92 3.2 Confirmation of the neutrophil hypoxic response ................................................... 93 3.3 The effect of different priming agonists on hypoxic NE release ............................. 94 3.4 Investigation of the mechanism of NE release under hypoxia ............................... 97 3.4.1 The effect of hypoxia on NET production ........................................................... 97 3.4.2 The effect of PI3K signalling modulation on the release of NE under hypoxia ... 98 3.4.2.1 The effect of PI3K inhibitors on the release of NE from human neutrophils under hypoxia ......................................................................................................... 100 3.4.2.2 The effect of PI3K isoform mutations in murine neutrophils on NE release under hypoxia ......................................................................................................... 103 3.5 The effect of hypoxia on neutrophil-induced pulmonary artery endothelial cell dysfunction .................................................................................................................... 105 3.5.1 The effect of neutrophil supernatants on endothelial cell activation ................. 105 3.5.2 The effect of hypoxia on neutrophil-induced endothelial cell detachment ........ 106 3.5.3 The effect of hypoxia on neutrophil-induced endothelial cell death .................. 108 3.6 Discussion .......................................................................................................... 113 4 Effect of hypoxia on the neutrophil secretome ............................................................ 120 4.1 Introduction ......................................................................................................... 120 4.2 2D Difference Gel Electrophoresis ...................................................................... 122 4.3 Optimisation of neutrophil supernatant generation for TMT-labelled MS ............. 125 4.4 Characterisation of the normoxic versus hypoxic neutrophil secretome by 10-plex tandem mass tag-labelled mass spectrometry (TMT-MS) .............................................. 132 4.5 Biochemical validation of differentially regulated proteins identified by TMT-MS . 137 4.6 The role of NDMV secretion in the differential protein release from neutrophils under hypoxia ................................................................................................................ 142 4.6.1 Quantification of NDMV release from normoxic and hypoxic neutrophils ......... 142 4.6.2 Evaluation of cyclophilin A protein content of NDMVs from normoxic and hypoxic neutrophils ................................................................................................................. 143 4.7 Discussion .......................................................................................................... 147 5 Effect of hypoxia on neutrophil function in exacerbating COPD patients .................... 156 5.1 Introduction ......................................................................................................... 156 5.2 Investigation of COPD versus healthy neutrophil secretion of proteins upregulated in the hypoxic secretome ............................................................................................... 158 5.2.1 The effect of hypoxia on the release of NE and MPO from COPD versus healthy neutrophils ................................................................................................................. 158 5.2.2 The effect of hypoxia on the release of resistin, NGAL and cyclophilin A from COPD versus healthy neutrophils .............................................................................. 161 5.3 COPD versus healthy plasma content of proteins upregulated in the hypoxic neutrophil secretome ..................................................................................................... 164 5.4 The COPD versus healthy plasma content of microvesicles ............................... 166 5.5 Discussion .......................................................................................................... 166 6 Discussion ................................................................................................................. 173 6.1 Overview ............................................................................................................. 173 6.2 Discussion of results ........................................................................................... 174 6.2.1 Differential effects of priming agents on hypoxic degranulation ....................... 174 6.2.2 Role of PI3K signalling in hypoxic degranulation ............................................. 175 6.2.3 Possible mechanisms of endothelial damage by hypoxic supernatants ........... 178 6.2.4 Proteomic analysis of the neutrophil secretome............................................... 180 6.2.5 Potential mechanisms of “differential degranulation” ....................................... 181 6.2.6 Enhanced release of cytoplasmic proteins ....................................................... 181 6.2.7 Translation to COPD and beyond .................................................................... 182 6.3 Future research avenues .................................................................................... 185 6.3.1 Investigating the role of hypoxia-upregulated proteins in mediating clinically relevant endothelial damage ...................................................................................... 185 6.3.2 Investigating how hypoxia impacts neutrophil intracellular signalling and granule trafficking pathways to enhance histotoxic protein release ......................................... 187 6.4 Conclusions ........................................................................................................ 188 7 Appendices ................................................................................................................ 190 7.1 10-plex TMT-MS and data analysis parameters .................................................. 190 7.1.1 LC-MS/MS ....................................................................................................... 190 7.1.2 Data analysis ................................................................................................... 191 7.2 Proteins identified by 10-plex TMT-MS ............................................................... 192 7.3 Publications arising from this thesis .................................................................... 199 7.3.1 Papers ............................................................................................................. 199 7.3.2 Abstracts ......................................................................................................... 199 8 Bibliography ............................................................................................................... 201 List of Figures Figure 1.1: Neutrophil signalling pathways involved in degranulation .................................. 20 Figure 1.2: Control of differential granule exocytosis ........................................................... 24 Figure 1.3: Neutrophil functions under conditions of normoxia versus hypoxia .................... 37 Figure 2.1: Isolation of neutrophils from human whole blood ............................................... 54 Figure 2.2: Expression of total AKT by western blot ............................................................ 57 Figure 2.3: The ability of monoclonal versus polyclonal antibodies to detect pAKT and AKT by western blot .................................................................................................................... 58 Figure 2.4: The effect of different buffers and lysate preparation methods on pAKT and AKT protein detection by western blot ......................................................................................... 60 Figure 2.5: Analysis of media pH, pO2 and pCO2 ................................................................ 62 Figure 2.6: Gating strategy and quantification of neutrophil shape change by flow cytometry ........................................................................................................................................... 63 Figure 2.7: Assessment of neutrophil apoptosis by morphology .......................................... 64 Figure 2.8: Gating strategy for assessment of neutrophil apoptosis by flow cytometry ........ 66 Figure 2.9: The effect of an NE-selective inhibitor on the detection of supernatant NE activity ........................................................................................................................................... 68 Figure 2.10: The effect of priming agonists and an NE-selective inhibitor on the detection of murine supernatant NE activity ........................................................................................... 69 Figure 2.11: Gating strategy for assessment of HPAEC apoptosis by flow cytometry: baseline and compensation................................................................................................. 77 Figure 2.12: Gating strategy for assessment of HPAEC apoptosis by flow cytometry: supernatant treatment ......................................................................................................... 78 Figure 2.13: 2D DIGE analysis by Decyder 2D DIA software .............................................. 81 Figure 2.14: Gating strategy for isolated NDMV quantification by flow cytometry ................ 87 Figure 2.15: Gating strategy for plasma microvesicle quantification by flow cytometry ........ 90 Figure 3.1: The effect of hypoxia on neutrophil apoptosis ................................................... 95 Figure 3.2: The effect priming agonists and hypoxia on NE release .................................... 96 Figure 3.3: The effect of hypoxia on NET production .......................................................... 99 Figure 3.4: The effect of PI3K inhibition on the hypoxic regulation of AKT phosphorylation101 Figure 3.5 The effect of PI3K inhibition on the hypoxic regulation of NE release ............... 102 Figure 3.6: The effect of hypoxia on NE release from murine neutrophils with PI3Kδ or PI3Kγ mutations .......................................................................................................................... 104 Figure 3.7: The effect of hypoxia and neutrophil supernatants on endothelial-leukocyte ICAM-1 expression by flow cytometry ............................................................................... 107 Figure 3.8: The effect of neutrophil supernatants on HPAEC detachment by confocal microscopy ........................................................................................................................ 109 Figure 3.9: The effect of neutrophil supernatants on HPAEC survival by MTT assay ........ 111 Figure 3.10: The effect of neutrophil supernatants on HPAEC survival by flow cytometry . 112 Figure 4.1: The effect of buffer and precipitation on neutrophil supernatant protein content ......................................................................................................................................... 124 Figure 4.2: Analysis of hypoxic versus normoxic neutrophil supernatant protein content by 2D DIGE ........................................................................................................................... 126 Figure 4.3: The effect of protease inhibitors on neutrophil supernatant protein content .... 127 Figure 4.4: The effect of small molecule protease inhibitors on neutrophil supernatant protein content .............................................................................................................................. 129 Figure 4.5: The effect of small molecule protease inhibitors on neutrophil morphology ..... 130 Figure 4.6: Optimised protease inhibition strategy ............................................................ 131 Figure 4.7: Comparison of TCA-precipitation and spin column protein concentration ........ 132 Figure 4.8: Characterisation of the normoxic versus hypoxic neutrophil secretome by TMT- MS .................................................................................................................................... 136 Figure 4.9: The effect of hypoxia on resistin, NGAL and cyclophilin A release from neutrophils by ELISA......................................................................................................... 138 Figure 4.10: The effect of hypoxia on S100A8 and S100A9 release from neutrophils by ELISA ............................................................................................................................... 140 Figure 4.11: The effect of hypoxia on S100A9 release from neutrophils by western blot ... 141 Figure 4.12: Comparison of neutrophil isolation methods and the effect of hypoxia on NDMV release .............................................................................................................................. 144 Figure 4.13: Neutrophil-derived microvesicle and supernatant content of annexin A1 and cyclophilin A by western blot ............................................................................................. 145 Figure 4.14: Quantification of neutrophil-derived microvesicle content of annexin A1 and cyclophilin A by western blot ............................................................................................. 146 Figure 4.15: Visualisation of MVs from neutrophils and endothelial cells by electron microscopy ........................................................................................................................ 151 Figure 5.1: Quantification of healthy versus COPD neutrophil shape change by flow cytometry .......................................................................................................................... 159 Figure 5.2: The effect of hypoxia on NE and MPO release from healthy and COPD neutrophils ........................................................................................................................ 162 Figure 5.3: The effect of hypoxia on resistin and NGAL release from healthy and COPD neutrophils ........................................................................................................................ 162 Figure 5.4: The effect of hypoxia on cyclophilin A release from healthy and COPD neutrophils ........................................................................................................................ 163 Figure 5.5: The healthy versus COPD plasma content of Aα-Val360 and Aα-Val541 ............ 164 Figure 5.6: The healthy versus COPD plasma content of resistin, NGAL and cyclophilin A by ELISA ............................................................................................................................... 165 Figure 5.7: Healthy versus COPD plasma microvesicle content ........................................ 167 Figure 6.1: The effect of neutrophil supernatants on IL-1β ................................................ 179 List of Tables Table 1.1: Contents of neutrophil granules .......................................................................... 19 Table 2.1: Antibodies .......................................................................................................... 52 Table 4.1: Comparison of protein concentration methods ................................................. 123 Table 4.2: Proteins significantly increased in normoxia ..................................................... 134 Table 4.3: Proteins significantly increased in hypoxia........................................................ 135 Table 5.1: Clinical and demographic data ......................................................................... 160 Table 7.1: Proteins increased in normoxia ........................................................................ 192 Table 7.2: Proteins increased in hypoxia ........................................................................... 195 Table 7.3: Differentially regulated proteins identified by 10-plex TMT-MS compared with published secretomes and sepsis patient plasma ............................................................. 198 List of Abbreviations 2D DIGE Two dimensional difference gel electrophoresis α1AT α1 antitrypsin AEBSF 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride ALI Acute lung injury ATP Adenosine triphosphate ARDS Acute respiratory distress syndrome BALF Bronchoalveolar lavage fluid BSA Bovine serum albumin BMP9 Bone morphogenetic protein 9 Ca2+ Calcium ion CaCl2 Calcium chloride CAM Cell adhesion molecule CF Cystic fibrosis CGD Chronic granulomatous disease CHX Cycloheximide COPD Chronic obstructive pulmonary disease CRP C-reactive protein CVD Cardiovascular disease DAPI 4',6-diamidine-2'-phenylindole dihydrochloride DMB o-dianisidine dihydrochloride DMSO Dimethyl sulfoxide DTT Dithiothreitol EDTA Ethylenediamine tetraacetic acid EGTA Ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid ELISA Enzyme-linked immunosorbent assay ECM Extracellular matrix ER Endoplasmic reticulum FACS Fluorescence-activated cell sorting FC Fold change FCS Foetal calf serum FDR False discovery rate FIH Factor inhibiting HIF FITC Fluorescein isothiocyanate FMD Flow-mediated dilatation fMLP N-formyl-methionyl-leucyl-phenylalanine GAP GTPase-activating protein GDP Guanosine diphosphate GEF Guanine nucleotide exchange factor GM-CSF Granulocyte-macrophage colony-stimulating factor GPCR G-protein coupled receptor GTP Guanosine triphosphate H2O2 Hydrogen peroxide HCl Hydrochloric acid HIF Hypoxia-inducible factor HOCl Hypochlorous acid HPAEC Human pulmonary artery endothelial cell HPLC High performance liquid chromatography HRE Hypoxia response element HRP Horseradish peroxidase HUVEC Human umbilical vein endothelial cell IBD Inflammatory bowel disease ICAM-1 Intercellular adhesion molecule-1 IEF Isoelectric focusing I/R Ischaemia/reperfusion IL Interleukin IMDM Iscove's modified dulbecco's medium IP3 Inositol 1,4,5-trisphosphate ITAM Immunoreceptor tyrosine-based activation motif kDa Kilodalton kPa Kilopascal LPS Lipopolysaccharide LTB4 Leukotriene B4 Mg2+ Magnesium ion MMP Matrix metalloproteinase MPO Myeloperoxidase MS Mass spectrometry mTOR Mammalian target of rapamycin MTT 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide MV Microvesicle NaCl Sodium chloride NADP Nicotinamide adenine dinucleotide phosphate NDMV Neutrophil-derived microvesicle NE Neutrophil elastase NET Neutrophil extracellular trap NF-κB Nuclear factor-κB NGAL Neutrophil gelatinase-associated lipocalin OSA Obstructive sleep apnoea PAD4 Peptidylarginine deiminase 4 PAF Platelet activating factor PAFR PAF receptor PAGE Polyacrylamide gel electrophoresis PBMC Peripheral blood mononuclear cells PBS+/+ Phosphate-buffered saline with calcium/magnesium chloride PBS-/- Phosphate-buffered saline without calcium/magnesium chloride PCA Principal component analysis PET Positron emission tomography PFA Paraformaldehyde PH Pulmonary hypertension PHD Prolyl hydroxylase domain PI3K Phosphoinositide 3-kinase PIP2 Phosphatidylinositol-(4,5)-bisphosphate PIP3 Phosphatidylinositol-(3,4,5)-trisphosphate PLC Phospholipase C PMA Phorbol myristate acetate PMSF Phenylmethylsulfonyl fluoride pO2 Partial pressure of oxygen PPP Platelet-poor plasma PR3 Proteinase 3 PRP Platelet-rich plasma PVDF Polyvinylidene fluoride ROS Reactive oxygen species RA Rheumatoid arthritis RBC Red blood cells rTEM reverse trans-endothelial migration SDS Sodium dodecyl sulphate SNAP Synaptosome associated protein SNARE Soluble NSF-attachment protein receptor TB Tuberculosis TCA Trichloroacetic acid TGF Transforming growth factor TLR Toll-like receptor TMT-MS Tandem mass tag-labelled mass spectrometry TNFα Tumour necrosis factor-α TNFR TNF receptor VAMP Vesicle-associated membrane protein VEGF Vascular endothelial growth factor vHL von Hippel-Lindau WBC White blood cells Chapter 1 Introduction 13 1 Introduction 1.1 The innate immune system The innate immune system, comprising physical barriers, and chemical and cellular components, protects the body from invading pathogens. The anatomical barriers of the respiratory system include the muco-ciliary escalator, whereby micro-organisms trapped by mucus in the airways are propelled towards the pharynx by the co-ordinated action of cilia, and the epithelial surface lining, which provides a structural barrier and secretes a variety of host defence molecules, including defensins and lysozyme. Additionally, resident alveolar macrophages are the primary phagocytic cells of the uninflamed lower airways, scavenging inhaled pathogens and environmental particles at the interface between air and lung tissue. Once a pathogen breaches these initial barriers, the ensuing inflammatory response initiated by the release of cytokines and chemicals from injured cells recruits the professional cells of the innate immune system. 1.2 Neutrophils Neutrophils are largely absent from the healthy lung, which has resident defences including the muco-ciliary escalator, antimicrobial peptides and proteins produced by airway epithelial cells, and alveolar macrophages. If these are overcome, neutrophils are recruited from the circulation, rapidly recognising and eliminating pathogens via a number of cytotoxic mechanisms, and modulating both the innate and adaptive host inflammatory response. Neutrophils also contribute to the recruitment of other immune cells, thereby amplifying the inflammatory cascade. Their striking cytological structure sheds light on several effector functions and is reflected in nomenclature: their cytoplasm contains numerous bactericidal granules and hence they are often termed granulocytes, and the characteristic multi-lobed nucleus confers the deformability to squeeze through cell junctions, giving rise to the term ‘polymorphonuclear leukocyte’ or polymorph. Neutrophils comprise 50% to 70% of circulating white blood cells and are generated in the bone marrow at a rate of 1011 per day, increasing to up to 1012 per day at times of infection (1). Neutrophil homeostasis is maintained through a controlled balance of granulopoeisis within and release from the bone marrow, margination in intravascular pools (including the liver and spleen), tissue recruitment, and cell death and clearance (2). Neutrophil half-life is thought to be short, with a generally accepted survival of 8-12 h in the circulation and 1-2 days in the tissues, although this can be modulated by the inflammatory response (3). However, some recent studies have challenged this paradigm, indicating a circulatory neutrophil lifespan of up to several days (reviewed in (4)). 14 In order to prevent inappropriate triggering within the circulation, neutrophils are required to undergo a two-step activation process, whereby an initial “priming” stimulus (first described by Guthrie et al. (5)) generates a pre-activated (primed) phenotype, which results in a substantially enhanced cytotoxic response upon encountering an activating stimulus or environment, such as a bacterial focus of infection. Once in the vicinity of their prey, neutrophils employ myriad trapping and killing mechanisms, including organism engulfment by phagocytosis, production of reactive oxygen species (ROS), phagosome-fusion or exocytosis (degranulation) of granules containing multiple antimicrobial proteins and proteases, and release of neutrophil extracellular traps (NETs). Subsequently, neutrophils are cleared: apoptotic cells are predominantly removed via phagocytosis by macrophages (efferocytosis), which hastens the resolution of inflammation. Although neutrophils are extremely well adapted for their dual role of immune surveillance and the elimination of invading pathogens, damage to host tissue from excessive or persistent neutrophilic inflammation has been implicated in the pathogenesis of numerous diseases, including both acute (e.g. acute respiratory distress syndrome, ARDS) and chronic (e.g. chronic obstructive pulmonary disease, COPD) inflammatory lung diseases. 1.2.1 Effector functions of neutrophils 1.2.1.1 Priming and activation In health, circulating neutrophils are quiescent but upon exposure to a wide range of priming agents, including cytokines such as granulocyte-macrophage colony-stimulating factor (GM- CSF) and tumour necrosis factor α (TNFα), or phospholipid mediators such as platelet activating factor (PAF), they are able to transition from an inactive state to an adhesive, primed phenotype. Primed neutrophils rapidly become less deformable (6), which may promote their retention in vascular beds (7). Priming also induces rapid shedding of L-selectin (which mediates initial neutrophil rolling interactions with endothelium) and increases the surface expression and enhances the function of β2 integrin (CD11b/CD18, MAC-1), which binds endothelial surface cell adhesion molecules (CAMs) to mediate firm adhesion to the vessel wall (8). Together, these changes promote migration of neutrophils through the vascular endothelium to sites of infection and/or tissue injury. Critically, priming agonists also markedly potentiate the neutrophil’s response to subsequently encountered activating stimuli, such as bacterial formylated peptide (fMLP), allowing swift initiation and potent augmentation of pathogen seeking (chemotaxis), ensnaring (e.g. NET release) and killing (e.g. ROS production) mechanisms in order to enhance bactericidal capacity (reviewed in (9)). Individual priming agents, however, result in subtly different regulation of phenotypic changes. For example, TNFα- and PAF-priming of neutrophil β2 integrin upregulation and activation display 15 distinct kinetic and functional profiles, with TNFα producing a sustained β2 integrin activation but PAF-induced activity declining within minutes of agonist treatment (8). Several priming agents, e.g. GM-CSF, increase neutrophil longevity by delaying apoptosis, thus extending the inflammatory response but, in contrast, TNFα can exert either a pro- or anti-apoptotic effect, which is both concentration- and time-dependent (10). 1.2.1.2 Recruitment Circulating neutrophils constantly patrol the endothelium for signs of inflammation and infection. Stimulation of the endothelium by inflammatory mediators, such as TNFα, upregulates endothelial selectins (e.g. E-selectin) which bind their glycosylated ligands on neutrophils (e.g. PSGL-1). This induces tethering and rolling of neutrophils along the blood vessel wall such that, once the degree of interaction is sufficient, the rolling neutrophil becomes captured. Chemokines, such as IL-8, decorate the endothelial surface; chemokine receptor ligation initiates priming of arrested neutrophils to induce mobilisation of intra-cellular integrin stores and conformational changes of cell surface-expressed integrins, which can then bind endothelial CAMs, e.g. ICAM-1, an essential process for firm adhesion. In response to priming and endothelial adhesion, neutrophils develop a polarised morphology, with cytoskeletal F actin polymerisation at the leading edge, enabling directional movement and facilitating trans-endothelial migration (reviewed in (11)). Neutrophil migration towards sites of inflammation and infection, a process termed chemotaxis, is directed by chemoattractant concentration gradients. Chemoattractant control of neutrophil migration is complex as the in vivo milieu is a dynamic environment, comprising multiple chemoattractant signals in intricate concentration, spatial and temporal patterns. Successful homing requires neutrophils to display multiple receptors, including those for chemoattractants released by the endothelium (e.g. leukotriene B4, LTB4) and so called “end target” chemoattractants, released in the immediate vicinity of the infective or inflammatory stimulus (e.g. bacterial fMLP). Foxman et al. have shown that neutrophil chemotaxis is a multistep process with sequential navigation through a series of gradients (12). Initially thought to be predominantly phosphoinositide-3-kinase (PI3K)-dependent (13), it has now been demonstrated that there is an intracellular signalling hierarchy with end target chemoattractants signalling though p38 MAPK (14). The traditional neutrophil recruitment paradigm of tethering, rolling, firm adhesion, crawling and finally endothelial transmigration is context-dependent. In the lung, neutrophils leave the pulmonary circulation though capillaries, as opposed to the post-capillary venule exit route from the systemic circulation. As a consequence of the unique lung microvasculature (an extensive alveolar capillary network), pulmonary neutrophil sequestration is reliant, at least in 16 part, on rheological changes due to space constraints and, in this situation, integrin- independent pulmonary neutrophil accumulation has been shown (15). Two aspects of neutrophil behaviour described recently in animal models may contribute to neutrophil recruitment and to its resolution, although their significance in humans is currently uncertain. Firstly, neutrophils have the capacity to promote their own rapid and massive recruitment in ‘swarms’, largely choreographed by the release of the priming agent LTB4 (reviewed in (16)). This may be beneficial, walling off infection (17), or detrimental due to vascular occlusion (18). Secondly, movement of primed neutrophils from the extravascular tissue back into the vascular lumen has been described: reverse trans-endothelial cell migration (rTEM). Colom et al. demonstrated neutrophil rTEM in a murine model of ischaemia/reperfusion (I/R) injury, which was dependent on proteolytic cleavage of the endothelial junctional adhesion molecule JAM-C by the neutrophil granule protease neutrophil elastase (NE) and was associated with the development of a systemic inflammatory response secondary to the local sterile injury (19). Thus, rTEM may also contribute to endothelial injury and remote organ dysfunction. 1.2.1.3 Phagocytosis and Reactive Oxygen Species Neutrophils are professional phagocytes with the ability to recognise and rapidly ingest pathogens. The process is initiated by ligation of phagocytic receptors, either by endogenous components of the particle or by host-derived opsonins (antibodies or complement components). Target particles are engulfed into the phagosome, a plasma membrane-derived vacuole formed by extension of neutrophil pseudopods which surround and enclose the particle. Phagocytic receptor ligation initiates a phosphorylation cascade, allowing recruitment of multiple adaptor and signalling proteins to facilitate pseudopod extension by means of dynamic changes in the actin cytoskeleton (reviewed in (20)). When initially formed, phagosomes are not intrinsically antimicrobial. As they mature into phagolysosomes, microbicidal activity is acquired by vacuole fusion with endocytic compartments, including cytosolic granule populations, containing abundant proteases and antimicrobial peptides (21– 23) and subunits of NADPH oxidase, assembly of which enables ROS production (24). Generation of ROS through activation of the NADPH oxidase electron transport chain plays a critical role, killing many (though not all) pathogens. NADPH oxidase is an electron donor which reduces molecular di-oxygen to form superoxide anion, spontaneous or catalytically- driven dismutation of which yields a range of antimicrobial ROS (reviewed in (25)). The prevailing view of the role of ROS in bacterial killing has been the generation of highly toxic oxidants in the presence of myeloperoxidase (MPO), which then catalyses the oxidation of halides, including cytotoxic hypochlorous acid (HOCl). However, Segal and colleagues have 17 challenged this dogma, suggesting that the primary role of NADPH oxidase is electron delivery to the phagosome, with the compensatory cation influx alkalinising the vacuolar pH so that it is optimal for antibacterial protease activity (26). Controversially, as generation of HOCl by MPO has long been thought to be fundamental to phagosomal antimicrobial activity, there is minimal peroxidase and chlorinating activity of MPO at this alkaline pH. It may be that this mechanism is more important in situations outside the phagosome, such as an incompletely phagocytosed particle. Still, Green et al. maintain that HOCl is vital for ingested bacterial killing, demonstrating that decreased chlorination in the phagosome halved killing of S. aureus (27), and modifications of host proteins by HOCl may provide further active antimicrobial species, such as chloramines. Whether ROS act directly or indirectly, the importance of this mechanism for neutrophil-mediated pathogen handling is highlighted by chronic granulomatous disease (CGD), a rare genetic disorder caused by a defective NADPH oxidase complex. As a consequence, these patients suffer severe recurrent infections with fungi and several species of bacteria (28). 1.2.1.4 Granule formation and composition As the name “granulocyte” suggests, neutrophils contain various granule populations, enclosing multiple antimicrobial proteins and proteases, as well as a reserve of adhesion molecules and receptors. In order to effect pathogen killing, these granules can either fuse with the phagosome, releasing their cytotoxic contents into the phagocytic vacuole to destroy ingested micro-organisms, or with the plasma membrane, secreting their arsenal of proteins extracellularly, termed degranulation. The fate of granule fusion is determined by the potency of the activating stimulus. Granulopoiesis is temporally-regulated during neutrophil maturation, initiated as myeloblasts differentiate into promyelocytes. Although in reality they comprise a spectrum, each granule sub-type contains a broadly different protein cargo, and this differential protein composition has led to granule classification. The first to be formed are the azurophil (primary) granules, followed sequentially by the specific (secondary) and then gelatinase (tertiary) granules, and finally by the secretory vesicles. The prevailing hypothesis of granule heterogeneity is that of protein targeting by timing of biosynthesis, proposed by Borregaard et al., whereby concomitantly formed proteins are directed into granules together, and different granule sub- types are formed as protein expression changes during neutrophil maturation (29). This is exemplified by forced expression of neutrophil gelatinase-associated lipocalin (NGAL), (usually synthesised in metamyelocytes and targeted to specific granules) in the promyelocytic HL-60 cell line, which instead resulted in its packaging into azurophil granules (30). However, it has also been shown that active sorting of proteins to granules occurs; for example, patients 18 lacking the β3A subunit of adaptor protein 3 (AP3; Hermansky-Pudlak type 2 syndrome) have a dramatic reduction of azurophil granule NE content (31), although it is not clear whether this is due to a failure of localisation, insufficient retention or enhanced degradation of this protease. In depth characterisation of granule sub-types has been performed, with cell fractionation and proteomic techniques describing granule protein composition; this is relatively distinct between sub-types although there is a degree of overlap, particularly between specific and gelatinase granules (32,33) (see Table 1.1). MPO is the defining protein of azurophil granules, hence they are designated peroxidase-positive, with other major constituents including serine proteases (e.g. NE, proteinase 3 (PR3) and cathepsin G). Specific granules contain abundant lactoferrin and matrix metalloproteinase-8 (MMP-8, collagenase), whilst the exemplar protein of gelatinase granules is MMP-9 (gelatinase). Specific and gelatinase granules also contain the membrane subunits of NADPH oxidase (p22phox and gp91phox), which are delivered to the NADPH oxidase complex to enable ROS production. Secretory vesicles are smaller than granules and have an endocytic origin, containing plasma proteins as well as receptors which can be mobilised to the neutrophil surface, e.g. MAC-1 and the fMLP receptor. It is noteworthy that in addition to proteases, neutrophil granules also contain anti-proteases, such as α1 antitrypsin (α1AT) and secretory leukocyte protease inhibitor, which may act to restrain the pro-inflammatory effects of exocytosis (34,35). 1.2.1.5 Degranulation Neutrophil granule exocytosis can be initiated by a range of stimuli, e.g. fMLP, with granule release augmented by priming agents, e.g. PAF. Neutrophils express a variety of cell surface receptor classes, including Fc receptors, adhesion receptors (e.g. integrins), G protein coupled receptors (GPCRs), cytokine receptors (e.g. TNF receptors, TNFRs) and innate immune receptors (e.g. toll-like receptors, TLRs), responsible for the control of neutrophil effector functions (reviewed in (36)). Although neutrophil degranulation signalling pathways are not completely delineated, activation of integrins (37), Fc receptors (38) and GPCRs (39) has been implicated. Receptor ligation initiates intracellular signalling cascades; the precise signalling pattern likely depends on the range, concentration and context of external stimuli, and downstream signalling events may occur in parallel. However, these pathways converge on common degranulation effector mechanisms, principally cytoskeletal rearrangement, and granule/vesicle trafficking and membrane fusion (Figure 1.1). 19 Table 1.1: Contents of neutrophil granules Compiled from the following references: (32,33,40–42) Azurophil Granules Specific Granules Gelatinase Granules Matrix Matrix Matrix ▪ Acid β glycerophosphatase ▪ Acid mucopolysaccharide ▪ α1 antitrypsin ▪ α mannosidase ▪ Azurocidin ▪ Bacterial permeability increasing protein ▪ β glucuronidase ▪ β glycerophosphatase ▪ Cathepsins G and C ▪ Defensins ▪ Epididymal secretory protein E1 ▪ Leukocyte elastase inhibitor ▪ Lysozyme ▪ Myeloperoxidase ▪ Neutrophil elastase ▪ Proteinase 3 ▪ Resistin ▪ Sialidase ▪ 14-3-3 protein zeta/delta ▪ α1 antitrypsin ▪ Apolipoprotein B receptor ▪ β2 microglobulin ▪ Cathelicidin ▪ Collagenase (MMP8) ▪ Coronin 1A ▪ γ glutamyl hydrolase ▪ Folate receptor γ ▪ Gelatinase (MMP9) ▪ Heparanase ▪ Histaminase ▪ Haptoglobin ▪ ICAM 3 ▪ Integrin α M ▪ Lactoferrin ▪ Leukocyte specific protein 1 ▪ Lysozyme ▪ Metalloproteinase inhibitor 2 ▪ Moesin ▪ Neutrophil gelatinase- associated lipocalin (NGAL) ▪ Olfactomedin 4 ▪ Pentraxin 3 ▪ Resistin ▪ Secretory leukocyte protease inhibitor ▪ Sialidase ▪ Talin 1 ▪ 14-3-3 protein zeta/delta ▪ Acetyltransferase ▪ Apolipoprotein B receptor ▪ Arginase 1 ▪ β2 microglobulin ▪ Cathelicidin ▪ Coronin 1A ▪ Gelatinase (MMP9) ▪ Heparanase ▪ ICAM 3 ▪ Integrin β 2 ▪ Leukocyte specific protein 1 ▪ Lysozyme ▪ Ficolin 1 ▪ Ras-related protein Rab 11B ▪ Ras GTPase activating like protein IQGAP1 ▪ Talin 1 ▪ Transcobalamin 1 Membrane Membrane Membrane CD63, CD68 Presenilin 1 V type H+ ATPase CD11b/CD18, CD66a, CD66b, CD15, CD45, CD177 Cytochrome b558 Fibronectin R, Laminin R, fMLP R, p55 TNF R, IL-10 R, Thrombospondin R Leukolysin NOX2 Rap 1, Rap 2 SCAMP SNAP-23 VAMP-2 CD11b/CD18, CD45, CD11c/CD18, CD66a, CD66b, CD177 Cytochrome b558 1-diacylglycerol lipase fMLP R Leukolysin NOX2 Rap 1, Rap 2 SCAMP SNAP 23 VAMP-2 V type H+ ATPase 20 Figure 1.1: Neutrophil signalling pathways involved in degranulation β2 integrins and Fc receptors, coupled to an immunoreceptor tyrosine-based activation motif (ITAM, which is phosphorylated by Src family kinases), signal through the Syk tyrosine kinase to activate downstream pathways resulting in cytoskeletal reorganisation (e.g. via Rac2) and a rise in intracellular calcium (Ca2+). GPCR ligation leads to dissociation of the heterotrimeric G protein into Gα and Gβγ subunits. Gα and/or the G adaptor protein, β arrestin, may activate Src family kinases. The Gβγ dimer triggers two parallel signalling events: 1) activation of phospholipase C (PLC) β hydrolyses PIP2, generating diacylglycerol (DAG) and IP3, which enables calcium release from the endoplasmic reticulum (ER) intracellular stores; 2) activation of PI3Kγ induces the production of PIP3 by phosphorylation of PIP2, signalling through the AKT signalling cascade and activating Rac2. Depletion of ER calcium is sensed, thus opening plasma membrane ion channels to allow influx of extracellular calcium. The ultimate result of all of these processes is to allow granule translocation and exocytosis. 1.2.1.6 Degranulation: receptors and intracellular signalling pathways Both integrins and Fc receptors signal through Src tyrosine kinases, with Hck, Fgr and Lyn being the most relevant family members in neutrophils. These Src tyrosine kinases are able to phosphorylate immunoreceptor tyrosine-based activation motifs (ITAMs) coupled to β2 integrins and Fc receptors, with subsequent signal transduction mediated by association of the Syk tyrosine kinase with phosphorylated ITAMs. Localisation of scaffolding proteins through activation of these tyrosine kinases promotes actin and microtubule polymerisation, recruitment of further signalling proteins such as the adaptor protein SLP-76, and a rise in 21 intracellular calcium (43,44), leading to granule exocytosis. Several key downstream pathways are activated, with degranulation mediated through phospholipase C (PLC) γ2, and Rho GTPase (e.g. Rac) signalling. Murine neutrophils lacking PLCγ2 (45), Syk (46) or both Hck and Fgr (47) could still adhere to a fibrinogen surface, but failed to degranulate in response to TNFα following β2 integrin or Fc receptor ligation. An additional tyrosine kinase, Pyk2, has been implicated in integrin-induced neutrophil degranulation, although deficiency did not produce as profound a defect in degranulation since its role is further downstream (48). These studies have demonstrated both detrimental (failure to clear pathogens (48)) and beneficial (protection from the development of arthritis (45)) consequences in murine in vivo models, both thought to result from impaired degranulation. GPCRs (including fMLP and PAF receptors) are heterotrimeric proteins which, upon activation, dissociate into Gα and Gβγ subunits; the majority of neutrophil signal transduction is thought to occur through the Gβγ fragment. Gβγ activates two parallel pathways: activation of phospholipase Cβ (PLCβ) generates inositol-trisphosphate (IP3) from phosphatidylinositol- (4,5)-bisphosphate (PIP2), resulting in release of calcium from intracellular stores, and activation of PI3Kγ generates phosphatidylinositol-(3,4,5)-trisphosphate (PIP3) from PIP2, with subsequent signalling through the AKT cascade. Both the elevated calcium (49) and AKT activation (50) lead to granule exocytosis, and pan-PI3K inhibition or genetic deletion of the AKT2 isoform have been shown to inhibit degranulation (50,51). Interestingly fMLP, which ligates GPCRs, also initiates Src kinase activation, and pharmacological inhibition or genetic deletion (Hck-/-Fgr-/-Lyn-/-) of Src tyrosine kinase activity prevented fMLP-induced azurophil and specific granule release in a p38 MAPK-dependent manner (52). The mechanism of GPCR activation of Src family kinases remains unclear but may be via the Gα subunit or β arrestins (GPCR-associated adaptor proteins) (53). Further receptor cross-talk is demonstrated by GPCR-dependent PIP3 regulation of the Rho GTPase Rac via its activator P-Rex1 (a guanine exchange factor, GEF) (54). Translocation and exocytosis of granules requires a rise in intracellular calcium, which occurs in a biphasic fashion, initially released from intracellular stores and then entering from the extracellular space as the intracellular stores become depleted (55). Although the exact effectors of calcium-dependent degranulation are currently poorly characterised, examples of proteins which are influenced by calcium transients and involved in degranulation include annexins (56) and the vesicle docking facilitator, Munc13-4 (57). Calcium also plays a role in neutrophil polarisation via regulation of AKT, Src family kinases and Rho GTPases, and this modulation of the cytoskeleton may represent another method of control of degranulation (58). 22 1.2.1.7 Degranulation: vesicle trafficking and fusion machinery Granule trafficking to the target phagosomal or plasma membrane, in conjunction with actin and microtubule dynamics, is predominantly controlled by Rac and Rab GTPases, which facilitate granule migration by orchestrating cytoskeletal rearrangement. Vesicle-membrane tethering, docking and fusion is then regulated by Munc family proteins and SNAREs (soluble NSF-attachment protein receptors) (reviewed in (59)). Rac and Rab GTPases are “molecular switches”, alternating between an inactive GDP-bound and an active GTP-bound form. Guanine exchange factors (GEFs) are GTPase activators, facilitating exchange of GDP for GTP, whereas GTPase-activating proteins (GAPs) actually inactivate GTPases by promoting their GTPase activity. Rac2 is particularly important in human neutrophils, mediating actin dynamics (and hence degranulation) through the formation of free actin filament ends, which regulates actin assembly (60). Functional characterisation of Rabs and their effector molecules is currently limited. However, one well- described example is granule-associated Rab27a, which can bind the effector molecule JFC1. JFC1 reduces the activity of GMIP (a RhoA-GAP), which facilitates actin depolymerisation around the granule, allowing access to the vesicle docking zone (61). Once the granule is able to access the membrane exocytic region, docking is mediated by Munc proteins, e.g. Munc 13-4, which is also a Rab27a effector and exhibits calcium- dependent control of SNARE complex formation (57). The SNARE complex requires interaction of vesicle (granule) proteins (v-SNAREs), which comprise various vesicle- associated membrane proteins (VAMPs), with target plasma membrane proteins (t-SNAREs), including syntaxins and synaptosome-associated proteins (SNAPs) in order to effect granule- membrane fusion (reviewed in (41)). 1.2.1.8 Degranulation: differential exocytosis Granule subtypes are not mobilised equally; granule exocytosis is regulated by both the external stimulus and the associated secretory machinery (Figure 1.2). Sequential control of granule subtypes is regulated by the strength of the degranulation stimulus. Tightly linked with the order of granule development, the last granules to be formed are the most easily mobilised. Regulation of exocytosis in this manner is logical, as the hierarchy of release correlates with granule protein function. Initial mobilisation of secretory vesicles is able to supply the neutrophil surface with the receptors required for transmigration without liberating histotoxic proteins, which could harm the vascular endothelium. Subsequently, secretion of gelatinase and specific granules allows release of MMPs, such as gelatinase and collagenase, which can degrade the extracellular matrix (ECM), enabling 23 neutrophil transit through host tissues towards sites of infection. Only once neutrophils have reached these inflammatory areas, where pathogen numbers and inflammatory stimuli are at the highest concentrations, is the toxic cargo of the azurophil granules unleashed. Azurophil granules, therefore, require the strongest stimulus for degranulation, which is dependent upon the concentration of intracellular free calcium (49): Sengeløv et al. showed an in vivo hierarchy of granule subtype mobilisation, with complete mobilisation of secretory vesicles, and 38%, 22% and 7% release of gelatinase, specific and azurophil granules, respectively, from neutrophils in skin blister exudate, dependent on intracellular calcium concentration (37), which was preserved on further stimulation with fMLP. The precise secretory machinery associated with a particular granule is also linked to its classification and function. Neutrophils display a non-redundant role for Rac2 in the secretion of azurophil granules: neutrophils from mice deficient in Rac2 showed absent azurophil granule exocytosis whereas specific and gelatinase granule release was unaffected (62). Equally, murine neutrophils deficient in Bcr and Abr (Rac-GAPs) had increased cytochalasin B/fMLP-induced Rac2 activity which led to elevated MPO and NE release, whilst lactoferrin and MMP-9 release was unaffected (63). In fact, only a subset of azurophil granules have the required secretory machinery to be released into the extracellular space. Subcellular fractionation and immunoelectron microscopy has shown that Rab27a is located predominantly on gelatinase and specific granules, with lesser localisation to azurophil granules (64). Monfregola et al. demonstrated that Rab27a was necessary for secretory azurophilic granules to fuse with the plasma membrane but in Rab27a deficient mice, azurophil granules lacking this protein were still able to fuse with the phagosome. However, Munc13-4 (present on all three granule subtypes) was required for both fusion mechanisms (65). Further regulation of differential granule fusion and mobilisation is mediated by Src tyrosine kinases and v-SNAREs. The Src family member, Hck, was found to be localised to azurophil granules and translocated toward the phagosome (66) whereas Fgr was associated with specific granules, with an increase in the plasma membrane fraction after fMLP-treatment (67). VAMP-2 and SNAP-23, both enriched on specific and gelatinase granules, directed fusion with the plasma membrane whereas VAMP-7, predominantly sited on azurophil granules, guided fusion with the phagosome as well as plasma membrane (41,68). Thus, a number of differential secretory control mechanisms result in the majority of azurophil granules being preferentially targeted to the phagosome rather than the surface membrane, unlike specific and gelatinase granules. Inadvertent leakage of azurophil granules from incompletely closed phagosome vacuoles can occur during “frustrated phagocytosis”, where there is an overwhelming excess of pathogens, or an organism or surface which is too large to engulf, 24 such as immune complex deposition on the vascular endothelium. . Here, release of histotoxic proteins can cause host tissue injury, e.g. NE-mediated vascular injury in thrombo- haemorrhagic vasculitis in response to complement C3 deposition (69). Importantly, priming can also overcome the degranulation ‘braking’ mechanisms and primed cells can release significant quantities of azurophil contents to the extracellular space. Whilst endogenous and concomitantly released anti-proteases may restrain damage, these protective mechanisms may be overpowered in certain environments, such as infected airways (e.g. in cystic fibrosis (CF) and bronchiectasis (70,71)). Figure 1.2: Control of differential granule exocytosis Gradated granule exocytosis is dependent on intracellular calcium (Ca2+) concentration. Secretory vesicles (SV) are mobilised most easily. Azurophil (AG), but not specific (SG) or gelatinase (GG), granule secretion is dependent on Rac2. The Src family kinase, Hck, guides AG for phagolysosome fusion whereas Fgr guides SG for plasma membrane fusion. Rab27a directs all granules for plasma membrane fusion; AG lacking Rab27a can still fuse with the phagolysosome. Munc 13-4 is required for both plasma membrane and phagolysosome fusion. SG and GG docking/fusion SNARE complexes comprise VAMP-2, SNAP-23 and syntaxins 4/6 whereas AG SNAREs comprise VAMP-7 and syntaxin 4. VAMP-7 can also direct AGs for phagolysosome fusion. The majority of AG fuse with the phagolysosome. 25 1.2.1.9 Neutrophil Extracellular Traps Neutrophil extracellular traps (NETs) were first described by Brinkmann et al. in 2004 (72) and are expulsions of decondensed chromatin, studded with antimicrobial proteins and proteases (including NE and MPO) into the extracellular space. NET release is stimulated by a range of pro-inflammatory mediators, e.g. IL-8 and S. aureus. The classical molecular pathway initiating NET release relies on ROS generation by the NADPH oxidase complex, and pharmacological inhibition of NADPH oxidase abrogates NET generation; similarly, neutrophils from CGD patients are unable to release NETs (73). NETosis, the process of NET formation, also relies on histone citrullination and NE, which are thought to accomplish chromatin decondensation by histone digestion. Neutrophils from mice lacking NE or peptidylarginine deiminase 4 (PAD4), which catalyses histone hyper-citrullination, were unable to produce NETs (74,75). ROS-independent early (within 10 minutes of the stimulus) NETosis has also been described, but is still dependent on PAD4 and NE (76). It has been hypothesised that NET production is a host defence mechanism, capable of trapping and killing micro-organisms, but this remains contentious. NETs have been shown to attach to various pathogens in vivo (77), although it has been suggested that this adherence may be utilised by some organisms to form biofilms (78). There is also conflicting evidence for a direct microbicidal effect of NETs: Brinkmann et al. reported that in the presence of cytochalasin D, phagocytosis was inhibited with preserved NET production but neutrophils were still able to kill S. aureus, an effect which was abolished with the addition of DNAse (72). In contrast, Menegazzi et al. showed no microbicidal effect of NETs, with entrapped S. aureus released alive on addition of DNAse (79). Differences in the extent of bacterial killing in these studies depended on the timing of DNAse addition (to dismantle NETs). NETosis was initially thought to be a form of neutrophil cell death distinct from apoptosis, not stimulating clearance by macrophages but rather chromatin disassembly by nucleases (80). However, situations where live neutrophils produce NETs and remain functional have been described: Pilsczek et al. showed release of NETs from live neutrophils through nuclear membrane blebbing and vesicle budding (81), and Yipp et al. visualised intact NETosing neutrophils with retained ability to move chemotactically towards and phagocytose live bacteria (82). Furthermore, as NET formation is an active process whereupon the distinct nuclear and granular architecture is rearranged, such that a number of granule-derived enzymes, including NE and MPO, are released extracellularly in association with chromatin, NETosis has the potential to cause host tissue injury due to high local concentrations of histotoxic proteases. Massive NET formation has been reported in several pulmonary diseases, including asthma, COPD, CF, respiratory syncytial virus bronchiolitis, influenza and bacterial pneumonia 26 (reviewed in (83)). For example, sputum NETs in COPD were correlated with disease severity and microbial dysbiosis, as well as reduced neutrophil phagocytic capacity (84). CF sputum contains abundant NETs (85) which increase sputum viscosity (86) and might additionally cause protease-dependent airway damage. NETs have also been associated with vascular pathology; for example NETs promote endothelial cell activation and increased thrombogenicity through the concerted action of IL-1α and cathepsin G (87). Intravascular NETs have been shown in several conditions, including sepsis, vasculitis and atherosclerosis, and may promote thrombosis in addition to endothelial activation (reviewed in (88)). Hence, the role of NETs in driving a range of diseases is under investigation. 1.2.1.10 Neutrophil-derived microvesicles Microvesicles (MVs) are small membrane-bound microparticle structures released by multiple cell types, including neutrophils, which have gained rapid recent interest as a novel intercellular communication network. MVs are 0.1 - 1 μm in diameter and are released from cell plasma membranes by an exocytic enzyme-regulated budding/blebbing process. MVs are distinct from even smaller extracellular vesicle structures, such as exosomes, which are pre- formed particles of less than 100 nm in diameter, generated by exocytosis of multi-vesicular bodies from the endocytic-lysosomal compartment (89). MVs commonly feature externalised phosphatidylserine, due to calcium stimulation of scramblase and floppase activity, and contain components derived from the parent cell, including RNA, lipids and protein. As well as detecting MVs by their exposed phosphatidylserine, e.g. by annexin V binding, membrane proteins can be used as markers of MV origin; for example, neutrophil-derived MVs (NDMVs) contain CD66b (90). Initially thought to be inert cellular debris, the potential for MVs to act as vehicles of intercellular communication has now been recognised. MVs can transfer their contents to other cells, including different cell types, altering their phenotype and with the potential to instigate activation, e.g. by mobilisation of cytokines or ligands. For example, Slater et al. showed that NDMVs could transfer MPO to intestinal epithelial cells, which impaired mucosal wound healing (91) and Salanova et al. demonstrated that platelet-derived MVs could transfer functional glycoprotein IIb/IIIa receptors to neutrophils (92). Moreover, release of active enzyme in NDMVs may play a role in the inflammatory response. For example: cleavage of adhesion molecules by NDMV-associated MMP-9 in vitro facilitated intestinal transepithelial neutrophil migration through disrupted intercellular junctions (93). NDMVs are a heterogeneous population, varying in protein composition depending on the method of generation. Differences in morphology and protein abundance have been observed, dependent on whether NDMVs were produced spontaneously or after stimulation of opsonin receptors (94). Furthermore, MVs released from neutrophils in fluid suspension had a distinct 27 proteome compared with those from neutrophils adhered to a human umbilical vein endothelial cell (HUVEC) monolayer, sharing only 50% of their content (95). As well as varying in composition, NDMVs generated by these differing methods also exerted diverse effects on endothelial cell gene expression in vitro, demonstrating both that they are biologically active and that the functional effect of MVs in vivo may depend on the precise conditions under which they are formed. This may explain the conflicting literature addressing the pro and anti- inflammatory properties of NDMVs (including (96–99), discussed in section 1.3.1.3). 1.2.1.11 Apoptosis and clearance Neutrophils are terminally differentiated and pre-programmed to undergo constitutive apoptosis, resulting in their short survival times. However, apoptosis can be delayed at sites of infection and inflammation by signals from both pathogens, e.g. lipopolysaccharide (LPS), and/or the host, e.g. GM-CSF (100). Apoptosis can be initiated through the extrinsic pathway, i.e. ligation of cell surface death receptors such as FAS, or the intrinsic pathway, instigated by intracellular oxidative stress leading to release of cytochrome C from mitochondria and consequent caspase activation (101). Electron microscopy studies have revealed neutrophils to have comparatively few mitochondria (102) and bioenergetic profiling has demonstrated a reliance on glycolytic respiration for energy production rather than mitochondrial oxidative phosphorylation (103), with the predominant function of these organelles in neutrophils thought to be regulation of cell death. Challenging this dogma, however, recent discoveries have suggested greater metabolic and energetic flexibility than previously thought (104). Apoptotic neutrophils signal for efferocytosis by changes in cell surface receptors and membrane composition, such as externalisation of phosphatidylserine. Their phagocytosis inhibits pro-inflammatory cytokine production by macrophages and limits the potential for host tissue damage by safe removal of potentially histotoxic neutrophil contents, accelerating the path towards resolution of inflammation. 1.2.1.12 Role of neutrophils in disease Neutrophils are superbly equipped to carry out their role of host defence, armed as they are with multiple pathogen containing and killing mechanisms, and the ability to orchestrate the immune response to microbial invasion. Although once thought to function purely as killing machines, we now appreciate that neutrophils can shape the immune landscape through communication with multiple cell types. For example, neutrophil-derived granule products, such as defensins and cathelicidins, allow communication with the adaptive immune system by inducing CD4+ and CD8+ T cell chemotaxis, in addition to their antimicrobial agenda (105). Under certain culture conditions, neutrophils can acquire functional properties of antigen presenting cells, expressing MHCII and co-stimulatory molecules (106). Although these 28 attributes confer a complex and plastic role for neutrophils, the crucial neutrophil function remains host defence against invading pathogens, well-illustrated by the increased susceptibility of patients with neutrophil defects such as Chediak Higashi syndrome (impaired degranulation) or CGD (impaired respiratory burst) to severe and recurrent infections (107). Neutrophils are designed to rapidly eliminate pathogens. However, if they do not encounter pathogens within a short timeframe, for example in the context of sterile inflammation where neutrophils adherent to inflamed endothelial surfaces may encounter circulating cytokines such as TNFα, large amounts of oxidants can be released (108). Coupled with granule exocytosis of histotoxic proteins and proteases, neutrophils have the capacity to cause extensive bystander tissue injury. Ideally, the release of toxic oxidants and proteases should be delayed until the cells reach a focus of inflammation/infection but primed neutrophils in the systemic circulation have been identified in disease states, including ARDS (109) and bacterial sepsis (110), potentially contributing to disease pathogenesis and endothelial dysfunction. It has been demonstrated in an in vivo human model that primed circulating neutrophils are preferentially retained in the pulmonary vasculature, which facilitates “de-priming” of these cells and their release back into the systemic circulation (7) but failure of this de-priming process could contribute not only to lung tissue damage but also to endothelial injury and remote organ dysfunction. Raised circulating levels of potentially histotoxic neutrophil-derived proteins have been found in inflammatory situations, including sepsis (111) and COPD (112– 114), where most (e.g. resistin and NGAL) but not all (e.g. cyclophilin A) are neutrophil granule components. Neutrophil proteins, such as MPO and NE, can also be externalised by the generation of NETs. As well as causing direct cytotoxicity, with the ability to induce epithelial and endothelial cell death (115), NETs have been implicated in remote organ injury (116) and even associated with the promotion of cancer metastases (117). Neutrophils present within tumours can develop a pro-tumorigenic phenotype, driven by transforming growth factor β (TGFβ) within the tumour microenvironment (118), and release of neutrophil matrix-degrading proteins, e.g. serine protease-activated MMPs, can promote tumour cell invasion (119). Hence, neutrophils are a double-edged sword, playing an essential role in host defence but also a pathogenic role in both acute and chronic inflammatory diseases. Neutrophil function must therefore be tightly controlled to prevent host tissue damage from inappropriate or excessive activation. 29 1.2.2 The impact of hypoxia on neutrophils 1.2.2.1 Relevance of hypoxia to neutrophils Neutrophils are generated within the bone marrow, which is a significantly hypoxic microenvironment; using two-photon phosphorescence lifetime microscopy of live mice, in vivo bone marrow oxygen tensions have been recorded as low as 1.3 kPa (120). Furthermore, staining with pimonidazole, a 2-nitroimidazole compound which indicates oxygen levels below 1.3 kPa, demonstrated sequestering of haematopoietic stem cells in the most hypoxic areas of the bone marrow architecture (121), and low oxygen tension favoured the maintenance of haematopoetic stem cells in vitro (122). Therefore, hypoxia appears to play a critical role in neutrophil development. Under physiological conditions, circulating neutrophils released from the bone marrow encounter a wide range of oxygen tensions, transiting rapidly from a pO2 of 13 kPa in main arteries, to 7 kPa in arterioles and 3-4 kPa in capillaries and venules (123). Due to the short oxygen diffusion capacity, normal tissue oxygenation can be even lower (physiological hypoxia), demonstrated in striated muscle (124), colonic epithelium (125) and the skin, despite being bathed in ambient oxygen (126). In pathological situations, systemic hypoxaemia, which can arise from various lung pathologies, such as fibrosis and emphysema, results in a further reduction of oxygen delivery to tissues. Moreover, significant amplification of this tissue hypoxia can occur in inflammation, infection or ischaemia (pathological hypoxia) due to damaged vasculature, compartmentalisation of infection, and high metabolic activity of pathogens and host cells. In addition to these well-described mechanisms, Lee et al. have described hypoxaemia secondary to LTB4-driven neutrophil swarming in a mouse fungal sepsis model, which resulted in capillaritis with neutrophils continuously lining alveolar walls, compounded by alveolar haemorrhage (18), and Campbell et al. have demonstrated depletion of local oxygen by transmigrating neutrophils in a murine model of colitis, which contributed to epithelial HIF stabilisation (127). Profound tissue hypoxia has been demonstrated in several pathological conditions by various in vitro and in vivo methods: by microelectrode measurement of ulcers (mean pO2 = 2.39 kPa) (128); by blood gas analysis of abscesses (median pO2 = 3.74 kPa) (129); by hypoxia inducible factor (HIF) staining (which is detected maximally at 0.5% oxygen (130)) in COPD (131), bronchitis (132) and inflammatory bowel disease (IBD) (133); by pimonidazole staining (indicating pO2 <1.3 kPa) of lung tissue in respiratory infection (134); by luminescence-based in vivo optical imaging in skin infection (mean pO2 = 2.66 kPa) (135) and by [18F]- fluoromisonidazole positron emission tomography (PET) imaging (indicating pO2 <1.3 kPa) in humans with pulmonary tuberculosis (TB) (136). Furthermore, hypoxia itself can cause 30 inflammation, as demonstrated by increased vascular leak in a mouse model of acute hypoxia (137), increased serum levels of IL-6 and CRP in healthy climbers at high altitude (138), and impaired resolution of pulmonary infection in hypoxic mice (139). Conversely, resolution of inflammation has been shown experimentally to correlate with an increase in oxygen tension, and supra-physiological oxygen levels can promote resolution of certain infections (85). 1.2.2.2 Sensing of hypoxia by neutrophils Neutrophils may be exposed to profound levels of hypoxia and, as the frontline mediators of host defence, need to be able to function successfully in such hypoxic environments. Neutrophil sensing and adaptation to surrounding oxygen tensions is dependent on the prolyl and asparaginyl hydroxylase enzymes, which control expression of the hypoxia-inducible transcription factor HIF. HIF is a heterodimeric protein, comprising a constitutively expressed β subunit and a hydroxylase-regulated α subunit; the predominant α subtype present in neutrophils is HIF-1 (142). Under normoxia, prolyl hydroxylase-containing enzymes (PHDs) direct HIF-1α for proteasomal degradation via the von Hippel Lindau (vHL) protein complex, whilst factor inhibiting HIF (FIH) inactivates HIF transcription by preventing binding of the transcription co-factor p300. PHDs and FIH display an absolute requirement for dioxygen, Fe(II) and 2-oxoglutarate. Under hypoxia, the lack of dioxygen reduces hydroxylase activity, allowing stabilisation of HIF-1α, which then translocates to the nucleus and binds HIFβ. The HIF heterodimer binds hypoxia response elements (HREs), thereby regulating the neutrophil’s response to local oxygen tensions by promoting the transcription of hypoxia-responsive genes. Bioenergetic profiling has demonstrated that neutrophils rely predominantly on non- oxidative glycolytic respiration for ATP production, and neutrophils deficient in HIF-1α have reduced ATP levels (143). It has also been shown that pro-inflammatory signals can stabilise HIF; for example: Peyssonnaux et al. demonstrated that HIF-1α expression was induced by LPS in a TLR4-dependent manner in macrophages (144). The induction of HIF-1α mRNA by TLR-induced nuclear factor kappa-light-chain-enhancer of activated B cells (NFκB) signalling is likely important in these pro-inflammatory environments, with a further increase in NFκB regulated by HIF-1α under hypoxic conditions (145). However, HIF-driven hypoxia-adaptation processes depend on transcription of effectors, and neutrophils may need more rapid adaptability. In other cell lines, hypoxia regulated the phosphorylation status of the translational control protein mTOR (mammalian target of rapamycin) and its effectors (146), and regulated T-plastin-mediated membrane trafficking (147), both in a HIF-independent manner. In neutrophils, hypoxia has been shown to reduce ROS production (148) and increase degranulation (149) in a HIF-independent fashion (see sections 1.2.2.4 and 1.2.2.5). 31 1.2.2.3 Effect of hypoxia on neutrophil recruitment Neutrophil firm adhesion to endothelium prior to transmigration is mediated predominantly by the interaction of neutrophil β2 integrins (e.g. MAC-1) with endothelial ligands, such as ICAM- 1. The literature supports a hypoxia-mediated increase in β2 integrin expression (148,150– 153) with only one study, which allowed re-oxygenation following neutrophil isolation, finding no difference (154). Reports of ICAM-1 expression in response to hypoxia are conflicting: Antonova et al. showed no difference in ICAM-1 levels of endothelial cells cultured under hypoxic conditions (155), whereas Yoon et al. demonstrated increased ICAM-1 in a murine hindlimb ischaemia model. Such discrepancies may reflect species differences or the in vitro vs in vivo experimental approaches. Importantly, as a more integrated outcome, several studies have shown increased neutrophil adhesion to the endothelium under hypoxia (152,156,157). Along with the apparent increase in neutrophil adhesion, studies of neutrophil transmigration in the hypoxic environment have shown it to be enhanced. Hypoxia increased neutrophil transmigration in vitro in a model of intestinal I/R, which was β2 integrin and CD47- dependent (158), and in vivo in rodent models of acute systemic hypoxia, assessing mesenteric endothelial neutrophil transmigration by intravital microscopy (159), or by MPO quantification in multiple organs (160). Investigation of the effect of hypoxia on chemotaxis has also produced contradictory results. Extravasated neutrophils undergo shape change to a polarised morphology, which is essential for directional movement towards sites of infection and inflammation. McGovern et al. showed no change in IL-8-induced human neutrophil shape change, or chemotaxis towards IL-8, fMLP or LPS under hypoxia (148), and, similarly, Peyssonnaux et al. found no difference in chemotaxis towards fMLP through an endothelial cell monolayer between wildtype, HIF-1α- deficient or vHL-deficient murine neutrophils (161). Rotstein et al. showed reduced chemotaxis of human neutrophils towards fMLP and zymosan-activated serum under hypoxia in a gel migration assay (162) whereas, in contrast, Ma et al. demonstrated increased polarisation of differentiated human neutrophil-like HL-60 cells after exposure to hypoxia (163), and Wang and Liu showed enhanced chemotaxis of neutrophils isolated from hypoxic volunteers towards fMLP (153). Discrepancies here are likely due to variation in cell type studied, true hypoxia vs manipulation of HIF-1α, and re-oxygenation of cells isolated from hypoxic donors prior to in vitro assays. Studies of neutrophil recruitment, reflecting a composite outcome of adhesion, transmigration and chemotaxis, are more likely to replicate the true physiological environment, with its myriad signals. However, reports of neutrophil recruitment under hypoxia are again inconsistent and context-dependent. In a murine model of chemical irritant-induced cutaneous inflammation, 32 myeloid HIF-1α deletion abrogated neutrophil tissue infiltration and vHL deletion promoted infiltration (143). Conversely, pharmacological HIF-1α stabilisation either diminished neutrophil recruitment (in murine uropathogenic Escherichia coli bladder infection) (164) or had no impact (in murine S. aureus skin infection) (165), perhaps because the skin is already hypoxic as noted above. Similarly, in a mouse model of group A Streptococcus cutaneous infection, wildtype, HIF-1α-null or vHL-null neutrophils showed no difference in recruitment up to 24 h (161). Likewise, in a mouse model of Pseudomonas aeruginosa keratitis, HIF-1α siRNA knockdown showed no difference in neutrophil recruitment at 24 h, although recruitment was increased at day 5 (166). Conflict between these studies may be explained by variations in time, inducing agent (i.e. infectious vs non-infectious) and tissue (hypoxic or non-hypoxic) site, and also suggests that HIF manipulation does not equate precisely to true hypoxia. At present, there is no clear understanding of how hypoxia influences neutrophil recruitment and variations between studies indicate that the effect is likely to be context-dependent. 1.2.2.4 Effect of hypoxia on phagocytosis and ROS generation The majority of studies suggest that phagocytosis by neutrophils is increased under hypoxia. Whole blood neutrophils collected from volunteers exposed to acute hypoxia showed enhanced phagocytosis of zymosan (167) and isolated neutrophils from acutely hypoxaemic volunteers demonstrated increased phagocytosis of E. coli, together with enhanced basal or E. coli-stimulated expression of opsonic and complement receptors (153), which are important for pathogen recognition. Similarly, phagocytosis of zymosan by neutrophils isolated from rabbits after experimental acute ischaemia (168) and phagocytosis of E. coli by neutrophils isolated from hypoxic pre-conditioned rats (169) were both increased. Following hypoxic exposure, neutrophil isolation and subsequent in vitro assays were performed in atmospheric oxygen; however, the Simms group demonstrated increased expression of most opsonic receptors and phagocytosis of S. aureus and E. coli, with in vitro assays performed under hypoxia in an air-tight chamber (170,171). The same group initially saw reduced expression of certain Fcγ neutrophil opsonic receptors under hypoxia but the hypoxic increase could be restored by the RDGS-binding epitope of fibronectin (172). Furthermore, integrin signalling via cross-linked VLA-5 and VLA-6 receptors on both fluid-phase and adherent neutrophils increased FcγIIIbR and MAC-1 expression under hypoxia (173). Indeed, modulation by matrix proteins in low oxygen tensions may more accurately reflect the in vivo environment. Further evidence that hypoxia can increase phagocytosis is that neutrophils isolated from patients with loss-of-function vHL mutations (which mimic hypoxia by impeding degradation of HIF) displayed enhanced phagocytosis of Streptococcus pneumoniae under normoxia which could be further enhanced under hypoxia (174). In contrast, our group also found that, although isolated neutrophils from healthy volunteers incubated under hypoxia had enhanced CD11b 33 expression, phagocytosis of S. pneumoniae was preserved but not increased (148). Two further studies have shown no effect of hypoxia on phagocytosis: Almzaiel et al. demonstrated no difference in phagocytosis of S. aureus under hypoxia with an increase in phagocytosis under hyperoxia, although these assays were performed with HL60 cells (175); and Berger et al. showed no change in phagocytosis of P. aeruginosa by neutrophils treated with HIF-1α inhibitor 17-DMAG, although this compound also caused a decrease in apoptosis and increase in necrosis which is inconsistent with the established literature on HIF-1α modulation of neutrophil apoptosis, suggesting a possible toxic effect (166). Phagocytosis and bacterial killing are tightly linked to ROS production, although the microbicidal activity of ROS is highly context-dependent and varies with bacterial species. Data regarding the impact of hypoxia on ROS generation are once again conflicting. When cell isolation and in vitro experiments were performed under normoxia, neutrophils isolated from hypoxic human donors or mice showed increased ROS production (153,154,169). This increase might reflect the re-oxygenation process, and potential clinical correlates include obstructive sleep apnoea (OSA) (176) and I/R injury (177). However, when in vitro assays were conducted under hypoxic conditions, neutrophils isolated from normoxic donors demonstrated decreased intracellular and extracellular superoxide production, which could be restored by re-oxygenation (148). In contrast, genetic and pharmacological HIF-1α manipulation did not affect ROS release (161,165). As modulation of HIF-1α signalling did not recapitulate data generated under true hypoxia, it seems logical that ROS production depends on oxygen availability, and that the decrease seen under hypoxic conditions is likely due to a fundamental lack of molecular oxygen. Using assessment of S. aureus bacterial killing as a surrogate for ROS production, as S. aureus killing is predominantly ROS-dependent (178), hypoxia impaired bacterial killing both in vitro (148,179) and in vivo (skin infection and pneumonia models (140,180)). Data generated from our laboratory support the hypothesis that impaired ROS production and consequent S. aureus killing under hypoxia is due to lack of molecular oxygen, as hypoxia did not change the expression of NADPH oxidase subunits, and addition of pyocyanin (which oxidises intracellular NADPH) did not increase ROS generation under hypoxia (148). Overall, the literature to date suggests that hypoxia increases phagocytosis by neutrophils but reduces ROS production, with a subsequent defect in ROS-dependent bacterial killing. 1.2.2.5 Effect of hypoxia on degranulation Although few studies have examined the effect of hypoxia on neutrophil degranulation, the consensus is that hypoxia augments granule exocytosis. Data generated from our laboratory demonstrated that release of the azurophilic granule proteins, NE and MPO, was increased 34 under hypoxia, measuring supernatant NE protein content (148), and NE and MPO activity (149), as well as release of the specific granule protein lactoferrin, and the gelatinase granule protein MMP-9. This effect occurred rapidly (within 2 h), was not prevented by cycloheximide, and could not be recapitulated by pharmacological HIF-1α stabilisation, together suggesting a HIF-independent mechanism. Similarly, neutrophils isolated from healthy volunteers subjected to acute hypoxia showed increased NE release ex vivo (154), and isolated neutrophils incubated with Mycobacterium tuberculosis-infected monocytes released more NE and the specific granule protein, MMP-8, under hypoxia. In addition to evidence of an acute HIF-independent increase in degranulation, several studies have suggested HIF-dependent changes in neutrophil granule protein mRNA or protein content. Peyssonnaux et al. showed an increase in murine CRAMP (analogous to the human specific granule protein cathelicidin/LL-37) mRNA and protein content under hypoxia, which was similarly increased in vHL-deficient cells and absent in HIF-1α-deficient cells. Additionally, both NE and cathepsin G activity was increased vHL-null in murine neutrophils, with the opposite true of HIF-1α-null cells (161). Likewise, HIF-1a siRNA knockdown reduced protein levels of beta defensins and CRAMP in a mouse model of pseudomonal infection (166), and pharmacological stabilisation of HIF-1α revealed upregulation of genes encoding LL-37 in human neutrophils (181). However, a HIF-independent control of granule exocytosis would allow quicker adaptation in the rapidly changing inflammatory environments encountered by neutrophil “responders”. HIF- dependent mechanisms may be relevant to neutrophils encountering prolonged hypoxia; circulating neutrophil granule content may be modulated in chronically hypoxaemic patients, thus contributing further to granule protein secretion. As noted above, the majority of granule biosynthesis occurs in the bone marrow, which is constitutively profoundly hypoxic, hence the impact of systemic hypoxia on this process is currently uncertain. It has been proposed that increased release of granule contents under hypoxia is detrimental, for example driving cavity formation and matrix degradation in TB (136,182). However, Christoffersson et al. showed that hypoxia-induced vascular endothelial growth factor (VEGF)- A signalling in a mouse model of avascular transplanted pancreatic islets recruited a subset of MMP-9hi neutrophils, which were instrumental in allowing revascularisation and implantation of transplanted tissue, with impaired angiogenesis in MMP-9-deficient mice (183). This demonstrates a potential beneficial role for increased neutrophil degranulation in vivo, enhancing phagocyte access to sites of inflammation. 1.2.2.6 Effect of hypoxia on NET production There has been limited investigation into the impact of hypoxia on NET production with variable results reported, depending on whether assays were performed under true hypoxia 35 or with HIF manipulation. McInturff et al. demonstrated reduced NETosis and extracellular bacterial killing in response to pharmacological and genetic HIF-1α knockdown (184). Similarly, Zinkernagel et al. showed increased bacterial killing with pharmacological stabilisation of HIF-1α, which was maintained in the presence of a phagocytosis inhibitor but abrogated by the addition of DNAse, although there was no observed effect on NET production (165). However, although Vӧllger et al. showed increased NETosis with pharmacological HIF- 1α stabilisation in a ROS-dependent manner (185), the same group showed a decrease in NET production under true hypoxia (186), consistent with data from our laboratory (149). As NETosis is predominantly dependent on NADPH generation of ROS (73), and hence reliant on oxygen availability, it is logical that NETosis under true hypoxia would be reduced, although it appears that HIF-1α signalling also has a role to play; the complex in vivo interplay of these linked but opposing influences remains to be elucidated. 1.2.2.7 Effect of hypoxia on neutrophil-derived microvesicle generation To date, few studies have looked at NDMV release in the context of hypoxia, with no studies investigating the direct effect of hypoxic neutrophil incubation on MV generation. Chen et al. showed that volunteers exercising under hypoxic conditions had increased circulating NDMVs and proposed that this contributes to increased vascular thrombotic risk in acute hypoxia by MV-mediated thrombin generation (187). Tremblay et al. demonstrated an increase in circulating NDMVs at high altitude (3800 m, approximately 13% oxygen), which correlated with a reduction in brachial flow-mediated dilatation (FMD) (188), although Ayers et al. did not find any difference in NDMV numbers at moderate altitude (2590 m, approximately 15% oxygen) compared with sea level (21% oxygen) (189). In a disease setting, Priou et al. found increased levels of NDMVs in patients with OSA, which correlated with the degree of cyclical nocturnal hypoxia, though re-oxygenation may be as relevant here as the level of hypoxia; furthermore, treatment of endothelial cells with MVs isolated from OSA patients with an increased oxygen- desaturation index (more than ten desaturation episodes per hour) caused increased adhesion molecule expression, which the authors propose may be a mechanism responsible for promoting endothelial dysfunction (190). 1.2.2.8 Effect of hypoxia on neutrophil apoptosis Along with pro-inflammatory cytokines, hypoxia is able to modulate the neutrophil apoptotic threshold, delaying constitutive neutrophil apoptosis in a concentration-dependent and reversible manner through HIF-1α mediation of NFκB signalling. This hypoxic survival effect can be prevented by the NFκB inhibitors gliotoxin and parthenolide, and is abrogated in HIF- 1α-deficient murine neutrophils, which also demonstrate decreased NFκB activity (145). Further evidence for control of survival by hypoxia and/or HIF includes: delayed apoptosis 36 exhibited by neutrophils from patients with loss-of-function mutations in vHL, which usually targets HIF-1α for degradation (174); delayed resolution of inflammation in a zebrafish tail transection model due to a reduction of neutrophil apoptosis and retention of neutrophils at the site of injury in response to genetic or pharmacological stabilisation of HIF-1α (191); and prolonged ex vivo neutrophil survival from healthy subjects exposed to acute hypoxia (68% O2) (154). The hypoxic survival effect was also apparent in neutrophils isolated from patients with OSA, which is characterised by intermittent hypoxia/reoxygenation (192). Hypoxia increases β2 integrin protein expression under the control of HIF-1 at the transcriptional level (151) which potentially also contributes to the hypoxic survival effect as β2 integrins have been implicated in the regulation of neutrophil apoptosis in addition to their role in adhesion. Whether the effect is pro or anti-apoptotic depends critically on the inflammatory environment; β2 integrin clustering or activation with endothelial ligands, such as ICAM-1, delays apoptosis through AKT and MAPK-ERK signalling, whilst β2 integrin activation in the presence of death-inducing agonists, such as TNF-α, can potentiate apoptosis, with a reduction in pro-survival AKT signalling (193). In addition to HIF-1α, two further mediators of hypoxic neutrophil survival have been identified: MIP-1β is a chemokine secreted by hypoxic granulocytes which confers survival when co- incubated with normoxic neutrophils (145), and PHD3, whose expression is upregulated in hypoxia, prolongs neutrophil survival independent of HIF-1α by control of pro-apoptotic SIVA- 1. This translated to a PHD3-deficiency phenotype of accelerated neutrophil apoptosis and enhanced resolution of inflammation in murine models of acute lung injury (ALI) and colitis (194). Conversely, increased lethality was demonstrated in a PHD3 knockout murine model of abdominal sepsis, although this was thought to be predominantly due to enhanced pro- inflammatory activity of PHD3-deficient macrophages (195). 1.2.2.9 Overall effect of hypoxia on neutrophil functions Overall, hypoxia appears to enhance neutrophil phagocytosis but impair killing of certain bacteria (e.g. S. aureus) due to a reduction in ROS production, increased degranulation and delayed apoptosis. The effect of hypoxia on NET generation is less clear but it is likely reduced. Thus, the augmented extracellular release of histotoxic granule products, the potential for increased pathogen survival (and escape), and the persistence of neutrophilic inflammation in hypoxic conditions markedly increases the capacity of this destructive neutrophil phenotype to cause substantial damage to host tissue (Figure 1.3). 37 Figure 1.3: Neutrophil functions under conditions of normoxia versus hypoxia Under normoxia, bacteria such as S. aureus are ingested into the phagosome. Pathogen killing is effected by the release of cytotoxic granule proteins and proteases, accompanied by ROS (O2-) generated from molecular oxygen via NADPH oxidase, into the phagosome. Neutrophils may either undergo NETosis (further trapping pathogens) or apoptosis and clearance. Under hypoxia, phagocytosis is maintained or enhanced but ROS-dependent killing is reduced due to the lack of oxygen availability, with the potential for pathogen escape. Increased extracellular release of toxic granule contents and prolonged neutrophil survival increases the potential for host tissue damage. Figure reproduced with permission from Lodge et. al, Hypoxic regulation of neutrophil function and consequences for Staphylococcus aureus infection, Microbes and Infection, 2017 (196). 38 1.2.3 The role of neutrophils in COPD COPD is a progressive inflammatory lung disease due to inhalational exposures, predominantly cigarette smoking, diagnosed by fixed airflow obstruction and characterised by intermittent exacerbations. Neutrophilic inflammation is a principal feature of COPD and correlates with disease severity, persisting despite cessation of smoking (197), even in the absence of an infectious driver (198). Patients with COPD suffer recurrent infection-driven exacerbations, which are a leading cause of mortality and functional decline, suggesting that there is impaired ability of these neutrophils to combat infection (199). Furthermore, in circumstances of prolonged or heightened inflammation when soluble mediators are present, or during frustrated phagocytosis, stimuli can promote neutrophil degranulation, releasing extracellular histotoxic antimicrobial proteins and proteases which can exacerbate host tissue damage. Complicating this picture, there is considerable heterogeneity between patients with COPD, and the diagnosis encompasses both chronic bronchitis (persistent cough with daily mucus production) and emphysematous (alveolar destruction and air space enlargement) phenotypes. There is abundant evidence that release of histotoxic neutrophil proteins contributes to tissue injury, and hence disease pathogenesis in COPD. NE has long been incriminated: NE is capable of causing lung epithelial damage, degrading multiple components of the ECM and increasing mucus secretion (200); NE induces the pathological features of COPD in animal models (201) and NE-deficient mice are protected from cigarette smoke-induced emphysema (202); NE is elevated in sputum from COPD patients, compared with smoking controls without airway obstruction, and correlates with decline in lung function (203); and plasma concentrations of the NE-specific fibrinogen cleavage product Aα-Val360, providing a “footprint” of NE activity in vivo, correlate with COPD severity (204). However, translation of NE inhibitors as therapy for patients with acute or chronic neutrophilic lung disease has proved inconclusive to date, perhaps reflecting the more complex array of neutrophil-secreted proteins with the capacity for damage. The proteinase PR3 has been more recently implicated in this setting. Although it displays a lower rate of elastolysis than NE, PR3 is more abundant in neutrophils and has increased activity in COPD sputum, with a further increase during exacerbations (205). Additionally, COPD sputum contains elevated levels of the neutrophil granule protein MPO (206) and metalloproteinases MMP-8 and MMP-9 (207). MPO, which promotes oxidative damage, correlates with lung function decline, and MPO inhibition prevents acute lung inflammation and emphysema progression in animal models of both short and long term cigarette smoke exposure (208,209). MMPs regulate ECM turnover, enabling neutrophil migration, but also have the capacity to cause substantial ECM 39 degradation and damage (210). MMP-8 and -9 sputum and plasma levels correlate with lung function decline and progression of emphysema in COPD patients (211) and broad spectrum MMP inhibitors have been shown to attenuate emphysema progression in some animal models (212). One theory of COPD progression, supported by the COPD-like pathology that occurs in α1AT deficiency, is that it reflects an imbalance of protease and anti-protease activity. Even in the presence of active anti-proteases, proteases such as NE are able to diffuse at least 10 μm into the tissue before sequestration, and NE in complex with α-2-macroglobulin retains its proteolytic activity (213). Moreover, high concentrations of neutrophil proteases at the point of release from the granule can exceed the concentration of surrounding anti-proteases in the local microenvironment, “quantum proteolysis” (214), and NE bound to the neutrophil surface remains catalytically active yet resistant to anti-protease inhibition (215). Hence, there is abundant evidence suggesting that neutrophil degranulation releases a range of histotoxic agents which contribute significantly to lung tissue injury in COPD. Abnormalities have been reported in neutrophils from COPD patients. Sapey and colleagues demonstrated that such neutrophils exhibit aberrant chemotaxis in vitro, with increased migratory speed but impaired accuracy towards a range of chemo-attractants; these abnormalities were consistent across all severities of COPD but were not found in healthy smokers or in patients with α1AT deficiency (216). They postulated that this abnormal migration may contribute to surrounding host tissue damage in COPD, but also report that this response may be amenable to correction by manipulating intracellular signalling pathways; improved chemotaxis was shown after treatment of COPD patient neutrophils in vitro with a pan-PI3K inhibitor (216) or with the 3-hydroxy-3-methylglutaryl coenzyme-A reductase inhibitor simvastatin (217), which has previously been shown to increase the expression of PTEN, a phosphatase which antagonises the PI3K-AKT signalling pathway (218). Although at the early stages of exploration, proteomic analysis of neutrophils from COPD patients compared with either sulphur mustard-exposed patients (exposure during warfare with long- term pro-inflammatory consequences) or healthy controls has revealed distinct COPD neutrophil profiles (219–221). However, it is important to note that due to the aforementioned heterogeneity within the COPD diagnosis, it is unlikely that there is a single entity comprising the “COPD neutrophil”. It is probable that COPD patient neutrophils display variable abnormalities and, at the other end of the spectrum, some patients may have normal neutrophil function, although these cells will still be functioning within a persistently inflamed and hypoxic environment. 40 Patients with severe COPD are frequently systemically hypoxic. However, more profound tissue hypoxia exists in areas of inflammation, infection and microcirculatory impairment, independent of systemic oxygen tensions, demonstrated in airway tissue (136,222) and atherosclerotic vasculature (223,224). Thus, even in mild COPD, neutrophils in disease- associated tissues and microcirculatory beds are exposed to severe hypoxia, which will be compounded by systemic hypoxia. As detailed in section 1.2.2, hypoxia delays neutrophil apoptosis, prolongs their presence at sites of injury, enhances the release of antimicrobial (and potentially histotoxic) products and, despite promotion of phagocytosis, the reduction in ROS production under true hypoxia can impair bacterial killing. As such, hypoxia promotes a destructive neutrophil phenotype with compromised ability to combat pathogens but enhanced capacity for host tissue damage and delayed resolution of inflammation. In addition to progressive lung damage, COPD is associated with endothelial dysfunction and accelerated cardiovascular disease (CVD), even after adjusting for shared risk factors, including smoking (225,226). Patients with COPD are at substantially increased risk of future acute arterio-vascular events, such as myocardial infarction and stroke (227), and CVD is a leading cause of morbidity and mortality in this group (228,229). Analysis of large datasets (including nearly 300,000 patients) and meta-analysis of 29 separate studies has shown at least a two-fold increased risk of having CVD in COPD patients compared with non-COPD patients or the general population (230), with an even higher risk (e.g. up to seven-fold for peripheral arterial disease) in younger patient cohorts (35 – 54 years) (231) and following a severe COPD exacerbation (232). Evidence is accumulating that inflammation, vascular tissue damage (for example, by neutrophil-derived proteases) and oxidative stress are important biological mechanisms linking COPD and CVD (233), with increased arterial wall stiffness and markers of systemic inflammation, such as CRP, identified in COPD patients compared with control smokers without airflow obstruction (234). 1.3 Endothelial cells The endothelial blood vessel lining is responsible for the regulation of vascular tone, smooth muscle cell migration, cellular adhesion and resistance to thrombosis. Endothelial dysfunction, frequently occurring as a maladaptive response to pathological stimuli, is a failure of the endothelium to perform these physiological roles. It leads to increased vascular smooth muscle tone, upregulated expression of CAMs (including the leukocyte-endothelial adhesion molecule ICAM-1), increased thrombosis (in part due to endothelial cell death) and compromised barrier function (reviewed in (235)). Diminished barrier function, and hence increased vascular permeability, occurs in a range of inflammatory diseases and in sepsis, with associated organ dysfunction. Malfunction of these endothelial processes also initiates 41 and promotes atherosclerosis and atherothrombosis, resulting in cardiovascular complications in advanced disease. However, endothelial dysfunction manifests much earlier and can be detected in the absence of symptoms, e.g. by ultrasound assessment of arterial diameter in response to endothelium-mediated shear stress-induced vasodilatation (FMD) or measurement of circulating biomarkers, such as soluble CAMs (236). In addition to traditional risk factors, such as hypertension, dyslipidaemia and smoking, inflammation has been linked to the progression of CVD; multiple chronic inflammatory diseases have been associated with increased cardiovascular morbidity and mortality, including rheumatoid arthritis (RA), IBD and COPD (225,235), although the mechanism(s) of immune-mediated endothelial dysfunction remain to be fully elucidated. 1.3.1 The role of neutrophils in endothelial dysfunction and cardiovascular disease Neutrophils have emerged in recent years as one of the key players in atherogenesis. They aggravate endothelial dysfunction, accumulate in atherosclerotic lesions, and neutrophil- derived proteases promote plaque instability and trans-endothelial migration. Myriad studies support an association of increased neutrophil/lymphocyte ratio with increased arterial stiffness (237), and an inverse correlation of the apnoea-hypopnoea index (indicating intermittent hypoxia) with FMD has been demonstrated in OSA (238). Primed circulating neutrophils have been identified in hyperlipidaemic patients (239), and hypercholesterolaemia-induced neutrophilia correlated with the extent of atherosclerotic lesion formation, which was reduced in neutropaenic mice (240). 1.3.1.1 Direct neutrophil-endothelial interactions Dysregulated neutrophil-endothelial interaction has been shown to impair endothelial barrier function. Co-culture of neutrophils with pulmonary vascular endothelial cell monolayers increased endothelial release of angiopoietin-2, a mediator which promotes blood vessel destabilisation and is elevated in ALI; depletion of peripheral neutrophils in a murine haemorrhage shock model reduced plasma and lung angiopoietin-2, and reduced vascular leak and indices of lung tissue injury (241). Further, Gill et al. demonstrated neutrophil- mediated microvascular endothelial apoptotic cell death, with resultant increased vascular leak, in a mouse sepsis model which could be ameliorated by peripheral neutrophil depletion or antibody blockade of CD18-binding (242). Interestingly, neutrophils isolated from septic patients, particularly those with ARDS, reduced endothelial barrier integrity to a greater extent than healthy control neutrophils when applied to venous monolayers (243). Neutrophil activation in the microcirculation can cause capillary malperfusion due to external compression of the microvasculature from accumulated fluid in the interstitial compartment. 42 This phenomenon of “capillary no-reflow” has been demonstrated in several I/R models, and could be attenuated by induction of neutropaenia or blockade of CD18-, ICAM-1- and P- selectin-dependent neutrophil-endothelial adhesion (244). Neutrophils and endothelial cells in this setting will be exposed to a profoundly hypoxic milieu, with the potential to affect the responses of both cell types and their interactions. P-selectin is contained within endothelial Weibel-Palade bodies, release of which could be primed by hypoxia with subsequent enhanced neutrophil recruitment; in a murine cardiac transplantation model, P-selectin-null hearts had reduced graft neutrophil infiltration and increased graft survival when compared with wildtype (245). Moreover, small interfering RNA (siRNA) silencing of multiple CAMs (ICAM-1, ICAM-2, VCAM-1, E-selectin, P-selectin) decreased neutrophil recruitment into ischaemic myocardium and reduced post-MI neutrophil recruitment into atherosclerotic lesions in mice (246). Intriguingly, transfer of active MPO (which can perpetuate inflammation and tissue injury through pro-inflammatory cytokine and ROS release) has been demonstrated from neutrophils to endothelial cells, which was β2 integrin-dependent but degranulation-independent (247). Substantial neutrophil accumulation within atherosclerotic plaques has been demonstrated (248), localised adjacent to areas of endothelial cell apoptosis (249) and associated with rupture-prone lesions (250). Intravital microscopy has allowed visualisation of in vivo thrombosis, revealing a key role for neutrophils, which accumulated prior to platelets at sites of endothelial injury and were critical for thrombus formation following laser-induced injury in mice (251). 1.3.1.2 Interaction of endothelium with secreted neutrophil products Neutrophils secrete a variety of potentially pro-atherogenic factors which have been implicated in endothelial dysfunction, including cytokines, ROS, and antimicrobial proteins and proteases, released by either granule exocytosis or NET generation. Activated neutrophils release mediators such as TNFα, which induces actin stress fibre formation and increased permeability via phosphorylation of cytoskeletal proteins, and IL-8, which increases PI3Kγ- mediated vascular permeability via phosphorylation and internalisation of the endothelial junctional protein VE-cadherin (252). Interestingly, the release of TNFα from neutrophils is increased by inhibition of GSK-3, a known negative regulator of HIF-1 (253). Neutrophil- generated ROS can contribute to endothelial barrier dysfunction by a number of mechanisms, including disruption of junctional proteins, cytoskeletal reorganisation into stress fibres, activation of permeability-enhancing signalling pathways and promotion of neutrophil- endothelial adhesion (254). Neutrophil granule proteins may contribute to endothelial dysfunction, and their release is augmented in the inflammatory setting. The extent of this damage may be even greater in 43 COPD if neutrophils display impaired directionality, hence increasing the capacity for collateral damage. Extracellular release of MPO and hydrogen peroxide from neutrophils catalyses ROS production and these products are capable of oxidative modification of low density lipoproteins; both MPO and markers of MPO-catalysed halogenation have been identified in atherosclerotic lesions (255), and inhibition of the peroxidation and halogenation activity of MPO in vitro reduced endothelial oxidative stress and permeability (256). Furthermore, levels of plasma low density lipoprotein 3-chlorotyrosine (a specific footprint of MPO-HOCl activity) were increased in patients with CVD, and MPO has emerged as a prognostic marker in patients with acute coronary syndrome, heart failure and peripheral vascular disease (257). NE mediates endothelial dysfunction and atherogenesis in several ways: NE treatment of pulmonary artery endothelial cells caused apoptosis in vitro (258), pre-treatment with elastase inhibitors prevented I/R-induced neutrophil adherence and extravasation in feline mesenteric venules (259), and NE has been detected in atherosclerotic plaques, where it can activate both other neutrophil proteases, e.g. MMP-9, and pro-inflammatory endothelial cytokines, e.g. IL-1β (260,261). Moreover, NE cleavage of the macrophage receptor CD163 resulted in reduced clearance of haemoglobin, which favours atherosclerotic plaque destabilisation (262). Genetic and pharmacological inhibition of NE decreased experimental murine atherosclerosis (263), and higher circulating NE levels in patients with angina are associated with an increased risk of future cardiovascular events (264). Similarly, MMP-9 has been shown to cause apoptosis of cultured endothelial cells (265), has been detected in atherosclerotic plaques (260), and blockade of MMP-9 reduced atherosclerosis in mice (265). Higher circulating MMP- 9 levels are associated with increased aortic stiffness (266) and MMP-8 and -9 were implicated in aortic rupture in a case report of S. aureus pneumonia (267). Neutrophils also contribute to the progression of atherosclerosis by recruitment of other immune cells: deposition of mouse neutrophil CRAMP (homologue of human LL-37) on the inflamed endothelium of large arteries promoted adhesion of both neutrophils and monocytes in vivo (268). Several granule proteins, including MPO, NE and CRAMP/LL-37, are associated with NETs, which have been shown to promote endothelial cytotoxicity (measured by lactate dehydrogenase release) in vitro in a histone- and MPO-dependent manner (115). NETs have been observed in pulmonary capillaries ex vivo from septic mice and were able to bind to vascular endothelium and platelet- leukocyte aggregates in vivo, although the significance of this for endothelial damage or thrombus formation was unclear (269). However, elevated levels of circulating DNA and chromatin have been associated with coronary atherosclerosis and future cardiac events in humans (270), and pharmacological prevention of NETosis improved endothelium-dependent relaxation in mouse models of obesity and systemic lupus erythematosus (271,272), and could reduce atherosclerotic burden and arterial thrombosis in a murine photochemical injury model (273). 44 1.3.1.3 Interaction of endothelium with neutrophil-derived microvesicles NDMVs are present in small amounts in physiological conditions but have been reported to be robustly increased during inflammation: elevated circulating NDMVs have been detected in vasculitis, pneumonia, sepsis and atherosclerosis, and also in the bronchoalveolar lavage fluid (BALF) of ARDS patients (reviewed in (274)). However, whether their role is pro- or anti- inflammatory remains unclear. Proponents of pro-inflammatory NDMV effects include Mesri et al., who showed that MVs from phorbol myristate acetate- (PMA) or fMLP-treated neutrophils induced endothelial cell activation in vitro, upregulating IL-6 release (96); Pitanga et al., who showed that calcium ionophore-treated neutrophils generated MVs containing active MPO which caused endothelial cell injury with loss of cell membrane integrity and morphological damage, such as cell blebbing (97); Pluskota et al. who demonstrated that NDMVs could induce platelet P- selectin expression, which was pro-thrombotic (275); Nolan et al. who showed that neutrophils treated with a nitric oxide synthase inhibitor generated L-selectin- and PSGL-1-expressing MVs, which enhanced neutrophil trans-endothelial migration in vitro (276); and Hong et al., who demonstrated that children with active ANCA-positive vasculitis had more serum NDMVs than those with inactive disease. In vitro, ANCA isolated from these patients could stimulate the release of MVs from primed neutrophils and these NDMVs could activate endothelial cells, increasing ICAM-1, IL-6 and IL-8 expression (277). In contrast, others have produced evidence for anti-inflammatory effects: Lim et al. demonstrated that extravasating neutrophils in the mouse cremaster muscle deposit CD18- positive MVs, contributing to the maintenance of endothelial barrier integrity after neutrophil transmigration (278); Gasser et al. showed that MVs from fMLP or C5a-treated neutrophils increased the release of anti-inflammatory TGFβ1 from human monocyte-derived macrophages and reduced LPS-stimulated TNFα and IL-8 secretion (99); and Dalli et al. showed that MVs from fMLP or PAF-treated neutrophils contained annexin A1 which mediated reduced neutrophil adhesion to a HUVEC monolayer in vitro and, when injected intravenously, decreased neutrophil recruitment to an IL-1β inflamed murine air pouch (98). Dalli et al. additionally showed that NDMVs have different contents and, hence, functional properties, depending on whether they were produced from in vitro neutrophils in suspension or adherent to an endothelial monolayer. Overall, it is clear that the effect of NDMVs is highly context- dependent and may depend on the preparation/detection method as well as the stimulus and location. 45 1.3.2 The role of neutrophils in endothelial dysfunction and cardiovascular disease in COPD Multiple studies provide evidence for an increased burden of both endothelial dysfunction/subclinical CVD (226) and overt CVD (227,228) in COPD independent of traditional risk factors, including smoking. Inflammation is thought to play a key role: COPD patients with biomarkers of persistent systemic inflammation have increased all-cause mortality (279); COPD patients display excessive aortic inflammation compared with healthy ex-smokers (280); the risk of MI and ischaemic stroke is substantially increased following an acute exacerbation of COPD (232,281), a situation where there may be both systemic and local hypoxia and primed circulating neutrophils; leukocyte count is an independent predictor of impaired FMD in COPD (234); multi-analyte profiling has revealed an association of markers of neutrophilic inflammation and endothelial dysfunction (serum CRP, MPO and VEGF) with COPD severity (282); and primed peripheral blood neutrophils exhibiting delayed apoptosis have been identified ex vivo from COPD patients (283,284). Elevated circulating levels of neutrophil proteins and proteases have been observed in COPD. For example, serum MMP-9 is increased in COPD patients compared with healthy controls (285) and correlates with systolic hypertension and arterial stiffness (266), and increased serum MPO levels are associated with worse cardiovascular outcomes in COPD patients (286). Several studies have demonstrated an increase in the expression of neutrophil adhesion molecules, such as MAC-1, in COPD (287,288) although the effect on neutrophil trans-endothelial migration is less certain: Woolhouse et al. showed enhanced adhesion and transmigration of COPD neutrophils compared with healthy controls and smokers without COPD (289) whereas Mackarel et al. demonstrated that, despite increased adhesion, there was reduced transmigration of neutrophils isolated from patients with either COPD, CF/non- CF bronchiectasis or pneumonia (290). However, kinetic studies have shown increased retention of radiolabelled neutrophils in the pulmonary circulation in smokers compared with non-smokers (291). Furthermore, pulmonary microvascular blood flow was reduced in COPD, even in patients with mild disease and was distinct from emphysematous areas (292). Recent unpublished in vivo data from our laboratory show increased accumulation of radiolabelled neutrophils in the lungs of clinically stable COPD patients, comparable to the accumulation seen in healthy volunteers treated with inhaled LPS (personal communication, Dr Nicola Newman). This delayed transit time may provide an increased opportunity for neutrophil transmigration and, hence, endothelial and pulmonary damage. Although trans-endothelial passage usually causes only transient barrier disturbance, profound junctional disruption can occur under pathological circumstances, such as COPD, where circulating and migrating neutrophils are already primed to release their cytotoxic cargo. 46 Chronic and progressive hypoxia is a fundamental feature of COPD and has been linked to endothelial dysfunction. Hypoxia induced an endothelial response similar to that seen in systemic inflammation, with upregulation of inflammatory mediators, such as PAF, and increased expression of cell adhesion molecules (293). Additionally, hypoxia potentiated endothelial injury in a murine systemic LPS-mediated inflammatory model of high altitude cerebral oedema (294). The degree of hypoxia correlated with CRP, a systemic marker of inflammation, in COPD patients (295), and endothelial dysfunction of ex vivo pulmonary arteries correlated with the partial pressure of arterial oxygen in COPD patients undergoing lung transplantation (296). Furthermore, pulmonary hypertension (PH) is a common complication of COPD; it is characterised by pulmonary vascular remodelling and intimal obstructive proliferation of distal pulmonary arteries eventually leading to right heart failure, and the main driving factors are thought to be chronic hypoxia and inflammation. The estimated prevalence of PH in COPD is variable, and has been reported between 23% and 50% depending on the study population and diagnostic methods used; however, multiple studies have shown that the presence of PH has a negative impact on survival in COPD patients (reviewed in (297)). Neutrophilic inflammation has been implicated in the pathogenesis of PH in COPD: mice subjected to intra-tracheal instillation of elastase developed emphysema and PH (298); elastase inhibition could reverse hypoxia-induced PH in rats (299); mice deficient in MPO were protected from hypoxia-induced PH, and increased plasma NE and MPO were detected in PH patients, with high MPO levels associated with increased mortality (300). The only treatment demonstrated to have a beneficial survival effect for patients with co-existent COPD and PH is long term oxygen therapy, which has also been shown to partially reduce pulmonary artery pressure in these patients (301), although its effect on inflammation remains unclear. 1.4 Summary Neutrophils are armed with an array of effective antimicrobial weapons, including cytotoxic granule proteins and proteases. As infected and inflamed tissues can be profoundly hypoxic, neutrophils are required to work efficiently under these conditions. However, hypoxia can synergise with inflammatory signals, such as cytokines, to promote a highly destructive neutrophil phenotype, which not only enhances pathogen killing mechanisms but also increases the capacity for bystander tissue injury. Direct and indirect neutrophil interactions with the endothelium have been implicated in promoting endothelial dysfunction and contributing to the development of atherosclerosis in this scenario. Neutrophilic inflammation under hypoxic conditions characterises diseases such as COPD, which is associated with multiple systemic complications as a consequence of endothelial dysfunction, including CVD and PH. Neutrophils from COPD patients, as well as exhibiting intrinsic migratory 47 abnormalities, may contribute to tissue damage via enhanced release of granule proteins (e.g. NE and MPO), a signal which may be augmented by hypoxia. However, the precise mechanisms and agents responsible for neutrophil-mediated tissue damage in these hypoxic environments are unknown, particularly with respect to the vascular manifestations, and will therefore be the focus of my thesis. 1.5 Hypothesis and aims 1.5.1 Hypothesis Hypoxia synergises with inflammatory cytokines to engender a destructive neutrophil phenotype with enhanced release of histotoxic proteins, which contributes to endothelial dysfunction in COPD. 1.5.2 Specific aims Chapter 3 1) To investigate the mechanism of enhanced degranulation by hypoxic neutrophils 2) To establish the impact of hypoxia on neutrophil-mediated endothelial cell damage Chapter 4 3) To quantitatively identify proteins in the hypoxic versus normoxic neutrophil secretome which might mediate endothelial injury 4) To explore the mechanism(s) of differential protein release from neutrophils under hypoxia versus normoxia Chapter 5 5) To examine whether neutrophils or plasma from exacerbating COPD patients exhibit a “hypoxic signature” which may contribute to cardiovascular morbidity Chapter 2 Materials and Methods 49 2 Materials and methods 2.1 Materials Amersham Biosciences (Buckinghamshire, UK): Percoll®, dextran 500 (mw 500,000, dissolved in sterile 0.9% saline (6% w/v) and stored at 4ºC) APEXBIO (Texas,USA): PI3Kδ inhibitor (CAL-101) Baxter Healthcare (Berkshire, UK): Sterile 0.9% sodium chloride (NaCl) solution BD Biosciences (Embodegen, Belgium): Falcon polypropylene tubes (50, 15 and 14 ml), 50 ml syringes, Cellfix™, Annexin V-FITC/PI apoptosis detection kit I BioCytex (Marseille, France): Megamix beads Biolegend (London, UK): LEGEND MAX™ Human MRP8/14 (S100A8/A9, calprotectin) ELISA kit with pre-coated plate Bio-Rad (California, USA): DC colorimetric protein assay, Quick Start™ Bradford Protein Assay, Precision Plus Protein™ Dual Color Standard, PROTEAN® II electrophoresis system, BosterBio (California, USA): Human thioredoxin PicoKine™ ELISA kit with pre-coated plate Cloud-Clone Corp. (Texas, USA): Human S100A9 ELISA kit with pre-coated plate Corning (New York, USA): 100mm sterile tissue culture petri dishes, T75 sterile tissue culture flasks, 5ml round bottom polystyrene FACS tube (Falcon®), 70µm cell strainer (Falcon®) Elabscience (Texas, USA): Human cyclophilin A ELISA kit with pre-coated plate (assay range 1.25 ng/ml – 80 ng/ml) GE Healthcare (Buckinghamshire, UK): ECL (Enhanced Chemiluminescence) system, Vivaspin 2 MWCO 3000 concentrator columns, cyanine 3/5 dyes, IPG buffer, IEF gel strip (Immobiline™ Drystrip pH 3-10), 600 Ruby Standard Dual-Cooled Vertical Gel Unit Gibco (Maryland, USA): trypsin 0.05%, sterile distilled water, Iscove's Modified Dulbecco's Medium (IMDM) Greiner Bio-One Ltd (Gloucestershire, UK): ELISA plate (flat bottomed, high binding), 6-,12- and 96-well sterile culture plates Hospira Venisystems (Berkshire, UK): sterile 19 gauge Butterfly needles Invitrogen (Paisley, UK): rhodamine phalloidin F actin stain, TrypLE select enzyme (1X) 50 Life Technologies (Paisley, UK): Enzchek® elastase assay kit, penicillin (10 μg/ml), streptomycin (25 μg/ml) Lifespan Biosciences (Seattle, USA): Human S100A8 ELISA kit with pre-coated plate Lonza (Slough, UK): primary human pulmonary artery endothelial cells, EGM™-2 Bulletkit™, foetal calf serum (FCS) Martindale Pharmaceuticals Ltd (Essex, UK): sterile calcium chloride (CaCl2), sterile sodium citrate Merck Ltd. (Nottingham, UK): May-Grünwald-Giemsa stain, sodium thiosulphate pentahydrate, luminol Miltenyi Biotec (Surrey, UK): MACS mouse neutrophil isolation kit, QuadroMACS™ Separator, 3ml LS columns National Diagnostics (Nottingham, UK): Protogel, stacking buffer and resolving buffer Oxford Biosystems (Oxford, UK): Human cyclophilin A ELISA kit with pre-coated plate (assay range 0.39 ng/ml – 25 ng/ml) PanReac Applichem (Barcelona, Spain): 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) PeproTech (London, UK): recombinant murine tumour necrosis factor α (TNFα), recombinant murine granulocyte macrophage colony-stimulating factor (GM-CSF) Premier Foods (St Albans, UK): Marvel milk powder Promocell (Heidelberg, Germany): Endothelial cell growth medium MV kit Radiometer (Copenhagen, Denmark): ABL80 Basic automated blood gas analyser qualichek calibration kit RayBiotech (Georgia, USA): human resistin and NGAL ELISA kits with pre-coated plates R&D Systems (Oxfordshire, UK): recombinant human GM-CSF, recombinant human TNFα Sarstedt (Leicester, UK): S-Monovette® venous blood collection tubes (4.9ml serum and 9ml EDTA K3) 51 Sigma-Aldrich, (Dorset, UK): Eppendorf (Safe-Lock microfuge tubes, PCR clean, 1.5 ml and 2 ml), Pasteur pipettes, phosphate buffered saline (PBS) with (PBS+/+) or without (PBS-/-) calcium chloride and magnesium chloride, Histopaque®-1077, cytochalasin B from Drechlera dematioidea, N-formyl-methionyl-leucyl-phenylalanine (fMLP), PI3Kγ inhibitor (AS605240), PI3Kδ inhibitor (IC87114), α1 antitrypsin from human plasma, dimethyl sulfoxide (DMSO), glycine, sodium dodecyl sulphate (SDS), Trizma (tris(hydroxymethyl)aminomethane, tris base), ammonium persulfate (APS), tetramethylethylenediamine (TEMED), dithiothreitol (DTT), bromophenol blue, tissue culture grade bovine serum albumin (BSA), Tween-20, polyvinylidene fluoride (PVDF) membrane, Triton X-100, 3-[4,5-dimethylthiazol-2-yl]-2,5- diphenyl tetrazolium bromide (MTT), trichloroacetic acid (TCA), paraformaldehyde (PFA), sodium azide 0.1 M solution, ethanol, formaldehyde, silver nitrate, Ponceau S, ammonium acetate, ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), ethylenediamine tetraacetatic acid (EDTA), leupeptin, pepstatin A, aprotinin, sivelestat, sodium orthovanadate, sodium chloride, tetramisole, sodium pyrophosphate, sodium fluoride, β glycerophosphate, phenylmethylsulfonyl fluoride (PMSF), RIPA buffer, cOmplete™ mini EDTA-free protease inhibitor cocktail tablets (Roche), hydrogen peroxide (H2O2), o-dianisidine dihydrochloride (DMB), horseradish peroxidase type VI Spherotech (Illinois, USA): SPHERO™ Accucount blank particles 2.0-2.4µm ThermoFisher Scientific (Massachusetts, USA): Prolong gold antifade mountant with DAPI, acetone, NuPAGE pre-cast 1.0 mm 12% 12 well polyacrylamide gels, NuPAGE MES SDS running buffer, XCell SurelockTM Mini-Cell electrophoresis system, sodium carbonate decahydrate, glycerol, acetic acid, Nalgene™ high speed centrifuge tubes (30ml), Pierce™ 660nm protein assay, PagerulerTM Plus Prestained Protein Ladder, Mark12™ Unstained Standard, Restore PLUS western blot stripping buffer, TMT10-plex isobaric label reagents, SYTOX™ green nucleic acid stain Tocris Bioscience (Bristol, UK): Platelet Activating Factor (PAF) VWR International (Leicestershire, UK): DPX mountant, methanol, 0.5ml tubes PCR clean, propan-2-ol (isopropanol) 52 Table 2.1: Antibodies The antibodies shown in the table were obtained and used at the concentrations indicated. Incubation times are specified in individual western blot (WB) and fluorescence-activated cell sorting (flow cytometry, FACS) experiments. Primary Antibody Company ID Host species Species reactivity Mono/poly clonal Dilution Conjugate Application in thesis phospho- AKT (Ser473) Cell Signaling Technology #4060 rabbit human mono 1:1000 nil WB phospho- AKT (Ser473) Cell Signaling Technology #9271 rabbit human poly 1:1000 nil WB AKT Cell Signaling Technology #4691 rabbit human mono 1:1000 nil WB AKT Cell Signaling Technology #9272 rabbit human poly 1:1000 nil WB S100A9 Abcam ab63818 rabbit human poly 1:2000 nil WB cyclophilin A GeneTex GTX113520 rabbit human poly 1:1000 nil WB annexin A1 GeneTex GTX101070 rabbit human poly 1:2000 nil WB β actin Abcam ab8227 rabbit human mono 1:5000 nil WB CD54 (ICAM-1) BD Biosciences 559771 mouse human mono 1:8 APC FACS CD66b BD Biosciences 562940 mouse human mono 1:50 BV421 FACS CD41a BD Biosciences 560979 mouse human mono 1:50 PE FACS CD14 BD Biosciences 561709 mouse human mono 1:50 APC Cy7 FACS CD144 BD Biosciences 561566 mouse human mono 1:50 PerCP Cy5/5 FACS Secondary Antibody Company ID Host species Species reactivity Mono/poly clonal Dilution Conjugate Application in thesis IgG Dako P0448 goat rabbit poly 1:5000 horse radish peroxidase WB 53 2.2 Neutrophils 2.2.1 Neutrophil preparation 2.2.1.1 Neutrophil isolation from human whole blood Ethical permission for taking peripheral blood from healthy volunteers (REC reference 06/Q0108/281) was obtained from the Cambridge Local Research Ethics Committee. Neutrophils were isolated from whole blood by dextran sedimentation and centrifugation over discontinuous plasma-Percoll® gradients. All equipment and reagents were sterile, and all reagents used at room temperature unless otherwise stated. The isolation was conducted under sterile conditions in a laminar flow cell culture hood (Microflow Class II cabinet). All neutrophil centrifugation was undertaken using a Hettich Rotina 420R (Hettich Zentrifugen, Tuttlingen, Germany). 40-240 ml whole blood from healthy volunteers was collected by venepuncture into 50 ml syringes using 19 gauge butterfly needles. Blood was transferred gently into 50 ml polypropylene Falcon tubes containing 3.8% sodium citrate (1 ml per 10 ml blood). Whole blood was centrifuged at 300g, acceleration 9, brake 3 for 20 min. Platelet-rich plasma (PRP) supernatant was aspirated into new 50 ml tubes and centrifuged at 1400g, acceleration 9, brake 9 for 20 min. After centrifugation, the platelet-poor plasma (PPP) supernatant, was aspirated into new 50 ml tubes and the platelet pellet discarded. To make autologous serum, 10 ml PRP was aspirated into a separate sterile glass container prior to centrifugation and thoroughly mixed with 220 µl 10 mM CaCl2. The pellet from the initial centrifugation of whole blood, containing leukocytes (white blood cells, WBC) and erythrocytes (red blood cells, RBC), was sedimented with dextran to remove RBC: 2.5 ml 6% dextran (pre-warmed to 37ºC) per 10 ml pellet was added, made up to 50 ml total volume with sterile 0.9% NaCl (pre-warmed to 37ºC) and mixed gently. RBC were sedimented for 20-40 min until the RBC pellet reached its original (post-centrifugation) volume. After sedimentation, the WBC-containing top layer was aspirated and the sedimented RBC discarded. WBC were centrifuged at 260g, acceleration 9, brake 9 for 5 min. The supernatant was discarded and the WBC-containing pellet was re- suspended gently in 2 ml PPP. Percoll® (pre-cooled to 4ºC) was diluted in PPP to make 42% and 51% gradients. 2 ml WBC in PPP were transferred to 15 ml polypropylene tubes. Firstly, the 42% Percoll® gradient and, secondly, the 51% gradient, were underlaid using a glass Pasteur pipette. Gradients were centrifuged at 150g, acceleration 1, brake 0 for 14 min (Figure 2.1A). The top layer of cells at the 42%/PPP interface, containing peripheral blood mononuclear cells (PBMCs, Figure 2.1B), was aspirated and discarded. The lower layer of cells at the 51%/42% interface, containing granulocytes (Figure 2.1C), was carefully aspirated using a plastic Pasteur pipette and re-suspended gently in the remaining PPP. The 54 granulocytes were centrifuged at 256g, acceleration 9, brake 9 for 5 min. The supernatant was discarded and the cell pellet gently re-suspended. Cells were washed in 50 ml PBS-/- (pre- warmed to 37ºC) by centrifuging at 256g, acceleration 9, brake 9 for 5 min. Prior to centrifugation,100 µl were removed by P200 Gilson using a cell saver pipette tip to make the purity slide cytospin. 20 µl were removed for haemocytometer cell count. The cells were then washed in 50 ml PBS+/+ (pre-warmed to 37ºC) by centrifuging at 256g, acceleration 9, brake 9 for 5 min. The supernatant was discarded and the cell pellet gently re-suspended to be used in further experiments. By this method the efficiency of neutrophil recovery was >80%, and the neutrophil isolate was >95% pure with less than 1% monocyte contamination. Cytospins to count neutrophil purity were made by centrifugation in a Shandon Cytospin 3 centrifuge, 300 rpm for 3 min. To assess purity, the cells were fixed in methanol and subsequently stained with May-Grünwald-Giemsa (Figure 2.1B&C). Cells were counted using light microscopy (Olympus CX31). Figure 2.1: Isolation of neutrophils from human whole blood A: Photograph of plasma-Percoll® gradient following centrifugation. Mononuclear cells comprise the upper cell layer between plasma and 42% Percoll®; granulocytes comprise the lower cell layer at the interface of 42% and 51% Percoll®. B&C: Representative photomicrograph of cytospins, stained with May-Grünwald-Giemsa, from the upper mononuclear (B) and lower granulocyte (C) cell layers (x40 magnification, scale bar indicates 10µm). 55 2.2.1.2 Neutrophil isolation from murine bone marrow Mouse strains: E1020K heterozygote or D910A homozygote mice bred on a C57BL/6J background with strain-matched wildtype controls were kindly provided by Professor Klaus Okkenhaug (Babraham Institute, Cambridge, Home Office License Numbers PPL 80/2248 and 70/7661). PI3Kγ-/- mice bred on a C57BL/6 E129 background with strain-matched wildtype controls were kindly provided by Dr Len Stephens (Babraham Institute, Cambridge, PPL Home Office License Number PPL 70/8100). Additional C57BL/6J wildtype mice were kindly provided by Professor Nicholas Morrell (University of Cambridge, Home Office License Number PPL 70/8850). All mice were aged between 8-15 weeks. Neutrophils were isolated from murine femoral bone marrow by negative immunomagnetic selection using a MACS neutrophil isolation kit. All equipment and reagents were used at 4ºC throughout. Bone marrow was obtained by flushing femurs with MACS buffer (PBS-/-, pH 7.2, 0.5% BSA, 2 mM ethylenediamine tetraacetatic acid (EDTA)) through a 70 µm sterile cell strainer. Total cell number was determined with a haemocytometer. The cell suspension was centrifuged at 300g for 10 min and the supernatant discarded. The cell pellet was re- suspended in MACS buffer (25*108 cells/ml), mixed with neutrophil biotin-antibody cocktail (1 µl per 1*106 cells, proprietary recipe) and incubated for 10 min at 4ºC. Cells were washed in 20 ml MACS buffer (300g, 10 min) and the supernatant discarded. The cell pellet was re- suspended in MACS buffer (12.5*108 cells/ml), mixed with anti-biotin microbeads (2 µl per 1*106 cells) and incubated for 15 min at 4ºC. Cells were washed in 20 ml MACS buffer (300g, 10 min) and the supernatant discarded. The cell pellet was re-suspended in MACS buffer (2*108 cells/ml) and subjected to magnetic separation using a QuadroMACS™ Separator and 3 ml LS columns. Columns were initially rinsed with 3 ml MACS buffer. The cell suspension was then applied and the neutrophil-rich flow-through collected in a fresh 15 ml polypropylene tube. Columns were washed thrice with 3 ml MACS buffer and the filtrate collected. 20 µl neutrophil suspension was removed for haemocytometer cell count and 100 µl removed to make the purity slide cytospin before centrifuging the cell suspension (300g, 5 min). The supernatant was discarded and the cell pellet washed in 50 ml serum-free PBS-/- (300g, 5 min). The supernatant was discarded and the cell pellet washed in 50 ml serum-free PBS+/+ (300g, 5 min). The supernatant was discarded and the cell pellet gently re-suspended to be used in further experiments. 2.2.1.3 Generation of neutrophil supernatants Human neutrophil supernatants were generated as previously published (149), with the modifications described below. Freshly isolated neutrophils were re-suspended in normoxic or hypoxic IMDM at a concentration of 11.1*106/ml. Aliquots of 270 µl (3*106 neutrophils) were 56 transferred to 2 ml Eppendorfs and incubated under normoxia or hypoxia in a thermomixer (37ºC, 450 rpm) for 4 h. After 4 h, cells were treated with GM-CSF (10 ng/ml, 30 min, final concentration), TNFα (20 ng/ml, 30 min, final concentration) or PAF (1 µM, 5 min, final concentration), and subsequently fMLP (100 nM, 10 min, final concentration), or IMDM control, with a final volume of 1 ml. After treatment, samples were centrifuged (13,000 rpm, 10 s, MSE Micro Centaur microfuge) to pellet cells, and supernatants were harvested. Unless otherwise stated, human neutrophil supernatants for all experiments were generated by this method. Murine neutrophil supernatants were generated similarly. Freshly isolated murine bone marrow neutrophils were re-suspended in normoxic or hypoxic IMDM at a concentration of 11.1*106/ml - 18.5*106/ml. Aliquots of 270 µl (3-5*106 neutrophils, depending on total number isolated from murine bone marrow) were transferred to 2 ml Eppendorfs and incubated under normoxia or hypoxia in a thermomixer (37ºC, 450 rpm) for 4 h. Neutrophil numbers were kept consistent between conditions and genotypes within individual experiments. After 4 h, cells were treated with murine GM-CSF (10 ng/ml, 30 min), murine TNFα (10 ng/ml, 30 min), or cytochalasin B (5 µg/ml, 5 min), and subsequently fMLP (10 µM, 10 min), or IMDM control, with a final volume of 400 µl. Cells were pelleted (13,000 rpm, 10 s, MSE Micro Centaur microfuge) and supernatants harvested. 2.2.1.4 Optimisation of neutrophil cell lysate generation It is challenging to generate neutrophil lysates as disruption to cell and internal membranes upon addition of lysis buffer can result in significant protein degradation following liberation of the abundant proteases and phosphatases contained within the granules. This was apparent when performing western blotting (see Table 2.1 for all antibody details) for AKT (#9272) and phosphorylated-AKT(ser473) (pAKT, #9271). Initially, lysates were prepared by adding hypotonic lysis buffer (10 mM Tris pH 7.8, 1.5 mM EDTA, 10 mM KCl, 500 µM dithiothreitol (DTT), 1 mM sodium orthovanadate, 2 mM tetramisole, one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet 1 tablet/4 ml) to cell pellets on wet ice and vortexing every 15 min for a total of 60 min before centrifugation (3100g, 5 min) to remove debris, in accordance with the protocol previously established in our laboratory. Using this lysate preparation method, probing for pAKT (#9271) and AKT (#9272) by western blotting (see section 2.2.5.3 for method) demonstrated extremely variable AKT protein bands compared with β actin loading control (Figure 2.2A), and complete inability to visualise any protein bands for pAKT. Further investigation revealed that AKT was being cleaved into fragments of a lower molecular weight, as the protein band for AKT should be seen at 60 kDa (Figure 2.2B). 57 Figure 2.2: Expression of total AKT by western blot Isolated neutrophils were incubated under normoxia (21% O2, N) or hypoxia (0.8% O2, H) for 4 h. Cells were treated with GM-CSF (10 ng/ml, 15 min, GM) or IMDM control (C) in the presence or absence of PI3K inhibitors (IC87114, 3 µM, δ; AS605240, 3 µM, γ). Cell lysates were prepared with the addition of hypotonic lysis buffer on wet ice and subjected to SDS- PAGE; western blotting was performed with A&B: anti-AKT (#9272) and A: anti-β actin. Firstly, I wished to investigate the ability of these antibodies to detect AKT and pAKT proteins, using lysates from other cell types as a positive control. I therefore trialled polyclonal anti- pAKT (#9271) against monoclonal anti-pAKT (#4060, kindly provided by Professor Klaus Okkenhaug), and polyclonal anti-AKT (#9272) against monoclonal anti-AKT (#4691) (Table 2.1). Lysates from PDGF-stimulated smooth muscle cells (kindly provided by Dr Paul Upton, University of Cambridge) and unstimulated or anti-IgM stimulated B cells (kindly provided by Dr Valentina Carbonaro, Babraham Institute) were used as positive controls (Figure 2.3). This test demonstrated that monoclonal anti-pAKT (#4060, Figure 2.3B) and polyclonal anti-AKT (#9272, Figure 2.3C) were better able to detect pAKT and AKT protein, and that bands of correct molecular weight were seen in the positive control samples but degraded protein was seen in my samples. These antibodies were therefore used for all future pAKT/AKT western blot experiments. 58 Figure 2.3: The ability of monoclonal versus polyclonal antibodies to detect pAKT and AKT by western blot Isolated neutrophils were incubated under normoxia (N) or hypoxia (H) for 4 h and treated with GM-CSF (10 ng/ml, 15 min). Neutrophil lysates were prepared with the addition of hypotonic lysis buffer on wet ice. Lysates from PDGF-stimulated smooth muscle cells (P) at 1, 5 and 10 min, and lysates from unstimulated (-) or anti-IgM-stimulated (+) B cells (BC) were used as controls. All lysates were subjected to SDS-PAGE; western blotting was performed with A: polyclonal anti-human pAKT (#9271), B: monoclonal anti-human pAKT (#4060), C: polyclonal anti-human AKT (#9272) and D: monoclonal anti-human AKT (#4691). 59 Secondly, I wished to investigate the capacity of different buffers and lysate preparation methods to inhibit protein degradation in my samples. I therefore trialled three different lysis buffers: Buffer 1) 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 20 mM EDTA, 20 mM ethylene glycol-bis(β- aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), 2.5 mM sodium pyrophosphate, 1 mM β glycerophosphate, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM sodium fluoride, 1% Triton-X 100, 10 µg/ml leupeptin, 10 µg/ml aprotinin and 10 µg/ml pepstatin A Buffer 2) Radioimmunoprecipitation assay (RIPA) buffer, 20 mM EDTA, 20 mM EGTA, 1 mM PMSF, 1 mM sodium fluoride, 1% Triton-X 100, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml pepstatin A and one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet (1 tablet/4 ml buffer) Buffer 3) 250 mM Tris-HCl pH 6.8, 20% glycerol, 4% SDS and one cOmplete™ mini EDTA- free protease inhibitor cocktail tablet (1 tablet/4 ml buffer) and three different preparation methods: 1) Addition of lysis buffer to cell pellet at room temperature, vortexing (30 s) every 15 min for 60 min 2) Addition of lysis buffer to cell pellet on wet ice, vortexing (30 s) every 15 min for 60 min on wet ice 3) Snap freezing of cell pellet immediately on dry ice, addition of lysis buffer to frozen pellet, warming and vortexing (30 s) before boiling with agitation (99ºC, 550 rpm, 10 min) and freezing again on dry ice This trial demonstrated the best protein recovery with buffer 3 in combination with snap freezing of the neutrophil pellet on dry ice prior to lysis buffer addition (Figure 2.4A). This method was employed for all further lysate preparation and gave good protein recovery for both pAKT (#4060, Figure 2.4B) and AKT (#9272, Figure 2.4C). 60 Figure 2.4: The effect of different buffers and lysate preparation methods on pAKT and AKT protein detection by western blot Isolated neutrophils were incubated under normoxia (N) or hypoxia (H) for 4 h. Cells were treated with GM-CSF (10 ng/ml, 15 min, GM) or IMDM control (C) in the presence or absence of PI3K inhibitors (IC87114, 3 µM, δ; AS605240, 3 µM, γ). A: Cell lysates were prepared with the addition of Buffer 1 (20 mM Tris-HCl pH 7.5, 150 mM NaCl, 20mM EDTA, 20 mM EGTA, 2.5 mM sodium pyrophosphate, 1 mM β glycerophosphate, 1 mM PMSF, 1 mM sodium fluoride, 1% Triton-X 100, 10 µg/ml leupeptin, 10 µg/ml aprotinin and 10 µg/ml pepstatin A), Buffer 2 (RIPA buffer plus 20mM EDTA, 20 mM EGTA, 1 mM PMSF, 1 mM sodium fluoride, 1% Triton-X 100, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml pepstatin A and one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet, 1 tablet/4 ml buffer) or Buffer 3 (250mM Tris-HCl pH 6.8, 20% glycerol, 4% SDS and one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet, 1 tablet/4 ml buffer) at room temperature (RT), on wet ice (wet) or dry ice (dry). B&C: Cell lysates were prepared with the addition of Buffer 3 on dry ice. Lysates were subjected to SDS-PAGE. Western blotting was performed with A&C: anti-AKT (#9272) and B: anti-pAKT (#4060). 61 2.2.1.5 Generation of neutrophil cell lysates Freshly isolated neutrophils were re-suspended in normoxic or hypoxic IMDM at a concentration of 11.1*106/ml. Aliquots of 900 µl (10*106 neutrophils) were transferred to 2 ml Eppendorfs. For some experiments, PI3K inhibitors were added prior to incubation: AS605240 (PI3Kγ-selective inhibitor, 3 µM) or IC87114 (PI3Kδ-selective inhibitor, 3 µM). Cells were incubated under normoxia or hypoxia in a thermomixer (37ºC, 450 rpm) for 4 h. After incubation, cells were treated with GM-CSF (10 ng/ml, 15 min), PAF (1 μM, 5 min) or IMDM control, with or without subsequent fMLP (100 nM, 10 min). After treatment, samples were centrifuged (13,000 rpm, 10 s, MSE Micro Centaur microfuge) to pellet cells, and supernatants were harvested. Cell pellets were snap frozen immediately on dry ice and 200 µl lysis buffer were added (250 mM Tris-HCl pH 6.8, 20% glycerol, 4% SDS, and one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet, 1 tablet/4 ml buffer). Samples were warmed, vortexed (30 s), boiled with agitation (99ºC, 550 rpm, 10 min) and then centrifuged (3100g, 5 min) to remove debris. Aliquots of 50 µl cell lysate were stored at -80ºC. 2.2.2 Working under hypoxia Hypoxic experiments were conducted in a SCI-tive Dual hypoxia workstation (Baker Ruskinn) (Figure 2.5A). Levels of O2 and CO2, set via a programmable touch screen control, were maintained by gas feeds of 100% N2, 100% CO2 and compressed air. The chamber is accessed by three glove ports with attached sleeves, which maintain an airtight seal around the user’s arms. To introduce an item into the chamber, it was placed into the central airlock which was flushed with nitrogen for 160 seconds to reach a level of 1.0% O2. This enabled introduction of items into the chamber’s central workspace without re-oxygenation. The pre-set working levels for hypoxia were 0.8% O2 and 5.0% CO2, which had been previously optimised for IMDM media (302). The pH of normoxic and hypoxic IMDM was measured over 5 h to ensure no pH difference during the experimental incubation timecourse (Figure 2.5B). All equipment and reagents were equilibrated in hypoxia for at least 3 h prior to commencing the experiment. Unless otherwise stated, the levels for all experiments carried out under hypoxia were 0.8% O2 and 5.0% CO2. The % O2 within the workstation could be assessed in two ways: by interrogation of the workstation trend log, which records actual internal % O2 every minute, and by attaching a calibrated oxygen sensor to a port on the side of the workstation, which accessed the inner workspace. However, as the workstation controls the oxygen tension within the chamber, rather than the oxygen tension experienced by the cells re-suspended in hypoxic media, it was also important to check the gaseous partial pressures 62 of O2 and CO2 in the media. This was achieved with an ABL80 Basic automated blood gas analyser. Due to the expense and limited lifespan of the calibration and QC solutions, the pH, pO2 and pCO2 of hypoxic equilibrated media was periodically checked at regular intervals to ensure that the required levels were being maintained (Figure 2.5C). Correlation of media gas analysis values with workstation O2 assessments confirmed that the required level of hypoxia was achieved and stable. The workstation was maintained at 37ºC and 70% humidity throughout experiments. Figure 2.5: Analysis of media pH, pO2 and pCO2 IMDM was incubated under normoxia (room air, 50 ml Falcon, bead bath at 37ºC) or hypoxia (0.8% O2 and 5.0% CO2, 100 mm diameter petri dish, hypoxia workstation at 37ºC). A: photograph showing the SCI-tive Dual hypoxia workstation. B: pH was measured hourly under normoxia or hypoxia using a benchtop pH meter. C: After 3 h hypoxic incubation, samples of IMDM were taken using a 1 ml syringe with an airtight cap and immediately assessed for pH, pCO2 and pO2 using an ABL 80 blood gas analyser. Data represent mean ± SEM, B: n=3, C: n=10. 2.2.3 Neutrophil functional assays 2.2.3.1 Assessment of shape change Freshly isolated neutrophils were re-suspended in PBS+/+ at a concentration of 5*106/ml. 90 µl aliquots of cells were transferred to 2 ml Eppendorfs in a thermomixer at (37ºC, 450 rpm). Cell were treated with fMLP (100 nM) or PBS+/+ control. Cells were fixed at baseline (control) or after 30 min (control and fMLP) by adding 250 µl Cellfix™ and placing on ice for 1 min. 90 µl from each sample were transferred to separate 14 ml polypropylene round-bottomed FACS tubes and 500 µl ice-cold PBS-/- added. Shape change was analysed by flow cytometry (BD FACSCanto II), gating on single granulocytes, and assessing change in forward scatter (Figure 2.6A-C). In response to a chemoattractant stimulus, activated (shape changed) 63 neutrophils become flattened and elongated, due to remodelling of the actin cytoskeleton, which increases their forward scatter profile. Figure 2.6D shows that basal neutrophils were not shape-changed but there was a significant increase in forward scatter correlating with an increase in % shape changed cells with fMLP (p<0.0001), confirming that I was able to isolate un-primed responsive neutrophils. Figure 2.6: Gating strategy and quantification of neutrophil shape change by flow cytometry Neutrophils were treated with fMLP (100 nM) or PBS+/+ (control). Cells were fixed at baseline or after 30 min. Shape change was assessed by flow cytometric analysis of forward scatter (BD FACSCanto II). A-C: Representative flow plots of A: control 0 min, B: control 30 min and C: fMLP-treated 30 min forward scatter area (FSC-A) vs side scatter area (SSC-A), gating on neutrophils, with corresponding histograms of FSC-A below. D: Quantification of shape change expressed as FSC-A positive cells as % of total singlet neutrophil population. Results represent mean ± SEM, n=5, ****=p<0.0001, one way ANOVA, Tukey’s multiple comparisons. 64 2.2.3.2 Assessment of apoptosis by morphology Freshly isolated neutrophils were re-suspended in normoxic or hypoxic IMDM, containing 10% autologous serum and 0.5% penicillin/streptomycin (50 U/ml), to give a final concentration of 5*106/ml. 135 µl aliquots of cells were added to a 96 well plate and treated with GM-CSF (1 ng/ml), PAF (1 μM) or IMDM control at the start of incubation. Cells were incubated at 37ºC under normoxia or hypoxia for 20 h. After incubation, neutrophil apoptosis was quantified by morphology, and/or flow cytometry. For morphologic assessment, neutrophils were harvested by gently pipetting up and down using cell saver tips. Cytospins were made by centrifugation in a Shandon cytospin 3 centrifuge, 300 rpm for 3 min. Cells were fixed in methanol (4 min) and stained with May- Grünwald-Giemsa. Morphology was examined by oil-immersion light microscopy (Olympus CX31, 100x magnification). Apoptotic neutrophils have a distinctive morphologic appearance: condensation of chromatin leads to loss of the multi-lobed nucleus (darkly staining pyknotic nuclei), and there is cell shrinkage and vacuolation of the cytoplasm (Figure 2.7). A minimum of 300 cells per slide in randomly selected fields of view were counted in order to quantify neutrophil apoptosis. Figure 2.7: Assessment of neutrophil apoptosis by morphology Cytospins were stained with May-Grünwald-Giemsa and examined at 100x magnification under oil immersion. Photomicrograph show A: non-apoptotic and B: apoptotic neutrophils, with cartoons representing progression of nuclear morphological changes. 65 2.2.3.3 Assessment of apoptosis by flow cytometry For flow cytometric assessment, neutrophils were harvested by gently pipetting up and down using cell saver tips and transferred to cooled FACS tubes. All reagents were pre-cooled to 4ºC and kept on ice. Cells were centrifuged at 4ºC, 256g, acceleration 9, brake 9 for 5 min. Supernatants were discarded and cell pellets gently re-suspended in 100 µl 1x binding buffer (1:10 dilution 10x stock binding buffer) containing 5% FITC-AnV and 5% PI. For compensation, cells were re-suspended in 100 µl 1x binding buffer (unstained), 100µl 1x binding buffer with 5% AnV (AnV single stained) or 100 µl 1x binding buffer with 5% PI (PI single stained). Samples were vortexed and stained on ice for 20 min, protected from light. The reaction was then quenched by adding 400 µl 1x binding buffer to each tube. Flow cytometry (BD FACSCanto II) was performed using the blue (488 nm excitation; 530/30 filter) and red (633 nm; 660/20 filter) lasers. Annexin V binds phosphatidylserine which is externalised on the membrane of apoptotic cells; PI is a nuclear stain which can penetrate necrotic cells, allowing assessment of membrane integrity. The gating strategy is shown in Figure 2.8: AnV-/PI- staining indicated viable non-apoptotic cells, AnV+/PI- staining indicated early apoptotic cells and AnV+/PI+ staining indicated late apoptotic or necrotic cells. Unstained and single-stained samples were used to set compensation. Results were analysed using FlowJo version 10 software. 2.2.3.4 Assessment of neutrophil reactive oxygen species production Extracellular ROS production was quantified by chemiluminescence. The chemiluminescent probe luminol is oxidized by ROS in the presence of an exogenously applied peroxidase catalyst, such as horseradish peroxidase (HRP), yielding light emission which can be measured with a luminometer. As HRP cannot cross the neutrophil cell membrane, the luminol/HRP-dependent chemiluminescence detects extracellular ROS production. Freshly isolated neutrophils were re-suspended in PBS+/+ at a concentration of 5*106/ml. Aliquots of 180 µl were transferred to 2 ml Eppendorfs and incubated in a thermomixer (37ºC, 450 rpm). Cells were treated with TNFα (20 ng/ml, 30 min), PAF (1 µM, 5 min) or PBS control before addition of luminol (1 µM, 3 min). Luminol-treated cells were transferred into an opaque 96- well plate containing HRP (62.5 U/ml) and chemiluminescence measured immediately with a luminometer. The luminometer injector was primed prior to ROS analysis and cells in all wells were treated with fMLP (100 nM) at baseline via the injector. ROS kinetics were measured for 10 min and extracellular ROS production was expressed as raw peak height in relative light units (RLU). Data were generated with the assistance of Dr Arlette Vassallo. 66 Figure 2.8: Gating strategy for assessment of neutrophil apoptosis by flow cytometry Neutrophils were incubated under normoxia (A&C) or hypoxia (B&D) for 20 h. Cells were treated at baseline with GM-CSF (1 ng/ml, C&D) or IMDM control (A&B). After 20 h, cells were stained with FITC-AnV and PI and apoptosis assessed by flow cytometry (BD FACSCanto II). A-D: Representative flow plots of FITC-AnV (y axis) and PI (x axis) staining. Q1 shows AnV+/PI- (apoptotic) cells, Q2 shows AnV+/PI+ (late apoptotic/necrotic) cells, Q3 shows AnV- /PI+ (necrotic) cells (though most necrotic cells are AnV+), Q4 shows AnV-/PI- (viable) cells. Cells per quadrant as % of total neutrophils are indicated. 2.2.3.5 Assessment of neutrophil extracellular trap production Freshly isolated neutrophils were re-suspended at 1x106/ml in normoxic or hypoxic IMDM, containing the cell impermeable nucleic acid stain SYTOX™ green (5 μM). This established method allows objective quantification of extracellular DNA, which equates to NETosis provided the cells are not necrotic (e.g. (303)). Aliquots of 200 µl (2*105) cells were transferred to a 96 well plate. At baseline, three wells were permeabilised with 0.5% Triton-X and fluorescence absorbance (indicating total cellular DNA content) measured after 30 min (peak signal). Remaining cells were treated at the beginning of incubation with GM-CSF (10 ng/ml), PAF (1 µM) or IMDM control, and fluorescence absorbance (indicating extracellular DNA) 67 measured hourly. For some experiments, cells were treated after 4 h normoxic/hypoxic incubation with GM-CSF (10 ng/ml, 30 min), and subsequently fMLP (100 nM, 10 min). Here, fluorescence absorbance was measured at baseline, 4 h and after treatment. As a positive control, cells were treated with PMA (20 nM) at baseline and incubated under normoxia for 4 h. SYTOX green fluorescence was measured with a Victor 3 multilabel plate reader (Perkin Elmer), reading fluorescence absorbance at 485/535 nm. NET production was expressed as extracellular DNA as % of total baseline DNA, calculated by subtracting fluorescence at baseline from the fluorescence at each time point and then dividing by the fluorescence values of lysed cells. All experimental values were standardized to total DNA. 2.2.4 Neutrophil degranulation assays 2.2.4.1 Assessment of neutrophil elastase release Supernatant NE activity was quantified by Enzchek® elastase activity assay, which measures the ability of NE to cleave non-fluorescent (quenched) BODIPY labelled-DQ-elastin substrate, into fluorescent fragments. DQ-elastin substrate (reconstituted in distilled water (dH2O) to 1 mg/ml) was diluted 1:20 in 1x reaction buffer and 100 µl added to required wells of a 96 well enzyme-linked immunosorbent assay (ELISA) plate. 100 µl undiluted sample were added to DQ-elastin wells in duplicate. 100 µl 1x reaction buffer or IMDM were added as negative control. Whether the observed proteolytic activity was the result of elastase was determined by addition of the selective NE inhibitor N-Methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone (Figure 2.9), supplied with the kit (10 mg/ml in DMSO). 50 µl elastase inhibitor (10 µM), or 1x reaction buffer control, were added to required wells of a 96 well ELISA plate. DQ-elastin substrate (1 mg/ml in dH2O) was diluted 1:10 in 1x reaction buffer and 50 µl added to each well. 100 µl undiluted supernatant sample were added to DQ-elastin wells in duplicate. 100 µl 1x reaction buffer or IMDM were added as negative control. 100 µl porcine pancreatic elastase, supplied with the kit, (PPE, 0.2 U/ml) was added as positive control. For all experiments, plates were incubated in the dark at room temperature and fluorescence measured at 485/535 nm by Victor 3 multilabel fluorescence plate reader (Perkin Elmer) at exactly 30 min. Background fluorescence in negative control wells was averaged and subtracted from each sample reading. 2.2.4.2 Optimisation of the NE activity assay for murine neutrophils Supernatants were assessed for NE activity by Enzchek® elastase activity assay, exactly as described in section 2.2.4.1. Compatibility of the kit with murine NE was tested by the addition of the selective elastase inhibitor N-Methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone, 10mg/ml in DMSO), using the protocol for determining NE inhibition exactly as described in section 2.2.4.1. Addition of the NE-selective inhibitor abrogated the NE response (Figure 68 2.10A), indicating that the Enzchek® kit was able to detect murine NE activity. In order to maximise NE release from murine neutrophils, several priming agonists were examined, with supernatants generated under normoxia only as cell numbers were limited. Cytochalasin B was the only priming agent to enable NE release from murine neutrophils substantially above baseline control (Figure 2.10A) so was used for all further experiments with these cells. As the intention of using murine neutrophils was to interrogate the contribution of PI3K signalling to the hypoxic uplift of NE release, the ability of hypoxia to affect NE release from cytochalasin B and fMLP-treated murine neutrophils was tested. Hypoxia increased NE release from cytochalasin B and fMLP-treated neutrophils (Figure 2.10B) although, as the NE activity was lower than that seen in human neutrophil supernatants, the assay plate was incubated overnight at 37ºC. This allowed optimisation of enzyme activity and, as the Enzchek® assay is continuous, kinetic data could be obtained at multiple timepoints. For subsequent experiments, a 20 h timepoint was chosen to take fluorescence readings. Figure 2.9: The effect of an NE-selective inhibitor on the detection of supernatant NE activity Human neutrophils were re-suspended in IMDM under normoxic or hypoxic (0.8% O2 and 5% CO2) conditions. Neutrophils (11.1*106/ml) were treated with PAF (1 μM, 5 min) and fMLP (100 nM, 10 min), or IMDM control, in the presence or absence of NE-selective inhibitor N- Methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone (10 mg/ml). Cells were pelleted and supernatant NE activity measured by Enzchek® assay. Results represent mean ± SEM, n=3; * = p<0.05, ** = p<0.01, two way ANOVA, Sidak’s multiple comparisons test. 69 Figure 2.10: The effect of priming agonists and an NE-selective inhibitor on the detection of murine supernatant NE activity Murine neutrophils were re-suspended in IMDM under normoxia or hypoxia. Neutrophils (18.5*106/ml) were treated with A: murine TNFα (10 ng/ml, 30 min), murine GM-CSF (10 ng/ml, 30 min) or A&B: cytochalasin B (5 µg/ml, 5 min), and subsequently fMLP (10 µM, 10 min), or IMDM control, in the presence or absence of NE-selective inhibitor N-Methoxysuccinyl-Ala- Ala-Pro-Val-chloromethyl ketone (10 mg/ml). Supernatant NE activity was measured by Enzchek® assay at 20 h. A: Results represent mean, n=1; B: Results represent mean ± SEM, n=2. 2.2.4.3 Assessment of human NE release with PI3Kinase inhibition Freshly isolated neutrophils were re-suspended in normoxic or hypoxic IMDM at a concentration of 11.1*106/ml. Aliquots of 270 µl (3*106 neutrophils) were transferred to 2 ml Eppendorfs. In order to investigate the effects of inhibition of PI3K signalling on neutrophil function, PI3K inhibitors were added prior to incubation: AS605240 (PI3Kγ-selective inhibitor, 3 µM), IC87114 (PI3Kδ-selective inhibitor, 3 µM) or CAL-101 (PI3Kδ-selective inhibitor, 100 nM). PI3K inhibitors were used at concentrations previously published (304) or at concentrations optimised in functional neutrophil assays in our laboratory to allow isoform selectivity. Cells were then incubated for 4 h under normoxia or hypoxia, and treated with PAF (1 µM, 5 min), and subsequently fMLP (100 nM, 10 min), or IMDM control. Supernatants were harvested and NE activity was quantified as described in section 2.2.4.1. 2.2.4.4 Assessment of NE release from murine neutrophils with PI3Kinase mutations To further investigate the role of the PI3Kδ isoform in the hypoxic regulation of NE release, neutrophils were isolated from mice with an activating (E1020K) or kinase-dead (D910A) mutation of PI3Kδ (kindly provided by Professor Klaus Okkenhaug, Babraham Institute). To further investigate the role of the PI3Kγ isoform in the hypoxic regulation of NE release, 70 neutrophils were isolated from mice deficient in PI3Kγ (PI3Kγ-/-) (kindly provided by Dr Len Stephens, Babraham Institute). Femurs were obtained from E1020K heterozygote or D910A homozygote C57BL/6J mice, or PI3Kγ-/- C57BL/6 E129 mice, aged between 8-15 weeks. Each experiment was conducted alongside the appropriate strain-matched wildtype controls. Neutrophils were isolated from murine femoral bone marrow by negative immunomagnetic selection, exactly as described in section 2.2.1.2 and supernatants generated as described in section 2.2.1.3, treating cells with cytochalasin B (5 µg/ml, 5 min) and fMLP (10 µM, 10 min), conditions optimised previously (section 2.2.4.2). Cells were pelleted (13,000 rpm, 10 s, MSE Micro Centaur microfuge) and supernatants carefully harvested. Supernatants were assessed for NE activity by Enzchek® elastase activity assay, as described in section 2.2.4.1. 2.2.4.5 Assessment of MPO release Supernatant MPO activity was quantified by measurement of the hydrogen peroxide (H202)- dependent oxidation of o-dianisidine dihydrochloride (DMB) by MPO. Neutrophil supernatants were generated as described in section 2.2.1.3. In addition, to allow assessment of total cellular MPO content, 3*106 pelleted neutrophils were re-suspended in 200 µl PBS-/-, lysed with 3 µl Triton X-100 (10%), vortexed, and the resulting lysates harvested. To measure MPO activity, the following reagents were added to 50 µl of cell supernatants/lysates: 300 µl BSA (0.025% in PBS+/+), 250 µl PBS-/- (0.1 M, pH 6.2), 50 µl H202 (0.01% in dH2O) and 50 µl DMB (1.25 mg/ml in dH2O). Samples were incubated at 37ºC for 15 min before the addition of 50 µl sodium azide (0.1% in dH2O) on ice to stop the reaction. Duplicate 200 µl aliquots per sample were transferred to a 96 well plate and absorbance read immediately at 450 nm using a spectrophotometer (Biorad). MPO activity was expressed as % of total MPO from lysed cells. 2.2.5 Protein detection methods 2.2.5.1 Sample preparation for gel electrophoresis Samples for SDS polyacrylamide gel electrophoresis (SDS-PAGE) were prepared from neutrophil lysates or supernatants. Neutrophil lysate protein concentration was determined by Bio-Rad DC colorimetric protein assay (compatible with up to 10% SDS), which is based on the reaction of protein with an alkaline copper tartrate solution, and the subsequent reduction of Folin reagent. The reduced species, proportionate to the protein concentration in the sample, have a characteristic blue colour which can be measured by reading absorbance. 20 µl sample was added to 25 µl copper tartrate solution in a 96 well plate, and then incubated with 200 µl Folin reagent for 15 min. Absorbance was measured at 595nm using a spectrophotometer (Biorad) and protein 71 concentration determined from a standard curve generated by known concentrations of BSA, enabling equal protein loading per well. Protein concentrations of supernatants were not determined prior to sample loading as the overall protein concentrations were low and it was expected that supernatants from hypoxic vs normoxic stimulated neutrophils would contain more protein due to enhanced degranulation. Also, there was no protein which was known to be consistent between normoxic and hypoxic conditions. Therefore, BSA (250 µg/ml) was added to each neat supernatant sample to allow correction for loading by Ponceau S staining following SDS-PAGE, with equal volumes from equal original cell numbers loaded onto gels. For some experiments, neutrophil supernatants were precipitated in 1.5 ml Eppendorfs with 10% TCA on ice for 1 h. Samples were then centrifuged (20,000g,15 min) and supernatants carefully removed with a P1000 Gilson pipette. Samples were washed with ice-cold acetone and centrifuged (20,000g, 10 min). Supernatants were carefully removed, the pellets were re- suspended in 20 μl 50 mM Tris-HCl, pH 8.4, and vortexed, and proteins were allowed to dissolve for 15 min. Reducing buffer (10 mM DTT, 250 mM Tris-HCl pH 6.8, 40% glycerol, 4% SDS, 0.02% bromophenol blue) was added to all lysate/supernatant samples (1:4). Samples were boiled with agitation (99ºC, 550 rpm, 10 min) and were then ready for SDS-PAGE. 2.2.5.2 SDS-PAGE 12 or 15% resolving gels and 4% stacking gels were cast with Protogel (30% (w/v) acrylamide: 0.8% (w/v) bisacrylamide) and polymerisation was catalysed by the addition of 0.1% ammonium persulfate (APS) and 0.1% tetramethylethylenediamine (TEMED). After the resolving gel had set, 4% stacking gel was poured above, with the insertion of 10 or 15 well combs to create wells. SDS-PAGE was performed using the PROTEAN® II electrophoresis system. For some experiments, 12 well pre-cast 12% polyacrylamide gels were used using the XCell SurelockTM Mini-Cell electrophoresis system. Samples (prepared as described in section 2.2.5.1) were loaded into wells. Using the PROTEAN® II system, samples were run through the stacking gel (100 V, 15 min) and subsequently through the resolving gel (150 V, 1 h) in running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS). When using the SurelockTM system, samples were run through the pre- cast gel (180 V, 40 min) in NuPAGE® MES running buffer (50 mM Tris base, 1 mM EDTA, 0.1% SDS, 50 mM 2-(N-morpholino)ethanesulfonic acid). One lane was used for molecular weight markers, according to the molecular weight of the protein studied or gel staining method to be used (either Precision Plus Protein™ Dual Color Standard or PagerulerTM Plus 72 Prestained Protein Ladder for western blot, or Mark12™ Unstained Standard for silver gels). Following electrophoresis gels were gently removed and either silver stained or proteins transferred to membranes for western blotting. 2.2.5.3 Western blotting All steps were performed at room temperature unless otherwise stated. Gels were transferred onto methanol soaked polyvinylidene fluoride (PVDF) membranes (75 V, 2 h) in ice-cold transfer buffer (24 mM Trizma base, 193 mM glycine, 10% methanol) using the Trans-blot system. After transfer, PVDF membranes were placed on a shaker (30-40 rpm) and washed twice in PBS-Tween (PBS-/- with 0.1% Tween-20, PBS-T) for 1 min, taking care to keep membranes moist at all times. For supernatant samples with BSA added as a standard, membranes were stained with Ponceau S for 20 min, washed once with dH2O for 1 min and then scanned, which enabled correction for sample loading by analysis of BSA band densitometry (66.5 kDa). Subsequently, all membranes were blocked in 5% BSA (for subsequent probing with pAKT and AKT antibodies) or 5% milk (for subsequent probing with all other primary antibodies) in PBS-T (30-40 rpm, 1 h). After blocking, the primary antibody (see Table 2.1 for dilutions) was incubated with membranes in either 2.5% BSA (for pAKT and AKT) or 2.5% milk (for all other primary antibodies) overnight (30-40 rpm, 4ºC). The following day, membranes were washed thrice (PBS-T, 10 min) and then incubated with HRP-conjugated goat anti-rabbit secondary antibody (1:5000) in either 2.5% BSA or 2.5% milk (the same agent used for the primary antibody, 30- 40 rpm, 1h). After incubation, membranes were washed thrice (PBS-T, 10 min), removed from the shaker and covered with ECL (Enhanced Chemiluminescence detection reagents 1 and 2 mixed 1:1) for 1 min to detect HRP-tagged proteins. Protein bands were detected by chemiluminescence, either by using a gel documentation (Gel Doc) automatic imaging system with GeneSys control software (Syngene, Cambridge, UK) or by exposure to auto radiographic film in the dark (1 s to 20 min exposure). For cell lysate samples, membranes were then incubated with stripping buffer (30-40 rpm, 15 min), re-blocked in 5% milk (30-40 rpm, 1 h) and probed for β actin (1:5000 in 2.5% milk, 30-40 rpm, 1 h) to confirm equal sample loading. Subsequent washes, secondary antibody probing and chemiluminescent detection was performed exactly as described above. Protein band densitometry analysis was performed using ImageJ software. Densitometry values for the protein of interest were obtained by dividing by densitometry values for the loading controls (BSA for supernatants, β actin for lysates). 73 2.2.5.4 Silver staining Following SDS-PAGE, gels were placed on a shaker (30-40 rpm) and washed twice (dH2O, 10 min). Gels were then fixed (30% ethanol, 10% acetic acid, 30 min), washed thrice (30% ethanol, 10 min) and sensitized in thiosulphate solution (370 µg/ml in dH2O, 1 min). Gels were washed twice (dH2O, 1 min) and silver stained (0.2% silver nitrate, 0.03% formaldehyde, 30 min) before washing once (dH2O, 1 min). Gels were developed until protein bands were visible (6% sodium carbonate decahydrate, 5 µg/ml thiosulphate solution, 0.02% formaldehyde) and the reaction terminated with the addition of 10% acetic acid. Finally, gels were washed in dH2O to remove background and scanned. 2.3 Endothelial Cells 2.3.1 Endothelial cell culture Primary human pulmonary arterial endothelial cells (HPAECs, Lonza) were used to investigate the potentially damaging effects of hypoxic neutrophils. Cell stocks were checked periodically for mycoplasma. Cell culture was conducted under sterile conditions in a laminar flow cell culture hood (Microflow Class II cabinet). All endothelial centrifugation was undertaken using a 5810R centrifuge (Eppendorf, Hamburg, Germany). HPAECs were cultured in 75 cm2 sterile cell culture flasks (T75) in EGM-2 medium with 0.1% gentamicin/amphotericin B and 10% foetal calf serum (FCS). Cell aliquots of 1*106 cells were removed from liquid nitrogen storage, defrosted and added to a flask with 15 ml EGM-2 medium supplemented with 10% FCS. After 24 h the medium was replaced to remove any DMSO present in the freezing solution. Cells were visually checked daily for signs of infection and medium was replaced every 48 h. Once cell monolayers had reached confluence, the medium was aspirated and the cell layer washed with 5 ml sterile PBS. 4 ml 1x trypsin was added to the flask and incubated at 37ºC for 3-5 min. Subsequently, flasks were gently tapped, and trypsinised cells viewed under an inverted microscope to ensure they had detached. 8 ml FCS-containing media were added to quench the trypsin reaction. Re-suspended cells were then transferred to a sterile 50 ml polypropylene Falcon tube and two 10 µl aliquots removed in order to count cells on a haemocytometer. Cells were centrifuged at 300g, acceleration 9, brake 9 for 5 min and the supernatant discarded. Pelleted cells were gently re-suspended and either transferred to sterile tissue culture plates for experiments, seeded into a new T75 flask in EGM-2 with 10% FCS to continue passaging, or frozen in freezing media (2*106 cells/ml) and stored in liquid nitrogen to maintain stocks. 74 2.3.2 Endothelial cell functional assays 2.3.2.1 Assessment of endothelial ICAM-1 expression by flow cytometry To assess the capacity of neutrophil supernatants to alter endothelial-leukocyte adhesion molecule expression, HPAEC expression of ICAM-1, which mediates firm adhesion of neutrophils to endothelial cells, was measured by flow cytometry. HPAECs were used at passage 10-12 to ensure consistency. 100,000 cells in 1 ml EGM-2 medium supplemented with 10% FCS were added to required wells of a sterile 12 well plate and grown to confluence (24-48 h) in either normoxia or hypoxia (0.8% O2). Once cell monolayers were confluent, media was aspirated and 1 ml supernatants from normoxic vs hypoxic, PAF&fMLP-treated vs control neutrophils, were added to required wells in the presence of 2% human autologous serum. The presence of serum is required to allow upregulation of ICAM-1 in this in vitro system (personal communication, Dr Sarah Appleby). HPAECs were incubated with neutrophil supernatants at 37ºC under normoxia or hypoxia for 24 h. After 24 h incubation, HPAECs were washed briefly in PBS-/- and 1x trypsin added. Cells were incubated for 5 min at 37ºC under normoxia or hypoxia as before. After incubation, the plate was gently tapped and the trypsin reaction quenched by addition of twice volume neutralising EGM-2 with 10% FCS. Cells were collected by gently pipetting up and down, and transferred to a 96 well plate. Cells were pelleted by centrifugation at 400g, acceleration 9, brake 9, 5 min (necessitating removal from hypoxia to access the centrifuge). Media was subsequently discarded and cells immediately fixed (1% paraformaldehyde (PFA) in PBS-/-, 10 min). After fixing, cells were washed in PBS-/- by centrifugation at 400g, acceleration 9, brake 9 for 5 min and then re-suspended in blocking buffer (5% FBS in 0.5% PBS-BSA, 10 min, 4ºC). Cells were stained with APC-anti-CD54 (anti-ICAM-1, Table 2.1), isotype control or PBS-/- (unstained). The plate was incubated for 30 min in the dark at 4ºC. After incubation, cells were washed in PBS-/- by centrifugation at 400g, acceleration 9, brake 9 for 5 min. Supernatants were discarded, pelleted cells re-suspended in staining buffer (0.09% sodium azide in 1% FCS-PBS) and stored in the dark at 4ºC until flow cytometric analysis was carried out. ICAM-1 expression was measured by flow cytometry (BD FACSCanto II) using the red (633 nm excitation; 660/20 filter) laser. Cells were gated on singlet HPAECs, and change in MFI was measured. APC-stained ICAM-1 positive cells, identified by comparison with the gated isotype control population, were expressed as % of the total singlet HPAEC population. Results were analysed using FlowJo version 10 software. 75 2.3.3 Endothelial cell viability assays 2.3.3.1 Assessment of cell detachment by immunofluorescence To assess the ability of neutrophil supernatants to cause endothelial cell detachment, HPAECs were stained with rhodamine-phalloidin (binds F-actin) and DAPI (4',6-diamidine-2'- phenylindole dihydrochloride: stains DNA in the nucleus). HPAECs were used at passage 10- 12 for consistency. 30,000 cells in 200 µl EGM-2 with 10% FCS were added to required wells of a sterile 96 well optical plate and grown to confluence (24 h). In order to be able to assess protease-mediated damage and the potential for rescue by α1AT, this experiment was conducted in serum-free conditions as serum contains multiple anti-proteases. Hence, neutrophil supernatants were diluted in endothelial specific media to prevent damage due to lack of growth factors over a 24 h incubation period. 100 µl supernatants from normoxic vs hypoxic, PAF&fMLP-treated vs control neutrophils were incubated in the presence or absence of α1AT (46 µg/ml, 10 min) and diluted 1:1 in serum-free EGM-2. Media was aspirated from HPAECs and 200 µl neutrophil supernatants (prepared as above) were added to wells in duplicate and the plate incubated at 37ºC. After 24 h the samples/media were aspirated and cells fixed in 3.5% PFA in PBS-/- (20 min). Cells were washed briefly thrice in PBS-/- and permeabilised in 0.5% Triton X-100 in PBS+/+ (10 min). Cells were washed thrice in PBS-/- and non-specific staining was blocked by 0.5% BSA-PBS (30 min). Cells were washed thrice in PBS-/- and cells incubated with rhodamine phalloidin (1:200 in 0.1% BSA-PBS, 30 min). Cells were washed thrice with PBS-/- and mounted in ProLong Gold antifade reagent with DAPI. Plates were kept in the dark at 4ºC until imaging with a Leica Sp5 confocal microscope. Three random fields of view were imaged in a central position for duplicate wells. Cell detachment was quantified with Image J and expressed as % detachment of whole field of view. 2.3.3.2 Assessment of cell viability by MTT assay The MTT toxicity assay allows the spectrophotometric quantification of cell viability. The tetrazolium salt 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) is reduced to purple formazan by NADPH-dependent oxidoreductase enzymes in living but not dead cells. HPAECs were used at passage 10-12 for consistency. For the same reasons as above, this experiment was conducted in serum-free conditions with neutrophil supernatants diluted in EGM-2 media. 30,000 cells in 200 µl EGM-2 with 10% FCS were added to required wells of a sterile 96 well plate and grown to confluence (24 h). 100 µl supernatants from normoxic vs hypoxic, PAF&fMLP-treated vs control neutrophils were incubated in the presence or absence of α1AT (46 µg/ml, 10 min) and diluted 1:1 in serum-free EGM-2. Media was aspirated from HPAECs and 200 µl neutrophil supernatants/EGM-2 were added to wells in duplicate. EGM- 76 2/IMDM (1:1), EGM-2 alone, or IMDM alone were added to control wells and the plate incubated at 37ºC. After 24 or 48 h, samples/media were aspirated and 100 µl MTT (500 µg/ml in IMDM) per well added for 2 h. Subsequently, the MTT solution was aspirated and purple formazan crystals were dissolved in 100 µl isopropanol. Absorbance was read immediately at 550 nm using a spectrophotometer (Biorad), correlating directly with the number of viable cells per well. 2.3.3.3 Assessment of apoptosis by flow cytometry The ability of neutrophil supernatants to cause endothelial cell apoptosis was measured by flow cytometric analysis of AnV and PI staining. For this experiment, HPAECs were used at passage 7 as the previous endothelial cell stocks had been exhausted and Promocell media was used instead of EGM-2 as stocks from Lonza were unavailable due to manufacturing issues. 100,000 cells in 1 ml Promocell media with 5% FCS were added to required wells of a sterile 12 well plate and grown to confluence. As this assay is very sensitive, conditions were optimised to avoid induction of apoptosis by cell manipulation: cells were quiesced overnight to synchronise the cell cycle, thereby removing variability which might arise from growth signalling in different cell cycle phases, and TrypLE select enzyme was used to detach cells as it is reported to be more gentle than trypsin (personal communication, Dr Andrew Cowburn). Prior to treatment, HPAEC were quiesced overnight in media containing antibiotics and 0.1% FCS but no growth factors. Media was aspirated from HPAECs and 800 µl supernatants from normoxic vs hypoxic, PAF&fMLP-treated vs control neutrophils were added to wells. Promocell media alone (where the majority of cells were expected to be viable) or Promocell media with TNFα (10 ng/ml) and cycloheximide (20 µg/ml, CHX) (where the majority of cells were expected to be apoptotic) were added to control wells and the plate incubated at 37ºC for 6 h. After treatment, supernatants/media were removed into sterile 50 ml Falcon tubes and cells in the 12 well plate were detached with 1x TrypLE select enzyme before adding to the corresponding Falcon. Wells were washed with Promocell media and added to corresponding Falcons. Cells were pelleted by centrifugation (300g, 5 min), media aspirated and cells re- suspended in 1x binding buffer supplied with the annexin V-FITC apoptosis detection kit (1*106 cells/ml). 50 µl cells from each sample were transferred into a separate well to be used for compensation. 100 µl sample was stained with FITC-AnV (1:20) and PI (1:20) for 15 min in the dark before quenching with 200 µl 1x binding buffer. FITC-AnV or PI single stained, and unstained cells were used for baseline and compensation. Flow cytometry (BD LSRFortessa) was performed using the blue (488 nm excitation; 530/30 filter) and yellow (561 nm excitation; 610/20 filter) lasers. Gating and analysis strategies are shown in Figure 2.11 and Figure 2.12. 77 Figure 2.11: Gating strategy for assessment of HPAEC apoptosis by flow cytometry: baseline and compensation HPAEC were incubated for 6 h with either supernatants from normoxic vs hypoxic, PAF/fMLP- treated vs unstimulated neutrophils, Promocell media alone, or Promocell media containing TNFα (10 ng/ml) and CHX (20 µg/ml). Mixed HPAEC from all conditions were A: unstained or B: single stained with FITC-AnV. HPAEC treated with C: Promocell media alone or D: Promocell media with TNFα and CHX were stained with FITC-AnV and PI. Staining was assessed by flow cytometry (BD LSRFortessa). A-D: Representative flow plots of FITC-AnV (y axis) and PI (x axis) staining. A&B: unstained and single-stained cells were used to set gates. Using this gating strategy, the majority of cells treated with Promocell media alone (C) were AnV-/PI- (viable) and the majority of cells treated with Promocell media containing TNFα and CHX (D) were AnV+/PI- (apoptotic). Cells per quadrant as % of total HPAEC are indicated. 78 Figure 2.12: Gating strategy for assessment of HPAEC apoptosis by flow cytometry: supernatant treatment HPAEC were incubated with supernatants from normoxic (A&C) or hypoxic (B&D) neutrophils treated with PAF (1 µM) and fMLP (100 nM, C&D) or IMDM control (A&B). After 6 h, HPAEC were stained with FITC-AnV and PI and apoptosis assessed by flow cytometry (BD LSRFortessa). A-D: Representative flow plots of FITC-AnV (y axis) and PI (x axis) staining. Q1 shows AnV+/PI- (apoptotic) cells, Q2 shows AnV+/PI+ (late apoptotic/necrotic) cells, Q3 shows AnV-/PI+ (necrotic) cells (though most necrotic cells are AnV+), Q4 shows AnV-/PI- (viable) cells. Cells per quadrant as % of total HPAEC are indicated. 79 2.4 Proteomics 2.4.1 Generation of neutrophil supernatants for proteomics In order to perform a screen of the proteins released by normoxic vs hypoxic neutrophils, supernatants were generated to quantitatively compare the normoxic vs hypoxic proteomes. Neutrophil supernatants contain several active proteases with the ability to cleave other proteins present in the sample. Hence, I carefully and thoroughly optimised the anti-protease strategy to ensure that the supernatants contained sufficient protein for proteomic analysis and that other proteins of interest were not degraded within this experimental system. Previous studies employing proteomic approaches to examine neutrophil supernatants have either used no (documented) protease inhibitors (305,306), EGTA alone (307,308), or a combination of leupeptin, 4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF), pepstatin A and EDTA (309). As my anti-protease protocol is therefore completely novel, this optimisation is described in section 4.3 of the results. The final optimised methodology, used to prepare samples for 10-plex TMT-MS, is described here. Freshly isolated neutrophils were re-suspended at a concentration of 11.1*106/ml in normoxic or hypoxic IMDM containing either a combination of EDTA (1 mM) and the NE-selective inhibitor sivelestat (10 µM), or no protease inhibitors (in order to be able to test supernatant NE activity in parallel). Aliquots of 270 µl (3*106 neutrophils) were transferred to 2 ml Eppendorfs and incubated under normoxia or hypoxia in a thermomixer (37ºC, 450 rpm) for 4 h. Cells were then treated with PAF (1 µM, 5 min) and subsequently fMLP (100 nM, 10 min), or IMDM control, with a final reaction volume of 1 ml. Both PAF and fMLP were diluted in IMDM with or without protease inhibitors (EDTA and sivelestat), as indicated by the previous incubation, to ensure that final concentrations of the inhibitors were constant throughout. Cells were pelleted (13,000 rpm, 10 s, MSE Micro Centaur microfuge) and the supernatants carefully harvested. 2.4.2 Two dimensional difference gel electrophoresis Two dimensional difference gel electrophoresis (2D DIGE) was performed as a pilot experiment to inform plans for mass spectrometry. This technique separates proteins in the first dimension by pH, using isoelectric focusing (IEF), and subsequently in the second dimension by molecular weight, using SDS-PAGE. Gel fluorescence imaging allows detection of changes in protein abundance between fluorescently-labelled samples. Supernatants from normoxic vs hypoxic PAF and fMLP-treated neutrophils were generated as described in section 2.4.1, although for this experiment all cells were re-suspended in IMDM containing the protease inhibitors EDTA (2 mM), EGTA (2 mM), sivelestat (1 µM) and AEBSF 80 (1 mM), and inhibitors were also added at the specified concentrations after 2 h and 4 h incubation. AEBSF was omitted in later experiments as it was found to induce cell death (see section 4.3, Figure 4.5). Samples were precipitated overnight with 5x volume ammonium acetate (0.1 M in 100% methanol, -80ºC) as TCA-precipitation was not compatible with 2D DIGE. Following precipitation, samples were transferred to 1.5 ml Eppendorfs, centrifuged (20,000g, 10 min, room temperature) and the supernatant discarded. Pellets were washed twice with 80% 0.1 M ammonium acetate (20,000g, 2 min) and then washed once with acetone (20,000g, 2 min). The supernatant was discarded and the pellets air-dried for 5 min. Protein pellets were re-suspended in urea buffer (6 M urea, 2 M thiourea, 4% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulfonate (CHAPS), 5 mM magnesium acetate, 10 mM Tris pH 8.5). Pellets from 24 technical replicates per condition were combined to ensure sufficient protein for labelling. Samples are described in this section as “normoxia activated” or “hypoxia activated”, referring to oxygenation status and PAF/fMLP-treatment of the cells. 2D DIGE was performed by Renata Feret (Gel Analysis Technician, Cambridge Centre for Proteomics (CCP)). The protein content of each sample was quantified by Quick Start™ Bradford Protein Assay, which is compatible with the urea buffer. Equal amounts (20 µg) of both samples were labelled with cyanine fluorescent dyes as any disparity in protein amount between samples may result in less abundant proteins not being labelled. “Normoxia activated” was labelled with Cy5 (250 pM/µl), and “hypoxia activated” with Cy3 (250 pM/µl) for 30 min in the dark before the reaction was quenched with lysine (10 mM) for 10 min in the dark. Both samples were combined and sample buffer (8 M urea, 4% CHAPS, 2% DTT, 2% immobilised pH gradient (IPG) buffer [pH 3-10 non linear (NL)]) added for 15 min in the dark. For IEF, the combined sample was applied to a rehydration tray in rehydration solution (8 M urea, 4% CHAPS, 0.2% DTT, 1% IPG buffer [pH 3-10 NL], 0.002% bromophenol blue) and the IEF gel strip (Immobiline™ Drystrip pH 3-10 NL, 24 cm) placed on top to absorb the sample. The strip was covered in mineral oil overnight to prevent drying. The following day, the strip was then placed in a PROTEAN® i12™ IEF system with running parameters as follows: 50 V, 12 h; 500 V 1 h; 1000 V 1 h; 8000 V, 6 h. Following IEF, the strip was removed, incubated with reducing equilibration buffer (6 M urea, 75 mM Tris pH 8.8, 30% glycerol, 2% SDS, 1% DTT) for 15 min and then incubated with alkylating equilibration buffer (6 M urea, 75 mM Tris pH 8.8, 30% glycerol, 2% SDS, 2.5% iodoacetamide) for 15 min, before being subjected to SDS-PAGE using a 600 Ruby Standard Dual-Cooled Vertical Gel Unit with running parameters as follows: 20 mA, 15 min; 40 mA until samples reach the bottom of the gel. The gel was removed and scanned with a Typhoon 9400 Variable Mode Imager. Fluorescence parameters were: Cy5 (normoxia activated) 633 nm laser excitation (red), 670 nm emission; Cy3 (hypoxia activated) 532 nm laser excitation (green), 580 nm emission. 81 Scanned images were processed for optimal pixel intensity (ImageQuant TL software, GE Healthcare) and analysed with DeCyder 2D software (version 5.2, GE Healthcare). The Differential In-gel Analysis (DIA) module allows protein spot detection and quantitiation of images produced from the same gel. Protein spots were automatically detected by the software and manually checked for appearance and correct matching. The fold change threshold between spots, i.e. the magnitude of the spot volume ratio between the two gels, was set at 2.0. Figure 2.13 shows an example of manual spot selection where a protein is present in only one image. Figure 2.13: 2D DIGE analysis by Decyder 2D DIA software Cy3 and Cy5 fluorescent images from the scanned 2D DIGE were analysed by Decyder 2D DIA. Protein spots were automatically selected by the software, indicated by green (unchanged), blue (increased) or red (decreased) boundaries. The protein spot selected (yellow) is shown below the gel as a localised three-dimensional representation. Here, a peak at the selected spot is present in the left-hand side gel (Cy3, hypoxia activated) but not in the right-hand side gel (Cy5, normoxia activated). Representative image from n=1. 82 2.4.3 10-plex tandem mass tag-labelled mass spectrometry Based on the results of 2D DIGE, which showed a number of obvious changes in protein abundance between normoxia and hypoxia activated samples, I proceeded to 10-plex tandem mass tag (TMT)-labelled mass spectrometry (MS), which enables protein identification and quantification. In this process up to ten protein samples can be labelled with different isotopic mass tags. These tags have an identical chemical structure and nominal mass but a variable substituted isotope, which allows quantitative data to be obtained: the peptide fragmentation process generates mass reporter ions with a mono-isotopic mass difference from the cleaved tags, and each tag is unique for one labelled sample, therefore allowing quantification of relative protein expression (310). Prior to performing the planned TMT-labelled MS experiment for n=10 samples, I undertook a pilot experiment to inform the feasibility of TMT-labelling. One supernatant from hypoxic PAF and fMLP-treated neutrophils was generated in the presence of EDTA and sivelestat, exactly as described in section 2.4.1, and TCA-precipitated, exactly as described in section 2.2.5.1. Sample processing was performed by Renata Feret (CCP). Protein content was quantified by Pierce™ 660nm protein assay, which is compatible with phenol red (present in IMDM media) up to 0.5 mg/ml. The total sample (4.7 µg) was reduced with the addition of tris-(2- carboxyethyl)phosphine (TCEP, 10 mM, 1 h, 55ºC). Free cysteine residues were alkylated with iodoacetamide (17 mM) for 30 min in the dark before samples were precipitated overnight with acetone (600 µl, -20ºC). Following precipitation, the sample was centrifuged (8000g, 10 min, 4ºC) and the supernatant carefully discarded. The protein pellet was re-suspended in triethyl ammonium bicarbonate (TEAB, 50 mM) and digested into peptides overnight with MS- grade trypsin (2.5 µg, 37ºC). The peptide sample was then reduced to dryness with a vacuum concentrator and subjected to unlabelled MS/MS (120 min), performed by Dr Mike Deery (Proteomics Facility Manager, CCP). MS data were analysed by Dr Marco Chiapello (Research Data Technician, CCP). Raw data were searched against UniProt Human reference proteome (Proteome ID: UP000005640) using Mascot 2.6 (Matrix Science) and Proteome Discoverer™ version 2.1 (ThermoFisher Scientific). Due to the relatively low complexity of the sample, with a total of 561 proteins identified, TMT-labelling was thought to be achievable with a minimum of 25 µg protein per sample. Whole blood was collected from n=5 healthy volunteers in order to prepare ten samples for 10-plex TMT-labelled MS (normoxia and hypoxia activated from each donor). Following experiments to optimise the protease inhibitor strategy and other conditions (section 4.3), supernatants from normoxic vs hypoxic PAF and fMLP-treated neutrophils were generated exactly as described in section 2.4.1. Supernatants from 23 replicates per condition per donor 83 were pooled in order to generate sufficient protein for TMT-labelling. For each experiment, one replicate was incubated without protease inhibitors and assessed for NE release by Enzchek® assay to ensure a hypoxic uplift of NE release. Supernatants from the remaining 23 replicates per condition were combined for each donor. As protein precipitation (see section 4.3) was found to lead to inaccurate/variable protein recovery, protein concentration Vivaspin columns were used. The pooled supernatant sample was applied to the top chamber of the spin column, after which centrifugation forced the solvent through a semi-permeable membrane into a lower chamber, leaving the concentrated protein sample in the upper chamber. Pooled samples were sequentially centrifuged (4000g, 4ºC) through 2 ml spin columns with a molecular weight cut-off of 3 kDa. Each sample was buffer exchanged with 50 mM Tris, pH 8.4 (1:1 sample:buffer). Once the total buffer-exchanged 23 ml sample had been centrifuged through the spin column, the concentrated protein was re-suspended in 100 µl buffer (50 mM Tris, pH 8.4). Samples were then processed by Renata Feret (CCP). Protein content was quantified by Pierce™ 660nm protein assay. To ensure equal labelling, 27.68 µg (the amount of the least abundant sample) of each sample was prepared for labelling with TMT-10-plex isobaric label reagents. Samples were adjusted to 100 µl in dissolution buffer (100 mM TEAB) and then reduced, alkylated, precipitated and trypsin-digested exactly as described above for the pilot experiment. Following digestion, each of the ten peptide samples was labelled with a unique 10-plex isobaric tag (4 µl, 1 h) as follows: normoxia (N) donor 1, tag 130N; hypoxia (H) donor 1, tag 126; N2, tag 127C; H2, tag 131; N3, tag 129C; H3, tag 128N; N4, tag 129N; H4, tag 130C; N5, tag 128C; H5, tag 127N. Labelling was quenched with 5% hydroxylamine for 15 min and the samples were then combined and reduced to dryness with a vacuum concentrator. In order to remove the salts and buffers used during preparation, which may interfere with ionisation, the dried sample was subjected to C18 solid-phase extraction using Sep-Pak cartridges, eluting with 75% acetonitrile/0.5% acetic acid, before further vacuum concentration to dryness. Given the relatively low number of proteins per sample (as assessed by previous gel electrophoresis and the pilot unlabelled MS/MS), fractionation was not performed. MS/MS was performed by Dr Mike Deery (CCP) using an Orbitrap Fusion™ Lumos™ Tribrid™ mass spectrometer (ThermoFisher Scientific). MS data were analysed by Dr Marco Chiapello (CCP). TMT-MS and data analysis methods and parameters were provided by CCP, and are described in Appendix 7.1. Statistical analysis was performed by Dr Marco Chiapello using R software. 84 2.4.4 Quantification of supernatant resistin, NGAL, cyclophilin A, S100A9, S100A8 and S100A8/A9 by ELISA Resistin, NGAL, cyclophilin A, S100A9 homodimer, S100A8 homodimer, and S100A8/A9 heterodimer content of supernatants from normoxic vs hypoxic PAF&fMLP-treated neutrophils were quantified by commercial sandwich-based ELISA with pre-coated plates. Preliminary experiments determined the appropriate dilutions to ensure readings fell within the standard curve. To serve as an example method, the protocol for resistin will be detailed. All steps were undertaken at room temperature. The supplied plate was pre-coated with anti-human resistin capture antibody; samples (100 μl, 1:10 dilution in assay diluent supplied with the kit) and standards (1.4-400 pg/ml, diluted in assay diluent supplied with the kit) were incubated in a 96-well plate for 2.5 h with gentle shaking. Wells were then washed four times (wash buffer supplied with kit), and incubated with biotinylated resistin detection antibody (100 μl) for 1 h. After four washes, HRP-conjugated streptavidin (100 μl), which binds biotin with high affinity, was added for 45 min. After four washes, 3,3’,5,5’-tetramethylbenzidine (TMB) substrate solution (100 μl) was added and the plate incubated for 30 min in the dark. Wells containing resistin turned blue with a colour intensity proportional to the resistin concentration in the original sample. The reaction was terminated by the addition of a stop solution, which turned the blue wells yellow, and the absorbance at 450 nm immediately measured with a spectrophotometer. A standard curve was generated by plotting absorbance against the corresponding concentrations of human resistin standard. The concentration of resistin present in the samples (which were run in duplicate concurrently with the standards) was determined from the standard curve by interpolation. ELISAs for NGAL, cyclophilin A, S100A8 homodimer, S100A9 homodimer and S100A8/A9 heterodimer were all based on the same sandwich principle and steps (similar to those described above) were conducted exactly as per manufacturers’ instructions. 2.5 Microvesicles 2.5.1 Microvesicle isolation from Histopaque®-1077-prepared neutrophils The study of NDMVs is challenging and requires dedicated detection systems that were not readily available in the University of Cambridge Department of Medicine. Hence, I collaborated with Dr Victoria Ridger (Senior Lecturer in Vascular Biology, University of Sheffield), who has considerable expertise in MV isolation and detection techniques (311), and an established protocol for NDMV isolation currently ongoing in her laboratory (276). Previous optimisation of this protocol has shown that optimal NDMV yield was obtained from neutrophils isolated by Histopaque®-1077 gradients. Therefore, as preliminary experiments attempting to isolate 85 NDMVs from plasma-Percoll®-isolated neutrophils had produced low numbers, I employed the Histopaque®-1077 neutrophil isolation method in order to maximise NDMV yield. NDMV isolation experiments were performed at the University of Sheffield. Ethical permission for taking peripheral blood from healthy volunteers (REC reference SMBRER310) was obtained from the Sheffield Research Ethics Committee. All equipment and reagents were sterile, and all reagents used at room temperature unless otherwise stated. The isolation was conducted under sterile conditions in a laminar flow cell culture hood (Microflow Class II cabinet). All neutrophil centrifugation was undertaken using a 5810R centrifuge (Eppendorf, Hamburg, Germany). For neutrophil isolation from whole blood by Histopaque®-1077 gradients, blood was collected from healthy donors into polypropylene tubes containing 3.8% sodium citrate, as described in section 2.2. Whole blood was centrifuged at 260g, acceleration 5, brake 0 for 20 min. The PRP supernatant was aspirated and discarded. The cell pellet, containing WBC and RBC, was sedimented with dextran to remove RBC: 6 ml 6% dextran was added, made up to 50 ml total volume with sterile 0.9% NaCl and mixed gently. RBC were sedimented for 20-40 min until the RBC pellet reached its original (post centrifugation) volume. After sedimentation, the WBC- containing top layer was aspirated into a new 50 ml tube, made up to 35 ml with sterile saline and very gently overlaid onto 16 ml Histopaque®-1077 (density 1.077 g/ml) using a 10 ml stripette with the pipette controller set to the lowest speed. The WBC/Histopaque®-1077 gradient was centrifuged at 400g, acceleration 5, brake 0 for 25 min. After centrifugation the supernatant, containing PBMCs, was aspirated and discarded. The pellet, containing granulocytes and red blood cells, was subjected to RBC lysis: the pellet was gently re- suspended, made up to 20 ml with 0.2% NaCl solution, mixed gently for 30 s, made up to 50 ml with 1.6% NaCl solution and inverted once. Following RBC lysis, the cells were centrifuged at 250g, acceleration 5, break 3 for 7 min. The supernatant was discarded and the neutrophil- rich cell pellet gently re-suspended. Cells were washed in 50 ml PBS-/- by centrifuging at 250g, acceleration 5, break 3 for 7 min. Prior to centrifugation,100 µl were removed by P200 Gilson using a cell saver pipette tip to make the purity slide cytospin. 20 µl were removed for haemocytometer cell count. The cells were then washed in 50 ml PBS+/+ by centrifuging at 250g, acceleration 5, break 3 for 7 min. The supernatant was discarded and the cell pellet gently re-suspended to be used in further experiments. To isolate NDMVs, fresh Histopaque®-1077-prepared neutrophils and, for comparison, plasma-Percoll®-prepared neutrophils (isolated concurrently from the same donor in exactly the same manner) were re-suspended in normoxic or hypoxic IMDM at a concentration of 1*107/ml. Aliquots of 1 ml (1*107 neutrophils) were transferred to 2 ml Eppendorfs and 86 incubated under normoxia or hypoxia in a thermomixer (37ºC, 450 rpm) for 4 h. As it was otherwise difficult to obtain sufficient NDMVs for quantification, the standard neutrophil treatment condition used in the Ridger laboratory (10 μM fMLP, 1 h) was adopted. Hence, after 3 h normoxic or hypoxic incubation, neutrophils were treated with fMLP (10 µM, 1 h), or IMDM as control. After treatment, samples were centrifuged at 500g, 4ºC for 5 min to pellet cells. Cell pellets were discarded. Cell-free supernatants were harvested and further centrifuged at 500g, 4ºC for 5 min. Any remaining pellets were discarded and supernatants were carefully harvested and centrifuged at 20,000g, 4ºC for 30 min to pellet MVs. MV-free supernatants were harvested, and both supernatants and MV pellets were either used straight away or stored at -20ºC for future experiments (as storage at -20ºC has been shown by the Ridger group not to have a detrimental effect on NDMV integrity). 2.5.2 Quantification of neutrophil-derived microvesicles by flow cytometry Quantification of isolated NDMVs by flow cytometry was undertaken at the University of Sheffield with the assistance of Merete Long (PhD Student, Laboratory of Dr Ridger, University of Sheffield). MV pellets were re-suspended in 190 µl sterile filtered PBS-/- with 10 µl (10,000) counting beads (SPHERO™ Accucount blank particles 2.0-2.4µm, 1*106/ml). Flow cytometry was performed (BD LSRFortessa) using an established protocol and gating strategy: the NDMV analysis region was standardised using Megamix beads, a mix of fluorescent beads of varied diameters (0.5, 1 and 3 µm), which allowed the MV collection gate to be set (Figure 2.14A). This gate was used for all NDMV experiments and is regularly calibrated. The addition of counting beads to the samples allowed quantification of NDMVs (Figure 2.14B). The stopping gate was set to count 1,000 beads (i.e. one tenth of the total beads in the sample). The number of NDMVs counted by flow cytometry was multiplied by 10 to give the total number of MVs in the pellet (i.e. the number of NDMVs produced by 107 neutrophils). 2.5.3 Preparation of microvesicle lysates for western blotting NDMV pellets were transferred from -20ºC storage to dry ice, and 25 µl lysis buffer added (250 mM Tris-HCl pH 6.8, 20% glycerol, 4% SDS and one cOmplete™ mini EDTA-free protease inhibitor cocktail tablet, 1 tablet/4 ml buffer). Samples were warmed, vortexed (30 s), boiled with agitation (99ºC, 550 rpm, 10 min) and returned to dry ice. Reducing buffer (10 mM DTT, 100 mM Tris-HCl pH 6.8, 40% glycerol, 4% SDS, 0.02% bromophenol blue) was added to re- warmed samples (1:4), which were subsequently boiled with agitation (99ºC, 550 rpm, 10 min). SDS-PAGE was performed exactly as described in section 2.2.5.2 and western blotting performed exactly as described in section 2.2.5.3, with the antibodies noted in Table 2.1. For some experiments TCA-precipitated (see section 2.2.5.1) MV-free neutrophil supernatants were also subjected to SDS-PAGE and western blotting. 87 Figure 2.14: Gating strategy for isolated NDMV quantification by flow cytometry MVs isolated from fMLP (10 μM)-treated normoxic vs hypoxic neutrophils were quantified by flow cytometry (BD LSRFortessa). A: Representative flow cytometry plots indicating the NDMV gate set by Megamix size beads of 0.5, 1 and 3 μm. B: Representative flow cytometry plots showing gates for isolated NDMVs and counting beads 2.6 COPD study 2.6.1 COPD patient recruitment Ethical approval for this study, involving identifying and taking blood from COPD patients during exacerbations, had already been obtained from the Cambridge Local Research Ethics Committee (REC reference 08/H0308/281) and was not designed specifically for this study. Patients were recruited from an existing ongoing cohort. Exacerbating COPD patients were recruited from Cambridge University Hospitals NHS Foundation Trust (Addenbrooke’s Hospital) by myself, using electronic patient medical records (Epic software). Inclusion criteria were: age < 80 years, a physician diagnosis of COPD, a minimum of 10 pack-years smoking history, and a current diagnosis of an exacerbation of COPD (infective or non-infective) by the 88 admitting medical team. Exclusion criteria were: concurrent non-respiratory infection, immunodeficiency, anaemia at the time of blood sampling, untreated malignancy, current smoker > 5 cigarettes per day, long term oral corticosteroid use, and consolidation on the admission chest radiograph. Pragmatically, the decision was taken to include patients who had started treatment with antibiotics or oral corticosteroids as initial attempts to exclude such patients resulted in very low recruitment. Following recruitment, blood was drawn within 24 h of admission. Peripheral venous blood was taken from all patients to obtain plasma (9 ml) and serum (4.9 ml), as described in section 2.6.2. From a subset of patients, 40 ml blood was also taken for neutrophil isolation, as described in section 2.2. Demographic and clinical data were collected (Table 5.1). Age- (within 10 years) and sex-matched healthy volunteers were recruited for identical venous blood sampling. However, due to the unpredictable recruitment of COPD patients, older age and short timeframe for venepuncture (less than 24 h), it was not possible to obtain plasma/serum or perform neutrophil isolation from healthy volunteers in parallel at the same time. 2.6.2 Obtaining plasma and serum from whole blood All plasma/serum centrifugation was undertaken using a 5810R centrifuge (Eppendorf, Hamburg, Germany). Peripheral venous blood was collected from healthy volunteers or COPD patients. To obtain plasma, whole blood was collected into sterile 9 ml S-Monovette® venous blood sampling collection tubes containing EDTA. To obtain serum, whole blood was collected into sterile 4.9 ml S-Monovette® venous blood sampling collection tubes containing polystyrene beads coated with silicate clotting activator. Blood collection tubes were centrifuged (3000g, acceleration 9, brake 9, 10 min) and plasma/serum aspirated into 2 ml Eppendorfs for storage in 500 µl aliquots at -80ºC. Plasma samples used for MV analysis were frozen at -20ºC. 2.6.3 Quantification of Aα-Val360 and Aα-Val541 Direct measurement of plasma or serum protease, e.g. NE, activity is difficult as the released enzyme is rapidly bound to its substrate or to serum protease inhibitors, such as α1AT. NE ELISAs do not always distinguish between free and bound enzyme, and enzyme-inhibitor complex assays only detect inactivated NE. An alternative approach was therefore employed, using a patented assay (US patent No. 6124107) to detect a NE-specific fibrinogen cleavage product in plasma, providing a “footprint” of NE activity. NE cleaves fibrinogen at multiple sites; cleavage at the Aα(Val360-Ser361) site generates a small C-terminal fragment and hence exposes an Aα-Val360 neo-epitope, which is associated with the stable large protein fragment. The assay is a sandwich-based ELISA of Aα-Val360, utilising europium-conjugated IgG as the readout. A similar unpatented assay was employed to measure a specific fibrinogen 89 breakdown product of the protease, PR3, which exposes an Aα-Val541 neo-epitope. Access to the Aα-Val360 and Aα-Val541 assays was kindly allowed by Professor Robert Stockley (Director of the Lung Immunobiochemical Research Group, University of Birmingham), and the assays were performed by Paul Newby (Respiratory Sciences Laboratory Manager, University of Birmingham). Plasma content of Aα-Val360 and Aα-Val541 was measured in plasma samples from healthy volunteers and COPD patients. 2.6.4 Quantification of plasma microvesicles by flow cytometry Plasma was diluted 1:5 in sterile filtered PBS and samples were incubated with an antibody mix (1:50 final concentration) of BV421-anti-CD66b, PE-anti-CD41a, APC-anti-CD14, PerCP Cy5/5-anti-CD144 and (Table 2.1) on ice for 45 min on a rocker. Samples were centrifuged (20,000g, 30 min) to pellet MVs and supernatants were carefully removed. MV pellets were re-suspended in 490 µl sterile filtered PBS with 10 µl counting beads. The MV gate was set using standardised size calibration beads (Figure 2.15A), and 1,000 counting beads were measured per sample. Fluorescence minus one negative controls (i.e. four samples each containing a combination of three fluorochromes without the fluorochrome that was being measured) and unstained controls were used to set up compensation. Flow cytometric analysis (BD LSRFortessa) of the plasma content of CD66b+ (neutrophil parent), CD41a+ (platelet parent), CD14+ (monocyte parent), and CD144+ (endothelial cell parent) MVs using the violet (355 nm excitation; 450/50 filter), blue (488 nm excitation; 575/26 and 695/40 filters) and red (633 nm excitation; 780/60 filter) lasers was performed with the assistance of Ben Ward and Merete Long (University of Sheffield) using previously optimised protocols. The gating strategy is shown in Figure 2.15. 2.7 Statistics Results are reported as mean ± standard error of the mean (SEM) from (n) independent experiments. Statistical analysis was performed on data generated from n≥3 experiments using GraphPad Prism version 7 software. A p value of less than 0.05 was considered significant. For experiments comparing two groups, normally distributed data were analysed by paired or unpaired t-test, and non-normally distributed unpaired data were analysed by Mann-Whitney test. Normality was assessed by Shapiro-Wilk normality test if n<8 or D’Agostino-Pearson normality test if n≥8. For experiments comparing more than two groups, data were analysed by one way analysis of variance (ANOVA) with Tukey’s post hoc correction for multiple comparisons. For experiments comparing two variables, e.g. treatment (stimulated vs unstimulated) and oxygenation (normoxia vs hypoxia), data were analysed by two way ANOVA with Sidak’s post hoc correction for multiple comparisons. For data generated from 10-plex TMT-MS, principal component analysis (PCA) and paired t-test analysis were 90 performed by Dr Marco Chiapello; p values were adjusted by the Benjamini-Hochberg false discovery rate correction for multiple comparisons (see section 7.1.2). An estimate of the required sample size to detect a significant difference between groups for the COPD study was calculated using the NE activity data for PAF and fMLP-treated neutrophils from healthy volunteers. Based on the sample size calculation: n > [1+(1/κ)] [[(Zα+Zβ)/δ]σ]2 where κ = ratio between the two group sample sizes i.e. 1, Zα = 1.96 and Zβ = 0.84, σ = common standard deviation, and δ = difference between the two group means, a sample size of 16 would detect a fold change of 2 with 80% power (power = 1-β, β = 0.2) and 2-tailed 95% confidence (α = 0.05). Figure 2.15: Gating strategy for plasma microvesicle quantification by flow cytometry Plasma from COPD patients and age/sex-matched healthy controls was stained for BV421- anti-CD66b, PE-anti-CD41a, APC-anti-CD14, and PerCP Cy5/5-anti-CD144 and plasma MVs quantified by flow cytometry (BD LSRFortessa). A: Representative flow cytometry plots showing gates for plasma MVs and counting beads. B: Representative flow cytometry plots showing gating strategy for plasma CD66b+ (neutrophil), CD41a+ (platelet), CD14+ (monocyte) and CD144+ (endothelial cell) MVs. Chapter 3 Results: Effect of hypoxia on NE release and neutrophil-mediated endothelial dysfunction 92 3 Effect of hypoxia on NE release and neutrophil-mediated endothelial dysfunction 3.1 Introduction Neutrophils are generated within the bone marrow, a significantly hypoxic microenvironment, with in vivo measurements in mice demonstrating local oxygen tension as low as 1.3kPa (120). The physiological hypoxia experienced by circulating (123) and transmigrated neutrophils in healthy tissues (125,126,312) can be further amplified in pathological conditions, such as organ inflammation (313) or ischaemia (314). Hypoxia is known to modulate neutrophil function; published effects include delayed constitutive apoptosis (145), reduced respiratory burst with resultant impairment of ROS-dependent bacterial killing of S. aureus (148), augmented degranulation (149), and most (although not all) reports demonstrate increased phagocytosis (170,174). Previous work in our laboratory has shown that a sustained and consistent cell culture medium oxygen tension of ~3kPa can be established using a sealed hypoxia workstation (Baker Ruskinn), delivering 0.8% oxygen within the chamber. This oxygen tension was able to stabilise HIF-1α in neutrophils (demonstrated by increased stabilisation of HIF-1α protein (145) and transcription of the HIF-1α target BNIP (315)), and is also biologically relevant. HIF staining has been demonstrated in colonic biopsies from patients with IBD (133) and bronchial biopsies from patients with COPD (131), although it is known that HIF can be stabilised by inflammatory stimuli as well as by hypoxia (316). I wished to extend the previous observation made in our laboratory by Dr Kim Hoenderdos that neutrophil degranulation from GM-CSF and fMLP-treated neutrophils is enhanced by hypoxia independent of HIF-1α (315), by investigating degranulation responses for a range of physiologically relevant priming agonists under the same hypoxic conditions. In addition, I aimed to explore the mechanism of hypoxia-enhanced neutrophil degranulation. Published data from our laboratory have implicated PI3K signalling in the modulation of degranulation by hypoxia in GM-CSF-primed neutrophils (149). I sought to further investigate this finding, both by exploring the contribution of PI3K signalling in hypoxic neutrophils treated with different agonists and by utilising neutrophils from transgenic mice with loss- or gain-of-function in relevant PI3K isoforms. I additionally investigated whether enhanced granule protein release under hypoxia might reflect increased generation of NETs, which are extracellular expulsions of decondensed chromatin decorated with granule proteins, including NE and MPO (72). Neutrophil granule proteases such as NE and PR3 have been implicated in the pathogenesis of chronic inflammatory diseases, including COPD (201,203) and vasculitis (317). Previous 93 published work from our laboratory has demonstrated that supernatants derived from neutrophils stimulated under hypoxic vs normoxic conditions cause more damage to epithelial cells, in a protease-dependent manner (149). As well as local tissue damage caused by adjacent neutrophil degranulation, intravascular degranulation has also been demonstrated, for example in sepsis (318) and vasculitis (319). Furthermore, evidence of increased circulating NE has been identified in pathologies associated with local or systemic hypoxia, including experimental I/R injury (320) and in patients experiencing exacerbations of COPD (detected indirectly, by measuring an elastase breakdown product) (204). Chronic inflammatory diseases (234,321,322) and hypoxia (188,323,324) are associated with endothelial dysfunction, and neutrophils can cause endothelial damage by a variety of mechanisms, including ROS production, NETosis and release of proteases (254). I therefore wished to investigate whether hypoxia was able to potentiate neutrophil-mediated endothelial activation and/or injury by examining the effects of supernatants generated from hypoxic neutrophils on human pulmonary artery endothelial cell adhesion molecule expression and survival/apoptosis. The specific aims of the work presented in this chapter are: 1. To investigate whether different priming agonists are able to increase stimulated NE release under hypoxia 2. To examine the mechanism of enhanced NE release under hypoxia 3. To compare the effects of normoxic and hypoxic neutrophil supernatants on pulmonary artery endothelial cells 3.2 Confirmation of the neutrophil hypoxic response Hypoxia is known to delay constitutive neutrophil apoptosis in a HIF-1α-dependent manner (145). To demonstrate that the hypoxia workstation was maintaining the desired level of biologically relevant hypoxia, neutrophil apoptosis was examined to show that these previously published results could be recapitulated. Isolated neutrophils were incubated under normoxia (21% O2, 5% CO2, 37ºC) or hypoxia (0.8% O2, 5% CO2, 37ºC) for 20 h, in the presence or absence of GM-CSF (1 ng/ml). Quantification of neutrophil apoptosis by morphology revealed a decrease under hypoxia compared with normoxia, in cells cultured in medium alone (31.6 vs 53.4%, p=0.004, Figure 3.1A); hypoxia did not significantly augment the GM-CSF-mediated survival effect. Flow cytometric measurement of neutrophil apoptosis (with the AnV+/PI- population identified as early apoptotic cells) was in agreement with the morphological findings (Figure 3.1B). 94 Assessment of apoptosis by two independent methods therefore showed that the hypoxia workstation was maintaining a biologically relevant level of hypoxia which could stabilise HIF- 1α, consistent with published literature. Hypoxia-driven delayed neutrophil apoptosis could compound the results of functional analyses. To establish that the hypoxic survival effect was not responsible for any differences seen under the planned experimental conditions, apoptosis was also assessed following a 4 h incubation period in the absence of serum (another potential experimental confounder as apoptosis assays are generally conducted in the presence of serum). Very low levels of apoptosis were seen under all conditions at 4 h as expected. Morphological assessment of neutrophil cytospins demonstrated no significant difference in apoptosis between normoxia and hypoxia in untreated, GM-CSF/fMLP or PAF/fMLP-treated cells, with all treatments applied precisely as planned for future functional and biochemical assays (Figure 3.1C). Thus, the hypoxia workstation delivers hypoxia but, at the selected 4 h time points, differential survival should not confound my assessment of degranulation. 3.3 The effect of different priming agonists on hypoxic NE release Our group has previously shown that hypoxia significantly increased NE release from activated neutrophils. This was demonstrated by ELISA for cytochalasin B and fMLP-treated cells (148) and by activity assay for GM-CSF and fMLP-treated cells (315). The latter experiment has greater physiological relevance, both in terms of stimulus-selection and of the measurement of NE activity, given that neutrophils are also capable of releasing the protease inhibitor α1AT upon degranulation (34). I wished to extend these observations to a wider range of physiologically relevant priming agents, assessing the effect of hypoxia on NE release by activity assay. I chose TNFα and PAF as well-established, disease-relevant priming agents that signal via TNF-receptors and GPCRs respectively. Neutrophils were incubated under normoxia or hypoxia for 4 h, primed with GM-CSF (10 ng/ml, 30 min), TNFα (20 ng/ml, 30 min) or PAF (1 µM, 5 min), and subsequently activated with fMLP (100 nM, 10 min). Supernatant NE activity was assessed by Enzchek® elastase activity assay, which measures the ability of elastase to cleave a non-fluorescent substrate into fluorescent fragments. Consistent with previous data, there was limited NE release from unstimulated (IMDM-treated) cells under normoxic or hypoxic conditions, but hypoxia significantly increased NE activity in supernatants from GM-CSF and fMLP-treated neutrophils (1.36*104 ± 4.3*103 arbitrary units (AU) vs 4.18*104 ± 7.2*103 AU, p<0.0001, Figure 3.2A). 95 Figure 3.1: The effect of hypoxia on neutrophil apoptosis Neutrophils were isolated by discontinuous plasma-Percoll® gradients and incubated under normoxia (21% O2) or hypoxia (0.8% O2) for 20 h in IMDM containing 10% autologous serum (A&B) or for 4 h in IMDM without serum (C). Cells were treated at baseline with A&B: GM-CSF (1 ng/ml) or IMDM control or C: after 4 h with GM-CSF (10 ng/ml, 30 min) or PAF (1 μM, 5 min), and subsequently fMLP (100 nM, 10 min), or IMDM control. A&C: Cytopsins were stained with May-Grünwald-Giemsa for morphological assessment by light microscopy. B: pelleted cells were stained with FITC-AnV and PI for flow cytometric assessment. Results represent mean ± SEM; A: n=6, B: n=2, C: n=3-4; ** = p<0.01, *** = p<0.001, two way ANOVA, Sidak’s multiple comparisons test. 96 Figure 3.2: The effect priming agonists and hypoxia on NE release A-E: Neutrophils were isolated by discontinuous plasma-Percoll® gradients and incubated under normoxia or hypoxia. After 4 h, cells were treated with A. GM-CSF (10 ng/ml, 30 min) B&E. PAF (1 µM, 5 min) C-E. TNFα (20 ng/ml, 30 min) and subsequently fMLP (100 nM, 10 min) or IMDM control. Supernatant NE activity was measured at 30 min by Enzchek® assay. F: Isolated neutrophils were incubated under normoxia and treated with TNFα (20 ng/ml, 30 min) or PBS control before addition of luminol (1 µM, 3 min), HRP (62.5 units/ml) and ROS production analysed in a luminometer following injection of fMLP (100 nM). Data are presented as raw peak height in relative light units (RLU). Results represent A-C&E: mean ± SEM or D&F: absolute values. A: n=12, B: n=12, C: n=12 (inset: subset of 2 individual donors labelled A and B), D: n=12, E: n=3 (subset of C&D), F: n=2 (subset of C-E); **** = p<0.0001, two way ANOVA, Sidak’s multiple comparisons test. 97 Hypoxia similarly augmented NE release from PAF and fMLP-treated cells (5.04*104 ± 1.21*104 AU vs 12.27*104 ± 2.32*104 AU, p<0.0001, Figure 3.2B). Surprisingly, hypoxia did not lead to a significant increase in NE release from TNFα and fMLP-treated cells (4.85*104 ± 1.33*104 AU vs 4.7*104 ± 9.34*103 AU, Figure 3.2C). However, when individual responses were examined separately, it became apparent that hypoxia can increase NE release from TNFα and fMLP-treated cells from some, but not all donors (Figure 3.2D); indeed a variable donor response was observed when neutrophils from two different volunteers were incubated and treated simultaneously, ensuring identical treatment and levels of hypoxia (Figure 3.2C inset). Further, in a subset of donors whose neutrophils were primed simultaneously with either TNFα or PAF, hypoxia consistently increased NE release from PAF and fMLP-treated but not TNFα and fMLP-treated cells (Figure 3.2E). However, these neutrophils were responsive to TNFα as they were able to produce ROS in response to TNFα-priming under normoxic conditions (Figure 3.2F, data generated with the assistance of Dr Arlette Vassallo). Overall, these data demonstrate that the hypoxic augmentation of NE release is dependent on the specific priming agonist rather than equating to a non-specific “priming” effect. I therefore wished to further investigate the mechanism of upregulated degranulation under hypoxia. 3.4 Investigation of the mechanism of NE release under hypoxia 3.4.1 The effect of hypoxia on NET production NETs comprise externalised chromatin beaded with antimicrobial proteins (72) and are thought to be a host defence mechanism which facilitates the trapping and killing of extracellular pathogens, although they also confer the potential for host tissue damage. Pharmacological stabilisation of HIF-1α has been shown to increase NETosis (184,185); conversely, pharmacological and genetic HIF-1α knockdown decreased NET production (184). I therefore hypothesized that enhanced NE release from neutrophils under hypoxia was, at least in part, due to an increase in NETosis. Isolated neutrophils were re-suspended in normoxic or hypoxic IMDM, containing the cell impermeable nucleic acid stain SYTOX™ green. To explore the effect of the cytokines alone, cells were treated at the start of incubation with GM-CSF (10 ng/ml) or PAF (1 µM), and fluorescence absorbance (indicating extracellular DNA as a surrogate for NETosis) measured hourly. To assess NET production under the exact experimental conditions employed for measurement of NE activity (section 3.3), cells were treated after 4 h normoxic vs hypoxic incubation with GM-CSF (10 ng/ml, 30 min), and subsequently fMLP (100 nM, 10 min). Fluorescence absorbance was measured at baseline, 4 h and after treatment. PMA (20 nM), 98 a pharmacological stimulus, was used as a positive control as it is known to induce massive NETosis. Results are expressed as percentage of total cellular DNA content, which was established by permeabilising cells with 0.5% Triton-X. There was no significant difference in NET release in response to hypoxia alone or in combination with GM-CSF (Figure 3.3A), PAF (Figure 3.3B) or GM-CSF and fMLP (Figure 3.3C). PMA induced dramatic NETosis as previously published, demonstrating that the assay was able to detect this process, and which was significantly increased compared with all other treatments at 4 h (p<0.01, Figure 3.3C). Altogether, these data do not support a role for NETosis in the enhanced NE release under hypoxia. Therefore, I next considered the signalling pathways known to regulate classical neutrophil degranulation, in particular the PI3K signalling pathway, a key mediator of azurophil granule release. 3.4.2 The effect of PI3K signalling modulation on the release of NE under hypoxia A number of signalling pathways involved in degranulation and potentially responsible for the hypoxic uplift of NE release from GM-CSF and fMLP-treated neutrophils have previously been investigated in our laboratory: Dr Kim Hoenderdos showed that phospholipase C and the generation of calcium transients are not central to this response (149). In view of the known role of PI3K and in particular PI3Kγ in priming of neutrophil degranulation, I next explored the role of this signalling pathway in the hypoxic uplift of degranulation, initially using phosphorylation of AKT as a readout of PI3K activity, together with isoform-selective PI3K inhibitors. Inhibition of PI3K signalling with the pan-PI3K inhibitor LY294002 completely abrogated the hypoxic uplift of NE release from GM-CSF/fMLP-treated neutrophils, although substantial inhibition of the normoxic response was also observed (149). Further dissection of the PI3K signalling pathway demonstrated that inhibition of the PI3Kγ isoform with AS605240 could also largely abolish the hypoxic uplift of NE release but that inhibition of the PI3Kδ isoform with IC87114 could not (149). I therefore hypothesized that the enhanced NE release from neutrophils under hypoxia was PI3K-dependent, and more specifically, PI3Kγ- dependent. 99 Figure 3.3: The effect of hypoxia on NET production Neutrophils were isolated by discontinuous plasma-Percoll® gradients and incubated under normoxia or hypoxia. A&B: Cells were treated at baseline with A. GM-CSF (10 ng/ml) B. PAF (1 µM), or IMDM control. NET production was quantified hourly by fluorescence absorbance with Sytox Green (5 μM). C: Following 4 h incubation, cells were treated with GM-CSF (10 ng/ml, 30 min) and subsequently fMLP (100 nM, 10 min) or IMDM control. NET production was quantified at baseline, 4 h and after treatment. Cells incubated under normoxia were treated at baseline with PMA (20 nM) and NET production quantified at baseline and 4 h. A- C: NETosis is expressed as extracellular DNA as % of total DNA from triton-X (0.5%) lysed neutrophils. Results represent mean ± SEM; A: n=4, B-C: n=3; ** = p<0.01; two way ANOVA, Sidak’s multiple comparisons test. 100 3.4.2.1 The effect of PI3K inhibitors on the release of NE from human neutrophils under hypoxia Working with Dr Kim Hoenderdos and using hypoxia in conjunction with GM-CSF priming to align with her previous dataset, I generated neutrophil lysates (as described in section 2.2.1.5) from isolated neutrophils (10*106/ml) re-suspended in normoxic or hypoxic IMDM and treated with PI3K inhibitors prior to incubation, either AS605240 (PI3Kγ-selective inhibitor, 3 µM), or IC87114 (PI3Kδ-selective inhibitor, 3 µM). Prior experiments undertaken by Dr Hoenderdos indicated that fMLP induced marked AKT phosphorylation that was not further enhanced by hypoxia (149), and hence this was not further explored. Cell lysates were subjected to SDS-PAGE and probed for pAKT. Stripped membranes were subsequently probed for total AKT and β-actin. In the setting of hypoxia, AKT phosphorylation in response to GM-CSF and fMLP was moderately enhanced relative to normoxia (fold change from normoxic control: 8.42 ± 1.9 vs 13.37 ± 2.11, p=0.029, Figure 3.4A), although this effect was variable between different donors (Figure 3.4B). The PI3Kγ-selective inhibitor completely abrogated the phosphorylation of AKT in both normoxic (fold change: 8.42 ± 1.9 vs 1.34 ± 0.55, p=0.0094, Figure 3.4A) and hypoxic neutrophils (fold change: 13.37 ± 2.11 vs 1.52 ± 1.1, p<0.0001, Figure 3.4A), whilst the PI3Kδ inhibitor did not. Given that GM-CSF signals via a tyrosine kinase-coupled receptor, it was surprising that inhibition of PI3Kγ abolished the primed and hypoxic-primed phosphorylation of AKT, whilst the use of a PI3Kδ inhibitor had little impact on the hypoxic response. Since the above results suggested possible involvement of PI3Kγ in mediating the hypoxic uplift of NE in GM-CSF-primed cells, I decided to pursue this further using PAF, which in my hands gave more consistent and robust effects on NE release, and which signals via a GPCR more likely to link to PI3Kγ. Initially, I investigated whether NE release in PAF-primed hypoxic neutrophils displayed the same profile of inhibition with isoform-selective PI3K inhibitors as did GM-CSF. Given the uncertain contribution of PI3Kδ inhibition, previously showing a minor reduction of the hypoxic uplift of NE release (149), I utilised two separate PI3Kδ-selective inhibitors, namely IC87114 and CAL-101 (the latter is a licensed drug, Idelalisib, used to treat haematological malignancy) to examine this effect more thoroughly. Prior to normoxic/hypoxic incubation, cells were treated with AS605240 (PI3Kγ-selective inhibitor, 3 µM), IC87114 (PI3Kδ-selective inhibitor, 3 µM) or CAL-101 (PI3Kδ-selective inhibitor, 100 nM). After 4 h, neutrophils were treated with PAF (1 µM, 5 min) and subsequently fMLP (100nM, 10 min), or IMDM control. NE release was measured by Enzchek® assay. Similar to the pattern observed with GM-CSF-primed neutrophils, PI3Kγ inhibition significantly reduced NE release from PAF and fMLP-treated neutrophils under hypoxia (4.96*103 ± 1.26*103 AU vs 5.75*104 ± 6.77*103 AU, p<0.0001, Figure 3.5) whereas there was no difference in NE release for these conditions 101 Figure 3.4: The effect of PI3K inhibition on the hypoxic regulation of AKT phosphorylation Isolated neutrophils were incubated under normoxia (N) or hypoxia (H) for 4 h. Cells were treated with GM-CSF (10 ng/ml, 15 min) or IMDM control in the presence or absence of PI3K inhibitors (IC87114, 3 µM; AS605240, 3 µM). Cell lysates were subjected to SDS-PAGE; western blotting was performed with anti-pAKT(Ser473), anti-AKT and anti-β actin. A: fold change in pAKT was quantified by densitometry using ImageJ. B: representative images from 2 donors. Results represent mean ± SEM; n=7; # is the difference in pAKT in comparison with unstimulated normoxic neutrophils in the absence of PI3K inhibitors; * = p<0.05, ** = p<0.01, ### = p<0.001, **** = p<0.0001; two way ANOVA, Sidak’s multiple comparisons test. 102 under normoxia, showing that PI3Kγ inhibition completely abolished the hypoxic augmentation of NE release. PI3Kδ inhibition with CAL-101, compared with no inhibitor treatment, had a similar but less marked effect on the magnitude of NE release from hypoxic neutrophils (3.29*104 ± 5.65*103 AU vs 5.75*104 ± 6.77*103 AU, p=0.0101, Figure 3.5) but IC87114 did not significantly reduce NE release under hypoxia (3.91*104 ± 1.02*104 AU vs 5.75*104 ± 6.77*103 AU, p=0.0701, Figure 3.5). In contrast to PI3Kγ inhibition, there was a trend towards increased NE release under hypoxia compared with normoxia in the setting of PI3Kδ inhibition but this increase did not reach statistical significance with either inhibitor. PI3Kγ inhibition also prevented the significant increase in NE release from unstimulated vs stimulated normoxic cells; however, whilst CAL-101 had a similar effect, IC87114 did not reduce stimulated NE release under hypoxia. This might reflect either off-target effects of CAL-101 or failure of IC87114 to fully inhibit PI3Kδ. Altogether, these experiments show that PI3Kγ is essential for PAF-primed fMLP-stimulated degranulation in both normoxia and hypoxia, with possible combinations of both PI3Kγ and PI3Kδ isoforms to the hypoxic uplift effect. To further clarify these contributions, I decided to take advantage of available transgenic mouse strains with absent or enhanced PI3Kγ and PI3Kδ activity. Figure 3.5 The effect of PI3K inhibition on the hypoxic regulation of NE release Isolated neutrophils were incubated under normoxia or hypoxia in the presence or absence of IC87114 (3 µM), CAL-101 (100 nM) or AS605240 (3 µM). After 4 h, cells were treated with PAF (1 µM, 5 min) and subsequently fMLP (100 nM, 10 min), or IMDM control. Supernatant NE activity was measured at 30 min by Enzchek® assay. Results represent mean ± SEM; n=4-6; # is the difference in NE release in comparison with unstimulated normoxic neutrophils in the absence of PI3K inhibitors; * or # = p<0.05, **** = p<0.0001; two way ANOVA, Sidak’s multiple comparisons test. 103 3.4.2.2 The effect of PI3K isoform mutations in murine neutrophils on NE release under hypoxia The above experiments suggested involvement of PI3Kγ and perhaps PI3Kδ but with a couple of uncertainties – namely, whether the apparent impact of the PI3Kδ inhibitors might reflect off-target effects, and the possibility that PI3Kγ inhibition was simply inhibiting degranulation rather than specifically the hypoxic uplift of degranulation. I therefore investigated PI3K modulation of NE release in neutrophils from transgenic mice with abolished or enhanced activity of these isoforms. To assess the role of PI3Kδ, neutrophils were isolated from the femoral bone marrow of C57BL/6J wildtype mice, or C57BL/6J mice with an activating (E1020K heterozygote (325)) or kinase-dead (D910A homozygote (326)) mutation of PI3Kδ (obtained from Professor Klaus Okkenhaug, Babraham Institute), using negative immunomagnetic selection, exactly as described in section 2.2.1.2. To assess PI3Kγ, neutrophils were isolated from the femoral bone marrow of C57BL/6 E129 wildtype mice, or C57BL/6 E129 mice lacking the catalytic subunit of PI3Kγ (p110γ-/- (327)) (obtained from Dr Len Stephens, Babraham Institute). Murine neutrophils primed with PAF or murine GM-CSF and stimulated with fMLP did not generate significant extracellular NE activity (Figure 2.10A). Cells were therefore primed with cytochalasin B to allow optimal stimulation of NE release (see section 2.2.4.2). Murine neutrophils (3-5*106) were incubated under normoxia or hypoxia for 4 h before treatment with cytochalasin B (5 µg/ml, 5 min) and subsequently fMLP (10 µM, 10 min), or IMDM control. Supernatants were assessed for NE activity by Enzchek® elastase activity assay. As observed in similar experiments with human neutrophils (315), the release of NE from cytochalasin B and fMLP-treated wildtype murine neutrophils was significantly enhanced under hypoxia, compared with normoxia (2.66*105 ± 5.07*104 AU vs 1.49*105 ± 1.47*104 AU, p=0.0176, Figure 3.6A). The release of NE from cytochalasin B and fMLP-treated murine neutrophils with activating or kinase-dead PI3Kδ mutations was no different from wildtype, and the hypoxic enhancement of NE release was wholly maintained (Figure 3.6B), although no statistical analysis could be performed on n=2 experiments; further mice were unfortunately not available to perform an additional experiment. However, the hypoxic increase in NE release from murine neutrophils deficient in PI3Kγ was completely abolished (1.53*105 ± 8.08*103 AU vs 1.85*105 ± 1.16*104 AU for wildtype mice, p=0.0058; 1.53*105 ± 1.17*104 AU vs 1.53*105 ± 9.06*103 AU for PI3Kγ-/- mice, p>0.9999, Figure 3.6C), although the ability of cytochalasin B/fMLP-treated PI3Kγ-/- neutrophils to degranulate under normoxia was preserved. 104 Figure 3.6: The effect of hypoxia on NE release from murine neutrophils with PI3Kδ or PI3Kγ mutations Femoral bone marrow neutrophils were isolated from A: wildtype B: E1020K, D910A and wildtype or C: p110γ-/- and wildtype mice by negative immunomagnetic selection. A-C: Isolated neutrophils were incubated under normoxia or hypoxia for 4 h, before treatment with cytochalasin B (5 µg/ml, 5 min) and subsequently fMLP (10 µM, 10 min), or IMDM control. Supernatant NE activity was measured at 20 h by Enzchek® assay. Results represent mean ± SEM; A: n=9, B: 3 mice per genotype per experiment, n=2 independent experiments, C: 3- 4 mice per genotype per experiment, n=5 independent experiments; * = p<0.05, ** = p<0.01, two way ANOVA, Sidak’s multiple comparisons test. 105 Despite documented differences between human and murine neutrophils regarding the roles of PI3Kγ and PI3Kδ isoforms (328), these results support my results from human cells indicating a non-redundant role for PI3Kγ but not PI3Kδ in the hypoxic augmentation of NE release. I therefore moved on to explore the consequences of this enhanced secretion on endothelial cells. 3.5 The effect of hypoxia on neutrophil-induced pulmonary artery endothelial cell dysfunction Neutrophil inflammatory disorders, such as COPD, are characterised by bystander tissue damage and are often also associated with systemic endothelial dysfunction. COPD patients have increased arterial wall stiffness and markers of systemic inflammation (for example CRP) when compared with control smokers without airflow obstruction (234), and furthermore have considerably more cardiovascular comorbidities, even when corrected for shared risk factors, such as smoking (225,227). I therefore hypothesized that hypoxia synergises with systemic neutrophil priming to promote a destructive neutrophil phenotype with the capacity to cause endothelial cell damage and/or activation, which is biologically relevant to endothelial dysfunction observed in patients with chronic inflammatory disease. To investigate this, I incubated supernatants derived from normoxic/hypoxic PAF and fMLP-treated neutrophils with primary HPAEC monolayers and assessed cell activation and detachment/death. 3.5.1 The effect of neutrophil supernatants on endothelial cell activation Neutrophil adhesion molecules interact with endothelial ligands to mediate adhesion and transmigration. Neutrophil integrins (e.g. MAC-1) interact with endothelial immunoglobulin-like adhesion molecules, in particular ICAM-1, to mediate firm adhesion and neutrophil trans- endothelial migration. Several studies have shown increased neutrophil adhesion to the endothelium under hypoxia (152,156,157), with one study showing reversal after re- oxygenation (156). Since endothelial cells ‘display’ neutrophil priming agents such as PAF on their surface (329), and since hypoxia is a feature of the sluggish microcirculation seen in sepsis/infection, this process could contribute both to endothelial dysfunction and persistent tissue neutrophilia seen in chronic inflammatory disease. I therefore wished to extend my study of the effects of normoxic and hypoxic neutrophil supernatants to determine their effects on endothelial cell ICAM-1 expression. To assess endothelial cell activation in the presence of neutrophil supernatants, ICAM-1 expression was measured. HPAECs (passage 10-12) were grown to confluence and incubated under normoxia (21% O2) or hypoxia (0.8% O2) with undiluted supernatants generated from normoxic vs hypoxic, PAF/fMLP-treated or control (IMDM) neutrophils for 24 106 h in the presence of 2% human serum. Following supernatant exposure, HPAECs were trypsinised, fixed (1% PFA), blocked (5% FBS in 0.5% PBS-BSA) and stained with APC- conjugated anti-ICAM-1, and MFI was quantified by flow cytometry. Supernatants from PAF and fMLP-treated hypoxic neutrophils applied to normoxic HPAECs led to a significant increase in ICAM-1 expression when compared to the equivalent supernatants from normoxic neutrophils (MFI: 92.97 ± 10.88 vs 52.53 ± 7.32, p=0.037, Figure 3.7A). Surprisingly, the expression of ICAM-1 by hypoxic HPAECs treated with hypoxic supernatants was significantly less than normoxic HPAECs treated with hypoxic supernatants (MFI: 34.6 ± 6.17 vs 92.97 ± 10.88, p=0.0101, Figure 3.7A) and indeed did not differ from normoxic HPAECs treated with normoxic supernatant. Results were similar whether total MFI or the percentage of cells positive for APC (ICAM-1) was assessed (Figure 3.7B). These data demonstrate opposing roles for hypoxia in the regulation of ICAM-1 expression, acting both indirectly, through its effect on secretion from stimulated neutrophils, and also directly, by blunting the ability of endothelial cells to upregulate ICAM-1 expression in response to hypoxic supernatants. The overall effect may be context-dependent (further discussed in section 3.6). I next proceeded to examine the ability of hypoxia to elicit neutrophil-mediated endothelial cell damage/death. 3.5.2 The effect of hypoxia on neutrophil-induced endothelial cell detachment To investigate the ability of neutrophil supernatants to cause endothelial cell detachment, HPAECs were stained with rhodamine-phalloidin (F-actin) and DAPI (nucleus), and assessed using confocal microscopy. Confluent HPAECs were incubated with supernatants from normoxic vs hypoxic PAF and fMLP-treated vs control (IMDM) neutrophils. As we have previously demonstrated that neutrophil supernatant-induced epithelial cell detachment can be rescued by the addition of the serine protease inhibitor α1AT (149), HPAECs were also incubated with supernatants in the presence or absence of α1AT (46 µg/ml), added to supernatants 10 min prior to HPAEC treatment. Since human serum contains abundant α1AT it was not used in these experiments; instead, supernatants were diluted 1:1 in serum-free EGM-2 (endothelial cell-specific media) to avert detrimental effects due to lack of growth factors over the 24 h treatment period. After treatment, HPAEC monolayers were fixed and stained as described in section 2.3.3.1. Three randomly selected fields from duplicate wells were imaged (Leica Sp5 confocal microscope, Figure 3.8A). Cell detachment was quantified with ImageJ and expressed as % detachment of the whole field of view. 107 Figure 3.7: The effect of hypoxia and neutrophil supernatants on endothelial- leukocyte ICAM-1 expression by flow cytometry Isolated neutrophils were incubated under normoxia or hypoxia. After 4 h incubation, neutrophils were treated with PAF and fMLP, or IMDM control exactly as described previously. After treatment, neutrophils were pelleted and supernatants harvested. HPAECs were grown to confluence in a 12 well plate and exposed to normoxia (21% O2) or hypoxia (0.8% O2) for 24 h prior to treatment with neutrophil supernatants (SN) in the presence of 2% human serum under normoxic or hypoxic conditions as indicated. After 24 h, supernatants were aspirated and HPAECs pelleted and stained with APC-conjugated anti-ICAM-1 prior to flow cytometric analysis. Data are presented as A: median fluorescence intensity or B: APC-positive HPAECs (compared with isotype control) as % of total single cells. Results represent mean ± SEM, n=3, * = p<0.05, ** = p<0.01, two way ANOVA, Sidak’s multiple comparisons test. 108 Supernatants from hypoxic PAF and fMLP-treated neutrophils caused significantly more HPAEC detachment than supernatants from control neutrophils (55.04 ± 13.71% vs 25.88 ± 9.25%, p=0.0007) and from normoxic stimulated neutrophils (55.04 ± 13.71% vs 37.31 ± 13.23%, p=0.0206, Figure 3.8B). There was no significant difference in HPAEC detachment resulting from treatment with normoxic control vs stimulated neutrophil supernatants. The inclusion of α1AT during incubation prevented the detachment caused by supernatants from both normoxic (19.5 ± 8.75% vs 37.31 ± 13.23%, p=0.0299) and hypoxic (23.16 ± 8.66% vs 55.04 ± 13.71%, p=0.0003) stimulated neutrophils (Figure 3.8B). These data demonstrate that hypoxia increases the capacity of PAF and fMLP-treated neutrophils to cause HPAEC detachment in a protease-dependent manner. 3.5.3 The effect of hypoxia on neutrophil-induced endothelial cell death It is possible that supernatant proteases were able to lift HPAECs from the plate surface by cleaving adhesion molecules, rather than detachment resulting from supernatant-induced cell damage/death. Hence, the detachment assay (section 3.5.2) would be giving a surrogate readout of supernatant protease concentration rather than direct damage per se. Therefore, the ability of neutrophil supernatants to cause endothelial cell death was explored by MTT toxicity assay, which allows the spectrophotometric quantification of cell viability, and by flow cytometric analysis of apoptosis. Confluent HPAECs (passage 10-12) were incubated with supernatants generated from normoxic vs hypoxic, PAF and fMLP -treated vs control (IMDM) neutrophils in the presence or absence of α1AT (46 µg/ml) exactly as described above (section 3.5.2), diluted 1:1 in serum- free endothelial medium EGM-2. After 24 or 48 h, samples/media were aspirated and survival (expressed as fold-change compared to normoxic control supernatants) was assessed by the MTT assay (Section 2.3.3.2). After 24 h, there was a modest but significant decrease in HPAEC survival between cells treated with hypoxic supernatants from control (IMDM) vs PAF/fMLP-treated cells (fold change: 1.03 ± 0.05 vs 0.79 ± 0.03, p=0.036, Figure 3.9A), but no significant difference between the equivalent normoxic control and stimulated samples (fold change: 0.9 ± 0.06 vs 0.79 ± 0.03, p=0.28, Figure 3.9A). In the presence of α1AT, the reduction in HPAEC survival with hypoxic supernatant treatment was not apparent, suggesting a protective effect of serine protease inhibition. Since these effects were relatively modest, I extended the incubation time to 48 h. 109 Figure 3.8: The effect of neutrophil supernatants on HPAEC detachment by confocal microscopy Supernatants from normoxic (N) vs hypoxic (H) PAF/fMLP (P) or control (IMDM)-treated neutrophils were prepared exactly as previously. Confluent HPAEC were incubated with supernatants diluted 1:1 with serum-free EGM-2 media for 24 h in the presence or absence of α1AT (46 µg/ml). HPAEC were fixed and stained with rhodamine-phalloidin and DAPI. A: representative confocal images. B: quantification of cell detachment using ImageJ, expressed as % detachment of whole field of view. Results represent mean ± SEM; n=5; # is the difference in detachment in comparison with hypoxic PAF/fMLP-treated HPAEC in the absence of α1AT; * = p<0.05, *** or ### = p<0.001; two way ANOVA, Sidak’s multiple comparisons test. 110 Treatment with supernatants from PAF and fMLP-treated neutrophils for 48 h demonstrated a significant reduction in HPAEC survival between cells treated with supernatants from hypoxic vs normoxic stimulated neutrophils (fold change: 0.76 ± 0.04 vs 0.93 ± 0.04, p=0.0031, Figure 3.9B). Although the inclusion of α1AT gave some protection, there remained a trend (which approached but did not reach significance) to reduced HPAEC survival when treated with hypoxic vs normoxic stimulated supernatants (fold change: 0.75 ± 0.06 vs 0.92 ± 0.09, p=0.055, Figure 3.9B). Of note, HPAEC treated with supernatants from hypoxic stimulated vs unstimulated neutrophils had significantly decreased survival, both in the absence (p<0.0001) or presence of α1AT (p=0.0007). These data show that HPAEC death induced by supernatants from hypoxic PAF and fMLP-treated neutrophils after 48 h is partially but not completely rescued by serine protease inhibition. Overall, the results from MTT assessment of HPAEC survival indicate that supernatants from hypoxic PAF-primed activated neutrophils cause more endothelial cell death, which is at least partly protease-dependent. However, the MTT assay does not quantify the survival/death of detached cells nor does it determine the mode of cell death. In order to explore the nature of cell death further, I investigated the ability of neutrophil supernatants to induce HPAEC apoptosis or necrosis using flow cytometry. Since apoptosis detection by this method precedes cell death, I employed a shorter (6 h) time point. As the assay is very sensitive, conditions were optimised to avoid induction of apoptosis by cell manipulation. Confluent HPAECs (passage 7) were incubated with undiluted supernatants generated from normoxic vs hypoxic, PAF and fMLP-treated vs control (IMDM) neutrophils. After 6 h cells were stained with FITC-AnV and PI. Viable, apoptotic and necrotic cells were quantified by flow cytometry: AnV-/PI- staining indicated viable non-apoptotic cells, AnV+/PI- staining indicated early apoptotic cells and AnV+/PI+ staining indicated late apoptotic or necrotic cells. The gating strategy is shown in Figure 2.12. Consistent with my detachment and MTT survival data, supernatants from hypoxic PAF and fMLP-treated neutrophils significantly decreased HPAEC viability when compared with supernatants from control neutrophils (27.55 ± 7.1% vs 33.48 ± 8.48% AnV-/PI-, p=0.0328) and from normoxic stimulated neutrophils (27.55 ± 7.1% vs 44.68 ± 2.08% AnV-/PI-, p=0.0016, Figure 3.10A), with an associated increase in AnV+/PI- apoptotic cells (Figure 3.10B). There was no significant difference in viable HPAEC resulting from treatment with normoxic control vs stimulated neutrophil supernatants. Interestingly, the same hypoxic decrease in HPAEC viability observed with stimulated neutrophil supernatants was seen after exposure to supernatants from unstimulated neutrophils (33.48 ± 8.48% vs 47.25 ± 4.79% AnV-/PI-, p=0.003, Figure 3.10A). These results demonstrate that, in contrast to the detachment and 111 MTT survival data, hypoxia increases the capacity of both unstimulated and PAF/fMLP-treated neutrophils to cause HPAEC apoptosis. Figure 3.9: The effect of neutrophil supernatants on HPAEC survival by MTT assay Isolated neutrophils were incubated under normoxia or hypoxia. After 4 h incubation, neutrophils were treated with PAF and fMLP, or IMDM control. After treatment, neutrophils were pelleted and supernatants harvested. HPAEC were grown to confluence in a 96 well plate and exposed to neutrophil supernatants diluted 1:1 with EGM-2 media (serum free) for A: 24 h or B: 48 h, in the presence or absence of α1AT (46 µg/ml). Supernatants were aspirated and cell survival quantified by MTT assay. Cell survival is expressed relative to HPAEC treated with normoxic control supernatant. Results represent mean ± SEM; A: n=4-7, B: n=6-12; * = p<0.05, ** = p<0.01, *** = p<0.001, **** = p<0.0001, two way ANOVA, Sidak’s multiple comparisons test. 112 Figure 3.10: The effect of neutrophil supernatants on HPAEC survival by flow cytometry Isolated neutrophils were incubated under normoxia or hypoxia. After 4 h incubation, neutrophils were treated with PAF and fMLP, or IMDM control. After treatment, neutrophils were pelleted and supernatants harvested. HPAEC were grown to confluence in a 12 well plate and exposed to neutrophil supernatants for 6 h. Supernatants were aspirated and pelleted HPAEC stained with FITC-AnV and PI for flow cytometric assessment. A: analysis of viable (AnV-PI-) HPAEC as % of total population. B: viable (AnV-PI-), apoptotic (AnV+PI-) and necrotic (AnV+PI+) HPAEC as % of total single cell population. Results represent mean ± SEM; n=4; * = p<0.05, ** = p<0.01, two way ANOVA, Sidak’s multiple comparisons test. 113 3.6 Discussion Areas of infection and inflammation can be profoundly hypoxic, as demonstrated in numerous settings in vitro and in vivo, including COPD and lung infection (131,134,136). The vasculature can also be hypoxic, demonstrated in vivo in rabbit atherosclerotic aortae (330) and in human atherosclerotic carotid arteries (331). Patients with lung disease and impaired gas exchange (e.g. COPD) are frequently systemically hypoxaemic. Systemic hypoxia increased leukocyte- endothelial adhesion in rat cremaster venules (152) and hypoxia has been shown in multiple studies to increase circulating pro-inflammatory cytokines, such as TNFα and IL-6 (138,332). Conversely, myeloid HIF-1α deletion decreased circulating pro-inflammatory cytokines, including TNFα and IL-1β, in a murine LPS sepsis model (333). Neutrophil priming, e.g. by circulating cytokines, is regarded as a prerequisite for neutrophil-mediated cellular damage; primed circulating neutrophils have been identified in stable (334,335) and exacerbating (283) COPD patients, and also in patients with ARDS, correlating with oxygenation status (7). We therefore hypothesized that systemic neutrophil priming synergises with hypoxia to promote a destructive neutrophil phenotype with the capacity to cause endothelial cell damage and/or activation, which is biologically relevant to endothelial dysfunction observed in patients with chronic inflammatory disease. Initially, I explored the role of hypoxia in promoting degranulation from TNFα and PAF-primed neutrophils in addition to the published interaction with GM-CSF, to determine whether hypoxia synergises with all or just a subset of priming agents. TNFα and PAF are established neutrophil priming agents with relevance to lung diseases associated with both endothelial dysfunction and local and systemic hypoxia. TNFα has been implicated in the pathogenesis of ALI (336). Several studies have detected increased sputum (337) and circulating (338) TNFα in COPD patients, with one study showing an inverse correlation of plasma TNFα with the arterial partial pressure of oxygen (339). Furthermore, serum TNFα is one of a panel of six biomarkers identified by the ECLIPSE study as associated with increased exacerbation frequency and mortality in COPD patients (279); however, this and other studies (221) suggest that TNFα may be a marker of smoking rather than COPD, at least during the stable state. Sputum TNFα seems more consistently elevated during acute exacerbations (340,341) and it seems likely that TNFα plays a role in COPD pathogenesis as TNF receptor (TNFR) deficient mice are partially protected from cigarette-induced emphysema and lung inflammation (342). PAF, a pro-inflammatory phospholipid, plays an important role in ALI: in a model of acid aspiration-induced lung injury, mice deficient for the PAF receptor (PAFR) were protected, whereas mice overexpressing PAFR died rapidly with profound lung inflammation, pulmonary oedema and impaired oxygenation (343). PAF has also been shown to contribute to the pathogenesis of experimental cigarette smoke-induced COPD, where PAFR antagonism 114 could attenuate the development of emphysema (221), and is elevated in systemic conditions associated with hypoxia and endothelial injury, for example in I/R injury (344) and acute pancreatitis (345). Finally, plasma levels of PAF were elevated in patients with sepsis and even further increased in those with multiple organ dysfunction (346), which is indicative of endothelial dysfunction. Possible explanations for the observed inconsistent response of TNFα-primed NE release under hypoxia include donor variability and oxygen sensitivity of exocytosis signalling pathways. It has been observed in our laboratory that the magnitude of the respiratory burst in response to TNFα-priming varies considerably between individuals (personal communication, Professor Edwin Chilvers). Further, polymorphisms in the TNFA gene have been associated with more severe COPD and enhanced neutrophil migration (347). It is also possible that the concentration of TNFα used was maximal for some subjects; however, in experiments performed in parallel on the same cells, PAF induced more NE release than TNFα in the context of hypoxia, making this explanation less likely. Neutrophils contain both TNFR1 and TNFR2, which might allow donor-dependent differential signalling in the context of degranulation. In my hands, hypoxia reliably synergised with PAF priming of fMLP-induced NE release, and the magnitude of this response was greater than that seen with GM-CSF. I therefore adopted PAF as the key mediator for further studies in this thesis. Since NETosis could plausibly explain extracellular NE release and Branitzki-Heinemann et al. had previously observed enhanced NETosis with pharmacological HIF1α stabilisation (185), I assessed the impact of true hypoxia (rather than modulation of HIF) on NET formation, using the same experimental conditions as for my degranulation assays. In agreement with a previous report, I in fact saw a non-significant reduction of NET production under true hypoxia (186), although the baseline value under normoxia was already very low. NETosis is partially dependent on NADPH production of ROS; neutrophils from patients with CGD (who have impaired or absent NADPH oxidase function), display diminished NET release that is restored by pharmacological induction of ROS (73). The requirement of ROS (and hence molecular oxygen) likely explains the difference between HIF stabilisation and hypoxia on NET release. Published work from our laboratory, focused on GM-CSF-primed neutrophils (including my data shown in Figure 3.4), suggested an important role for PI3Kγ (but not PLC or calcium transients) in mediating the hypoxic uplift of degranulation. PAF ligates a GPCR, whose βγ subunit can directly activate PI3Kγ, hence it was logical to explore this pathway in the current work by examining the impact of hypoxia on PAF-primed neutrophil degranulation. PI3Kγ mediates several neutrophil functions, although most of these have not been assessed in the context of hypoxia. Neutrophils from mice lacking functional PI3Kγ have impaired oxidative 115 burst, migration and recruitment (13), and the early generation of PIP3 in response to fMLP stimulation for both human peripheral blood neutrophils and murine bone marrow-derived neutrophils is PI3Kγ-dependent (328). Both PI3Kγ and δ are fundamental to neutrophil function, although their roles are often temporally distinct and context-dependent (348). Previous reports have suggested that there is considerable cross-talk between Class IA (α/β/δ) and IB (γ) PI3Ks. For example, mouse neutrophils lacking functional PI3Kγ have a marked reduction in chemotaxis but it has also been shown that PI3Kδ activity plays a role in directional chemotaxis (349), and in some situations, such as GM-CSF-mediated prolonged neutrophil survival, there is complete functional redundancy between isoforms (304). Despite documented differences between human and murine neutrophils regarding the roles of PI3Kγ and PI3Kδ isoforms (328), my combined data from human cells treated with isoform-selective inhibitors and cells from transgenic mice support a non-redundant role for PI3Kγ in the hypoxic augmentation of NE release. However, the mechanism(s) by which hypoxia might modulate PI3Kγ activity are unclear and warrant further investigation. Studies in different cell types have demonstrated that hypoxia upregulates both PI3Kγ expression (350) and AKT phosphorylation (351); however, previous published data that cycloheximide does not prevent the hypoxic enhancement of degranulation (149) argues against increased PI3Kγ expression mediating this effect. Of note, in a mouse model of I/R, Alloatti et al. showed that PI3Kγ signalling mediated PAF-induced NO release and depression of cardiac contractility, and that PI3Kγ-null hearts had improved post-ischaemic recovery (352). Although I have focused on PI3K signalling, a number of pathways and signalling molecules are important in regulating neutrophil degranulation, including the p38 and ERK pathways, SRC-kinases and phospholipase D (353), as well as intracellular trafficking and small GTPase activation. Whilst a detailed exploration of these pathways in the setting of hypoxia would have been of considerable interest, the time constraints of a PhD precluded my exploring all of these potential avenues. Whilst I focused instead on establishing the potential biological relevance of my findings, prospective plans for further investigation into the hypoxic regulation of degranulation are further discussed in Section 6.3. In addition to local tissue damage, chronic inflammatory diseases, such as COPD, are associated with systemic endothelial dysfunction and excess cardiovascular morbidity and mortality (225,227). Similarly, activation of neutrophils with anti-MPO or anti-PR3 vasculitic autoantibodies has been shown to injure vascular endothelial cells (354), and patients with vasculitis have increased prevalence of atherosclerosis (322). Neutrophil granule contents have also been shown to induce endothelial activation or damage directly. Neutrophils can transfer enzymatically active MPO to endothelial cells, promoting vascular inflammation (247); treatment of bovine pulmonary artery endothelial cells with NE or PR3 induced apoptosis 116 (258); NE, PR3 and MMP-9 were all able to cleave the proteolytically activated GPCR proteinase receptor 1 (PAR1), which induced endothelial cell stress fibre formation and apoptosis (265,355); and reduction of neutrophil MMP-9 by siRNA or pharmacological inhibition reduced murine atherosclerosis in vivo. Since hypoxia has been shown to enhance the release of NE, MPO, and MMP9 (149), it seemed plausible that hypoxic neutrophils would have enhanced capacity to lead to endothelial cell dysfunction and/or damage. Based on the above reports, I explored the impact of hypoxic neutrophil supernatants on HPAEC adhesion molecules, detachment, and death/apoptosis. I found that hypoxic neutrophil supernatants led to increased ICAM-1 expression of HPAECs that had been cultured as monolayers in normoxic settings. Since neutrophils circulate between the arterial and venous circulations, cells exposed to hypoxia will encounter normoxic endothelial cells. This would be more marked in the setting of I/R injury, when vessel occlusion may ‘trap’ cells in a hypoxic environment before perfusion is restored by physiological processes or therapeutic interventions (e.g. stent placement). Additionally, in the setting of local or systemic infection or inflammation, neutrophils may be delayed in the microvascular circulation. Whilst I have used higher concentrations of PAF than are usually present in the circulation, inflamed endothelial cells express PAF (329) and hence the local exposure of adherent neutrophils may be significant. Furthermore, multiple inflammatory mediators circulate in the setting of infection and inflammation, and these may have additive effects, although I have not explored this possibility. The effect of hypoxia in attenuating the increase in endothelial ICAM-1 seen in response to treatment with supernatants from hypoxic PAF and fMLP-stimulated neutrophils is somewhat surprising and the mechanism for this has not been explored, although it is possible that the hypoxic exposure had a detrimental effect on endothelial cells which was not obvious by light microscopic visualisation. In the literature, reports of endothelial ICAM-1 expression in response to hypoxia are conflicting: Antonova et al. showed no difference in endothelial cell ICAM-1 levels cultured under hypoxic conditions (155), whereas Yoon et al. demonstrated increased ICAM-1 in a murine hindlimb ischaemia model (356) and Pinsky et al. demonstrated hypoxia-induced exocytosis of Weibal-Palade bodies (storage granules containing pre-formed P-selectin) from endothelial cells in a rodent model of cardiac ischaemia, which may mediate enhanced neutrophil adhesion (245). Such discrepancies may reflect species differences, in vitro vs in vivo experimental approaches or the degree and duration of hypoxia. Existing evidence suggests possible synergy between hypoxia and PAF signalling in mediating neutrophil-induced endothelial dysfunction: hypoxia increased ICAM-1-mediated neutrophil adherence to HUVECs in a PAF-dependent manner in vitro (293) and, similarly, 117 hypoxia increased neutrophil adhesion to porcine coronary endothelium in a PAF-dependent manner, although ICAM-1 was not assessed (357); PAF receptor antagonism was cardioprotective following myocardial I/R in mice (344); and blockade of ICAM-1 or the PAF receptor reduced neutrophil-endothelial adhesion and transmigration in a mouse model of mesenteric I/R injury (358). These observations suggest it would be relevant to further explore the impact of hypoxia on HPAEC expression of ICAM-1; of interest, a preliminary experiment suggested that the addition to hypoxic endothelial cells of supernatants derived from hypoxic stimulated neutrophils primed with GM-CSF rather than PAF in fact led to an increase in ICAM- 1 expression. This intriguing result needs to be further defined; it may suggest that the secretome of hypoxic and primed neutrophils varies according to the specific priming agent. This would align with and add to data presented in the following chapter, analysing the hypoxic neutrophil secretome in more detail. As noted above, numerous studies have demonstrated neutrophil-mediated endothelial damage. My data reveal that supernatants from hypoxic neutrophils cause more endothelial cell detachment and death in vitro than do matched supernatants generated from normoxic cells. Aligning well with my results, Vercellotti et al. showed that PAF and fMLP-treated neutrophils caused detachment of cultured human endothelial cells (359), Koshio et al. demonstrated endothelial apoptosis due to degranulation from cytochalasin B and fMLP- treated neutrophils (360) and Jerke et al. showed endothelial cytoskeletal architecture disruption from ANCA-stimulated neutrophil supernatants (361). Despite neutrophil-mediated endothelial cell detachment and apoptosis being well established, these studies were all conducted under conditions of ambient oxygen; the novelty of my data derives from the inclusion of hypoxia, which has relevance to a range of pathological situations as noted above. Neutrophil proteases have been implicated in mediating endothelial damage, for example in hypoxia-induced PH, with some protection afforded by in vivo serine protease inhibition (362). It is intriguing that in my experiments, endothelial injury could not be completely reversed by α1AT, indicating a small but significant protease-independent damage component. It is possible that the decreased survival recalcitrant to α1AT, observed at 48 but not 24 h, was partly due to protease-independent cell detachment, particularly given I used serum-free conditions to exclude the α1AT present in serum. Additionally, the degree of endothelial death did not fully correlate with NE activity in the applied supernatants, for example supernatants from hypoxic unstimulated neutrophils caused a similar degree of endothelial cell apoptosis as supernatants from hypoxic stimulated neutrophils. It would be informative to establish whether protease inhibition is able to abrogate the increase in endothelial cell apoptosis caused by supernatants from both stimulated and unstimulated hypoxic neutrophils, although time constraints prevented further exploration of the mechanism of this injury during my PhD. 118 However, when taken together, these results suggest that agents in addition to proteases contribute to neutrophil-mediated endothelial dysfunction. This is entirely plausible as neutrophil granules have many potentially damaging proteins that do not have protease activity. I therefore wished to interrogate the hypoxic neutrophil secretome in order to identify which factors may be responsible for the enhanced endothelial damage. Chapter 4 Results: Effect of hypoxia on the neutrophil secretome 120 4 Effect of hypoxia on the neutrophil secretome 4.1 Introduction Proteomics is the analysis of the entire protein complement (proteome) of a biological system, such as a cell or biological fluid, in a defined setting. As the proteome is dynamic and potentially modulated by multiple factors (e.g. source and associated environmental stimuli), it is important to interpret proteomic studies in the context of precise study conditions. Having identified that supernatants from hypoxic neutrophils cause extensive endothelial cell damage, which is not completely prevented by serine protease inhibition (section 3.5), I wished to perform an independent new quantitative proteomic characterisation of the hypoxic vs normoxic neutrophil secretome (secreted proteins). This would be entirely novel, would identify differentially regulated proteins with the potential to contribute to this supernatant- induced endothelial dysfunction, and might also yield insights into the mechanism(s) by which hypoxia regulates protein secretion. Separation techniques for complex protein/peptide samples in proteomic studies include gel electrophoresis, in one (separation by molecular weight) or two (separation by pH and molecular weight) dimensions, and high performance liquid chromatography (HPLC: separation by variable interactions with adsorbent material within a column), with subsequent identification by mass spectrometric (MS) analysis. Simultaneous labelling, e.g. by 2D difference gel electrophoresis (2D DIGE) or multiplex tandem mass tag mass spectrometry (TMT-MS), allows quantification of relative protein abundance between concurrently analysed samples. 2D DIGE enables a fairly rapid global view of the proteome, with differential protein abundance quantified by comparing two fluorescent dyes within the same gel. Differences in sample protein abundance from more than one experiment can be quantified by inter-gel analysis. However, this method can be laborious and expensive if MS is subsequently used to identify proteins present in individual spots, and there are limitations of 2D DIGE in displaying the complete range of proteins within a sample. Furthermore, spot fusion or co- migration can lead to ambiguous identification, and quantification is difficult if proteins are present in multiple isoforms. Multiplex TMT-MS involves labelling of individual protease- digested samples with unique isobaric tags, which give rise to a cleavage product with a specific mass-to-charge ratio for each sample. In this manner, up to 10 samples can be subjected to MS simultaneously, eliminating the technical variability arising from multiple MS runs. The abundance of all identified proteins can then be quantified and compared for each sample. However, it is time-consuming and requires a substantial amount of protein per sample. Since 2D DIGE requires less protein, I chose to perform an initial 2D DIGE study to give an overview of the normoxic vs hypoxic secretome, enabling rapid demonstration of any 121 obvious differences in protein abundance, before proceeding to protein identification by 10- plex TMT-MS for characterisation of the neutrophil secretome. In the context of my experimental setup, proteomic analysis posed two significant challenges. Firstly, I needed to ensure that secreted neutrophil proteases did not degrade other proteins within the supernatant prior to examination. This is because MS peptide identification is dependent on the predicted molecular weight of peptides generated by digestion of the intact full length protein with a specific protease (usually trypsin). Therefore, samples partially degraded by other proteases present would not give rise to the expected peptides and would hinder protein target identification. As it was therefore important to prevent digestion of proteins within the sample by other abundant secreted proteases, such as NE, I carefully optimised the anti-protease strategy for neutrophil supernatant generation. Secondly, I needed to generate supernatant samples with sufficient protein content (a minimum of 25 μg) and concentration (≥2 mg/ml) to allow TMT labelling. As the neutrophil supernatant samples were too dilute (1-4 μg/ml) to be used in the proteomic experiment directly, I needed to optimise the protein concentration method, maximising efficiency, accuracy, consistent protein recovery between samples and compatibility with the subsequent proteomic analysis. Although proteomics is a powerful investigative tool, analysis of the vast amount of data produced by tandem MS is not straightforward. A critical step is that of protein inference, i.e. assembling peptides identified from tandem MS spectra into protein lists. Not only can ambiguity arise from peptide sequences which are shared between different proteins or protein isoforms, but comparison of database search engines and protein inference software has demonstrated variable results (363). As it is therefore vital to validate proteomic data by independent replication of the experiment using a different experimental strategy, I selected a number of differentially regulated protein candidates, identified by TMT-MS, to take forward for validation by ELISA. Analysis of the pattern of significantly upregulated proteins in the hypoxic supernatants provided by the proteomic study demonstrated that these proteins did not entirely segregate with discrete granule populations, suggesting that an alternative secretion mechanism was involved. Neutrophils are known to produce MVs, which contain small packages of proteins, and secretome proteins may derive from MV release as well as degranulation. Therefore, I investigated the impact of hypoxia on the release of NDMVs and whether this mechanism was, at least in part, responsible for the differential release under hypoxia vs normoxia of proteins identified in the proteomic study. 122 The specific aims of the work presented in this chapter are: 1. To provide a preliminary overview of the differential protein release under hypoxia vs normoxia using 2D DIGE. 2. To optimise the anti-protease and protein concentration strategies for neutrophil supernatant generation for TMT-MS proteomic analysis. 3. To quantitatively identify proteins differentially released by neutrophils under hypoxia vs normoxia using 10-plex TMT-MS. 4. To validate the proteomic results by alternative biochemical methods for selected differentially regulated proteins. 5. To explore possible mechanisms of enhanced protein secretion from hypoxic neutrophils. 4.2 2D Difference Gel Electrophoresis Before undertaking expensive and complex whole secretome mass spectrometric analysis, a pilot 2D DIGE study without MS protein identification, which required less sample, was performed to confirm that I could detect differential protein release between hypoxic and normoxic samples. Because the supernatants are relatively dilute, a concentration step was required. For 2D DIGE, proteins must be natively charged in order to undergo fluorescent labelling and isoelectric focusing (IEF). Therefore, negatively-charged SDS buffer and conventional protein- denaturing TCA precipitation could not be used (see Table 4.1). Ammonium acetate precipitation and a urea-based buffer were therefore trialled alongside the TCA precipitation/SDS buffer method to test whether they achieved similar protein recovery. Furthermore, small molecule protease inhibitors were used rather than inhibitors that are proteins themselves, such as α1AT and leupeptin, to avoid the latter obscuring other supernatant proteins in the 2D DIGE analysis. Hence, neutrophils were incubated under normoxia or hypoxia in the presence of the protease inhibitors EDTA (2 mM), EGTA (2 mM), the NE-selective inhibitor sivelestat (1 µM) and serine protease inhibitor AEBSF (1 mM). Inhibitors were added at these concentrations at baseline, 2 h and 4 h. Cells were then treated with PAF (1 μM, 5 min) and fMLP (100 nM, 10 min). Supernatants were precipitated with either TCA (section 2.2.5.1) or ammonium acetate (section 2.4.2). Protein pellets from TCA- or ammonium acetate-precipitated supernatants were re-suspended in either SDS- or urea- based buffers, prior to SDS-PAGE and silver staining. Silver gel analysis demonstrated broadly similar protein recovery with both precipitation methods and both buffers. (Figure 4.1), 123 Table 4.1: Comparison of protein concentration methods Protein concentration method Advantages Disadvantages TCA-precipitation Good recovery when concentrating small sample volumes (less than 1 ml) and low protein amounts Fast and reproducible Can resuspend final sample in below 20 μl which is convenient for gel samples Requires multiple high speed centrifugation steps therefore not convenient for concentrating large sample volumes Denatures proteins, therefore not compatible with 2D DIGE Pellet often requires addition of detergent to dissolve into aqueous buffer Ammonium acetate- precipitation Non denaturing Pellet is more easily solubilised than TCA-precipitated pellet Requires multiple high speed centrifugation steps therefore not convenient for concentrating large sample volumes Requires large (5x) volumes of ammonium acetate which may lead to inaccurate recovery with large sample volumes Requires overnight precipitation Spin column Non denaturing Good for handling large volumes Difficult to get the final concentration below 100 μl without compromising the recovery, hence not good for routine SDS-PAGE during optimisation experiments Lower recovery when concentrating small sample volumes or low protein amounts Expensive relative to TCA and ammonium acetate precipitation methods 124 Therefore, neutrophil supernatants were prepared for 2D DIGE by ammonium acetate precipitation and re-suspension in urea buffer. Of note, there did not appear to be more protein in the hypoxic than normoxic samples and in some lanes (e.g. TCA precipitation/SDS buffer) there was a suggestion of lesser protein recovery from the hypoxic samples; this guided subsequent protease inhibitor strategies (section 4.3). Figure 4.1: The effect of buffer and precipitation on neutrophil supernatant protein content Neutrophils were re-suspended in IMDM under normoxia (N) or hypoxia (H) in the presence of EDTA (2 mM), EGTA (2 mM), AEBSF (1 mM) and sivelestat (1 μM). Protease inhibitors were also added to all samples at these concentrations after 2 h and 4 h incubation, giving a final concentration of 3 mM AEBSF and 3 µM sivelestat prior to stimulation. Cells were then treated with PAF (1 μM, 5 min) and fMLP (100 nM, 10 min) (P). Neutrophil supernatants were TCA- (T) or ammonium acetate- (A) precipitated and re-suspended in SDS- (S) or urea- (U) based buffer before SDS-PAGE and silver staining. Image from n=1 experiment. Neutrophil supernatants were prepared exactly as described in section 2.4.2. 2D DIGE was performed by Renata Feret (CCP). Equal amounts of both samples were labelled with cyanine fluorescent dyes; “normoxia activated” (i.e. generated from normoxic PAF and fMLP-treated cells) was labelled with Cy5, and “hypoxia activated” (i.e. generated from hypoxic PAF and fMLP-treated cells) labelled with Cy3. The two samples were subsequently mixed and subjected to IEF, followed by SDS-PAGE. Fluorescent images were generated, with 125 “normoxia activated” visualised as red (Figure 4.2A) and “hypoxia activated” visualised as green (Figure 4.2B). Overlaid scanned images (Figure 4.2C) were analysed with DeCyder 2D software. The fold change threshold of the volume ratio between automatically detected (and manually checked) protein spots in the two overlaid images was set at 2.0. Gel analysis indicated a number of proteins which were differentially abundant when comparing normoxic vs hypoxic supernatants (Figure 4.2C&D), suggesting that more detailed protein identification and quantification by TMT-MS would be both feasible and informative in showing differentially regulated protein release under hypoxia. 4.3 Optimisation of neutrophil supernatant generation for TMT-labelled MS Since neutrophil supernatants contain active proteases with the ability to cleave other proteins present in the sample which could thus compromise MS identification, anti-protease strategies were necessary, in particular to combat the enhanced protease release following hypoxic incubation. As noted above, the hypoxic samples prepared for 2D-DIGE did not contain a higher protein content than the normoxic equivalents as might be expected, hence further optimisation was undertaken. Initially, a combination of several protease inhibitors was employed. Freshly isolated neutrophils were re-suspended in normoxic or hypoxic IMDM containing either: no protease inhibitors; α1AT (46 µg/ml) and a protease inhibitor cocktail tablet (cOmplete™ mini EDTA-free protease inhibitor cocktail); or a combination of α1AT (46 µg/ml), EDTA (20 mM), EGTA (20 mM), leupeptin (12.5 µg/ml), pepstatin A (12.5 µg/ml) and a protease inhibitor cocktail tablet. Cells were incubated under normoxia or hypoxia for 4 h before treatment with PAF and fMLP, or IMDM control, exactly as previously described. Supernatants were TCA precipitated, subjected to SDS-PAGE and silver stained (see section 2.2.5). Although ammonium acetate precipitation was needed for 2D DIGE sample preparation, for the following anti-protease optimisation silver gel experiments, TCA precipitation was used as the method was more rapid and required lower volumes (see Table 4.1). Silver staining confirmed better protein recovery with the addition of protease inhibitors (Figure 4.3). Protein recovery was best with α1AT and a protease inhibitor cocktail tablet, rather than the combination of α1AT, EDTA, EGTA, leupeptin, pepstatin A and a protease inhibitor cocktail tablet. EDTA and EGTA were employed to inhibit MMPs but as they chelate divalent cations (the mechanism of MMP inhibition), lower protein recovery in their presence at these high concentrations (20 mM) likely reflects inhibition of degranulation due to chelation of extracellular calcium. 126 Figure 4.2: Analysis of hypoxic versus normoxic neutrophil supernatant protein content by 2D DIGE Isolated neutrophils were re-suspended in IMDM under normoxia or hypoxia in the presence of EDTA (2 mM), EGTA (2 mM), AEBSF (1 mM) and sivelestat (1 µM). Protease inhibitors were also added to all samples at these concentrations after 2 h and 4 h incubation. Cells were then treated with PAF and fMLP exactly as described previously. Ammonium acetate- precipitated supernatants were fluorescently labelled with red Cy5 (normoxia) or green Cy3 (hypoxia). Mixed samples were subjected to IEF and SDS-PAGE (2D DIGE) and fluorescent images were produced from the scanned gel. A. Fluorescent image of normoxic supernatant proteins using Cy5 parameters (633nm excitation, 670nm emission, red). B. Fluorescent image of hypoxic supernatant proteins using Cy3 parameters (532nm excitation, 580nm emission, green). C. Fluorescent Cy5 and Cy3 images were overlaid for analysis; green spots represent proteins increased in hypoxia, red spots represent proteins increased in normoxia, yellow spots represent unchanged proteins. D: Differential In-gel Analysis (DIA) software analysis of protein spots. Yellow boundaries indicate proteins with a volume ratio >2.0. Images from n=1 experiment. 127 Figure 4.3: The effect of protease inhibitors on neutrophil supernatant protein content Neutrophils were re-suspended in IMDM under normoxia (N) or hypoxia (H) in the presence or absence of protease inhibitors: either α1AT (46 µg/ml), EDTA (20 mM), EGTA (20 mM), leupeptin (12.5 µg/ml), pepstatin A (12.5 µg/ml) and a protease inhibitor cocktail tablet (1 tablet/4 ml) (ALL); α1AT (46 µg/ml) and a protease inhibitor tablet (AAT); or no protease inhibitors (NIL). After 4 h, neutrophils were treated with PAF and fMLP (P), or IMDM control (C). TCA-precipitated neutrophil supernatants were subjected to SDS-PAGE and silver stained. Representative image from n=2 experiments. Having demonstrated better supernatant protein recovery in the presence of protease inhibitors, small molecule protease inhibitors were trialled, to prevent inhibitors which are themselves proteins (α1AT, leupeptin and pepstatin A) obscuring supernatant proteins during proteomic analysis. The following inhibitors were used: EDTA, EGTA, sivelestat and AEBSF. The previous experiment had suggested reduced protein recovery in the presence of high concentrations of EDTA/EGTA, and EGTA chelation of extracellular calcium has been shown to reduce/inhibit neutrophil degranulation (149). Therefore, as IMDM contains 1.5 mM CaCl2 and 0.8 mM MgSO4, and chelation occurs in a 1:1 stochiometric ratio, EDTA and EGTA were used at 0.5 mM, in order to exert MMP inhibition whilst also allowing calcium-dependent granule exocytosis. Analysis of supernatant NE by Enzchek® assay confirmed the expected level of NE activity (i.e. degranulation) in the presence of EDTA and EGTA (both 0.5 mM), which was inhibited at higher concentrations (Figure 4.4A). 128 Isolated neutrophils were re-suspended in normoxic or hypoxic IMDM, containing a combination of EDTA (0.5 mM), EGTA (0.5 mM), sivelestat (10 µM) and AEBSF (2 mM). Higher concentrations of sivelestat and AEBSF were used here as it had been noted in previous experiments that protein recovery in hypoxic supernatants seemed less than in normoxic samples (Figure 4.1), thought to be due to inadequate protease inhibition under hypoxia. After 4 h, cells were treated with PAF and fMLP, or IMDM control. Neutrophil supernatants were TCA-precipitated, subjected to SDS-PAGE and silver stained. Surprisingly, despite increased sivelestat and AEBSF concentrations, silver staining now consistently demonstrated a marked reduction of protein in the hypoxic compared with normoxic supernatants, both under control and stimulated conditions, despite demonstration of equal protein loading across samples by western blotting for β-actin (Figure 4.4B). To further investigate this difference in protein recovery, cytospins of normoxic vs hypoxic PAF and fMLP-treated neutrophils, incubated with or without these small molecule protease inhibitors, were generated for morphological assessment of the neutrophil pellets. Substantial cell death was evident in the presence of protease inhibitors, particularly in the normoxic incubation conditions (Figure 4.5A). This suggested that the increased protein content in normoxic supernatants was due to cell death and subsequent protein release, rather than decreased secretion or lower protein recovery from the hypoxic samples. Cytospins of neutrophils exposed to AEBSF, EDTA, EGTA or sivelestat individually, revealed that AEBSF was toxic to neutrophils (Figure 4.5B). Although 2D DIGE was performed prior to the discovery of AEBSF-induced neutrophil toxicity, it was associated with a lesser difference in protein content between conditions and was still thought to be informative. However, I wished to eliminate the use of AEBSF before proceeding with TMT-MS. To avoid AEBSF-induced cell death, neutrophil supernatants were generated using sivelestat and EDTA only. EDTA alone was chosen, rather than the previous combination of EDTA and EGTA, because EGTA preferentially binds Ca2+ over Mg2+, with the potential for increased inhibition of degranulation. Neutrophils were re-suspended in normoxic or hypoxic IMDM containing either no protease inhibitors (for subsequent assessment of degranulation), or a combination of EDTA (1 mM), and sivelestat (10 µM). After 4 h, cells were treated with PAF and fMLP, or IMDM control. Neutrophil supernatants were TCA-precipitated, subjected to SDS-PAGE and silver staining. Silver staining (Figure 4.6A) demonstrated improved recovery of protein with this combination of protease inhibitors, particularly in the setting of hypoxia. Cytospins of pelleted cells re-suspended in IMDM confirmed viable cells (Figure 4.6B). Elastase assay confirmed inhibition of NE in the presence of EDTA and sivelestat (Figure 4.6C), whilst MPO assay confirmed that degranulation was not inhibited (Figure 4.6D). These optimised conditions were chosen to take forward for TMT-labelled MS. 129 Figure 4.4: The effect of small molecule protease inhibitors on neutrophil supernatant protein content Neutrophils were re-suspended in IMDM under normoxia (N) or hypoxia (H) in the presence of A: EDTA and EGTA at the concentrations indicated or B: a combination of EDTA (0.5 mM), EGTA (0.5 mM), sivelestat (10 µM) and AEBSF (2 mM). A: After 4 h all cells were treated with PAF and fMLP and supernatants were assessed for NE activity by Enzchek® assay. B: After 4 h, neutrophils were treated with PAF and fMLP (P), or IMDM control (C) and supernatants were subjected to SDS-PAGE and silver staining (upper panel). Corresponding cell pellets were lysed and probed for β-actin by western blotting (lower panel). A: n=2-3, B: Images represent n=3 experiments. 130 Figure 4.5: The effect of small molecule protease inhibitors on neutrophil morphology Neutrophil cytospins were stained with May-Grünwald-Giemsa. Cells were re-suspended in normoxic (N) or hypoxic (H) IMDM in the presence (+) or absence (-) of protease inhibitors (PI: EDTA (0.5 mM), EGTA (0.5 mM), sivelestat (10 µM) and AEBSF (2 mM)). After 4 h, neutrophils were treated with PAF and fMLP (P). A: incubation with/without PI. B: incubation with individual inhibitors as indicated. A&B: x40 magnification, arrows indicate examples of cell death. Representative images from n=1 experiment with 2-3 replicates per condition. 131 Figure 4.6: Optimised protease inhibition strategy Neutrophils were re-suspended in IMDM under normoxia (N) or hypoxia (H) in the presence (+) or absence (-) of sivelestat (10 µM) and EDTA (1 mM). After 4 h, cells treated with PAF and fMLP (P), or IMDM control (C). A: TCA-precipitated supernatants were subjected to SDS- PAGE and silver stained. B: Cytospin stained with May-Grünwald-Giemsa was generated from pelleted neutrophils incubated with sivelestat and EDTA; x40 magnification, scale bar represents 10 µm. C: Supernatant NE activity was measured by Enzchek® assay. D: MPO activity was measured by DMB-based MPO assay. A&B: representative images from n=2 experiments, C&D: Results represent mean ± SEM, n=3; *=p<0.05, **=p<0.01; two way ANOVA, Sidak’s multiple comparisons test 132 As TMT-MS required a minimum of 25 μg protein to allow effective sample labelling, the protein concentration method was evaluated. As each 1 ml supernatant sample contained between 1-2 μg protein, 24 replicates were prepared per condition (the maximum capacity of the thermomixer), with a resultant 24 ml produced. Since TCA and ammonium acetate precipitation methods are not ideal for handling large sample volumes (Table 4.1), spin columns were deemed the best option. A test sample of pooled neutrophil supernatant was spin-column concentrated and compared with TCA-precipitation by subsequent SDS-PAGE and silver staining. Spin column concentration showed similar protein recovery when compared with previous gels, which was consistent between replicate samples, whereas the TCA-precipitated pellet was not re-suspended properly (a recognised problem) (Figure 4.7). Hence, spin column protein concentration was used to prepare samples for TMT-MS. Figure 4.7: Comparison of TCA-precipitation and spin column protein concentration Supernatants from normoxic and hypoxic, PAF and fMLP-treated neutrophils (in the presence of sivelestat (10 µM) and EDTA (1 mM)) were pooled. The combined sample was either TCA- precipitated (500 μl) or concentrated using a spin column with a molecular weight cut off of 3 kDa (3 ml). The equivalent of 500 μl original pooled supernatant was loaded per well, subjected to SDS-PAGE and silver stained. Replicate samples generated from one spin column were compared. Representative image from n=2 experiments. 4.4 Characterisation of the normoxic versus hypoxic neutrophil secretome by 10-plex tandem mass tag-labelled mass spectrometry (TMT-MS) Informed by preliminary 2D DIGE analysis showing evidence of differential protein release under hypoxia vs normoxia and by the optimisation strategy outlined above, 10-plex TMT-MS was undertaken to fully characterise the hypoxic vs normoxic neutrophil secretome of five healthy donors simultaneously. Isolated neutrophils were re-suspended in IMDM under 133 normoxia or hypoxia in the presence of EDTA and sivelestat and treated with PAF and fMLP. For each experiment, cells in one 2 ml Eppendorf were incubated without protease inhibitors and assessed for NE release by Enzchek® assay to ensure a hypoxic uplift of elastase release. The remaining supernatants per condition were pooled in order to generate sufficient protein for TMT-MS labelling. The proteins in the pooled supernatant samples were concentrated using spin columns with a molecular weight cut-off of 3 kDa. Equal amounts of each sample were prepared for TMT-MS by Renata Feret (CCP). 10-plex TMT-MS was performed by Dr Mike Deery (CCP) and the raw data were analysed by Dr Marco Chiapello (CCP), with the parameters described in section 2.4.3 and appendix 7.1. Satisfactory TMT labelling efficiency (96.95%, measured by CCP) was achieved (see section 7.1.2). PCA (a statistical technique used to visualise data variance, section 7.1.2) indicated separation of the normoxic and hypoxic samples by PC1 (which accounts for the largest data variability, Figure 4.8A). A total of 1245 proteins were identified by MS, 717 of which were present in all 10 samples. Of these 717 proteins, 199 had both a false discovery rate (i.e. the proportion of false peptide spectrum matches, assessed by comparison with a decoy protein database, FDR) of <0.01, and were also differentially regulated between normoxia and hypoxia (see section 7.2, Table 7.1 (normoxia upregulated) and Table 7.2 (hypoxia upregulated)). 63 of these proteins were significantly changed (adjusted p value<0.05) between normoxia and hypoxia; 35 were increased in normoxia (Table 4.2) and 28 were increased in hypoxia (Table 4.3). The 199 differentially regulated proteins with FDR <0.01 are represented by volcano plot (Figure 4.8B). These data confirmed that the granule protein MPO (log2FC 1.64, adj. p value=0.006), previously shown by alternative biochemical assays to be increased in hypoxic neutrophil supernatants (149), was also increased in hypoxic supernatants when examined by proteomic methods, although NE (log2FC 0.68, adj. p value=0.25) was not significantly increased in this experimental setting (discussed in section 4.7). Additional granule proteins, including NGAL (log2FC 0.88, adj. p value=0.004) and resistin (log2FC 0.92, adj. p value=0.004), were also increased in hypoxic supernatants. Surprisingly, some proteins which were increased under hypoxia are predominantly cytoplasmic, e.g. cyclophilin A (log2FC 0.52, adj. p value=0.03), and although more cytoplasmic proteins were detected at higher levels in normoxic supernatants, some granule-associated proteins were increased in the normoxic relative to the hypoxic samples, e.g. leukocyte elastase inhibitor (log2FC 0.86, adj. p value=0.006) and leukocyte specific protein 1 (log2FC 0.91, adj. p value=0.01). Taken together, these data suggest that, whilst the release of the majority of granule proteins is increased in hypoxia, the differential protein release may not be purely mediated by enhanced degranulation and that alternative mechanisms warrant consideration. 134 Table 4.2: Proteins significantly increased in normoxia Proteins significantly increased in normoxia (adj. p value <0.05) which were present in all 10 samples with a FDR <0.01 are listed in order of the magnitude of the fold change (FC). The p value was adjusted by the Benjamini-Hochberg FDR correction for multiple comparisons. Accession Description Adj. p value FC Location MV presence Q5TCU8 Tropomyosin beta chain 0.001 11.855 CYT N P10599 Thioredoxin 0.017 3.909 CYT N E7EX29 14-3-3 protein zeta/delta 0.015 3.844 S/G Y P52566 Rho GDP-dissociation inhibitor 2 0.004 3.802 CYT Y P08670 Vimentin 0.006 3.766 CYT Y O00299 Chloride intracellular channel protein 1 0.030 2.918 ? Y P11021 78 kDa glucose-regulated protein 0.009 2.873 CYT N E9PK25 Cofilin-1 0.023 2.753 CYT Y E7EMB3 Calmodulin 0.026 2.488 ? Y P62993 Growth factor receptor-bound protein 2 0.004 2.478 CYT N P20700 Lamin-B1 0.004 2.435 NUC Y Q32MZ4 Leucine-rich repeat flightless-interacting protein 1 0.004 2.337 NUC/ CYT N P06737 Glycogen phosphorylase, liver form 0.025 2.217 ? Y P32942 Intercellular adhesion molecule 3 0.015 2.149 S/G Y Q9Y490 Talin-1 0.015 2.116 S/G Y O15144 Actin-related protein 2/3 complex subunit 2 0.031 2.086 CYT Y P06702 Protein S100-A9 0.013 2.084 CYT Y P52209 6-phosphogluconate dehydrogenase, decarboxylating 0.013 1.950 CYT Y P26038 Moesin 0.019 1.878 S Y P33241 Leukocyte-specific protein 1 0.013 1.878 S/G Y P18206 Vinculin 0.015 1.816 CYT Y P30740 Leukocyte elastase inhibitor 0.006 1.813 A Y Q96C19 EF-hand domain-containing protein D2 0.027 1.794 ? Y A6NIZ1 Ras-related protein Rap-1b-like protein 0.009 1.749 SV Y P35579 Myosin-9 0.012 1.631 CYT Y Q29963 HLA class I histocompatibility antigen, Cw-6 alpha chain 0.039 1.627 PM N P61247 40S ribosomal protein S3a 0.028 1.528 ? N P02042 Hemoglobin subunit delta 0.049 1.492 ? Y E7EQR4 Ezrin 0.023 1.473 ? N P08133 Annexin A6 0.035 1.458 ? Y P31146 Coronin-1A 0.034 1.451 S/G Y Q15907 Ras-related protein Rab-11B 0.015 1.447 G N P46781 40S ribosomal protein S9 0.044 1.438 CYT N P46940 Ras GTPase-activating-like protein IQGAP1 0.049 1.411 G Y P39687 Acidic leucine-rich nuclear phosphoprotein 32 family member A 0.043 1.358 ? N 135 Table 4.3: Proteins significantly increased in hypoxia Proteins significantly increased in hypoxia (adj. p value <0.05) which were present in all 10 samples with a FDR <0.01 are listed in order of the magnitude of the fold change (FC). The p value was adjusted by the Benjamini-Hochberg FDR correction for multiple comparisons. Accession Description Adj. p value FC Location MV presence P62805 Histone H4 0.004 3.587 NUC Y P05164 Myeloperoxidase 0.006 3.107 A Y A6NC48 ADP-ribosyl cyclase/cyclic ADP- ribose hydrolase 2 0.025 2.468 SV Y O75083 WD repeat-containing protein 1 0.029 2.382 CYT Y Q92820 Gamma-glutamyl hydrolase 0.020 2.143 S Y P02788 Lactotransferrin 0.033 2.136 S Y A5A3E0 POTE ankyrin domain family member F 0.015 2.074 CYT N Q0VD83 Apolipoprotein B receptor 0.025 1.993 S/G Y P10124 Serglycin 0.012 1.904 CYT N Q9HD89 Resistin 0.004 1.898 A/S Y P07737 Profilin-1 0.015 1.890 CYT Y A0A087WXL1 Folate receptor gamma 0.029 1.877 S Y P11215 Integrin alpha-M 0.004 1.872 S Y P16035 Metalloproteinase inhibitor 2 0.034 1.853 G N X6R8F3 Neutrophil gelatinase-associated lipocalin 0.004 1.842 S Y V9GYM3 Apolipoprotein A-II 0.049 1.769 ? N G3V3D1 Epididymal secretory protein E1 0.007 1.765 A N P10153 Non-secretory ribonuclease 0.029 1.650 ? Y P04217 Alpha-1B-glycoprotein 0.015 1.594 ? Y P05107 Integrin beta-2 0.018 1.580 G Y P01024 Complement C3 0.032 1.556 ? Y P20061 Transcobalamin-1 0.042 1.492 G Y P78324 Tyrosine-protein phosphatase non-receptor type substrate 1 0.018 1.449 SV N J3KNB4 Cathelicidin antimicrobial peptide 0.039 1.446 S/G Y P62937 Peptidyl-prolyl cis-trans isomerase A/Cyclophilin A 0.035 1.438 CYT Y P30086 Phosphatidylethanolamine- binding protein 1 0.049 1.406 ? N A0A075B6H6 Ig kappa chain C region 0.030 1.358 ? Y A0A075B6K9 Ig lambda-2 chain C regions 0.049 1.305 ? N Tables 4.2 and 4.3: Location data were compiled using the following references: (32,33,42,306,318,364,365) and the Uniprot database (www.uniprot.org). Presence (Y) or absence (N) of proteins within NDMVs was confirmed by searching supplementary data tables from the following references: (95,366). Abbreviations: azurophil (A), specific (S), gelatinase (G) granules; secretory vesicles (SV); cytoplasm (CYT); nucleus (NUC); plasma membrane (PM). For some proteins, the location within neutrophils is currently uncertain/unknown (?). 136 Figure 4.8: Characterisation of the normoxic versus hypoxic neutrophil secretome by TMT-MS Isolated neutrophils were re-suspended in IMDM under normoxia or hypoxia in the presence of EDTA (1 mM) and sivelestat (10 µM) for 4 h and then treated with PAF and fMLP. Spin column-concentrated supernatants were trypsin-digested and individually labelled with unique 10-plex isobaric tags. The mixed sample was subjected to MS/MS. A: PCA of normoxic (N, blue) and hypoxic (H, red) supernatants, with normoxic and hypoxic samples for individual donors linked by dotted lines. Dashed line indicates separation of normoxic vs hypoxic samples by PC1. B: volcano plot representation of differential protein expression between paired normoxic and hypoxic supernatants. Vertical dotted line represents log2fold change (FC) of protein abundance = ±1. Horizontal dotted line represents a significance level of adj. p value=0.05; proteins above this line are significantly changed (adj. p value < 0.05). Selected proteins are labelled in red. B: n=5; paired t-test, p value adjusted by the Benjamini-Hochberg false discovery rate correction for multiple comparisons. 137 4.5 Biochemical validation of differentially regulated proteins identified by TMT-MS Based on a combination of the magnitude of the increase, novelty and biological plausibility for a role (either injurious or protective) in endothelial dysfunction, five candidate proteins identified by TMT-MS were taken forward for validation in alternative biochemical assays. These candidates comprised three proteins which were increased in hypoxic supernatants (resistin, NGAL and cyclophilin A (peptidyl-prolyl cis-trans isomerase A)), and two proteins which were increased in normoxic supernatants (thioredoxin and S100A9). Firstly, regarding proteins identified as increased in the hypoxic vs normoxic secretome by TMT-MS: resistin was increased 1.9-fold (adj. p value=0.004), NGAL was increased 1.8-fold (adj. p value=0.004) and cyclophilin A was increased 1.4-fold (adj. p value=0.035). These targets were taken forward for validation by ELISA, using freshly generated supernatants from different donors. Isolated neutrophils were incubated under normoxia or hypoxia for 4 h before treating with PAF and fMLP. Supernatants were assessed for resistin (predominantly in azurophil granules), NGAL (predominantly in specific granules) and cyclophilin A (a cytoplasmic protein) content by commercial ELISA. Hypoxia increased the release of resistin (232 ± 30.36 pg/ml vs 173.3 ± 19.15 pg/ml, p=0.0209, Figure 4.9A), NGAL (4.764 ng/ml ± 0.647 ng/ml vs 2.875 ± 0.469 ng/ml, p=0.0053, Figure 4.9C) and cyclophilin A (0.0082 ± 0.0025 OD450nm vs 0.0 OD450nm [all readings below the threshold of detection], p=0.0171, Figure 4.9E) from PAF and fMLP-treated neutrophils when compared with normoxia. These results were fully in accordance with the proteomic data. Secondly, regarding proteins identified as increased in the normoxic vs hypoxic secretome by TMT-MS: thioredoxin was increased 3.9-fold (adj. p value=0.017) and S100A9 was increased 2.1-fold (adj. p value=0.013). These targets were taken forward for validation by ELISA. However, it was not possible to detect thioredoxin in any supernatant by ELISA, as all readings were below the lowest value set by the standards (156 pg/ml; assay range 156 pg/ml – 10,000 pg/ml). This was despite trialling neat neutrophil supernatants, the protein-concentrated supernatants remaining from the TMT-MS experiment, and supernatants in combination with the non-denaturing detergent, Triton-X100 (in case membrane lipoproteins were preventing protein-antibody binding). 138 Figure 4.9: The effect of hypoxia on resistin, NGAL and cyclophilin A release from neutrophils by ELISA Isolated neutrophils were re-suspended in IMDM under normoxia or hypoxia for 4 h and then treated with PAF and fMLP. Supernatants were harvested and assessed for A&B: resistin, C&D: NGAL or E: cyclophilin A content by commercial ELISA. Data from A&C are represented in B&D to show individual paired normoxic and hypoxic samples for resistin (B) and NGAL (D) linked by dotted lines for each donor. Data are presented as A-D: supernatant protein concentration, E: absorbance at OD450nm minus baseline as ELISA readings for several samples were below the lowest value set by the standards (1.25 ng/ml; Elabscience ELISA assay range 1.25 ng/ml – 80 ng/ml). Results represent mean ± SEM; A: n=17, C&E: n=7; * = p<0.05, ** =p<0.01; paired t-test. 139 In contrast to the TMT-MS data, there was no significant difference between normoxia and hypoxia in the stimulated neutrophil supernatant content of S100A9 homodimer (6.606 ± 0.655 ng/ml vs 6.853 ± 0.681 ng/ml, p=0.6, Figure 4.10A) by ELISA. Since S100A9 preferentially forms a heterodimer with its binding partner S100A8 (which was not identified as differentially regulated by TMT-MS), both the heterodimer and S100A8 homodimer content of neutrophil supernatants were also quantified to see if this explained the lack of increase in S100A9 homodimer in normoxic samples. However, neither the S100A8A9 heterodimer (3.508 ± 0.221 µg/ml vs 3.262 ± 0.201 µg/ml, p=0.2, Figure 4.10B) nor the S100A8 homodimer (1.427 ± 0.243 ng/ml vs 1.289 ± 0.151 ng/ml, p=0.3, Figure 4.10C) were significantly different between normoxic vs hypoxic supernatants from PAF and fMLP-treated neutrophils by ELISA. Further examination by western blot of S100A9 content in supernatants from unstimulated or PAF and fMLP-treated neutrophils revealed that S100A9 release was significantly increased upon stimulation (5722 ± 1447 AU vs 13670 ± 1618 AU (normoxia), p<0.0001; 6528 ± 1619 AU vs 13825 ± 2094 AU (hypoxia), p=0.001, Figure 4.11B) but, again, there was no significant difference between normoxia and hypoxia (5722 ± 1447 AU vs 6528 ± 1619 AU (unstimulated), p=0.8; 13670 ± 1618 AU vs 13825 ± 2094 AU (PAF and fMLP), p=1, Figure 4.11B). These data do not align with those obtained from the proteomics study; possible reasons for this discrepancy are discussed in section 4.7. Given time constraints, I did not pursue additional studies of the secreted proteins identified by the proteomics study as increased in normoxia vs hypoxia. 140 Figure 4.10: The effect of hypoxia on S100A8 and S100A9 release from neutrophils by ELISA Isolated neutrophils were re-suspended in IMDM under normoxia or hypoxia for 4 h and then treated with PAF and fMLP. Supernatants were harvested and assessed for A: S100A9 homodimer, B: S100A8A9 heterodimer or C: S100A8 homodimer content by commercial ELISA. Data are presented as supernatant protein concentration. Results represent mean ± SEM; A&B: n=10, C: n=5, paired t-test. 141 Figure 4.11: The effect of hypoxia on S100A9 release from neutrophils by western blot Isolated neutrophils were re-suspended in IMDM under normoxia (N) or hypoxia (H) for 4 h and then treated with PAF and fMLP (P) or IMDM control (C). Supernatants were harvested and BSA (250 µg/ml) added to each sample. Equal volumes of supernatant were subjected to SDS-PAGE. Following protein transfer, the PVDF membrane was stained with Ponceau S, and western blotting was subsequently performed with anti-S100A9. A: representative western blot for S100A9 (upper panel) and corresponding Ponceau S-stained membrane (lower panel). B: Quantification of supernatant S100A9 by densitometry using Image J, corrected for loading by densitometry of Ponceau S-stained BSA. A: representative image from n=12 experiments. B: Results represent mean ± SEM; n=12; *** = p<0.001, **** = p<0.0001; two way ANOVA, Sidak’s multiple comparisons test. 142 4.6 The role of NDMV secretion in the differential protein release from neutrophils under hypoxia Having established that hypoxia enhances the release of proteins characteristic of neutrophil degranulation (section 3.3, reference (149)), it logically followed that a number of granule proteins, including NGAL and resistin, were increased in supernatants from hypoxic neutrophils. However, the apparent reduction of some granule proteins (e.g. leukocyte elastase inhibitor) and increase of a subset of cytoplasmic proteins (e.g. cyclophilin A) in hypoxic supernatants was intriguing. Neutrophils, among many other cell types, are able to release membrane-bound MVs, spherical structures less than 1 µm in diameter, which contain components derived from the parent cell, including proteins (274). Published proteomic data analysing NDMVs derived from normoxic cells has detected cyclophilin A, amongst other contents (94,95,366). I therefore hypothesized that the mechanism of enhanced release of cytoplasmic proteins from neutrophils under hypoxia was increased NDMV secretion. 4.6.1 Quantification of NDMV release from normoxic and hypoxic neutrophils To assess MV release from neutrophils under normoxia vs hypoxia, supernatants were subjected to sequential centrifugation, and NDMV content was measured by flow cytometry. Dr Victoria Ridger (University of Sheffield) has previously shown that optimal NDMV yields were obtained using Histopaque-1077® gradients to isolate neutrophils and a standard 1 h incubation with 10 μM fMLP (276). Neutrophils were isolated from whole blood using Histopaque®-1077 gradients (see section 2.5.1) and, concurrently from the same donor, by plasma-Percoll® gradients (since this method was used for the proteomic studies). Isolated neutrophils were re-suspended in normoxic or hypoxic IMDM (1*107/ml) and incubated for 1 or 4 h. In addition to the standard 4 h incubation, a 1 h timepoint was also chosen to make sure that any observed effects were not due to release of apoptotic bodies at the later time point and that the delicate NDMVs were not being lost following the more protracted incubation. Cells were treated with fMLP (10 µM) 1 h prior to the end of incubation. After treatment, cells were pelleted and supernatants sequentially centrifuged to pellet MVs (see section 2.5.1). Quantification of NDMVs was undertaken with the assistance of Merete Long (PhD student supervised by Dr Victoria Ridger, University of Sheffield), using established protocols (section 2.5.2). Pelleted NDMVs were re-suspended in sterile PBS with a known number of counting beads and quantified by flow cytometry. Although there was no significant difference in NDMV release between normoxia and hypoxia for either isolation method at 4 h, in fact NDMV release from Histopaque-1077® isolated neutrophils was higher in normoxia compared with hypoxia at the 1 h time point (NDMV count: 18340 ± 5109 vs 9616 ± 2795, p=0.0227, Figure 4.12A). These data refute the hypothesis that hypoxia increases NDMV release, although it must be acknowledged that the stimulation conditions were not identical. 143 Of note, in my hands the Histopaque-1077® isolated neutrophils were basally shape-changed (Figure 4.12B), suggesting priming/activation by the preparative method, but these cells also had a blunted response to PAF and fMLP-stimulated ROS production (Figure 4.12C, data generated with the assistance of Dr Arlette Vassallo). This may have affected the generation of MVs from Histopaque-1077® isolated neutrophils, and it is possible that the longer incubation period allowed reversion to a less activated (or de-primed) state. Importantly, these results emphasise that the plasma-Percoll® method isolates un-primed but responsive cells. 4.6.2 Evaluation of cyclophilin A protein content of NDMVs from normoxic and hypoxic neutrophils NDMVs have been shown to have myriad functions, with both pro- and anti-inflammatory effects (reviewed in (274)). Interestingly, studies of NDMV proteomes have identified distinct protein signatures which are dependent on the neutrophil stimulus and environment (94,95,366). Thus, I hypothesized that the protein composition of NDMVs is regulated by the normoxic or hypoxic environment, and that this controlled packaging contributes to the differential protein release. To investigate whether hypoxia modulates the incorporation of cyclophilin A into NDMVs, I performed western blotting for cyclophilin A and the established NDMV marker, annexin A1 (95,98). Having optimised the method using neutrophils isolated by Hisotpaque-1077® gradients, I performed western blotting on NDMV lysates from both Histopaque®-1077 and plasma-Percoll® isolated neutrophils, derived from the same donors in parallel, to ensure that results were not confounded by the neutrophil isolation method. Neutrophils were re- suspended in normoxic or hypoxic IMDM for 1 or 4 h (corresponding to the conditions previously chosen for flow cytometric analysis) and treated with fMLP 1 h prior to the end of incubation. Cells were then pelleted and the supernatants sequentially centrifuged to pellet MVs as before. After 4 h incubation, MV-deplete supernatants (generated by pelleting NDMVs in the final centrifugation step) were also harvested and then TCA-precipitated. NDMV lysates (section 2.5.3) and TCA-precipitated MV-deplete supernatants were subjected to SDS-PAGE, and western blotting performed for cyclophilin A and annexin A1 (Figure 4.13). Protein bands for annexin A1 (the double-band is standardly detected with this antibody) were visualised in NDMV lysates but not supernatants, demonstrating that the latter were not contaminated by NDMVs. Importantly, unlike annexin A1, cyclophilin A was detected in MV-deplete supernatants, suggesting that the release of cyclophilin A is at least partially independent of NDMV generation. 144 Figure 4.12: Comparison of neutrophil isolation methods and the effect of hypoxia on NDMV release Neutrophils were isolated using plasma-Percoll® or Histopaque-1077® gradients as indicated. A: Neutrophils were re-suspended in IMDM (1*107/ml) under normoxia or hypoxia. Cells were incubated for 1 or 4 h and treated with fMLP (10 µM) 1 h prior to the end of incubation. Cells were pelleted and supernatants sequentially centrifuged to pellet MVs. MV pellets were re- suspended in sterile PBS (190 µl) with counting beads (10 µl, 10,000 beads) and analysed by flow cytometry. The MV forward scatter vs side scatter gate was set using standardised size calibration beads, and 1,000 counting beads were measured per sample. B: Neutrophils were re-suspended in normoxic IMDM (5*106/ml) and treated with fMLP (100 nM) or PBS+/+ (control). Cells were fixed at baseline or after 30 min. Shape change was assessed by flow cytometric analysis of forward scatter (BD FACSCanto II). C: Neutrophils were re-suspended in normoxic PBS+/+ (5*106/ml) and treated with PAF (1 µM, 5 min) or PBS+/+ (unprimed) before addition of luminol (1 µM, 3 min) and HRP (62.5 units/ml). ROS production was analysed in a luminometer following injection of fMLP (100 nM). Data were generated with the assistance of Dr Arlette Vassallo. A&B: Results represent mean ± SEM; A: n=7, B: n=2-3. C: Image representative of n=2 experiments. A: *=p<0.05, two way ANOVA, Sidak’s multiple comparisons test. 145 There was no significant difference in NDMV annexin A1 content between normoxia and hypoxia for any condition (Figure 4.14A&B). Hypoxia, compared with normoxia, resulted in increased content of cyclophilin A in MV lysates derived from plasma-Percoll® isolated neutrophils at 1 h (28643 ± 16335 AU vs 16951 ± 15930 AU, p=0.0142, Figure 4.14D) but not at 4 h. There was no difference in cyclophilin A content for MV lysates derived from Histopaque®-1077 isolated neutrophils at either timepoint (Figure 4.14C). As annexin A1 is a surrogate for NDMV number, cyclophilin A densitometry was divided by the fold change in annexin A1 (relative to normoxic control at 1 h) to assess whether cyclophilin A was differentially packaged under normoxia vs hypoxia. However, there was no significant difference in cyclophilin A content between normoxia and hypoxia for any condition after this correction for NDMV number (Figure 4.14E&F). These data do not support the hypothesis that the differential release of cyclophilin A under hypoxia is due to increased NDMV protein content. Although an increase of cyclophilin A was observed at 1 h for MVs derived from plasma-Percoll® isolated neutrophils under hypoxia, there were no significant differences between normoxia and hypoxia at the 4 h (corresponding to the previous proteomic and ELISA assessment) timepoint. Additionally, there appeared to be more cyclophilin A in MV-free supernatants than NDMV lysates. Figure 4.13: Neutrophil-derived microvesicle and supernatant content of annexin A1 and cyclophilin A by western blot Neutrophils, re-suspended in normoxic (N) or hypoxic (H) IMDM (1*107/ml), were incubated for 1 or 4 h and treated with fMLP (10 µM) 1 h prior to the end of incubation. NDMVs were isolated at the 1 or 4 h timepoints as indicated by pelleting cells and sequentially centrifuging supernatants to pellet NDMVs. MV-deplete supernatants (SN) were also generated after the 4 h incubation. NDMV lysates and TCA-precipitated SN were subjected to SDS-PAGE, and western blotting was performed for annexin A1 and cyclophilin A. Representative images from n=4-5 (NDMV) and n=2 (SN) experiments. 146 Figure 4.14: Quantification of neutrophil-derived microvesicle content of annexin A1 and cyclophilin A by western blot Neutrophils were isolated by A,C,E: Histopaque-1077® gradients or B,D,F: plasma-Percoll® gradients. A-F: Neutrophils were re-suspended in IMDM (1*107/ml) under normoxia or hypoxia for 1 or 4 h and treated with fMLP (10 µM) 1 h prior to the end of incubation. Cells were pelleted and the supernatants sequentially centrifuged to pellet NDMVs. MV lysates were subjected to SDS-PAGE and probed for annexin A1 (A&B) and cyclophilin A (C&D) by western blotting. NDMV lysate annexin A1 and cyclophilin A content was quantified by Image J analysis of protein band densitometry. E&F: Cyclophilin A content was divided by the fold change in annexin A1 (relative to 1 h normoxia) to correct for NDMV number. Results represent mean ± SEM; A: n=6, B: n=5, C&E: n=5, D&E: n=4; * = p<0.05; two way ANOVA, Sidak’s multiple comparisons test. 147 4.7 Discussion The application of proteomics techniques to neutrophil biology has yielded detailed information on protein abundance in both healthy and disease states, with relevant functional implications. Studies of the whole cell proteome of healthy neutrophils after stimulation with PAF and fMLP (367), TNFα and/or GM-CSF (368), or ligation of the formyl peptide receptor like-1 receptor (a low affinity fMLP receptor) (369) have revealed multiple up- and down-regulated proteins when compared with quiescent neutrophils. Some granule proteins, e.g. MMP-9, were decreased in stimulated neutrophils, likely reflecting degranulation (369). Of relevance to my findings in Chapter 3 that hypoxia augments NE release from GM-CSF- but not TNFα-primed neutrophils, GM-CSF but not TNFα induced a discriminative neutrophil protein profile when compared with quiescent neutrophils by PCA (368). Studies of whole cell proteomes of circulating neutrophils from rats exposed to surgical trauma (370), and COPD (220) or trauma patients (371), compared with healthy controls, have similarly identified multiple differentially regulated proteins, with functional implications for cell survival, trafficking and activation. Interestingly, Langereis et al. found that the proteomic profile of stable COPD patient neutrophils did not match that of TNFα and/or GM-CSF-treated neutrophils from healthy controls using 2D DIGE, although protein identification was not performed in this study (220). The authors suggest that the disparity is due to the effect of an interplay of multiple inflammatory signals in COPD which does not reflect changes induced by treatment with single cytokines. Hypoxia (which was not assessed in this study) may also impact the neutrophil proteome, even in stable COPD. A few proteomic studies have examined the secretion profile of neutrophils. These studies comprise identification of proteins secreted by healthy isolated neutrophils treated with ionomycin (309), TNFα or cytochalasin B/fMLP (306), chromofungin or catestatin (antimicrobial peptides derived from Chromogranin A) (307), S. aureus leukotoxins (LukE/D) (308), or several strains of S. pyogenes (305,306). One study also examined the presence of known neutrophil-secreted proteins in the plasma of patients with bacterial sepsis compared with healthy controls (306). However, all of these experiments were performed under normoxic conditions, and no studies have investigated the effect of hypoxia on the neutrophil secretome. Since neutrophil supernatants contain multiple proteases which could hamper subsequent MS peptide identification, the anti-protease proteomics strategy was extensively optimised. The finding that AEBSF caused substantial neutrophil death was unexpected as, in contrast to PMSF, it is reported to be non-toxic. The product information data sheet recommends a working concentration of up to 2 mM, although data sheets from other companies suggest using a concentration of up to 0.25 mM in cell culture media. The literature on AEBSF-induced cytotoxicity is limited as AEBSF is normally employed as a protease inhibitor during the 148 preparation of cell lysates, hence short incubation times would normally be used, and a degree of cytotoxicity would likely not be apparent. However, one study has shown that AEBSF in solution at a concentration of >75 μM was highly cytotoxic to keratinocytes, with the authors suggesting that the mechanism was reactivity of the fluorosulfonyl group with molecules critical for cell survival (372). It is probable that the relatively high concentration of AEBSF and lengthy incubation times (> 4h, in order to counteract any proteases which may have been released from cells prior to stimulation, particularly under hypoxia) used for my initial experiments masked the subsequent difference between normoxic and hypoxic samples revealed by silver staining when using a lower concentration, which allowed a degree of hypoxic protection. Importantly, having recognised that AEBSF was toxic to neutrophils at the concentrations necessary for serine protease inhibition during supernatant generation, the anti-protease approach was adjusted to ensure neutrophil viability following inhibitor treatment. Having established a robust method for the preparation of neutrophil supernatants for TMT-MS, I have generated a novel and thorough description of the hypoxic vs normoxic neutrophil secretome. Previous published data from our laboratory (149) has shown an increase in the release of the granule proteins NE and MPO (azurophil granules), lactoferrin (specific granules) and MMP- 9 (gelatinase granules) from hypoxic neutrophils. Supporting these data, my secretome analysis showed a hypoxic upregulation of all of these proteins. However, the hypoxic increase in supernatant NE (p=0.25) and MMP-9 (p=0.08) content was not statistically significant (see Table 7.2). It is possible that the proteomic detection and analysis methods resulted in masking of some true positive identifications (373). Previous data for MMP-9 were generated from GM-CSF and fMLP-treated neutrophils, showing an approximate 3-fold increase in release under hypoxia by ELISA (149). It is possible that the discrepancy was due to the use of different priming agonists. Additionally, donor variability, as demonstrated by the ELISA data for normoxic vs hypoxic resistin release (Figure 4.9A), might necessitate an increased sample size to detect any difference; the use of five subjects was dictated by the ability to run 10 samples in parallel. I have previously shown a highly significant increase in supernatant NE activity from PAF and fMLP-treated neutrophils under hypoxia (Figure 3.2B). It may be that the activity assay is more sensitive than assessment of NE amount. Another potential explanation for discordance between the proteomic and activity assay data is that, rather than being entirely dependent on the amount of protein, NE activity may be influenced by other secreted proteins within the supernatant, such as leukocyte elastase inhibitor, which was decreased in hypoxia relative to normoxia by proteomic analysis. Although interaction of proteases and their physiological inhibitors is highly relevant in vivo, the use of sivelestat and EDTA was necessary in the setting of the proteomics study to prevent protein degradation by NE and MMPs prior to peptide analysis. However, neutrophils release other proteases which 149 could have been active in the proteomics study supernatants. PR3, another serine protease, is in fact inhibited by high (10 μM) concentrations of sivelestat (despite being NE-selective) as the half maximal inhibitory concentration (IC50) is 340 nM but the IC50 of sivelestat for cathepsin G is reported as >10μM. MPO, although not a protease, is able to oxidise certain amino acids, e.g. methionine, but these variations should not have influenced the TMT-MS results as variable oxidation modifications were built into the data analysis parameters. Overall, this discrepancy highlights the importance of performing additional biochemical assay confirmation. The specific targets identified by proteomic analysis which were selected for validation by ELISA were resistin, NGAL and cyclophilin A (increased in hypoxia), and thioredoxin and S100A9 (increased in normoxia). Both resistin (an adipokine) and NGAL (an iron trafficking protein) are neutrophil granule proteins. The majority of resistin is present in azurophil granules (318) whereas the NGAL is predominantly found in specific granules (33), although a degree of overlap occurs. Both proteins have been implicated in acute, e.g. sepsis (111), and chronic, e.g. RA and IBD (374,375), inflammatory conditions, and increased serum levels of both proteins were shown to correlate with the severity of sepsis (111). Although release of these proteins from hypoxic neutrophils has not previously been examined, increased resistin levels were present in the synovial fluid of patients with RA, a known hypoxic environment, when compared with osteoarthritis (374). Furthermore, increased serum resistin has been demonstrated in a rat model of chronic intermittent hypoxia (376) and increased circulating NGAL was detected in humans exposed to hypobaric hypoxia at high altitude (>5000 m) (377). My work demonstrated that both resistin and NGAL were increased in supernatants from hypoxic PAF and fMLP-treated neutrophils by proteomic assessment, and this hypoxic upregulation was confirmed by subsequent ELISA measurement in independently generated samples. These data provide further evidence of a hypoxic enhancement of degranulation, although time constraints prevented investigation of the role of PI3K signalling in modulating the release of these proteins under hypoxia to align with my previous data on NE. Published (normoxic) neutrophil secretomes, have shown active secretion of both resistin and NGAL in response to neutrophil treatment with various stimuli in vitro (including cytokines and bacterial products). Furthermore, resistin was also shown to be increased in septic patient vs healthy control plasma; NGAL, however, was found to be decreased (306) (see Table 7.3). This difference may reflect the heterogeneity of the sepsis-inducing agents in this study, which included Gram positive and negative bacteria and malaria parasites. Surprisingly, the proteins identified by proteomic assessment did not segregate precisely as, although the majority were increased in hypoxia, some granule proteins were relatively 150 increased in supernatants from normoxic cells. These data, although not confirmed by alternative biochemical assay due to time constraints, suggest that hypoxia does not globally increase degranulation. It is possible that hypoxia impacts movement of proteins between granules; one study has demonstrated variation in the granule co-localisation of resistin and azurocidin, which resulted in a differential secretion profile, dependent on the stimulus used (378). Another possibility is that hypoxia is able to direct selective granule release, akin to the “piecemeal” degranulation of eosinophils and mast cells (379), although this mechanism of secretion has not been described for neutrophils. These data are fascinating and worthy of further exploration; potential future research plans in this area are outlined in section 6.3. A likewise surprising finding was that, although the majority of cytoplasmic proteins were increased in normoxic supernatants, some cytoplasmic proteins, including cyclophilin A, were increased in hypoxic supernatants. This was unexpected, given that hypoxia promotes neutrophil survival, suggesting that the increase was not due to excess cell death. Cyclophilin A is an isomerase which regulates protein folding and trafficking. It has been shown to induce neutrophil chemotaxis and secretion of pro-inflammatory molecules from monocytes in vitro (380,381), and is associated with several inflammatory conditions, including sepsis (382), ulcerative colitis (383), ARDS (where it was more abundant in BALF from non-survivors) (384) and RA (where it was detected in the synovial fluid of patients with RA but not osteoarthritis) (385). Despite lacking a traditional leader export sequence, cyclophilin A has been shown to be actively secreted from a number of cell types, including macrophages in response to LPS (386) and cardiac myocytes in response to hypoxia/reoxygenation (387). There remains a paucity of literature on its secretion from neutrophils, although it has been detected in neutrophil secretomes (see Table 7.3). Given the cytoplasmic location of cyclophilin A in neutrophils, a release mechanism alternative to that of degranulation is intriguing. One study demonstrated that cyclophilin A release from vascular smooth muscle cells was via a vesicular transport mechanism, dependent on VAMP-2 (388). In neutrophils, cyclophilin A has been detected in NDMVs from healthy volunteers (95,366). No studies to date have looked directly at the release of MVs from hypoxic neutrophils. However, increased circulating NDMVs have been identified following complex insults that include a hypoxic component e.g. I/R and high altitude (188,389). Our initial hypothesis, therefore, was that cyclophilin A is secreted from neutrophils in MVs and that increased NDMV release contributed to enhanced (cytoplasmic) protein secretion under hypoxia. NDMVs were challenging to isolate. Initial trials aiming to generate NDMVs from plasma- Percoll® isolated PAF and fMLP-treated neutrophils were performed with the assistance of Dr Elisabet Ferrer (University of Cambridge), using an established ultracentrifugation MV isolation protocol (390). This method yielded very low numbers in comparison with 151 concurrently isolated MVs from endothelial cells, both by electron microscopy (Figure 4.15) and Nanosight analysis, where NDMV numbers were at the very limit of detection and could not be confidently quantified (data not shown). Figure 4.15: Visualisation of MVs from neutrophils and endothelial cells by electron microscopy Supernatants from 4*107 neutrophils (A) and 1*107 blood outgrowth endothelial cells (B) were sequentially centrifuged to pellet MVs. MV isolation was performed with the assistance of Dr Elisabet Ferrer (University of Cambridge). Fresh MV pellets, re-suspended in sterile filtered PBS, were negatively stained with uranyl acetate and visualised by transmission electron microscopy, with the assistance of Dr Jeremy Skepper (University of Cambridge). Scale bar 500 nm, arrows indicate examples of MVs. Representative images from n=3 experiments. Advice was therefore sought from Dr Victoria Ridger (University of Sheffield), and protocols well-established in her laboratory were used to optimise NDMV isolation. Hence, neutrophils were isolated using Histopaque-1077® gradients and a higher concentration of fMLP. However, the increase in yield was relatively minor and did not reach statistical significance. To ensure that the findings generated using the Histopaque-1077® method were not simply related to the cell isolation, particularly given my data showing basal neutrophil priming, I checked all key results using samples derived from my standard plasma-Percoll® cells. As flow cytometric analysis of NDMV number did not support a hypoxic increase in release, we then hypothesised that hypoxia may lead to differential packaging of NDMVs and thus impact protein release. Western blotting was performed to investigate whether hypoxia modulated the NDMV content of cyclophilin A. Although cyclophilin A was detected in NDMV lysates, when corrected for NDMV number using the MV marker annexin A1, there was no difference 152 in content between normoxic vs hypoxic NDMV. Furthermore, cyclophilin A (but not annexin A1) was detected in MV-free neutrophil supernatants, indicating that its release is partly NDMV-independent. Hence, the mechanism of enhanced cyclophilin A release from hypoxic neutrophils remains elusive; it would be interesting to investigate the role of PI3K modulation in this context, as well as inhibitors of vesicular transport. Interestingly, of all the proteins upregulated in hypoxic supernatants, histone H4 (a component of NETs) showed the largest fold increase. Although my investigation of NET release by measuring extracellular DNA did not reveal an increase under hypoxia, it is important to note that this is a surrogate marker of NETosis. Any DNA release from necrotic cells (which may be increased under normoxia as hypoxia is a survival signal) could obscure true differences in NET generation. Additionally, the literature to date is conflicting (see Section 1.2.2.6). My data suggest a possible “NET signature” in hypoxic supernatants; therefore, NETosis may contribute to protein release under hypoxia. This could be further investigated with more sensitive and specific assays, including ELISA or western blotting to measure citrullinated histones, or histones associated with NE or MPO. Another surprising finding was the presence of small numbers of blood proteins, such as a haemoglobin subunit (increased in normoxia) and immunoglobulins (increased in hypoxia). It is possible that the supernatants samples contained haemoglobin due to contamination from red blood cells during the neutrophil isolation process as a red cell lysis step was not used but this would not explain a significant difference between normoxia and hypoxia, given that the neutrophils were split equally between conditions from one original pooled sample. The effect of hypoxia on red blood cells is not well defined but it has been reported to induce increased cell deformability alongside membrane fragility (391), which would suggest an increase in hypoxia in contrast to my observations. The presence of the immunoglobulins, which are present in plasma, is perhaps more easily explained. Neutrophil secretory vesicles are intracellular storage granules formed by endocytosis and contain a number of plasma proteins, which may include immunoglobulins. The observed increase in hypoxia would therefore be consistent with enhanced degranulation. As well as investigating proteins increased in hypoxic supernatants, I also sought to validate proteins that were relatively increased in normoxic supernatants. Thioredoxin and S100A9 are both cytoplasmic proteins which have been shown to be actively secreted from other cell types (392,393) and, in the case of S100A9, from neutrophils (394). Both have been detected in normoxic neutrophil secretomes (see Table 7.3). Thioredoxin, a cytosolic disulphide reductase which functions as a ROS scavenger and regulator of redox-sensitive transcription factors has been shown to be anti-inflammatory, reducing neutrophil trans-endothelial migration and 153 protecting rabbits from I/R-induced ALI (395,396). S100A9 is a calcium-binding protein, preferentially forming a heterodimer (calprotectin) with its partner S100A8, and comprises approximately 45% of the neutrophil cytosol (397). There is debate as to whether S100A9 is pro- or anti-inflammatory. Proposed anti-inflammatory roles include: anti-oxidant free radical scavenging (398) and MMP inhibition (399); S100A9 was protective in a mouse model of LPS- induced ALI (400), whereas deficiency promoted lung inflammation in pneumonia model (401). However, the balance of the literature supports a pro-inflammatory role: S100A9 could stimulate neutrophil adhesion, degranulation and pro-inflammatory cytokine release in vitro (402–404), and S100A9-deficient mice were protected from mortality in murine models of endotoxin-induced shock and pneumonia (401,405). The two normoxic protein targets chosen for further analysis did not align with my proteomic data. The lower limit of detection of the ELISA used for thioredoxin was not sensitive enough to detect the protein in my supernatants, despite using the remainder of the concentrated supernatants from the proteomic study. Had time allowed, I would have generated a new set of protein-concentrated samples and trialled a more sensitive ELISA. However, as I have shown an impact on endothelial cells from unconcentrated supernatants (section 3.5), it seems unlikely in the context of this experimental setup that thioredoxin would exert a detectable effect, although it would also be important to test that this ELISA was compatible with detection of neutrophil supernatant proteins. In contrast to the proteomic data, measurement of S100A9 by both ELISA and western blot showed no difference between normoxia and hypoxia. As S100A9 comprises almost half of the cytosolic protein content, any small differences in cell death between normoxia and hypoxia (allowing protein release into the supernatant) may account for discrepant results. Although cytospins of neutrophil cell pellets were generated prior to proteomic assessment, which did not demonstrate any obvious morphological difference between normoxia and hypoxia, apoptosis was not formally assessed. Secretion of S100A9 from neutrophils has been shown to be dependent on an intact microtubule network (394); another possibility is that the presence of protease inhibitors in proteomic samples affected this pathway (e.g. through calcium chelation), thereby influencing S100A9 release. A further consideration is that of the technical ability of the ELISA to detect neutrophil supernatant proteins as the assay is only validated for serum and plasma samples. It is possible that proteases within the supernatant samples were able to cleave the binding antibody in the coated plate thereby impairing the capture/detection process. Future experiments could test the ELISA validity by the addition of a known quantity of recombinant protein to a supernatant sample to ensure that the expected amount is detected in this system. Similarly, this technique, in combination with a more sensitive assay, could be used to validate the results from my cyclophilin A ELISA (where protein content in the normoxic samples was 154 below the limit of detection) to ensure that the signal from the hypoxic supernatants was accurately reflecting protein content rather than due to interference from proteases within the supernatant samples. My characterisation of the normoxic vs hypoxic neutrophil secretome has important functional implications. The hypoxia-upregulated proteins resistin, NGAL and cyclophilin A have all been implicated in endothelial dysfunction. Resistin and cyclophilin A increased endothelial cell adhesion molecule expression in vitro (406–408), and some studies (e.g. (407)) have shown induction of endothelial cell apoptosis by cyclophilin A. Resistin also increased secretion of the potent vasoconstrictor endothelin 1 (408), and circulating levels correlated with endothelial cell adhesion molecule expression in sepsis (409) and with impaired FMD in normotensive individuals (410). Although evidence for direct endothelial injury by NGAL is lacking, it was shown to be increased in vulnerable atherosclerotic plaques ex vivo (411), as was cyclophilin A (406). Increased circulating NGAL has been detected in patients with coronary heart disease (412), and levels were associated with increased arterial stiffness and future cardiovascular events (413). High circulating levels of cyclophilin A have been observed in patients with type 2 diabetes mellitus in association with CVD (414) and cyclophilin A-deficient mice developed less severe atherosclerosis in a standard apolipoprotein E-deficient model. Hence, these proteins may be responsible for some of the endothelial dysfunction caused by hypoxic neutrophil supernatants and therefore represent potential novel therapeutic targets. Plans to establish the capacity of these proteins to cause endothelial damage in this context are outlined in section 6.3. Overall these proteomic data give a detailed characterisation of the normoxic vs hypoxic secretome, with biochemical validation of proteins which are upregulated under hypoxia and have the potential to cause endothelial damage. I wished to investigate these targets further in a disease-relevant patient group, namely patients experiencing exacerbations of COPD. Chapter 5 Results: Effect of hypoxia on neutrophil function in exacerbating COPD patients 156 5 Effect of hypoxia on neutrophil function in exacerbating COPD patients 5.1 Introduction Having established that hypoxia differentially regulates protein release from neutrophils and that a number of these hypoxia-upregulated proteins have histotoxic capacity, I wished to extend my observations to a clinically relevant cohort of patients. COPD is characterised by persistent airway neutrophilic inflammation, with increased numbers of neutrophils identified both in BALF and within the bronchial epithelium when compared to healthy controls (198). Moreover, the degree of airway neutrophilia in COPD correlates with disease severity (415). COPD patients are frequently hypoxaemic, particularly during exacerbations, but even in the absence of systemic hypoxia, profound local tissue hypoxia can exist in areas of infection and/or inflammation (136,222); stabilisation of HIF-1α, which increases exponentially below 6% oxygen, has been demonstrated in the bronchial epithelium of COPD patients, where neutrophilic inflammation is a key disease driver (131). Further to our published data showing injury to epithelial cells, I have demonstrated that hypoxic neutrophil supernatants also damage endothelial cells. COPD patients have increased endothelial dysfunction compared with healthy controls, even after correcting for shared risk factors, such as smoking (225). This manifests clinically as cardiovascular comorbidity, with an increased incidence of myocardial infarction and stroke in COPD patients (227). Several studies have shown that COPD patients have increased vascular inflammation and arterial stiffness (226,280), even when those with overt CVD are excluded (234). Therefore, as COPD encompasses chronic neutrophilic inflammation in the context of local and systemic hypoxia plus endothelial dysfunction, I wished to investigate whether neutrophils from COPD patients exhibited a “hypoxic signature”, focusing on my hypoxia-upregulated proteins of interest, and whether this was reflected in the plasma. Although multiple studies have looked at blood, sputum and BALF protein content in COPD (416), with a view towards identifying clinically useful biomarkers, only two studies have looked specifically at proteins secreted by COPD patient neutrophils: Blidberg et al. showed no difference in IL-8 or macrophage inflammatory protein (MIP)-1α release from TNFα-treated COPD vs. healthy neutrophils (417) whereas Baines et al. showed an increase in MMP-9 release from LPS-treated COPD neutrophils (418). Genetic (335) and proteomic (219,221) profiling of peripheral blood neutrophil lysates has demonstrated a number of differences between COPD neutrophils compared with healthy controls, including increased elastase activity alongside undetectable levels of leukocyte elastase inhibitor (Serpin B1). Regarding the protein targets I have identified as upregulated in the hypoxic neutrophil secretome, a role 157 for NE in the pathogenesis of COPD has long been established (419) but the contribution of MPO, resistin, NGAL and cyclophilin A is less clear, although increased sputum MPO and NGAL (420), and increased circulating resistin (113), NGAL (112) and cyclophilin A (114) have been detected in COPD patients. I therefore hypothesised that neutrophils from exacerbating COPD patients would release more NE, MPO, resistin, NGAL and cyclophilin A and that these proteins would be increased in COPD patient plasma. Despite my not demonstrating a difference in NDMV release under hypoxia compared with normoxia in vitro (section 4.6.1), studies have shown increased circulating NDMVs in the setting of altitude-associated hypoxia or following hypoxic exercise testing (188,421), increased NDMVs in the coronary sinus following percutaneous intervention in patients suffering from acute coronary syndrome (422), an association of NDMVs with subclinical atherosclerosis burden (423) and increased circulating exosomes (MVs <100 nm) in exacerbating COPD patients (424). Given this evidence and as it is possible that my results in section 4.6 do not accurately reflect the in vivo situation, I additionally decided to investigate the distribution of neutrophil-, monocyte-, endothelial cell- and platelet-derived MVs in COPD patient plasma, hypothesising that exacerbating COPD patients have more circulating NDMVs than healthy controls. The hypothesis driving the work in this chapter is that exacerbating COPD patients have increased release of hypoxia-upregulated histotoxic proteins from neutrophils. The specific aims of the work presented in this chapter are: 1. To compare the release of the hypoxia-upregulated protein targets identified by proteomics from neutrophils isolated from exacerbating COPD patients vs healthy controls 2. To compare the content of these hypoxia-upregulated protein targets in plasma from exacerbating COPD patients vs healthy controls 3. To examine the distribution of neutrophil-, monocyte-, endothelial cell- and platelet-derived MVs in plasma from exacerbating COPD patients vs healthy controls 158 5.2 Investigation of COPD versus healthy neutrophil secretion of proteins upregulated in the hypoxic secretome Exacerbating COPD patients were recruited from Cambridge University Hospitals NHS Foundation Trust. Patients who met the inclusion and exclusion criteria (see section 2.6.1) were identified using electronic patient medical records (Epic software) and written consent was obtained. Demographic and clinical data are listed in Table 5.1. Venous blood was drawn within 24 h of admission to obtain plasma and serum. For a subset of patients, blood was also drawn for neutrophil isolation. Age- (within 10 years) and sex-matched healthy controls had blood drawn to obtain plasma and serum, with or without neutrophil isolation. As COPD patients were older than the healthy volunteers in our database and prior notice for venepuncture was short (< 24 h), it was not possible to obtain blood from age- and sex- matched healthy controls at the same time as COPD patients. Also, given the unpredictable timing of COPD patient recruitment, the significant (80 ml) volume of blood required to obtain sufficient neutrophils for analysis, and the logistics of performing neutrophil isolations at short notice alongside other planned experiments, neutrophils were isolated from fewer COPD patients than healthy controls. Overall, plasma samples were obtained from n=12 COPD patients and n=14 healthy volunteers whereas neutrophil isolations, as a subset of these subjects, were performed for n=7 COPD patients and n=13 healthy volunteers. As the mean age of healthy volunteers was 66 years, they inevitably had some comorbidities. However, no healthy control had respiratory disease and five healthy controls did not have any past medical history. On admission, COPD patients were systemically hypoxic, with a mean oxygen saturation of 86%, although it was not possible to determine the duration of hypoxia prior to admission. As patients were treated with oxygen by the medical team to maintain saturations of 88 – 92%, the mean oxygen saturation at the time of blood sampling was 92%. Oxygen saturations were not recorded for healthy volunteers but, as stated, there was no past medical history of respiratory disease. 5.2.1 The effect of hypoxia on the release of NE and MPO from COPD versus healthy neutrophils Neutrophils were isolated from exacerbating COPD patients using plasma-Percoll® gradients. As I wished to examine neutrophil degranulation, it was first important to establish whether circulating neutrophils from COPD patients were primed or basally activated, since this might modulate degranulation responses. Shape change was measured by flow cytometry at baseline, or after 30 min in the presence or absence of fMLP (100 nM), using an established protocol from our laboratory (425). The small but significant difference in shape change for unstimulated neutrophils between baseline and 30 min likely reflects their incubation in PBS, which lacks the nutrients present in IMDM. Importantly, there was no significant difference in 159 shape change between COPD or healthy neutrophils for any condition, indicating that circulating COPD neutrophils in these patients were not basally primed or activated and were able to fully respond to an exogenous activating stimulus (fMLP: Figure 5.1). Out of the full cohort, shape change was not assessed for neutrophils isolated from two COPD patients and one healthy control, but the following neutrophil supernatant assays (sections 5.2.1 and 5.2.2) did include these subjects. Figure 5.1: Quantification of healthy versus COPD neutrophil shape change by flow cytometry Neutrophils, isolated from exacerbating COPD patients or healthy controls, were incubated in the presence or absence of fMLP (100 nM). Cells were fixed at baseline or after 30 min. Shape change was assessed by flow cytometric analysis of forward scatter. Results represent mean ± SEM; n=5 (COPD) n=12 (healthy); two way ANOVA, Sidak’s multiple comparisons test. 160 Table 5.1: Clinical and demographic data Data were collected from the Cambridge University Hospitals NHS Foundation Trust electronic patient medical record system (Epic software) at the time of recruitment and/or blood sampling. Abbreviations: diabetes mellitus (DM), irritable bowel syndrome (IBS), gastro- oesophageal reflux disease (GORD), benign prostatic hypertrophy (BPH), obstructive sleep apnoea (OSA), ischaemic heart disease (IHD), pulmonary embolism (PE), motor neuron disease (MND), atrial fibrillation (AF), abdominal aortic aneurysm (AAA), chronic kidney disease (CKD), congestive cardiac failure (CCF), white cell count (WCC), C reactive protein (CRP). Characteristic Healthy Volunteer N=14 COPD patient N=12 Age – mean year ± SD 66±5 70±7 Male sex – no. (%) 7 (50) 6 (50) Smoking history – mean pack year ± SD 5±12 58±33 Current smoker – no. (%) 0 (0) 3 (21) Current no. of cigarettes/day – mean ± SD 0 3±2.5 WCC – mean ± SD Neutrophil count – mean ± SD CRP – mean ± SD Not tested 10.15±4.5 8.0±4.3 20±21 O2 saturation on admission – mean % ± SD O2 saturation at venepuncture – mean % ± SD Not recorded 86.3±4.5 91.8±2.7 Comorbidities/ past medical history Type 2 DM, Migraine, IBS, GORD, Hypertension, Hypercholesterolaemia, Hypothyroidism, Gout, BPH Hyperthyroidism, Hiatus hernia, Bicuspid aortic valve, Hayfever Type 2 DM, Spinal stenosis, OSA, Hypertension, IHD, PE Obesity, Hypothyroidism Osteoporosis, Alcohol excess, MND, AF Hysterectomy, Cholecystectomy, Resected squamous cell lung cancer, Gallstones GORD, Glaucoma, AAA, Mild bronchiectasis, Treated bladder cancer, Depression, CKD, CCF, Pernicious anaemia 161 Next, I wished to investigate the release of NE and MPO from COPD vs healthy neutrophils in conditions of normal and reduced oxygen tension. Exactly as before, isolated neutrophils were incubated under normoxia or hypoxia for 4 h prior to treatment with PAF (1 µM, 5 min) and fMLP (100 nM, 10 min), or IMDM control. Supernatant NE activity was assessed by Enzchek® elastase activity assay, and MPO activity was assessed by H2O2-dependent oxidation of DMB. NE and MPO release from unstimulated neutrophils showed no difference between all conditions (Figure 5.2). Comparing PAF/fMLP-treated cells, hypoxia significantly increased the release of NE from COPD neutrophils (7.57*104 ± 1.82*104 AU) compared with both normoxic COPD neutrophils (4.17*104 ± 8.58*103 AU, p=0.0004) and, notably, with hypoxic healthy neutrophils (3.63*104 ± 9.46*103 AU, p=0.005, Figure 5.2A). There was no difference in NE release from normoxic COPD compared with normoxic healthy neutrophils (4.17*104 ± 8.58*103 AU vs 2.49*104 ± 4.1*103 AU, p=0.58). In comparison, although hypoxia, compared with normoxia, increased the release of MPO from both healthy (37.0 ± 6.06 % vs 20.32 ± 3.19 %, p<0.0001) and COPD (38.64 ± 5.83 % vs 24.33 ± 5.63 %, p=0.0036) stimulated neutrophils, there was no difference in MPO release between healthy and COPD neutrophils for any condition (Figure 5.2B). This was surprising since both NE and MPO are azurophilic granule proteins. However, it must be noted that, due to the logistics of performing both assays, the n numbers for these assays were different (with more healthy controls included for MPO measurement and more COPD patients included for NE assessment). 5.2.2 The effect of hypoxia on the release of resistin, NGAL and cyclophilin A from COPD versus healthy neutrophils Having established that NE release from neutrophils under hypoxia is further enhanced in COPD compared with cells obtained from with healthy volunteers, I then examined the release of the proteins resistin (predominantly present in azurophilic granules) and NGAL (predominantly present in specific granules). These proteins were identified as upregulated under hypoxia by proteomic examination of the neutrophil secretome, and subsequently validated biochemically by ELISA (see section 4.5). Isolated neutrophils were incubated under normoxia or hypoxia for 4 h prior to treatment with PAF and fMLP, or IMDM control. Supernatant resistin and NGAL content was assessed by ELISA. Again, resistin and NGAL release from unstimulated neutrophils did not differ between any of the conditions assessed (Figure 5.3). Comparing PAF/fMLP-treated cells, higher levels of resistin were detected in supernatants from hypoxic COPD neutrophils compared with hypoxic healthy neutrophils (423.4 ± 67.26 pg/ml vs 250.1 ± 40.95 pg/ml, p=0.0213, Figure 5.3A), although the difference in release between normoxia and hypoxia was not significant for either healthy or COPD neutrophils. Similarly, the detection of NGAL was increased in supernatants from hypoxic COPD neutrophils compared with hypoxic healthy neutrophils (9.1 ± 1.9 ng/ml vs 4.76 ± 0.65 162 ng/ml, p=0.0114, Figure 5.3B). Hypoxia vs normoxia increased the release of NGAL from COPD neutrophils (9.1 ± 1.9 ng/ml vs 6.0 ± 1.46 ng/ml, p=0.0091) but the difference between normoxia and hypoxia for healthy neutrophils was not significant (Figure 5.3B). Figure 5.2: The effect of hypoxia on NE and MPO release from healthy and COPD neutrophils Neutrophils, isolated from exacerbating COPD patients or healthy controls, were incubated under normoxia or hypoxia for 4 h before treatment with PAF and fMLP, or IMDM control. A: supernatant NE activity was measured by Enzchek® assay. B: supernatant MPO activity was measured by H202-dependent oxidation of DMB, with absorbance read at 450 nm and activity expressed as % total activity from triton-X (10%)-lysed neutrophils. Results represent mean ± SEM; A: n=6 (COPD), n=7 (healthy); B: n=4 (COPD), n=10 (healthy); ** = p<0.01, *** = p<0.001, **** = p<0.0001; two way ANOVA, Sidak’s multiple comparisons test. Figure 5.3: The effect of hypoxia on resistin and NGAL release from healthy and COPD neutrophils Neutrophils, isolated from exacerbating COPD patients or healthy controls, were incubated under normoxia or hypoxia for 4 h before treatment with PAF and fMLP, or IMDM control. Supernatant A: resistin and B: NGAL content were measured by commercial ELISA. Results represent mean ± SEM; A&B: n=3-4 (unstimulated neutrophils), n=5-7 (stimulated neutrophils); *= p<0.05, **=p<0.01; two way ANOVA, Sidak’s multiple comparisons test. 163 Having investigated the release of granule proteins by COPD neutrophils, I next wished to examine the secretion of the cytosolic protein cyclophilin A, which was similarly identified as upregulated under hypoxia by proteomic examination of the neutrophil secretome. Isolated neutrophils were incubated under normoxia or hypoxia for 4 h prior to treatment with PAF and fMLP, or IMDM control. Supernatant cyclophilin A content was assessed by ELISA (Elabscience). As the lower limit of detection of the ELISA (1.25 ng/ml) was not sensitive enough to allow interpolation of most absorbance readings, cyclophilin A content of neutrophil supernatants was measured by absorbance (OD450 – baseline). For PAF and fMLP-treated neutrophils, the release of cyclophilin A from COPD compared with healthy neutrophils was increased in both normoxia (0.021 ± 0.007 vs 0.0, p=0.029) and hypoxia (0.043 ± 0.011 vs 0.008 ± 0.003, p=0.0004). Further, hypoxia, compared with normoxia, increased the release of cyclophilin A from COPD neutrophils (0.043 ± 0.011 vs 0.021 ± 0.007, p=0.0009, Figure 5.4). Samples from the same subjects used for the NE assay were employed for ELISA measurement of resistin, NGAL and cyclophilin A, with the exception of one different healthy control and one fewer COPD patient. Figure 5.4: The effect of hypoxia on cyclophilin A release from healthy and COPD neutrophils Neutrophils, isolated from exacerbating COPD patients or healthy controls, were incubated under normoxia or hypoxia for 4 h before treatment with PAF and fMLP, or IMDM control. Supernatant cyclophilin A content was measured by commercial ELISA and is expressed as absorbance (OD450 – baseline) as several samples were below the lowest value set by the standards (1.25 ng/ml; Elabscience ELISA assay range 1.25 ng/ml – 80 ng/ml). Results represent mean ± SEM; n=5-7; * = p<0.05, *** = p<0.001; two way ANOVA, Sidak’s multiple comparisons test. 164 5.3 COPD versus healthy plasma content of proteins upregulated in the hypoxic neutrophil secretome Having established that the hypoxia-upregulated release of NE, resistin, NGAL and cyclophilin A was further augmented in supernatants from COPD vs healthy control-derived stimulated neutrophils, I wished to investigate whether these proteins were also increased in plasma from exacerbating COPD patients vs healthy volunteers. Venous blood was drawn from exacerbating COPD patients within 24 h of admission, or age and sex-matched healthy volunteers, into sterile EDTA K3 collection tubes. Plasma was obtained by centrifugation of blood collection tubes. The data described in this section were obtained from the entire cohort of COPD patients and healthy controls. Measurement of plasma protease activity, such as NE, is challenging as the activity is rapidly quenched by α1AT, and detection of NE by immunoassays may quantify both free and α1AT- bound inactive enzyme. Two assays, which detected NE- and PR3-specific cleavage products of fibrinogen (Aα-Val360 and Aα-Val541 respectively), were therefore employed in collaboration with Professor Robert Stockley (University of Birmingham), performed by Paul Newby. These novel assays provided a footprint of both NE and PR3 activity in plasma. The Aα-Val360 content of COPD plasma was significantly higher than healthy plasma, (12.81 ± 0.68 nM vs 9.91 ± 0.64 nM, p=0.0048, Figure 5.5A). Similarly, the Aα-Val541 content of COPD plasma was significantly higher than healthy plasma (22.73 ± 6.7 nM vs 9.15 ± 2.15 nM, p=0.0228, Figure 5.5B). These results indicate increased circulating NE and PR3 activity in exacerbating COPD patients, although the values for COPD and healthy control samples do overlap in each assay. Figure 5.5: The healthy versus COPD plasma content of Aα-Val360 and Aα-Val541 Plasma from exacerbating COPD patients or healthy controls was assessed for content of A: Aα-Val360 (NE-specific fibrinogen cleavage product) or B: Aα-Val541 (PR3-specific fibrinogen cleavage product). Data were provided by Paul Newby (Birmingham). Results represent mean ± SEM; n=12 (COPD), n=14 (healthy); A: unpaired t test, B: Mann-Whitney test. 165 Plasma resistin, NGAL and cyclophilin A content was assessed by ELISA. In view of the previous difficulty in quantitating cyclophilin A (Figure 5.4), a more sensitive ELISA (assay range 0.39 ng/ml – 25 ng/ml; Oxford Biosystems) was employed. There was no significant difference between the healthy and COPD plasma content of resistin (1.84 ± 0.21 ng/ml vs 2.54 ± 0.33 ng/ml, p= 0.13, Figure 5.6A), NGAL (9.34 ± 1.29 ng/ml vs 8.93 ± 1.45 ng/ml, p= 0.81, Figure 5.6B) or cyclophilin A (21.19 ± 6.77 ng/ml vs 16.92 ± 7.63 ng/ml, p= 0.33, Figure 5.6C). Figure 5.6: The healthy versus COPD plasma content of resistin, NGAL and cyclophilin A by ELISA Plasma from exacerbating COPD patients or healthy volunteers was assessed for A: resistin, B: NGAL or C: cyclophilin A content by commercial ELISA. Results represent mean ± SEM; A: n=12 (COPD), n=14 (healthy), B: n=8 (COPD), n=10 (healthy), C: n=12 (COPD), n=14 (healthy); Mann-Whitney test. 166 5.4 The COPD versus healthy plasma content of microvesicles Despite the fact that previous experiments did not demonstrate a hypoxic increase in NDMVs (section 4.6.1), it is noteworthy that 15 of the top 20 proteins upregulated in the hypoxic neutrophil secretome have been identified in NDMVs in the literature (Table 4.3). I therefore wished to explore whether there was an NDMV pro-inflammatory “signature” in plasma from exacerbating COPD patients by examining the relative distribution of neutrophil-, monocyte-, endothelial cell- and platelet-derived MVs. Flow cytometric analysis of plasma MV content was performed with the assistance of Merete Long (PhD student) in collaboration with Dr Victoria Ridger (Sheffield). MVs were stained with BV421 anti-CD66b, APC Cy7 anti-CD14, PerCP Cy5/5 anti-CD144 and PE anti CD41a in order to identify neutrophils, monocytes, endothelial cells or platelets respectively as the parent cells. Comparing plasma from exacerbating COPD patients vs healthy controls, there was no difference in the distribution of NDMVs (17.17 ± 4.75% vs 10.97 ± 2.96%, p=0.54, Figure 5.7A), monocyte-derived MVs (11.68 ± 3.83% vs 7.61 ± 2.5%, p=0.42, Figure 5.7B), endothelial cell-derived MVs (4.44 ± 0.62% vs 5.47 ± 1.46%, p=0.51, Figure 5.7C), or platelet-derived MVs (39.19 ± 7.31% vs 45.32 ± 10.99%, p=0.64, Figure 5.7D). There was also no difference in total MV number per µl plasma (2127 ± 230.9 vs 2734 ± 426.7, p=0.22, Figure 5.7E). 5.5 Discussion By translating my characterisation of the hypoxic neutrophil secretome to a patho- physiologically and clinically relevant patient group, I have established that peripheral blood neutrophils from COPD patients display a more marked (compared to age- and sex-matched controls) hypoxic uplift of NE, resistin, NGAL and cyclophilin A release in vitro. NE is well established in the pathogenesis of COPD, with the ability to cause experimental emphysema in animal models (201), and sputum levels correlating with COPD severity and decline in FEV1 (203). However, its role in endothelial dysfunction is less well recognised. Recently, and of great relevance to the work described in this chapter, murine genetic and pharmacological inhibition of NE has been shown to attenuate the development of atherosclerosis (263), and higher NE levels are associated with increased risk of cardiovascular events, even after adjusting for known risk factors (264). Neutrophils accumulate in atherosclerotic plaques, with higher numbers found in eroded or ruptured lesions ex vivo (248), and in vivo 18F-FMISO PET imaging of atherosclerotic rabbit aortae has demonstrated significant vascular hypoxia, corroborated by ex vivo immunohistochemistry (330). This locally hypoxic environment may increase NE release, which has the potential to promote plaque progression and instability: NE, detected in atherosclerotic lesions, was able to potentiate the processing and release of active IL-1β from coronary endothelial cells (261) and promote endothelial apoptosis (426). 167 Figure 5.7: Healthy versus COPD plasma microvesicle content Plasma from exacerbating COPD patients or healthy controls was stained with BV421 anti- CD66b (neutrophil, A), APC Cy7 anti-CD14 (monocyte, B), PerCP Cy5/5 anti-CD144 (endothelial cell, C) and PE anti-CD41a (platelet, D). MV pellets, obtained by centrifugation (20,000 g), were re-suspended in sterile PBS (490 µl) with counting beads (10 µl) and analysed by flow cytometry. The MV forward scatter vs side scatter gate was set using standardised size calibration beads, and 1,000 counting beads were measured per sample. The MV distribution per parent cell type is expressed as % total MVs (A-D); total MV count per µl plasma is shown in panel E. Results represent mean ± SEM; n=5 (healthy), n=6 (COPD); A: Mann-Whitney test, B-E: unpaired t test. 168 Increased circulating resistin levels have also been found in COPD (221,427), and were inversely correlated with FEV1, although increased levels were also found in asthmatics and smokers in this study (113). Multiple studies support a role for resistin in endothelial dysfunction: resistin over-expression promoted myocardial apoptosis, remodelling and left ventricular dysfunction in a rat model (428) and circulating resistin is associated with increased risk of MI, ischaemic stroke and heart failure (429–431). Increased circulating NGAL has been detected in COPD patients, correlating with more frequent exacerbations and hypoxaemia (112), although controversy surrounds its association with disease severity, and smoking may affect levels (432,433). NGAL has recently been implicated in airway remodelling in COPD (434) and, although causality is not proven, higher levels were detected in vulnerable carotid artery atherosclerotic plaques (411) and increased plasma levels were associated with mortality and major adverse cardiac events following ST-elevation MI (435). Only two studies have associated cyclophilin A with COPD: Zhang et al. recently showed increased serum levels in COPD patients (114) and Hu et al. detected increased levels in lung tissue from smokers, with a further increase in smokers with COPD (436). Extensive evidence supports the ability of cyclophilin A to cause endothelial activation and injury (reviewed in (437)), including the development of PH in a murine over-expression model (407). My data are the first description of enhanced neutrophilic secretion of these proteins under hypoxia in the context of COPD, and although I was not able to detect increased levels in the plasma of a cohort of exacerbating patients (see below), the literature suggests that these proteins have the potential to contribute to cardiovascular comorbidity. Due to time limitations, I was not able to examine the effects of these proteins, either alone or in combination, on endothelial cells directly; this will be important to establish and also whether inhibition of these proteins is protective. A limitation of this study was the pragmatic nature of patient recruitment which meant that I was unable to control the timing and use of treatments, such as antibiotics and steroids. A further challenge was presented by the variable aetiology of an acute exacerbation, e.g. viral vs. bacterial stimulus, as well as the heterogeneity in COPD diagnosis; the cause of an acute exacerbation may be unclear and identification of a culprit organism is rare. Patients with pneumonia, a pathological process distinct from that of an acute exacerbation, were excluded from the study. Interestingly, two COPD patients who had blood drawn prior to their chest radiograph, which subsequently showed consolidation indicating pneumonia, had basally shape changed neutrophils (and were excluded from further analysis). This validates the ability of the shape change assay to detect systemic neutrophil priming, and may explain why some studies have shown priming of circulating COPD patient neutrophils (283,334,438). It would be informative to compare the same COPD patients during exacerbation and in the stable 169 state to investigate whether any differences in neutrophil function are persistent or contributed to by acute inflammation. Attempts to recruit these patients for further studies in the stable state were hampered by frequent hospital admissions for exacerbations, highlighting the burden of disease. The mechanism for a further increase in protein release from COPD neutrophils under hypoxia remains to be elucidated. I did not observe any difference in shape change between healthy and (non-pneumonic) COPD neutrophils, demonstrating that COPD neutrophils were not basally primed or activated in my study. These data argue against a primed phenotype explaining my observed increase in the release of granule proteins from COPD neutrophils. One potential explanation for enhanced protein release is a change in protein or mRNA abundance in COPD neutrophils. Although I was not able to perform these experiments due to time constraints, this would be important to investigate. Gene array (335) and proteomic (219–221) studies of COPD vs healthy neutrophil lysates have shown differential mRNA transcript and protein expression. Interestingly, the proteome of peripheral blood COPD neutrophils was distinct from that of TNFα and/or GM-CSF-treated healthy neutrophils (220), and two diverse COPD proteomic phenotypes were identified, (221), suggesting a complex interplay between multiple inflammatory signals, which may include hypoxia. This differential mRNA/protein expression is reflected in transcriptomic and proteomic examination of COPD patient blood (439,440) and sputum (441) which, although not cell-type specific, has suggested a neutrophil signature. Although putative HREs have been identified in the promotor regions of NE and MPO (315), mRNA transcripts of granule proteins have not been found following neutrophil maturation (442) and our previous data showed no change in NE transcription or translation under hypoxia in the relevant timeframe of 4-5 h (149). Analysis of changes in rodent neutrophil lysate enzymes following abdominal surgery revealed an increase in the cytoplasmic protein, cyclophilin A (370), suggesting that inflammation can modulate neutrophil protein content, although it is not clear how this would lead to enhanced release of cytoplasmic proteins. However, it is possible that exposure to circulating inflammatory cytokines and systemic hypoxia may have impacted neutrophil granule protein content in COPD patients prior to ex vivo investigation, leading to enhanced release upon stimulation of degranulation. It would be valuable to perform a quantitative lysate and secretome proteomic screen of COPD vs healthy neutrophils. Enhanced protein secretion has been observed from other cell types in COPD: Kawayama et al. showed increased release of MMP-9, TNFα and IL-6 from LPS-treated COPD patient PBMCs (443). Similarly, augmented neutrophil protein secretion has been shown in other 170 disease states, e.g. increased LTB4 release from PAF-treated asthma patient neutrophils (444), which the authors propose was due to mobilisation of intracellular calcium, and, with direct relevance to my work, increased resistin release from cytochalasin B/fMLP-treated septic patient neutrophils (306). Mechanisms for these observations, however, are lacking. Examination of NE and MPO release from COPD vs healthy neutrophils in my study demonstrated a number of differences. COPD neutrophils showed an augmented hypoxic uplift of NE, which was not apparent from healthy control neutrophils in this experiment, whereas hypoxia increased MPO release similarly from both COPD and healthy neutrophils. Also, the difference between stimulated and unstimulated cells was much less for MPO than NE. This disparity could reflect differences in the assay performance: MPO activity is expressed as percentage of the total whereas the NE assay is a kinetic assay read at a specific timepoint to give an activity value in arbitrary units. The fact that NE and MPO release do not mirror each other is surprising as both are azurophilic granule proteins. Possible explanations, apart from assay differences, include: differential granule packaging, mobilisation of different granule populations, pre-programmed differences in protein content in diseased neutrophils, recruitment of a mechanism of protein release in addition to that of degranulation, or effects of other proteins present in the cells/supernatants, including possible degradation of MPO by enhanced NE activity. Due to time limitations, I was not able to examine NET release or the effect of PI3K inhibition on the hypoxic uplift of protein release from COPD neutrophils. It would be interesting to see if COPD neutrophils have a higher propensity for NETosis and whether the hypoxic uplift could be completely abrogated by PI3Kγ inhibition. Juss et al. found that the pro-survival phenotype observed in exacerbating COPD patient neutrophils ex vivo was resistant to pan-PI3K inactivation. However, Sapey and colleagues demonstrated that pan-PI3K inhibition of neutrophils from COPD patients corrected abnormal migration (216). Whether this aberrant signalling is a reflection of exposure of circulating neutrophils to systemic hypoxia and/or inflammation remains unclear. These studies suggest that the role of PI3K signalling in COPD neutrophils is highly context dependent. It will be important to investigate whether aberrant PI3K signalling contributes to my observed enhanced protein release under hypoxia, particularly with the advent of clinical PI3K inhibitors. Although I found no increase in NDMV release on hypoxic neutrophil incubation, elevated levels of plasma exosomes (424) and sputum NDMVs, which correlated with disease severity (445), have been found in COPD patients, and both granule proteins, including NE and MPO, and cytosolic proteins, including cyclophilin A, have been detected by proteomic examination of NDMVs (95,366) from healthy individuals. However, my data did not show any increase in 171 COPD plasma NDMV content compared with healthy controls, although the n number (6) was relatively small given the potential heterogeneity of these patient-derived samples, and there was a trend for an increased percentage of NDMVs in COPD plasma. However, taken alongside my earlier data, it does not seem likely that NDMVs play a significant role in the hypoxic uplift of protein secretion, either from healthy volunteers or exacerbating COPD patients. Surprisingly, my data did not show an increase in the COPD vs healthy plasma content of resistin, NGAL or cyclophilin A. It is possible that this was not the correct compartment to examine, although more likely that my study numbers were too small as published reports have shown that COPD patients have increased circulating resistin (113), NGAL (112) and cyclophilin A (114), including 26, 402 and 93 patients, respectively. I have, therefore, set up a collaboration with Dr Elizabeth Sapey (Birmingham) to examine the plasma content of these proteins in a larger, well-phenotyped cohort of COPD patients. Another potential explanation for my results is the secretion of these proteins from other cell types. Resistin is released from adipocytes in rodents, whereas release in humans is predominantly from immune cells, including those resident in adipose tissue, such as PBMCs and macrophages as well as neutrophils (446). Although the majority of NGAL is released from neutrophils, low levels are secreted from epithelial cells. Changes in circulating levels may therefore reflect more than once cellular source and obscure a plasma neutrophil signature. For example, circulating NGAL is increased during acute kidney injury, thought to be due to a combination of increased secretion from stressed renal tubular cells, increased neutrophil release and changes in protein reabsorption from urine due to renal damage (447). My data did show an increase in the COPD vs healthy plasma content of the fibrinogen breakdown products Aα-Val360 and Aα-Val541, indicating increased circulating NE and PR3 activity, respectively, in these patients. Previous studies using the Aα-Val360 assay have shown increased content in plasma from patients with α1AT deficiency (448) and COPD, which correlated with disease severity (204), although these were not performed in comparison with healthy controls. These data support my hypothesis that circulating neutrophils in exacerbating COPD patients release more NE, which is able to cause tissue damage, including damage to endothelial cells. Overall, I have demonstrated an enhanced hypoxic increase in the release of the histotoxic proteins NE, resistin, NGAL and cyclophilin A from exacerbating COPD patient neutrophils, and an increase in circulating NE and PR3 activity these patients. As such, I have identified potential novel therapeutic targets which may be responsible for tissue damage in COPD, including endothelial dysfunction. Chapter 6 Discussion 173 6 Discussion 6.1 Overview Neutrophils are recruited rapidly and in great numbers to sites of tissue damage, commonly caused by invading pathogens. Circulating neutrophils patrol the endothelium, sensing sites of cellular/tissue injury via a range of membrane receptors which respond to cytokines, chemokines, and damage or pathogen associated molecular patterns (DAMPs, PAMPs); these signals direct neutrophils to regions under attack, where they deploy their powerful armamentarium and release pro-inflammatory mediators to summon additional immune cells. Neutrophils also migrate to sites of sterile injury; necrotic cells can release mitochondrial formylated peptides, closely resembling bacterial formylated peptides which are potent neutrophil activators. In this manner, mitochondrial formyl peptides from dying cells directly activate neutrophils and facilitate chemotaxis towards damaged tissues, as demonstrated in murine sterile lung injury (449), and in the systemic inflammatory response to trauma in humans (450). This response may be a pre-emptive call for neutrophils to attend breached tissues at risk of microbial invasion or it may be that neutrophils remove cellular debris and promote damage resolution, enabling neo-vascularisation by release of matrix-degrading metalloproteinases (183). However, at sites of sterile or persistent injury, extracellular release of antimicrobial but histotoxic neutrophil proteins and proteases may induce substantial collateral tissue damage. This maladaptive immune response has been implicated in a number of inflammatory situations, such as I/R injury (451) and vasculitis (69). Inflamed and infected tissues are often profoundly hypoxic, and ideally neutrophils should work effectively in such low oxygen environments. In COPD, inflamed airways exhibit persistent neutrophilic inflammation and have been shown to be severely hypoxic, both in animal models (222) and bronchial biopsies from patients (131), notwithstanding their proximity to ambient air/oxygen. Despite airway neutrophilia, COPD patients may develop chronic bacterial airway colonisation, or suffer recurrent infection-driven exacerbations (452). This suggests that COPD neutrophils are not effective at protecting the host from infection and are in fact contributing to host tissue injury and disease progression. Patients with advanced COPD are often systemically hypoxaemic and, consequently, both tissue and circulating neutrophils are exposed to low oxygen levels. COPD patients also suffer excessive cardiovascular comorbidity even after correcting for common risk factors such as smoking, and CVD accounts for a substantial proportion of COPD patient deaths (up to 60%) (233). Laboratory studies of neutrophil function are almost always undertaken at ambient oxygen tensions, but it is important to recapitulate the pathological environment to understand how neutrophils operate in these hypoxic environments. Published data from our laboratory 174 demonstrate that hypoxic neutrophils fail to eliminate certain pathogens (148), and that supernatants from hypoxic vs normoxic neutrophils contain more granule proteases and cause more damage to bronchial epithelial cells (149). This suggests that hypoxic environments engender a destructive neutrophil phenotype, which contributes to the pathogenesis of inflammatory diseases such as COPD. The mechanisms responsible for endothelial dysfunction and development of athero-thrombotic disease in COPD remain unclear. However, several studies (for example (248)) have shown that neutrophils accumulate in atherosclerotic lesions, and that these regions are profoundly hypoxic (330,331). Hence, my PhD aimed to investigate the hypothesis that hypoxia promotes neutrophil-mediated endothelial damage in COPD. My work has extended the description and understanding of the detrimental neutrophil phenotype engendered by hypoxia: by increasing the repertoire of agonists that are relevant to hypoxic degranulation, by further delineating the mechanism of increased neutrophil degranulation under hypoxia, by demonstrating the capability of hypoxic neutrophils to damage lung endothelial cells in vitro, and by identifying additional potential histotoxic mediators of this damage, which are upregulated in hypoxic neutrophil supernatants. Translating these results to the clinical context of COPD, I have shown further upregulation of these histotoxic proteins in supernatants from hypoxic COPD neutrophils and detected a plasma signature of increased protease (both NE and PR3) activity in these patients. Furthermore, my proteomic examination of the neutrophil secretome has suggested a novel mechanism of selective protein secretion under hypoxia. My work will instigate future research projects and potentially identify novel therapeutic targets. These data could be extrapolated to a wide range of neutrophilic inflammatory diseases manifesting endothelial dysfunction, such as bronchiectasis, RA and IBD (235,453). 6.2 Discussion of results 6.2.1 Differential effects of priming agents on hypoxic degranulation In the work presented in this thesis, I have extended the observation that hypoxia markedly increased degranulation from GM-CSF-primed neutrophils (149,315) by showing that this hyper-secretory response is also present (in fact with a greater and more consistent effect) in PAF-primed cells, broadening the relevance to a wider range of inflammatory scenarios. For example, PAF has been implicated in the inflammatory response to I/R injury (where interruption of blood supply results in severe tissue hypoxia), promoting local neutrophil recruitment, oedema and endothelial dysfunction, and also priming circulating neutrophils to initiate distant neutrophil-endothelial interactions in multiple vascular beds, resulting in multi- organ dysfunction and circulatory collapse (454). In contrast, the “hypoxic uplift” of 175 degranulation was not consistently seen with TNFα-primed neutrophils. There are a number of possible explanations for this result. Firstly, the concentration of TNFα used might have been sufficient to induce a maximal response alone, and it would be interesting to explore the interaction of hypoxia with lower concentrations of this cytokine. Secondly, donor variability in the magnitude of the response to TNFα-priming, previously observed in our laboratory when measuring ROS production, may have obscured differences between normoxia and hypoxia. Thirdly, different agonists may be entraining different signalling pathways and hence different functional effects. Multiple signal transduction pathways induce the molecular events required for priming: GM-CSF ligates type 1 cytokine receptors, initiating signalling via recruitment and phosphorylation of Janus kinases (JAKs) and STAT molecules; PAF receptors are GPCR- linked, initiating a range of signalling pathways following dissociation of the Gα and Gβγ subunits; and TNFα can bind either TNF receptor 1 (TNFR1) or TNF receptor 2 (TNFR2), triggering pro-inflammatory pathways mediated by JNK and NFκB. The complex mechanisms of neutrophil priming and degranulation are still incompletely understood (353), partly due to the complicated and inter-linked intracellular signalling pathways involved, and partly due to the need to consider the numerous priming and activating agents present in inflammatory situations in vivo. For example, PAF (a priming agent) and fMLP (an activating agent) both ligate GPCRs; however their activation of downstream MAPK signalling involves different pathways, with PAF- but not fMLP-induced phosphorylation of ERK dependent on protein kinase C and extracellular calcium influx (455). Despite both TNFα and GM-CSF being priming agents, neutrophil transcriptomic analysis using RNA-sequencing has shown they induce substantially different gene expression profiles (456). Furthermore, whilst GM-CSF uniformly delays neutrophil apoptosis, TNFα may be pro- or anti-apoptotic, dependent on time and concentration, suggesting that priming by these agonists is not equivalent (10). In keeping with these observations, my data suggest that the increase in neutrophil degranulation under hypoxia is dependent on signalling pathways specific to the priming agonist rather than an overall “priming event”. 6.2.2 Role of PI3K signalling in hypoxic degranulation Published experiments from our laboratory indicated a role for PI3K signalling in hypoxic neutrophil degranulation, as pan- and γ-selective PI3K inhibitors, but not PI3Kδ inhibitors, abrogated the hypoxic uplift seen with GM-CSF primed neutrophils (149). Class I PI3Ks are heterodimers, signalling through downstream effectors such as AKT, to control a variety of processes, including priming of the neutrophil respiratory burst and degranulation. Class IA PI3Ks (α/β/δ) are usually activated via tyrosine kinase receptors, while Class IB PI3Ks (γ) are primarily activated by GPCR βγ subunits. Both PI3Kγ and δ are fundamental to neutrophil function (304). It is possible that the observed role of PI3Kγ in the hypoxic enhancement of 176 neutrophil degranulation is related to its role in (GPCR-linked) fMLP signalling. However, this seems unlikely as I did not observe the hypoxic uplift response for TNFα (despite subsequent fMLP treatment); in addition, previous experiments have shown that addition of PI3K inhibitors immediately prior to fMLP stimulation reduced degranulation under normoxia and hypoxia equally, with preserved hypoxic fold-upregulation (149), implying that hypoxia itself entrains PI3Kγ signalling pathways, which then mediate the augmented release of neutrophil granule contents stimulated by further priming/activation events. Previous reports have shown PIP3 accumulation occurs in neutrophils in response to GM-CSF treatment via phosphorylation of the Class IA PI3K p85 subunit (457,458), and it would be of interest to study p85 phosphorylation in the setting of hypoxia. At present, the link between hypoxia and PI3Kγ activity is unclear, particularly given the short time-frame required to establish the response and the lack of a role for the transcription factor HIF (10). Hypoxia does not increase neutrophil protein content of either p85 or the catalytic subunits of PI3Kγ and δ, although the levels of the PI3Kγ regulatory subunits p101 and p84 were not assessed (Dr Kim Hoenderdos, personal communication). Using γ- and δ-selective inhibitors, I have also demonstrated a non-redundant role for PI3Kγ in the hypoxic enhancement of degranulation from PAF-primed neutrophils, which is logical as PAF ligates GPCRs, triggering PI3Kγ activation via the dissociated Gβγ subunit. In order to confirm that the PI3Kγ-dependent hypoxic response seen in human neutrophils was not an off-target effect or due to a blanket inhibition of granule exocytosis, I confirmed the PI3Kγ- dependence of this effect using murine neutrophils lacking PI3Kγ. The mechanism by which hypoxia modulates PI3Kγ signalling is currently unclear and warrants further investigation, particularly in view of the lack of HIF-dependence. In a clear cell renal carcinoma cell line, Bienes-Martínez et al. demonstrated that hypoxia was able to augment AKT phosphorylation and downstream signals in a HIF-independent manner, which the authors suggested may be partly due to effects of low oxygen on PHDs (459). In neutrophils, Xu et al. showed that accumulation of ROS in apoptotic cells inhibited PI3Kγ activity (460); a reduction in ROS in the setting of hypoxia could allow enhanced PI3Kγ signalling and explain why we observed the hypoxic uplift effect under true hypoxia but could not recapitulate this with HIF mimetics (149). Another potential mechanism might be for hypoxia to modulate the negative regulation of PI3K activity provided by lipid phosphatases such as PTEN, which converts PIP3 to PIP2. Recent preliminary data generated by Professor Sarah Walmsley (personal communication) have indicated a decrease in the PTEN content of BALF neutrophils from hypoxic compared with normoxic LPS-challenged mice, although as noted above the short (4 hour) timecourse of my experiments and lack of HIF-dependence suggest that transcriptional regulation of protein synthesis is less likely. However, it is possible that hypoxia (de)stabilises proteins 177 involved in this and other signalling pathways independent of protein synthesis, e.g. by altering phosphorylation. Investigation of the effect of hypoxia on neutrophil signalling pathways using phospho-proteomics would form the basis of an exciting future research project. It is possible that hypoxia impacts actin cytoskeletal rearrangement to enhance degranulation responses. Hypoxia re-distributes F actin to form cap-like structures (149), and cytoskeletal rearrangement is linked to enhanced neutrophil degranulation (461). It is conceivable that hypoxic reorganisation of the cytoskeleton and/or plasma membrane facilitates access of PI3Kγ to its PIP2 phospholipid substrate. In a breast tumour model, mass spectrometric imaging of hypoxic regions identified upregulation of proteins involved in actin cytoskeleton regulation and PI3K-AKT signalling by network pathway analysis, with co-localization of various lipid species in hypoxic areas (462); however this study employed a HIF-1α-driven HRE-fluorescent protein construct to identify areas of hypoxia, suggesting these effects may be dependent on HIF-mediated transcription. In keeping with our finding that hypoxic degranulation is HIF-independent, Wottawa et al. demonstrated in a breast cancer cell line that the actin-bundling protein T-plastin was recruited to the plasma membrane, and mediated enhanced endocytic uptake in a manner independent of HIF transcription (147). Although not studied in hypoxic conditions, the T-plastin homologue, L-plastin, which is abundant in neutrophils, has been shown to be phosphorylated in response to fMLP treatment, dependent on PI3K, phospholipase D and/or protein kinase C activity (463). Finally, although it is the small GTPase Rac2 that mediates neutrophil actin dynamics and azurophil granule secretion, its homologue Rac1 was shown to be upstream of HIF-1α in human embryonic kidney cells, and Rac1 activity was increased by hypoxia within 30 minutes in a PI3K-dependent manner (464). In addition to effects on cytoskeletal reorganisation, hypoxia may modify downstream granule transport and docking mechanisms. For example, Rab27a (which in neutrophils directs granules to the plasma membrane for fusion) was upregulated and displayed increased activity following hepatocyte hypoxia/re-oxygenation in vitro and murine liver I/R injury in vivo; Rab27a silencing decreased liver neutrophil numbers and serum NE in this model (465). In pancreatic β cells, PI3K signalling promoted Rab27a-mediated actin bundling, which played a pivotal role in endocytosis-exocytosis coupling (466), and in murine neutrophils, reduced AKT phosphorylation was linked to downregulation of Rab27a with a resultant decrease in azurophil granule exocytosis in a tumour setting, which is often a hypoxic environment (467). Hence, there is a plausible mechanistic link between hypoxia, cytoskeleton regulation, plasma membrane and granule trafficking, and PI3K(γ) signalling. 178 PI3Kγ signalling has been implicated in NET formation (468,469). As neutrophil proteases, such as NE and MPO, are released in association with NETs, we considered whether increased NETosis might contribute to the hypoxic enhancement of granule protein release. In the literature, the effect of hypoxia on NET generation is controversial, with studies showing it to be both increased (within hypoxic liver tumours, (117)) or unchanged (neutrophils incubated under hypoxia (116)). It is possible that the more complex tumour environment provides additional cues for NETosis, to explain these differing results; both studies showed that products released by other hypoxic cells could stimulate NET release, which is likely relevant in vivo. It might be expected that ROS-dependent NETosis would be decreased under true hypoxia, since ROS generation will be limited by the lack of molecular oxygen; however, the absolute requirement for ROS has been challenged and is likely agonist-dependent (468). Similar to the results obtained by Nakazawa et al. (116), I did not see any difference in NET generation between normoxia and hypoxia using precisely the same experimental conditions as in my degranulation experiments, so it is unlikely that NETosis contributes significantly to the augmented release of neutrophil granule products under hypoxia in this setting. 6.2.3 Possible mechanisms of endothelial damage by hypoxic supernatants In keeping with my finding that hypoxic neutrophil supernatants contain increased active degranulation products, and our previous observation that these hypoxic supernatants caused extensive airway epithelial cell damage (149), I have shown that supernatants from hypoxic compared with normoxic neutrophils also cause substantially more damage (both detachment and cell death) to pulmonary artery endothelial cell monolayers. Whilst it is difficult to translate the effects of supernatants derived from large numbers of maximally activated neutrophils applied to endothelial cells for several hours with the situation in the circulation in vivo, it is noteworthy that individual neutrophils will be in direct contact with individual endothelial cells. Degranulation enhances the surface expression of NE and other proteases (470), and although I did not assess this directly, future work will address this possibility (see below). As well as direct cellular injury mediated by granular proteases, modification of other circulating mediators may be relevant in vivo. My supervisor, Dr Li, has demonstrated that NE is responsible for the proteolytic degradation of BMP9, an endogenous circulating factor which promotes endothelial integrity, and supernatants from hypoxic vs normoxic neutrophils resulted in significantly more NE-dependent BMP9 cleavage (unpublished data, manuscript in preparation). My additional preliminary unpublished data, generated with Kimberley Wiggins (PhD student, Laboratory of Dr Murray Clarke, University of Cambridge), showed that incubation with hypoxic vs normoxic neutrophil supernatants resulted in more cleavage of a precursor of the pro-inflammatory cytokine IL-1β into active fragments (Figure 6.1). Hence, 179 active NE can cause endothelial damage in a tissue culture setting and has the potential to initiate further pro-inflammatory events in vivo. Figure 6.1: The effect of neutrophil supernatants on IL-1β Supernatants from normoxic (N) vs hypoxic (H) PAF/fMLP (P) or IMDM (C)-treated neutrophils were incubated with pro-IL-1β for 2 h or overnight (ON). Western blotting was subsequently performed with anti-IL-1β. Image representative of n=1 experiment. Data generated with Kimberley Wiggins. However, since my results indicated that there was a serine protease-independent component to neutrophil-mediated endothelial damage (as cell death could not be completely abrogated by the addition of α1AT), I performed a proteomic examination of the normoxic vs hypoxic neutrophil secretome. Based on previous data generated in our laboratory showing differential secretion from eosinophils under hypoxia (471), it was thought that these data may also give mechanistic insights regarding neutrophil protein release under hypoxia. Proteomic analysis of the normoxic vs hypoxic secretome revealed a hypoxic enrichment of multiple potentially histotoxic proteins but surprisingly, whilst MPO was the most upregulated protein in the hypoxic vs normoxic secretome, the increase in NE as measured by proteomic analysis did not reach significance (adj. p value=0.25). This discrepancy might reflect the fact that a small change in the amount of this protease can lead to larger detectable changes in activity. An alternative explanation is a reduction in the concomitant release of inhibitory proteins; notably, leukocyte elastase inhibitor was significantly upregulated in normoxic vs hypoxic supernatants, suggesting an alteration in the protease-anti-protease balance in the 180 setting of hypoxia. Finally, the complex technical requirements for proteomics (including the limited number of samples that can be analysed simultaneously), the expense, and the requirement for multiple testing correction methods in data analysis may lead to a failure to detect true positives using this demanding methodology (373). 6.2.4 Proteomic analysis of the neutrophil secretome My proteomic dataset was very intriguing for two main reasons. Firstly, the hypoxia- upregulated histotoxic proteins did not segregate precisely with discrete granule populations, with upregulation of some granule-derived proteins in normoxic relative to hypoxic supernatants as well as vice versa. Secondly, although most cytoplasmic protein release was upregulated under normoxia, some cytoplasmic proteins, including cyclophilin A, were increased in hypoxic supernatants, even though hypoxia is a neutrophil survival signal, suggesting that they were being actively secreted. Together, these findings suggest that hypoxia may drive a novel neutrophil secretion mechanism, as opposed to simply an upregulation of classical degranulation. As cyclophilin A has been shown to be secreted from other cell types, e.g. cardiac myocytes in response to hypoxia (387), it may be that the secretion mechanism for this protein is unique. However, the cytoplasmic protein S100A9, which was upregulated under normoxia, has also been shown to be actively secreted (393). Although my characterisation of the hypoxic neutrophil secretome is entirely novel, a limited number of secretome analyses of normoxic neutrophils treated with various stimuli have been performed previously (305–308). Consistent with my findings, their results did show some variation in secretome content according to the inciting stimulus: Malmstrӧm et al. treated neutrophils with TNFα, heat-killed S. pyogenes, or cytochalasin B/fMLP and observed NGAL and MPO release in response to all three treatments but NE release only with cytochalasin B/fMLP or heat-killed S. pyogenes, and cathepsin G release only with cytochalasin B/fMLP (306). Snäll et al. visualised an increase in azurophil granule co-localisation of resistin and azurocidin upon treatment with a specific streptococcal strain, which was not evident on stimulation with LPS or fMLP. Furthermore, only fMLP-treatment, when compared with streptococcal M1 protein, LPS or PMA, resulted in a differential secretion profile of resistin and azurocidin (378). In addition, fusion of different granule subtypes has been observed: Bjӧrnsdottir et al. demonstrated fusion of specific and azurophil granules using lactoferrin and CD63 as markers, albeit in response to PMA, which is a non-physiological stimulus (472). These studies suggest that the inflammatory environment can influence the nature of granule subsets that are secreted, which may have a bearing on the precise proteins released during degranulation. However, the mechanism(s) by which these events might occur is unknown, and how hypoxia might access such mechanisms is likewise uncertain. 181 6.2.5 Potential mechanisms of “differential degranulation” The effect of chronic inflammation and hypoxia on granule formation and packaging is currently unknown. As granule subtype formation and protein expression is not entirely synchronised, there is a degree of granule subtype heterogeneity, for example there are defensin-rich and defensin-poor azurophilic granules (473). However, NE vs MPO discrepancy within azurophilic granules has not been described for healthy human neutrophils in this fashion, although there may be some insights from human disease and animal models. Divergence in granule packaging pathways has been revealed by mutations of AP3B1, encoding adaptor protein-3 (causing Hermansky-Pudlak syndrome), which resulted in reduced NE levels but normal MPO and PR3 levels (474). Moreover, azurophilic NE was absent in serglycin-deficient mice, with no effect on MPO, PR3 or cathepsin G (475). Intriguingly, serglycin was significantly upregulated in my hypoxic neutrophil secretome analysis (FC 1.9, adj. p value = 0.012). It is possible that granule formation during maturation may be modulated, for example in chronic inflammatory states, with the potential for serglycin to direct increased azurophil granule content of NE. However, this would not explain my results which were obtained by short-term incubation of terminally differentiated mature cells. A potential mechanism by which hypoxia might result in differential granule protein release is by movement of proteins into/between neutrophil granules in a manner analogous to the “piecemeal” degranulation described for eosinophils and mast cells, which allows release of selected granule-stored proteins. Eosinophils contain cationic proteins and cytokines as pre- formed pools within secretory granules, which can be released by classical granule fusion with the plasma membrane or by piecemeal degranulation, whereby selected proteins or cytokines are released via vesicle or vesiculo-tubular sequential emptying of granules. This process allows differential secretion which is stimulus-dependent (379). Published data from our laboratory have shown that hypoxia promoted IL-8 release but attenuated eosinophil-derived neurotoxin release from eosinophils, providing the first evidence of hypoxia selectively modulating granulocyte secretion (471); this finding is highly relevant to my work and might indicate a mechanism of hypoxic degranulation that is common to the granulocyte lineage. It would be informative to visualise the subcellular localisation of my hypoxia-regulated proteins to see if a similar mechanism is employed by neutrophils. 6.2.6 Enhanced release of cytoplasmic proteins Intrigued by the surprising observation that hypoxia upregulated the release of certain cytoplasmic proteins, I undertook a detailed review of the literature which revealed that the majority of my hypoxia-upregulated proteins have been found in NDMVs. I therefore explored the impact of hypoxia on NDMV release. As MVs enable multiple cell types to release 182 packages of cytoplasmic proteins and studies have suggested that hypoxia increases the release of NDMVs (187,188,190), this represented a plausible mechanism by which neutrophils may actively secrete cytoplasmic proteins under hypoxia. NDMVs were challenging to isolate in sufficient numbers for analysis, prompting a collaboration with Dr Victoria Ridger (University of Sheffield). However, even using her established protocols and working in her laboratory alongside her group members with expertise in these methods, I was unable to detect an increase in NDMVs under hypoxia. Consistent with this, I did not see any difference in NDMVs between COPD patient and healthy control plasma, which alleviated concerns that the cell and NDMV isolation techniques might mask differences between samples. Furthermore, there was no difference in NDMV content of cyclophilin A between normoxia and hypoxia, and cyclophilin A could also be detected in NDMV-free supernatants. Thus, my integrated dataset suggests that MV release does not significantly contribute to the increased secretion of cytoplasmic proteins from hypoxic neutrophils. Of note, hypoxia is a powerful neutrophil survival factor and I have shown no increase in cell death under these experimental conditions; if cell death were responsible for the increased detection then other cytoplasmic proteins would also have been upregulated in the hypoxic secretome. It would be interesting to explore whether cytoplasmic proteins are somehow transferred into granules for secretion or whether an alternative mechanism, such as reverse pinocytosis, is instigated. 6.2.7 Translation to COPD and beyond I wished to translate my proteomic findings to a clinically relevant patient group. As noted above, COPD is associated with airway neutrophilia and hypoxia, and with excessive cardiovascular morbidity. Evidence is emerging that inflammation plays a key role in both the progression of airway disease and development of CVD in COPD. I demonstrated a trend (which did not achieve statistical significance) to increased NE release from normoxic COPD vs healthy control neutrophils, which may reflect a pro-inflammatory or hypoxic signature in a subset of these patients. This observation could reflect the primed circulating neutrophil phenotype reported in exacerbating COPD patients (283) although I did not observe any difference in shape change between untreated COPD and healthy neutrophils. Importantly however, I have shown that neutrophils isolated from exacerbating COPD patients release even more NE when stimulated under hypoxia than do cells from healthy controls. Furthermore, the fold increase in NE release from normoxic vs hypoxic cells was greater for COPD than healthy controls, indicating that, rather than just starting from an increased baseline, hypoxia is acting to further augment degranulation in these cells. Extrapolating from my earlier results, this would confer the potential for causing increased endothelial injury. A previous study has shown increased NE activity in neutrophil lysates from COPD patients compared with healthy controls (219), although NE content was not measured biochemically 183 and proteomic examination also revealed reduced leukocyte elastase inhibitor (Serpin B1) so this may represent a protease/anti-protease imbalance rather than a true increase in NE levels. Still, this is physiologically relevant and, as previously noted, my proteomic screen confirmed higher levels of leukocyte elastase inhibitor in the normoxic compared with hypoxic neutrophil secretome (FC 1.8, adj. p value = 0.013). Reflecting these NE data, I showed an increase in resistin and NGAL in COPD vs healthy neutrophil supernatants under hypoxia, again demonstrating that the hypoxic upregulation of degranulation was increased in the disease population. Further relating the pattern of the proteomic data to COPD, I showed a similar hypoxic increase in the release of the cytoplasmic protein, cyclophilin A. Taken together, these results suggest that COPD neutrophils exhibit a hypoxic hyper-secretory phenotype and may therefore have enhanced capacity for tissue damage. These findings are particularly exciting as they may represent novel therapeutic targets. Although already associated in the literature with endothelial dysfunction in various in vitro and in vivo situations (406,408,413), it would still be important to establish whether these proteins cause direct endothelial injury. Despite seeing an increase in plasma breakdown products of both NE and PR3 activity in COPD, I did not observe a significant increase in resistin, NGAL or cyclophilin A in exacerbating COPD patient vs healthy control plasma. It is likely that this experiment was not sufficiently powered to detect small changes in these circulating neutrophil proteins (given my sample size calculation requirement of a minimum of 16 per group based on my previous NE activity data), or the results could have been impacted by the heterogeneity of the COPD patient cohort. Indeed, there is likely a spectrum of neutrophil abnormality within this diagnosis and it is important to note that changes at a population level may not represent consistent features of the disease. Plasma levels of these proteins have been shown to vary in the literature in an infection setting: plasma resistin was increased in patients with sepsis compared with critically ill non-infected patients, and in a mixed group of sepsis patients (with pathogens including S. pyogenes, S. pneumoniae, E. coli, Neisseria meningitidis and unknown pathogen), plasma MPO and resistin levels were increased overall whereas NGAL was decreased (306). Moreover, significantly more resistin was secreted from healthy neutrophils treated with Group A Streptococcus than either S. aureus or E. coli (378). It would be informative to pursue plasma levels of these proteins in a larger, well-phenotyped COPD patient group, using chemiluminescence-based technology, which allows increased sensitivity and a wider range of detection. 184 Despite multiple lines of evidence suggesting a prominent role for NE in COPD, phase 2 clinical trials of the small molecule NE inhibitor, AZD9668, showed no benefit in lung function, exacerbation incidence, respiratory symptoms or quality of life (476,477). However, these were relatively short-term studies undertaken after significant irreversible airway damage had likely already occurred, and neither study included cardiovascular endpoints, which my data suggest may be a promising avenue to examine. It has been proposed that the limited clinical success of NE inhibitors may be due to a combination of inadequate patient phenotype selection, the short study period, the inability to attain stoichiometric equivalent concentrations of inhibitor at neutrophil degranulation sites, instability of the inhibitor in vivo, or failure of inhibitors to access neutrophils adjacent to the ECM (478). A further reason may be that NE inhibition alone is not enough to prevent tissue damage, which would be consistent with my results obtained following neutrophil supernatant treatment of endothelial cells. Whilst it is difficult to equate the in vitro incubation of neutrophil supernatants and endothelial cells to the in vivo pathological setting, the fact that circulating neutrophil granule products, including resistin and NGAL, have been associated with worse cardiovascular outcomes (479,480) and my demonstration of increased protease activity in COPD vs healthy plasma, make the hypoxic uplift of degranulation a biologically plausible mechanism for promoting endothelial dysfunction in these patients, in combination with perturbations of circulating inflammatory mediators. It would be interesting to pursue my observations in a more sensitive and physiologically relevant system, for example assessing the effect of COPD vs healthy neutrophil supernatants on neutrophil-endothelial interactions in a flow adhesion system. What might be the translational potential of my work? Endothelial dysfunction is a poorly understood feature of COPD, associated with substantial cardiovascular comorbidity and mortality. Having demonstrated endothelial damage caused by hypoxic neutrophil supernatants, suggested a novel mechanism of enhanced neutrophil secretion under hypoxia, and identified specific histotoxic protein targets, my work has the potential to inform the development of disease-modifying therapeutics in COPD, which could be extrapolated to a wide range of neutrophilic inflammatory diseases underpinned by hypoxia and endothelial dysfunction. In the setting of respiratory medicine, it would be relevant to investigate these mechanisms in bronchiectasis; this condition is associated with marked neutrophilic airway inflammation and hypoxia in an animal model (222), with established correlations between airway NE and disease severity (70), and emerging evidence of disproportionate cardiovascular morbidity (453,481). My research supports a key role for NE in promoting cardiovascular comorbidity in COPD and my work could, therefore, prompt a focus on cardiovascular endpoints for small molecule NE inhibitors (and other anti-protease strategies) in clinical trials in COPD and other inflammatory diseases. 185 I have also identified PI3Kγ as a mediator of increased hypoxic neutrophil NE release. Increased PI3K signalling has been implicated in COPD: neutrophils isolated from COPD patients displayed impaired migratory accuracy (potentially causing bystander tissue damage) which could be improved by pan-PI3K inhibition (216); additionally, reduced PTEN protein expression, which correlated with the severity of airflow obstruction, and increased AKT phosphorylation were identified in peripheral lung tissue from COPD patients (482). Furthermore, pan-PI3K inhibition prevented formation and progression of murine abdominal aortic aneurysms (483), and genetic PI3Kγ deletion in mice reduced airway inflammation, hyper-responsiveness and remodelling in an asthma model (484) and reduced airway inflammation and structural lung damage in a CF model, with a mortality benefit (485). As several PI3K inhibitors are in Phase 3 clinical trials (mostly in the oncology domain), the exciting prospect of translation to COPD and beyond could be realised in the relatively near future. Since PI3Kγ is essential for optimal neutrophil function, long term systemic targeting is likely to impair host defence and increase infection susceptibility; potential options to mitigate such risks include the use of inhaled inhibitors to deliver compounds directly to the diseased airway, and perhaps limited short-term use of systemic inhibitors at high-risk times such as following exacerbations. More speculatively, identification of hypoxia-upregulated histotoxic proteins could inform the development of new targeted treatments, and further exploration of the hypoxic control of neutrophil secretion, e.g. using phospho-proteomics, may lead to new therapeutic opportunities for a wide range of inflammatory diseases. 6.3 Future research avenues As hypoxia drives neutrophil hyper-secretion of multiple histotoxic proteins and proteases, coupled with the fact that inhibition of NE alone has not had a major impact in clinical trials, it seems highly relevant to further explore the mechanisms underlying the establishment of this destructive neutrophil phenotype and to identify the most damaging effectors. Enhanced understanding of this phenomenon may lead to the identification of modifiable pathways/targets contributing to tissue injury in situations where inflammation and hypoxia co- exist. 6.3.1 Investigating the role of hypoxia-upregulated proteins in mediating clinically relevant endothelial damage Supernatants derived from hypoxic neutrophils contain a multitude of histotoxic proteins and proteases, and cause substantial endothelial cell damage. Having also shown that these hypoxia-upregulated proteins are even further increased in neutrophil supernatants from exacerbating COPD patients, it would be informative to examine the ability of these patient- derived supernatants to mediate endothelial injury, compared with supernatants derived from 186 patients with stable disease and those from healthy controls. As well as assessing endothelial damage in static culture, it would be interesting to progress these experiments into a more physiological flow-based adhesion system, measuring rolling, arrest and transmigration of neutrophils perfused over endothelial cells (pre-treated with COPD vs healthy supernatants) by time lapse imaging. Pulmonary neutrophil recruitment is associated with changes in vascular permeability in vivo (reviewed in (252)) and may further exacerbate endothelial damage. Moreover, the endothelial adhesion molecule ICAM-1 correlates with emphysema on CT imaging and lung function decline in COPD (486), and I have shown that supernatants from hypoxic neutrophils increase HPAEC ICAM-1 expression in static culture, all suggesting that supernatants from hypoxic COPD neutrophils would augment neutrophil adhesion and transmigration under flow and in the in vivo setting. As well as demonstration of aberrant neutrophil behaviours in COPD, pulmonary vascular endothelial cells from COPD patients may be intrinsically abnormal, having been shown to display markers of senescence, such as telomere shortening, which was linked with lung inflammation (487). Derivation of patient blood outgrowth endothelial cells from circulating endothelial progenitors in whole blood could be used to investigate whether there is increased susceptibility of these cells to neutrophil- mediated damage, which could also be combined with perfusion of patient neutrophils. Having identified NE, resistin, NGAL and cyclophilin A as candidate histotoxic proteins which are upregulated in hypoxic neutrophil supernatants (particularly in COPD), it will be important to examine the capacity of these proteins to cause direct endothelial injury. Further assessment of a more detailed plasma signature for these proteins in COPD vs healthy plasma could be investigated using multiplex electrochemiluminescence technology, which has a high sensitivity and wide detection range. Additionally, circulating markers of endothelial damage could be measured, such as angiopoietin-2, VEGF and sFlt-1 (soluble VEGF receptor) (241,488). The key experiments to establish the role of hypoxic neutrophil secretion products in mediating endothelial damage would help to establish whether inhibition of PI3Kγ signalling (or other yet-to-be determined mechanistic pathways) or depletion/inhibition of specific hypoxia-upregulated proteins alone or in combination, might be able to ameliorate endothelial injury. Initial in vivo translation of these data could involve trialling inhibitors of PI3Kγ or specific proteins in a chronic hypoxia (or COPD) rodent model and assessing both pulmonary inflammation and cardiovascular outcomes. Excitingly, several PI3K inhibitors are already in oncological clinical trials, paving the way for translation to inflammatory disorders. As noted above, the potential for benefit will need to be weighed against possible side effects, and mitigation strategies explored, such as targeted delivery of short-duration dosing. Although α1AT did not prevent all endothelial injury induced by hypoxic supernatants, it did provide 187 substantial protection, and α1AT augmentation therapy is in clinical use to treat patients deficient in this protein. Whilst the impact of this therapy on respiratory outcomes has not been overwhelming (489), there is little data on cardiovascular comorbidity and this would be plausible to investigate in animal models and in patients. Overall, the delineation of the nature and extent of hypoxic neutrophil-mediated endothelial damage in the clinical context of COPD, and further identification of the responsible factors and degranulation processes, has the potential to inform disease-modifying therapy in a broad range of hypoxic inflammatory lung diseases with enhanced cardiovascular risk. 6.3.2 Investigating how hypoxia impacts neutrophil intracellular signalling and granule trafficking pathways to enhance histotoxic protein release I have shown that PI3Kγ plays a prominent role in mediating enhanced NE release from neutrophils under hypoxia. However, the mechanism by which hypoxia enables enhanced PI3Kγ signalling and the responsible downstream pathways remain unclear, and it is likely that other signalling events are involved. An initial candidate approach could be employed: for example, assessing the hypoxic regulation of PI3Kγ or negative regulator, PTEN, expression by western blotting of neutrophil lysates or by measurement of downstream lipid mediators, such as PIP3. As a complementary approach, phospho-proteomics is a powerful technique which could be utilised to characterise the effect of hypoxia on neutrophil protein phosphorylation, allowing large scale bioinformatic network analysis but also with the potential to identify new phosphoprotein candidates. As the actin-bundling protein, T plastin, has been shown to regulate hypoxia-induced membrane trafficking in a HIF-independent fashion in other cell types (147), neutrophil L plastin, which is phosphorylated in response to fMLP- treatment, represents an example of a plausible phosphoprotein candidate which may be implicated in the hypoxic regulation of degranulation. Given my proteomic data demonstrating the upregulated release of both selected granule proteins and a small subset of cytosolic proteins under hypoxia, it would be valuable to understand the mechanism(s) controlling this differential release. My results suggest that this phenomenon is not due to NET or NDMV release, but previous data from our group has indicated an effect of hypoxia on modulating the neutrophil cytoskeleton, redistributing sub- cortical actin into a focal cap (315). It would be exciting, though challenging, to visualise the effect of hypoxia on neutrophil shape, subcellular structure and granule distribution using live cell imaging. New technology exists, which may facilitate these experiments, including the development of small transportable containers which maintain the temperature, humidity and oxygenation status of the hypoxia workstation. 188 Exploration of neutrophil intracellular vesicle trafficking, which is controlled by Rab family proteins, and vesicle membrane docking, which is mediated by interaction of SNAREs, may allow enhanced understanding of how neutrophils selectively upregulate protein release under hypoxia. Hypoxia can increase the expression and/or activity of Rab27a (which directs vesicles for plasma membrane fusion) and the vesicle-associated membrane protein VAMP- 2 (which complexes with membrane SNAREs) in liver and neuronal cells, respectively (465,490). Neutrophil total content of these proteins under hypoxia could be determined by western blotting, and activity by techniques such as pull-down assays for small GTPases. The plasma membrane content of SNAREs under hypoxia vs normoxia could be examined by plasma membrane biotinylation and quantification of protein abundance by TMT-MS, a technique which has been optimised for neutrophils in our laboratory. Subcellular localisation is also key: Rab27a is relatively lacking on neutrophil azurophilic granules, which are hence directed for phagosome rather than plasma membrane fusion. Examination of Rab localisation during normoxic and hypoxic incubation by immunofluorescence or electron microscopic analysis, with co-staining of the relevant granule populations, would enhance our understanding of the effect of hypoxia on the localisation of vesicle trafficking proteins. These techniques could also be employed to ascertain the effect of hypoxia on granule localisation of candidate hypoxia up- and down-regulated proteins (including those of cytoplasmic origin), predicting that hypoxia regulates their movement into/between vesicles in a manner analogous to the piecemeal degranulation described for eosinophils and mast cells, which subsequently impacts their release. These experiments have the potential to identify an entirely novel mechanism of neutrophil protein secretion. Overall, further investigation of how hypoxia impacts neutrophil signalling and granule trafficking pathways would not only add significantly to our scientific knowledge of this largely unexplored field but may highlight new research avenues, including identification of additional clinically relevant targetable pathways/mechanisms. 6.4 Conclusions Persistent and excessive neutrophilic activation has been implicated in inflammatory damage to bystander cells and tissues, contributing to the pathogenesis of both acute and chronic inflammatory diseases. Regions of infection and inflammation are profoundly hypoxic and, in diseases such as COPD, systemic hypoxia can develop. Hence, a greater understanding of how neutrophils function in physiologically and pathologically relevant low oxygen tensions is paramount in order to inform the development of new treatments to combat tissue injury and disease progression. In this thesis, I have extended our knowledge of the impact of hypoxia on neutrophil degranulation by showing that hypoxia enhances the release of damaging 189 granular proteases in an agonist- and PI3Kγ-dependent manner, and that supernatants from these hypoxic cells cause substantial injury to endothelial cells. Thus, my results suggest that this destructive neutrophil phenotype may promote endothelial dysfunction in a wide range of diseases where inflammation and hypoxia co-exist, potentially contributing to increased cardiovascular risk. As the endothelial damage from hypoxic neutrophil supernatants was not entirely protease-dependent, I performed an analysis of the normoxic vs hypoxic neutrophil secretome. This characterisation identified novel hypoxia-upregulated histotoxic protein targets, which have the potential to lead to the development of new treatment strategies. Translating these results to a clinically relevant patient cohort, I confirmed that the release of selected potentially histotoxic proteins was even further increased in supernatants from exacerbating COPD patient neutrophils under hypoxia. I hope that these data will create further interest in exploring inhibition of new histotoxic protein candidates and/or PI3Kγ signalling in this and related disease settings. Unexpected and intriguing proteomic results, demonstrating that differential protein secretion under hypoxia did not entirely segregate with distinct granule populations and that release of certain cytoplasmic proteins was consistently upregulated, generated further questions regarding hypoxic control of neutrophil secretion mechanisms. It would be exciting to explore the effect of hypoxia on signalling pathways, and granule/vesicle packaging and trafficking in neutrophils. These studies may provide a new array of therapeutic targets or mechanisms which could be exploited for patient benefit. 190 7 Appendices 7.1 10-plex TMT-MS and data analysis parameters Full parameters for 10-plex TMT-MS and data analysis, provided by Dr Mike Deery and Dr Marco Chiapello (CCP), are as follows. 7.1.1 LC-MS/MS The dried sample was resuspended in 30 μl of 0.1% formic acid and placed into a glass vial. 1 μl was injected by the HPLC autosampler and separated by the LC method detailed below. LC-MS/MS experiments were performed using a Dionex Ultimate 3000 RSLC nanoUPLC (Thermo Fisher Scientific) system and a Lumos Orbitrap mass spectrometer (Thermo Fisher Scientific). Peptides were loaded onto a pre-column (Thermo Scientific PepMap 100 C18, 5 mm particle size, 100 A pore size, 300 mm x 5 mm length) from the Ultimate 3000 auto- sampler with 0.1% formic acid for 3 min at a flow rate of 10 ml/min. After this period, the column valve was switched to allow elution of peptides from the pre-column onto the analytical column. Separation of peptides was performed by C18 reverse-phase chromatography at a flow rate of 300 nl/min using a reverse-phase nano Easy-spray column (Thermo Scientific PepMap C18, 2 mm particle size, 100 A pore size, 75 mm x 50 cm length). Solvent A was water + 0.1% formic acid and solvent B was 80% acetonitrile, 20% water + 0.1% formic acid. The linear gradient employed was 2-40% B in 93 minutes. The total LC run time was 120 min, including a high organic wash step and column re-equilibration. The eluted peptides from the C18 column LC eluant were sprayed into the mass spectrometer by means of an Easy-Spray source (Thermo Fisher Scientific). All m/z values of eluting peptide ions were measured in an Orbitrap mass analyzer, set at a resolution of 120,000 and were scanned between m/z 380-1500 Da. Data dependent MS/MS scans (top speed) were employed to automatically isolate and fragment precursor ions by collision-induced dissociation (CID, Normalised Collision Energy (NCE): 35%), which were analysed in the linear ion trap. Singly charged ions and ions with unassigned charge states were excluded from being selected for MS/MS and a dynamic exclusion window of 70 s was employed. The top ten most abundant fragment ions from each MS/MS event were then selected for a further stage of fragmentation by Synchronous Precursor Selection (SPS) MS 3 in the HCD (High energy Collisional Dissociation) high energy collision cell using HCD (NCE: 65%). The m/z values and relative abundances of each reporter ion and all fragments (mass range from 100- 500 Da) in each MS 3 step were measured in the Orbitrap analyser, which was set at a resolution of 60,000. This was performed in cycles of 10 MS 3 events before the Lumos 191 instrument reverted to scanning the m/z ratios of the intact peptide ions and the cycle continued. 7.1.2 Data analysis Proteome DiscovererTM version 2.1 (Thermo Fisher Scientific) and Mascot version 2.6 (Matrix Science) were used to process raw data files. Data were aligned with the UniProt Human reference proteome (Proteome ID: UP000005640) database, in addition to using the common repository of adventitious proteins (cRAP) version 1.0. Database searches were also performed using a sequence scrambled decoy database in order to yield a more robust false discovery rate (FDR). Protein identification allowed an MS tolerance of ± 20 parts per million (ppm) and an MS/MS tolerance of ± 0.6 Da, along with permission of up to two missed tryptic cleavages. Fixed modifications were carbamidomethylation of cysteine and TMT 10-plex at lysine and N-termini; variable modifications were oxidized methionine, deamidated aspartic acid and asparagine, and TMT10-plex at threonine, serine and methionine. Quantification was achieved by calculating the sum of centroided reporter ions within a ± 2 millimass unit (mmu) window around the expected m/z for each of the TMT reporter ions. For quantification, the integration window tolerance was set to 20 ppm. The confidence thresholds were set at a strict FDR of 0.01 and a relaxed FDR of 0.05. Finally, the following filters were applied during the Proteome Discoverer search: peptide score: 20, peptide rank: 1, peptide confidence: high. Labelling efficiency was calculated by setting TMT10-plex as a variable modification and expressing the number of peptides with the modification as a percentage of the total number of peptides. A labelling efficiency of ≥ 95% was considered acceptable. Principal component analysis (PCA) and paired t-test statistical analyses were performed. PCA was used to visualise data relatedness by mathematical transformation of the original (possibly correlated) variables into a set of linearly uncorrelated variables (principal components). The first principal component (PC1) had the largest possible variance (i.e. accounted for as much variability in the data as possible), with each succeeding component (e.g. PC2) having the highest remaining variance possible. T-test was performed using R limma package software, with p values adjusted by the Benjamini-Hochberg FDR correction for multiple comparisons. Using this method, the p-values were ranked from smallest (1) to largest (N). Each p-value was multiplied by N and divided by its assigned rank to give the adjusted p-values. This method provides a good balance between discovery of statistically significant proteins and limitation of false positive occurrences. 192 7.2 Proteins identified by 10-plex TMT-MS Table 7.1: Proteins increased in normoxia Proteins increased in normoxia which were present in all 10 samples with a FDR <0.01 are listed below in order of the magnitude of the fold change (FC). Accession Description log2FC adjusted p value FC Q5TCU8 Tropomyosin beta chain 3.567 0.001 11.855 cRAP106 Thioredoxin 1.967 0.017 3.909 E7EX29 14-3-3 protein zeta/delta 1.943 0.015 3.844 P52566 Rho GDP-dissociation inhibitor 2 1.927 0.004 3.802 P08670 Vimentin 1.913 0.006 3.766 O00299 Chloride intracellular channel protein 1 1.545 0.030 2.918 P11021 78 kDa glucose-regulated protein 1.523 0.009 2.873 E9PK25 Cofilin-1 1.461 0.023 2.753 E7EMB3 Calmodulin 1.315 0.026 2.488 P62993 Growth factor receptor-bound protein 2 1.309 0.004 2.478 P20700 Lamin-B1 1.284 0.004 2.435 Q32MZ4 Leucine-rich repeat flightless- interacting protein 1 1.224 0.004 2.337 P06737 Glycogen phosphorylase, liver form 1.149 0.025 2.217 P32942 Intercellular adhesion molecule 3 1.104 0.015 2.149 Q9Y490 Talin-1 1.081 0.015 2.116 O15144 Actin-related protein 2/3 complex subunit 2 1.061 0.031 2.086 P06702 Protein S100-A9 1.060 0.013 2.084 P60709 Actin, cytoplasmic 1 1.007 0.164 2.010 P52209 6-phosphogluconate dehydrogenase, decarboxylating 0.963 0.013 1.950 P26038 Moesin 0.909 0.019 1.878 P33241 Leukocyte-specific protein 1 0.909 0.013 1.878 P18206 Vinculin 0.861 0.015 1.816 P30740 Leukocyte elastase inhibitor 0.858 0.006 1.813 Q96C19 EF-hand domain-containing protein D2 0.843 0.027 1.794 A6NIZ1 Ras-related protein Rap-1b-like protein 0.807 0.009 1.749 P61160 Actin-related protein 2 0.775 0.178 1.711 F8VPF3 Myosin light polypeptide 6 0.759 0.069 1.693 P31946 14-3-3 protein beta/alpha 0.747 0.184 1.678 P35579 Myosin-9 0.706 0.012 1.631 A0A140T930 HLA class I histocompatibility antigen, Cw-6 alpha chain 0.702 0.039 1.627 S4R359 Heterogeneous nuclear ribonucleoprotein K 0.685 0.142 1.607 P80723 Brain acid soluble protein 1 0.677 0.460 1.598 P61247 40S ribosomal protein S3a 0.612 0.028 1.528 193 Q14847 LIM and SH3 domain protein 1 0.604 0.050 1.520 P07237 Protein disulfide-isomerase 0.597 0.071 1.512 P02042 Hemoglobin subunit delta 0.577 0.049 1.492 cRAP036 Hemoglobin subunit beta 0.575 0.159 1.490 P0C0S5 Histone H2A.Z 0.574 0.059 1.489 E7EQR4 Ezrin 0.558 0.023 1.473 P31949 Protein S100-A11 0.557 0.179 1.471 P08133 Annexin A6 0.544 0.035 1.458 P31146 Coronin-1A 0.537 0.034 1.451 Q15907 Ras-related protein Rab-11B 0.533 0.015 1.447 P46781 40S ribosomal protein S9 0.524 0.044 1.438 P16150 Leukosialin 0.514 0.054 1.428 P68036 Ubiquitin-conjugating enzyme E2 L3 0.506 0.090 1.420 J3QRS3 Myosin regulatory light chain 12A 0.501 0.141 1.415 P46940 Ras GTPase-activating-like protein IQGAP1 0.496 0.049 1.411 cRAP035 Hemoglobin subunit alpha 0.487 0.050 1.401 P12814 Alpha-actinin-1 0.481 0.099 1.395 P46976 Glycogenin-1 0.469 0.105 1.384 P09382 Galectin-1 0.467 0.108 1.382 P19878 Neutrophil cytosol factor 2 0.464 0.204 1.379 P50552 Vasodilator-stimulated phosphoprotein 0.461 0.062 1.377 P11413 Glucose-6-phosphate 1- dehydrogenase 0.460 0.120 1.376 A0A075B738 Platelet endothelial cell adhesion molecule 0.449 0.163 1.365 P39687 Acidic leucine-rich nuclear phosphoprotein 32 family member A 0.442 0.043 1.358 P07900 Heat shock protein HSP 90-alpha 0.412 0.089 1.330 P14618 Pyruvate kinase PKM 0.409 0.249 1.328 P28065 Proteasome subunit beta type-9 0.408 0.052 1.327 Q9H4G4 Golgi-associated plant pathogenesis- related protein 1 0.398 0.578 1.317 P16070 CD44 antigen 0.364 0.173 1.287 P49721 Proteasome subunit beta type-2 0.351 0.052 1.275 P25788 Proteasome subunit alpha type-3 0.334 0.050 1.260 Q9ULZ3 Apoptosis-associated speck-like protein containing a CARD 0.324 0.051 1.252 P25786 Proteasome subunit alpha type-1 0.319 0.298 1.247 P37837 Transaldolase 0.313 0.147 1.242 A0A0G2JIW1 Heat shock 70 kDa protein 1B 0.302 0.137 1.233 G3V5Z7 Proteasome subunit alpha type 0.282 0.119 1.216 cRAP019 Cathepsin G 0.271 0.646 1.207 P04083 Annexin A1 0.267 0.106 1.204 P28070 Proteasome subunit beta type-4 0.255 0.160 1.193 O14818 Proteasome subunit alpha type-7 0.254 0.113 1.192 O43707 Alpha-actinin-4 0.242 0.446 1.183 Q9UJ70 N-acetyl-D-glucosamine kinase 0.212 0.202 1.158 194 P21333 Filamin-A 0.208 0.266 1.155 P13796 Plastin-2 0.206 0.504 1.154 P78417 Glutathione S-transferase omega-1 0.196 0.470 1.146 P06703 Protein S100-A6 0.169 0.406 1.125 Q9H299 SH3 domain-binding glutamic acid- rich-like protein 3 0.169 0.702 1.124 P29401 Transketolase 0.166 0.278 1.122 P26447 Protein S100-A4 0.160 0.485 1.118 J3QQX2 Rho GDP-dissociation inhibitor 1 0.155 0.575 1.114 P06733 Alpha-enolase 0.152 0.455 1.111 P60174 Triosephosphate isomerase 0.152 0.575 1.111 A6NEL0 Non-histone chromosomal protein HMG-14 0.145 0.379 1.106 P51858 Hepatoma-derived growth factor 0.145 0.365 1.106 P20618 Proteasome subunit beta type-1 0.141 0.356 1.103 P11142 Heat shock cognate 71 kDa protein 0.140 0.562 1.102 O75368 SH3 domain-binding glutamic acid- rich-like protein 0.132 0.702 1.096 P40121 Macrophage-capping protein 0.131 0.491 1.095 P62979 Ubiquitin-40S ribosomal protein S27a 0.117 0.495 1.085 P04080 Cystatin-B 0.115 0.715 1.083 P15531 Nucleoside diphosphate kinase A 0.111 0.687 1.080 P06396 Gelsolin 0.093 0.522 1.067 Q01518 Adenylyl cyclase-associated protein 1 0.084 0.687 1.060 P05089 Arginase-1 0.066 0.767 1.047 J3KPS3 Fructose-bisphosphate aldolase 0.049 0.840 1.035 Q6IBS0 Twinfilin-2 0.028 0.933 1.020 P04040 Catalase 0.027 0.860 1.019 Q14019 Coactosin-like protein 0.025 0.933 1.017 P07195 L-lactate dehydrogenase B chain 0.017 0.930 1.012 195 Table 7.2: Proteins increased in hypoxia Proteins increased in hypoxia which were present in all 10 samples with a FDR <0.01 are listed below in order of the magnitude of the fold change (FC). Accession Description log2FC adjusted p value FC P62805 Histone H4 1.843 0.004 3.587 P05164 Myeloperoxidase 1.635 0.006 3.107 O75594 Peptidoglycan recognition protein 1 1.409 0.185 2.655 A6NC48 ADP-ribosyl cyclase/cyclic ADP-ribose hydrolase 2 1.303 0.025 2.468 O75083 WD repeat-containing protein 1 1.252 0.029 2.382 Q92820 Gamma-glutamyl hydrolase 1.100 0.020 2.143 cRAP109 Lactotransferrin 1.095 0.033 2.136 A5A3E0 POTE ankyrin domain family member F 1.052 0.015 2.074 Q0VD83 Apolipoprotein B receptor 0.995 0.025 1.993 P19652 Alpha-1-acid glycoprotein 2 0.937 0.233 1.915 P10124 Serglycin 0.929 0.012 1.904 Q9HD89 Resistin 0.925 0.004 1.898 P07737 Profilin-1 0.918 0.015 1.890 A0A087WXL1 Folate receptor gamma 0.908 0.029 1.877 P11215 Integrin alpha-M 0.904 0.004 1.872 P16035 Metalloproteinase inhibitor 2 0.890 0.034 1.853 X6R8F3 Neutrophil gelatinase-associated lipocalin 0.882 0.004 1.842 V9GYM3 Apolipoprotein A-II 0.823 0.049 1.769 G3V3D1 Epididymal secretory protein E1 0.820 0.007 1.765 P10153 Non-secretory ribonuclease 0.722 0.029 1.650 G5EA09 Syndecan binding protein 0.702 0.055 1.627 Q05315 Galectin-10 0.691 0.184 1.614 P08246 Neutrophil elastase 0.676 0.250 1.598 P04217 Alpha-1B-glycoprotein 0.673 0.015 1.594 P05107 Integrin beta-2 0.660 0.018 1.580 P28799 Granulins 0.644 0.083 1.563 P01024 Complement C3 0.638 0.032 1.556 A0A0A0MS08 Ig gamma-1 chain C region 0.611 0.050 1.528 A0A0G2JMY9 Leukocyte immunoglobulin-like receptor subfamily A member 3 0.598 0.178 1.514 J3KPA1 Cysteine-rich secretory protein 3 0.594 0.160 1.509 O00391 Sulfhydryl oxidase 1 0.589 0.050 1.504 P20061 Transcobalamin-1 0.578 0.042 1.492 cRAP002 Serum albumin 0.552 0.051 1.466 P78324 Tyrosine-protein phosphatase non- receptor type substrate 1 0.535 0.018 1.449 J3KNB4 Cathelicidin antimicrobial peptide 0.532 0.039 1.446 A0A087WW89 NA 0.525 0.050 1.439 196 P62937 Peptidyl-prolyl cis-trans isomerase A/Cyclophilin A 0.524 0.035 1.438 P14625 Endoplasmin 0.522 0.050 1.435 Q9UHB9 Signal recognition particle subunit SRP68 0.499 0.216 1.413 P30086 Phosphatidylethanolamine-binding protein 1 0.491 0.049 1.406 P02647 Apolipoprotein A-I 0.483 0.460 1.398 P15144 Aminopeptidase N 0.477 0.083 1.392 P20160 Azurocidin 0.475 0.081 1.390 P26022 Pentraxin-related protein PTX3 0.445 0.108 1.362 A0A075B6H6 Ig kappa chain C region 0.441 0.030 1.358 P14780 Matrix metalloproteinase-9 0.438 0.081 1.354 P61626 Lysozyme C 0.436 0.396 1.353 P09960 Leukotriene A-4 hydrolase 0.422 0.184 1.340 A0A0A0MTS2 Glucose-6-phosphate isomerase 0.419 0.353 1.337 P02763 Alpha-1-acid glycoprotein 1 0.415 0.460 1.333 P36222 Chitinase-3-like protein 1 0.389 0.233 1.309 P12429 Annexin A3 0.388 0.189 1.309 A0A075B6K9 Ig lambda-2 chain C regions 0.384 0.049 1.305 P23284 Peptidyl-prolyl cis-trans isomerase B 0.379 0.061 1.300 P22894 Neutrophil collagenase 0.378 0.081 1.299 O43493 Trans-Golgi network integral membrane protein 2 0.350 0.081 1.275 P41218 Myeloid cell nuclear differentiation antigen 0.348 0.387 1.273 P12724 Eosinophil cationic protein 0.338 0.549 1.264 P00558 Phosphoglycerate kinase 1 0.332 0.160 1.259 P40925 Malate dehydrogenase, cytoplasmic 0.327 0.081 1.255 A0A087WTS4 Peptidyl-prolyl cis-trans isomerase FKBP1A 0.327 0.178 1.254 U3KQP0 Brain acid soluble protein 1 0.324 0.561 1.252 A0A024R6I7 Alpha-1-antitrypsin 0.295 0.397 1.227 P02787 Serotransferrin 0.283 0.361 1.216 O15389 Sialic acid-binding Ig-like lectin 5 0.282 0.248 1.216 P14151 L-selectin 0.270 0.117 1.206 P00738 Haptoglobin 0.265 0.309 1.202 O00602 Ficolin-1 0.263 0.185 1.200 P02750 Leucine-rich alpha-2-glycoprotein 0.259 0.626 1.196 Q6UX06 Olfactomedin-4 0.256 0.581 1.194 Q13231 Chitotriosidase-1 0.246 0.565 1.186 P05204 Non-histone chromosomal protein HMG-17 0.244 0.160 1.184 P62942 Peptidyl-prolyl cis-trans isomerase FKBP1A 0.223 0.445 1.167 P27695 DNA-(apurinic or apyrimidinic site) lyase 0.213 0.396 1.159 cRAP034 Glutathione S-transferase P 0.205 0.565 1.153 P27797 Calreticulin 0.169 0.314 1.124 197 Q6P4A8 Phospholipase B-like 1 0.160 0.565 1.118 P04264 Keratin, type II cytoskeletal 1 0.131 0.737 1.095 Q96G03 Phosphoglucomutase-2 0.120 0.565 1.087 P31150 Rab GDP dissociation inhibitor alpha 0.103 0.695 1.074 Q01469 Fatty acid-binding protein, epidermal 0.102 0.695 1.073 P00338 L-lactate dehydrogenase A chain 0.101 0.745 1.073 Q31612 HLA class I histocompatibility antigen, B-73 alpha chain 0.097 0.615 1.069 P50395 Rab GDP dissociation inhibitor beta 0.093 0.606 1.067 P16083 Ribosyldihydronicotinamide dehydrogenase [quinone] 0.093 0.568 1.066 H7C2N1 Prothymosin alpha 0.086 0.581 1.061 Q13442 28 kDa heat- and acid-stable phosphoprotein 0.082 0.615 1.059 cRAP112 Trypsin 0.078 0.767 1.055 O95336 6-phosphogluconolactonase 0.073 0.738 1.052 Q9BRF8 Serine/threonine-protein phosphatase CPPED1 0.054 0.702 1.038 P00915 Carbonic anhydrase 1 0.034 0.898 1.024 P14174 Macrophage migration inhibitory factor 0.031 0.898 1.021 P00491 Purine nucleoside phosphorylase 0.030 0.891 1.021 A0A024RA52 Proteasome subunit alpha type 0.024 0.885 1.016 P00441 Superoxide dismutase [Cu-Zn] 0.014 0.935 1.010 P07910 Heterogeneous nuclear ribonucleoproteins C1/C2 0.011 0.965 1.008 P17213 Bactericidal permeability-increasing protein 0.001 0.997 1.001 198 Table 7.3: Differentially regulated proteins identified by 10-plex TMT-MS compared with published secretomes and sepsis patient plasma Differentially regulated proteins (as indicated), identified by my 10-plex TMT-MS, are listed below if also present in published (normoxic) neutrophil secretomes (A-F) or septic plasma (G-H), indicated by “Y” if present. Neutrophils in these published secretomes were stimulated with A: ionomycin (309); B: S. pyogenes supernatants (305); C: S. aureus leukotoxins (LukE/D) (308); D: chromofungin or catestatin (307); E: TNFα/cytochalasin B/heat-killed S pyogenes and fMLP (306). F-H: Blood was taken from three patients with sepsis secondary to S. pyogenes (erysipelas), E. coli (urosepsis), and N. meningitidis (meningococcal sepsis) (306). F: The secretome of ex vivo pooled isolated neutrophils stimulated with cytochalasin B and fMLP was examined. G-H: Measurement of the same proteins in pooled plasma showed a subset to be up-regulated (G) or down-regulated (H) when compared with healthy control plasma. Numbers in square brackets indicate the number of proteins identified in each study. Protein identification (10-plex TMT-MS) A [19] B [46] C [60] D [82] E [82] F [49] G [8] H [12] Normoxia significantly increased Protein S100-A9 Y Y Y Y Y Y Thioredoxin Y Y Cofilin-1 Y Y Rho GDP-dissociation inhibitor 2 Y Y Leukocyte-specific protein 1 Y Y 14-3-3 protein zeta/delta Y Vimentin Y Chloride intracellular channel protein 1 Y 78 kDa glucose-regulated protein Y Glycogen phosphorylase, liver form Y 6-phosphogluconate dehydrogenase, decarboxylating Y Leukocyte elastase inhibitor Y Myosin-9 Y Annexin A6 Y Hypoxia significantly increased Neutrophil gelatinase-associated lipocalin Y Y Y Y Y Y Y Myeloperoxidase Y Y Y Y Y Y Lactotransferrin Y Y Y Y Y Y Cathelicidin antimicrobial peptide Y Y Y Y Y Profilin-1 Y Y Y Y Non-secretory ribonuclease Y Y Y Resistin Y Y Y Peptidyl-prolyl cis-trans isomerase A/ Cyclophilin A Y Y Transcobalamin-1 Y Y WD repeat-containing protein 1 Y Integrin alpha-M Y 199 7.3 Publications arising from this thesis 7.3.1 Papers Lodge KM, Vassallo AM, Newby P, Suire S, Stark A-K, Long M, Stephens LR, Okkenhaug K, Ridger V, Stockley RA, Sapey E, Chilvers ER, Li W, Condliffe AM. Hypoxia differentially regulates histotoxic neutrophil protein release in COPD by a novel secretion mechanism. Manuscript in preparation Guo J*, Lodge K*, Newnham M*, Bunclark K, Toshner M, Morrell NW, Li W. Increased anti- elastase activity in idiopathic pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension. American Journal of Respiratory Cell and Molecular Biology 2018;59:652-655 (* denotes joint first author) Lodge KM, Thompson AAR, Chilvers ER, Condliffe AM. Hypoxic regulation of neutrophil function and consequences for Staphylococcus aureus infection. Microbes and Infection. 2017;19:166-176 Hoenderdos K, Lodge KM, Hirst RA, Chen C, Palazzo SG, Emerenciana A, Summers C, Angyal A, Porter L, Juss JK, O'Callaghan C, Chilvers ER, Condliffe AM. Hypoxia upregulates neutrophil degranulation and potential for tissue injury. Thorax. 2016;71:1030-1038 Appleby SL, Mitrofan CG, Crosby A, Hoenderdos K, Lodge K, Upton PD, Yates CM, Nash GB, Chilvers ER, Morrell NW. Bone Morphogenetic Protein 9 enhances lipopolysaccharide- induced leukocyte recruitment to the vascular endothelium. Journal of Immunology. 2016;197:3302-3314 7.3.2 Abstracts Hypoxia induces differential protein secretion from neutrophils to drive endothelial damage in COPD. November 2018. Clinical Academics in Training Annual Conference, Academy of Medical Sciences. Edinburgh, UK (poster discussion) Hypoxia drives a novel, destructive neutrophil phenotype. April 2018. Therapeutic Targeting of Hypoxia-Sensitive Pathways Keystone Symposium. Oxford, UK (poster) The hypoxic neutrophil secretome. November 2017. NECS (Newcastle, Edinburgh, Cambridge, Sheffield investigators) meeting. Sheffield, UK (oral presentation) Hypoxia drives neutrophil-mediated endothelial damage in COPD. December 2017. British Thoracic Society Winter Meeting. London, UK (oral presentation) 200 Hypoxia upregulates PI3Kinase-dependent neutrophil degranulation and neutrophil-mediated tissue injury. May 2017. American Thoracic Society. Washington D.C., USA (poster discussion) The impact of hypoxia on the neutrophil secretome. March 2017. Research in Progress Meeting, University of Cambridge Department of Medicine and East Anglian Thoracic Society Biannual Meeting. Cambridge, UK (oral presentation) Hypoxia upregulates PI3Kinase-dependent neutrophil degranulation and neutrophil-mediated tissue injury. December 2016. British Thoracic Society Winter Meeting. London, UK (oral presentation) Hypoxia upregulates PI3Kinase-dependent neutrophil degranulation and neutrophil-mediated tissue injury. September 2016. Society of Leukocyte Biology Annual Meeting and Neutrophil 2016. Verona, Italy (oral presentation and poster) The hypoxic neutrophil secretome and its effect on vascular endothelium. September 2015. Research in Progress Meeting, University of Cambridge Department of Medicine and East Anglian Thoracic Society Biannual Meeting. Cambridge, UK (oral presentation) 201 8 Bibliography 1. Dancey J, Deubelbeiss K, Harker L, Finch C. Neutrophil kinetics in man. J Clin Invest. 1976;58(3):705–15. 2. Summers C, Rankin SM, Condliffe AM, Singh N, Peters AM, Chilvers ER. Neutrophil kinetics in health and disease. Trends Immunol. 2010;31:318–24. 3. Mayadas TN, Cullere X, Lowell CA. The multifaceted functions of neutrophils. Annu Rev Pathol Mech Dis. 2014;9:181–218. 4. Tak T, Tesselaar K, Pillay J, Borghans JAM, Koenderman L. What’s your age again? Determination of human neutrophil half-lives revisited. J Leukoc Biol. 2013;94:595– 601. 5. Guthrie LA, McPhail LC, Henson PM, Johnston RB. Priming of neutrophils for enhanced release of oxygen metabolites by bacterial lipopolysaccharide. Evidence for increased activity of the superoxide-producing enzyme. J Exp Med. 1984;160:1656– 71. 6. Ekpenyong A, Töpfner N, Fiddler C, Herbig M, Li W, Cojoc G, et al. Mechanical deformation induces depriming of neutrophils. Sci Adv. 2017;3:e1602536. 7. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, et al. Pulmonary retention of primed neutrophils: A novel protective host response, which is impaired in the acute respiratory distress syndrome. Thorax. 2014;69:623–9. 8. Condliffe A, Chilvers ER, Haslett C, Dransfield I. Priming differentially regulates neutrophil adhesion molecule expression/function. Immunology. 1996;89:105–11. 9. Miralda I, Uriarte SM, McLeish KR. Multiple phenotypic changes define neutrophil priming. Front Cell Infect Microbiol. 2017;7:217. 10. Murray J, Barbara J, Dunkley S, Lopez A, Van Ostade X, Condliffe A, et al. Regulation of neutrophil apoptosis by tumor necrosis actor-α: Requirement for TNFR55 and TNFR75 for induction of apoptosis in vitro. Blood. 1997;90:2772–83. 11. Kolaczkowska E, Kubes P. Neutrophil recruitment and function in health and inflammation. Nat Rev Immunol. Nature Publishing Group; 2013;13:159–75. 12. Foxman EF, Campbell JJ, Butcher EC. Multistep navigation and the combinatorial control of leukocyte chemotaxis. J Cell Biol. 1997;139:1349–60. 13. Li Z, Jiang H, Xie W, Zhang Z, Smrcka A V, Wu D. Roles of PLC-β2 and -β3 and PI3Kγ in chemoattractant-mediated signal transduction. Science (80- ). 2000;287:1046–9. 14. Heit B, Tavener S, Raharjo E, Kubes P. An intracellular signaling hierarchy determines direction of migration in opposing chemotactic gradients. J Cell Biol. 2002;159(1):91–102. 15. Doerschuk CM. The role of CD18-mediated adhesion in neutrophil sequestration induced by infusion of activated plasma in rabbits. Am J Respir Cell Mol Biol. 1992;7:140–8. 202 16. Kienle K, Lammermann T. Neutrophil swarming: An essential process of the neutrophil tissue response. Immunol Rev. 2016;273:76–93. 17. Liese J, Rooijakkers SHM, van Strijp JAG, Novick RP, Dustin ML. Intravital two- photon microscopy of host-pathogen interactions in a mouse model of Staphylococcus aureus skin abscess formation. Cell Microbiol. 2013;15:891–909. 18. Lee EKS, Gillrie MR, Li L, Arnason JW, Kim JH, Babes L, et al. Leukotriene B4- mediated neutrophil recruitment causes pulmonary capillaritis during lethal fungal sepsis. Cell Host Microbe. 2018;23:121–33. 19. Colom B, Bodkin J V., Beyrau M, Woodfin A, Ody C, Rourke C, et al. Leukotriene B4- nNeutrophil elastase axis drives neutrophil reverse transendothelial cell migration in vivo. Immunity. The Authors; 2015;42:1075–86. 20. Flannagan RS, Jaumouille V, Grinstein S. The cell biology of phagocytosis. Annu Rev Pathol Mech Dis. 2012;7:61–98. 21. Fairn GD, Grinstein S. How nascent phagosomes mature to become phagolysosomes. Trends Immunol. 2012;33(8):397–405. 22. Levine AP, Duchen MR, Villiers S De, Rich PR, Segal AW. Alkalinity of Neutrophil Phagocytic Vacuoles Is Modulated by HVCN1 and Has Consequences for Myeloperoxidase Activity. PLoS One. 2015;10(4):e0125906. 23. Faure MC, Sulpice J, Delattre M, Lavielle M, Prigent M, Cuif M-H, et al. The recruitment of p47 phox and Rac2G12V at the phagosome is transient and phosphatidylserine dependent. Biol Cell. 2013;105:501–18. 24. Li XJ, Tian W, Stull ND, Grinstein S, Atkinson S, Dinauer MC. A fluorescently tagged C-terminal fragment of p47phox detects NADPH oxidase dynamics during phagocytosis. Mol Biol Cell. 2009;20:1520–32. 25. Quinn MT, Gauss KA. Structure and regulation of the neutrophil respiratory burst oxidase : comparison with nonphagocyte oxidases. J Leukoc Biol. 2004;76:760–81. 26. Reeves EP, Lu H, Jacobs HL, Messina CGM, Bolsover S, Gabella G, et al. Killing activity of neutrophils is mediated through activation of proteases by K flux. Nature. 2002;416:291–7. 27. Green JN, Kettle AJ, Winterbourn CC. Protein chlorination in neutrophil phagosomes and correlation with bacterial killing. Free Radic Biol Med. 2014;77:49–56. 28. Kuijpers T, Lutter R. Inflammation and repeated infections in CGD: Two sides of a coin. Cell Mol Life Sci. 2012;69:7–15. 29. Borregaard N, Sehested M, Nielsen BS, Sengeløv H, Kjeldsen L. Biosynthesis of granule proteins in normal human bone marrow cells. Gelatinase is a marker of terminal neutrophil differentiation. Blood. 1995;85:812–7. 30. Le Cabec V, Cowland JB, Calafat J, Borregaard N. Targeting of proteins to granule subsets is determined by timing and not by sorting: The specific granule protein NGAL is localized to azurophil granules when expressed in HL-60 cells. Proc Natl Acad Sci U S A. 1996;93:6454–7. 31. Fontana S, Parolini S, Vermi W, Booth S, Gallo F, Donini M, et al. Innate immunity defects in Hermansky-Pudlak type 2 syndrome. Blood. 2006;107:4857–64. 203 32. Lominadze G, Powell DW, Luerman GC, Link AJ, Ward RA, McLeish KR. Proteomic analysis of human neutrophil granules. Mol Cell proteomics. 2005;4:1503–21. 33. Rørvig S, Østergaard O, Heegaard NHH, Borregaard N. Proteome profiling of human neutrophil granule subsets, secretory vesicles, and cell membrane: correlation with transcriptome profiling of neutrophil precursors. J Leukoc Biol. 2013;94:711–21. 34. Clemmensen SN, Jacobsen LC, Rørvig S, Askaa B, Christenson K, Iversen M, et al. Alpha-1-antitrypsin is produced by human neutrophil granulocytes and their precursors and liberated during granule exocytosis. Eur J Haematol. 2011 Jun;86:517–30. 35. Jacobsen LC, Sorensen OE, Cowland JB, Borregaard N, Theilgaard-Monch K. The secretory leukocyte protease inhibitor (SLPI) and the secondary granule protein lactoferrin are synthesized in myelocytes, colocalize in subcellular fractions of neutrophils, and are coreleased by activated neutrophils. J Leukoc Biol. 2008;83:1155–64. 36. Futosi K, Fodor S, Mócsai A. Neutrophil cell surface receptors and their intracellular signal transduction pathways. Int Immunopharmacol. 2013;17:1185–97. 37. Sengelov H, Follin P, Kjeldsen L, Lollike K, Dahlgren C, Borregaard N. Mobilization of granules and secretory vesicles during in vivo exudation of human neutrophils. J Immunol. 1995;154:4157–65. 38. Huizinga TW, Dolman KM, van der Linden NJ, Kleijer M, Nuijens JH, von dem Borne AE, et al. Phosphatidylinositol-linked FcRIII mediates exocytosis of neutrophil granule proteins, but does not mediate initiation of the respiratory burst. J Immunol. 1990;144:1432–7. 39. Ito N, Yokomizo T, Sasaki T, Kurosu H, Penninger J, Kanaho Y, et al. Requirement of phosphatidylinositol 3-kinase activation and calcium influx for leukotriene B4-induced enzyme release. J Biol Chem. 2002;277:44898–904. 40. Cowland JB, Borregaard N. Granulopoiesis and granules of human neutrophils. Immunol Rev. 2016;273:11–28. 41. Mollinedo F. Human neutrophil granules and exocytosis molecular control. Immunologia. 2003;22:340–58. 42. Feuk-Lagerstedt E, Movitz C, Pellmé S, Dahlgren C, Karlsson A. Lipid raft proteome of the human neutrophil azurophil granule. Proteomics. 2007;7:194–205. 43. Ng-Sikorski J, Andersson R, Patarroyo M, Andersson T. Calcium signaling capacity of the CD11b/CD18 integrin on human neutrophils. Exp Cell Res. 1991;195:504–8. 44. Newbrough SA, Mocsai A, Clemens RA, Wu JN, Silverman MA, Singer AL, et al. SLP- 76 regulates Fcγ receptor and integrin signaling in neutrophils. Immunity. 2003;19:761–9. 45. Jakus Z, Simon E, Frommhold D, Sperandio M, Mócsai A. Critical role of phospholipase Cγ2 in integrin and Fc receptor-mediated neutrophil functions and the effector phase of autoimmune arthritis. J Exp Med. 2009;206:577–93. 46. Mócsai A, Zhou M, Meng F, Tybulewicz VL, Lowell CA. Syk is required for integrin signaling in neutrophils. Immunity. 2002;16:547–58. 204 47. Mócsai A, Ligeti E, Lowell CA, Berton G. Adhesion-dependent degranulation of neutrophils requires the Src family kinases Fgr and Hck. J Immunol. 1999;162:1120– 6. 48. Kamen LA, Schlessinger J, Lowell CA. Pyk2 is required for neutrophil degranulation and host defense responses to bacterial infection. J Immunol. 2011;186:1656–65. 49. Sengeløv H, Kjeldsen L, Borregaard N. Control of exocytosis in early neutrophil activation. J Immunol. 1993;150:1535–43. 50. Chen J, Tang H, Hay N, Xu J, Ye RD. Akt isoforms differentially regulate neutrophil functions. Blood. 2010;115:4237–46. 51. Nanamori M, Chen J, Du X, Ye RD. Regulation of leukocyte degranulation by cGMP- dependent protein kinase and phosphoinositide 3-kinase: Potential roles in phosphorylation of target membrane SNARE complex proteins in rat mast cells. J Immunol. 2007;178:416–27. 52. Mocsai A, Jakus Z, Vantus T, Berton G, Lowell CA, Ligeti E. Kinase pathways in chemoattractant-induced degranulation of neutrophils: The role of p38 Mitogen- Activated Protein Kinase activated by Src family kinases. J Immunol. 2000;164:4321– 31. 53. Barlic J, Andrews JD, Kelvin AA, Bosinger SE, DeVries ME, Xu L, et al. Regulation of tyrosine kinase activation and granule release through beta-arrestin by CXCR1. Nat Immunol. 2000;1:227–33. 54. Welch HCE, Coadwell WJ, Ellson CD, Ferguson GJ, Andrews SR, Erdjument- Bromage H, et al. P-Rex1, a PtdIns(3,4,5)P3- and Gβγ-regulated guanine-nucleotide exchange factor for Rac. Cell. 2002;108:809–21. 55. Clemens RA, Lowell CA. Store-operated calcium signaling in neutrophils. J Leukoc Biol. 2015;98:497–502. 56. Rosales JL, Ernst JD. Calcium-dependent neutrophil secretion: Characterization and regulation by annexins. J Immunol. 1997;159:6195–202. 57. Boswell KL, James DJ, Esquibel JM, Bruinsma S, Shirakawa R, Horiuchi H, et al. Munc 13-4 reconstitutes calcium-dependent SNARE-mediated membrane fusion. J Cell Biol. 2012;197:301–12. 58. Zou W, Meng X, Cai C, Zou M, Tang S, Chu X, et al. Store operated calcium entry (SOCE) plays a role in the polarization of neutrophil-like HL-60 cells by regulating the activation of Akt, Src, and Rho Family GTPases. Cell Physiol Biochem. 2012;30:221– 37. 59. Ramadass M, Catz SD. Molecular mechanisms regulating secretory organelles and endosomes in neutrophils and their implications for inflammation. Immunol Rev. 2016;273:249–65. 60. Chun XS, Magalhães MAO, Glogauer M. Rac1 and Rac2 differentially regulate actin free barbed end formation downstream of the fMLP receptor. J Cell Biol. 2007;179:239–45. 61. Johnson JL, Monfregola J, Napolitano G, Kiosses WB, Catz SD. Vesicular trafficking through cortical actin during exocytosis is regulated by the Rab27a effector JFC1/Slp1 and the RhoA-GTPase-activating protein Gem-interacting protein. Mol Biol Cell. 2012;23:1902–16. 205 62. Abdel-Latif D, Steward M, Macdonald DL, Francis GA, Dinauer MC, Lacy P. Rac2 is critical for neutrophil primary granule exocytosis. Blood. 2004;104:832–9. 63. Cunnick JM, Schmidhuber S, Chen G, Yu M, Yi S-J, Cho YJ, et al. Bcr and Abr cooperate in negatively regulating acute inflammatory responses. Mol Cell Biol. 2009;29:5742–50. 64. Herrero-Turrión MJ, Calafat J, Janssen H, Fukuda M, Mollinedo F. Rab27a regulates exocytosis of tertiary and specific granules in human neutrophils. J Immunol. 2008;181:3793–803. 65. Monfregola J, Johnson JL, Meijler MM, Napolitano G, Catz SD. MUNC13-4 protein regulates the oxidative response and is essential for phagosomal maturation and bacterial killing in neutrophils. J Biol Chem. 2012;287:44603–18. 66. Möhn H, Le Cabec V, Fischer S, Maridonneau-Parini I. The src-family protein-tyrosine kinase p59hck is located on the secretory granules in human neutrophils and translocates towards the phagosome during cell activation. Biochem J. 1995;309:657–65. 67. Gutkind JS, Robbins KC. Translocation of the FGR protein-tyrosine kinase as a consequence of neutrophil activation. Proc Natl Acad Sci. 1989;86:8783–7. 68. Mollinedo F, Calafat J, Janssen H, Martin-Martin B, Canchado J, Nabokina SM, et al. Combinatorial SNARE complexes modulate the secretion of cytoplasmic granules in human neutrophils. J Immunol. 2006;177:2831–41. 69. Hirahashi J, Mekala D, Van Ziffle J, Xiao L, Saffaripour S, Wagner DD, et al. Mac-1 signaling via Src-family and Syk kinases results in elastase-dependent thrombohemorrhagic vasculopathy. Immunity. 2006;25:271–83. 70. Chalmers JD, Moffitt KL, Suarez-Cuartin G, Sibila O, Finch S, Furrie E, et al. Neutrophil elastase activity is associated with exacerbations and lung function decline in bronchiectasis. Am J Respir Crit Care Med. 2017;195:1384–93. 71. William Vandivier R, Fadok VA, Hoffmann PR, Bratton DL, Penvari C, Brown KK, et al. Elastase-mediated phosphatidylserine receptor cleavage impairs apoptotic cell clearance in cystic fibrosis and bronchiectasis. J Clin Invest. 2002;109:661–70. 72. Brinkmann V, Reichard U, Goosmann C, Fauler B, Uhlemann Y, Weiss DS, et al. Neutrophil extracellular traps kill bacteria. Science. 2004;303(5663):1532–5. 73. Fuchs TA, Abed U, Goosmann C, Hurwitz R, Schulze I, Wahn V, et al. Novel cell death program leads to neutrophil extracellular traps. J Cell Biol. 2007;176:231–41. 74. Li P, Li M, Lindberg MR, Kennett MJ, Xiong N, Wang Y. PAD4 is essential for antibacterial innate immunity mediated by neutrophil extracellular traps. J Exp Med. 2010;207:1853–62. 75. Papayannopoulos V, Metzler KD, Hakkim A, Zychlinsky A. Neutrophil elastase and myeloperoxidase regulate the formation of neutrophil extracellular traps. J Cell Biol. 2010;191:677–91. 76. Rochael NC, Guimarães-Costa AB, Nascimento MTC, DeSouza-Vieira TS, Oliveira MP, Garcia e Souza LF, et al. Classical ROS-dependent and early/rapid ROS- independent release of neutrophil extracellular traps triggered by Leishmania parasites. Sci Rep. 2015;5:18302. 206 77. Clark SR, Ma AC, Tavener SA, Mcdonald B, Goodarzi Z, Kelly MM, et al. Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood. Nat Med. 2007;13(4):463–9. 78. Hong W, Juneau RA, Pang B, Swords WE. Neutrophil Extracellular Traps Promotes Nontypeable Haemophilus influenzae Persistence in the Chinchilla Model for Otitis Media. J Innate Immun. 2009;1:215–24. 79. Menegazzi R, Decleva E, Dri P. Killing by neutrophil extracellular traps: fact or folklore? Blood. 2012;119(5):1214–7. 80. Brinkmann V, Zychlinsky A. Beneficial suicide: why neutrophils die to make NETs. Nat Rev Microbiol. 2007;5(8):577–82. 81. Pilsczek FH, Salina D, Poon KKH, Fahey C, Yipp BG, Sibley CD, et al. A novel mechanism of rapid nuclear neutrophil extracellular trap formation in response to Staphylococcus aureus. J Immunol. 2010;185:7413–25. 82. Yipp BG, Petri B, Salina D, Jenne CN, Scott BN V, Zbytnuik LD, et al. Infection- induced NETosis is a dynamic process involving neutrophil multitasking in vivo. Nat Med. 2012;18:1386–93. 83. Porto BN, Stein RT. Neutrophil extracellular traps in pulmonary diseases: Too much of a good thing? Front Immunol. 2016;7:311. 84. Dicker AJ, Crichton ML, Pumphrey EG, Cassidy AJ, Suarez-Cuartin G, Sibila O, et al. Neutrophil extracellular traps are associated with disease severity and microbiota diversity in patients with chronic obstructive pulmonary disease. J Allergy Clin Immunol. 2018;141:117–27. 85. Manzenreiter R, Kienberger F, Marcos V, Schilcher K, Krautgartner WD, Obermayer A, et al. Ultrastructural characterization of cystic fibrosis sputum using atomic force and scanning electron microscopy. J Cyst Fibros. 2012;11:84–92. 86. Marcos V, Zhou-Suckow Z, Önder Yildirim A, Bohla A, Hector A, Vitkov L, et al. Free DNA in cystic fibrosis airway fluids correlates with airflow obstruction. Mediators Inflamm. 2015;2015:408935. 87. Folco EJ, Mawson TL, Vromman A, Bernardes-Souza B, Franck G, Persson O, et al. Neutrophil extracellular rraps induce endothelial cell activation and tissue factor production through interleukin-1α and cathepsin G. Arterioscler Thromb Vasc Biol. 2018;38:1901–12. 88. Kapoor S, Opneja A, Nayak L. The role of neutrophils in thrombosis. Thromb Res. 2018;170:87–96. 89. György B, Szabó TG, Pásztói M, Pál Z, Misják P, Aradi B, et al. Membrane vesicles, current state-of-the-art: emerging role of extracellular vesicles. Cell Mol Life Sci. 2011 Aug;68(16):2667–88. 90. Baj-Krzyworzeka M, Baran J, Szatanek R, Siedlar M. Application of flow cytometry in the studies of microparticles. In: Schmid I, editor. Flow cytometry - recent perspectives. IntechOpen; 2012. p. 203–36. 91. Slater TW, Finkielsztein A, Mascarenhas LA, Mehl LC, Butin-Israeli V, Sumagin R. Neutrophil microparticles deliver active myeloperoxidase to injured mucosa to inhibit epithelial wound healing. J Immunol. 2017;198:2886–97. 207 92. Salanova B, Choi M, Rolle S, Wellner M, Luft FC, Kettritz R. beta2-Integrins and Acquired Glycoprotein IIb/IIIa (GPIIb/IIIa) Receptors Cooperate in NF-KB Activation of Human Neutrophils. J Biol Chem. 2007;282(38):27960–9. 93. Butin-israeli V, Houser MC, Feng M, Thorp EB, Nusrat A, Parkos CA, et al. Deposition of microparticles by neutrophils onto inflamed epithelium: A new mechanism to disrupt epithelial intercellular adhesions and promote transepithelial migration. FASEB J. 2016;30:4007–20. 94. Lorincz AM, Schutte M, Timar CI, Veres DS, Kittel A, McLeish KR, et al. Functionally and morphologically distinct populations of extracellular vesicles produced by human neutrophilic granulocytes. J Leukoc Biol. 2015;98:583–9. 95. Dalli J, Montero-Melendez T, Norling L V, Yin X, Hinds C, Haskard D, et al. Heterogeneity in neutrophil microparticles reveals distinct proteome and functional properties. Mol Cell Proteomics. 2013;12:2205–19. 96. Mesri M, Altieri DC, Alerts E. Endothelial Cell Activation by Leukocyte Microparticles. J Immunol. 1998;161:4382–7. 97. Pitanga TN, de Aragão França L, Rocha VCJ, Meirelles T, Borges VM, Gonçalves MS, et al. Neutrophil-derived microparticles induce myeloperoxidase-mediated damage of vascular endothelial cells. BMC Cell Biol. 2014 Jan;15:21. 98. Dalli J, Norling L V, Renshaw D, Cooper D, Leung K-Y, Perretti M. Annexin 1 mediates the rapid anti-inflammatory effects of neutrophil-derived microparticles. Blood. 2008 Sep 15;112(6):2512–9. 99. Gasser O, Schifferli J a. Activated polymorphonuclear neutrophils disseminate anti- inflammatory microparticles by ectocytosis. Blood. 2004 Oct 15;104(8):2543–8. 100. Colotta F, Re F, Polentarutti N, Sozzani S, Mantovani A. Modulation of granulocyte survival and programmed cell death by cytokines and bacterial products. Blood. 1992;80(8):2012–20. 101. Murphy BM, O’Neill AJ, Adrain C, Watson RWG, Martin SJ. The Apoptosome Pathway to Caspase Activation in Primary Human Neutrophils Exhibits Dramatically Reduced Requirements for Cytochrome c. J Exp Med. 2003;197(5):625–32. 102. Maianski NA, Geissler J, Srinivasula SM, Alnemri ES, Roos D, Kuijpers TW. Functional characterization of mitochondria in neutrophils: a role restricted to apoptosis. Cell Death Differ. 2004;11:143–53. 103. Chacko BK, Kramer PA, Ravi S, Johnson MS, Hardy RW, Ballinger SW, et al. Methods for defining distinct bioenergetic profiles in platelets, lymphocytes, monocytes, and neutrophils, and the oxidative burst from human blood. Lab Invest. Nature Publishing Group; 2013;93:690–700. 104. Thompson AAR, Dickinson RS, Murphy F, Thomson JP, Marriott HM, Tavares A, et al. Hypoxia determines survival outcomes of bacterial infection through HIF-1alpha dependent re-programming of leukocyte metabolism. Sci Immunol. 2017;2:eaal2861. 105. Chertov O, Michiel DF, Xu L, Wang JM, Tani K, Murphy WJ, et al. Identification of defensin-1, defensin-2, and CAP37/azurocidin as T-cell chemoattractant proteins released from Interleukin-8-stimulated neutrophils. J Biol Chem. 1996;271(6):2935– 40. 208 106. Takashima A, Yao Y. Neutrophil plasticity: acquisition of phenotype and functionality of antigen-presenting cell. J Leukoc Biol. 2015;98:489–96. 107. Lakshman R, Finn A. Neutrophil disorders and their management. J Clin Pathol. 2001;54:7–19. 108. Nathan CF. Neutrophil activation on biological surfaces. Massive secretion of hydrogen peroxide in response to products of macrophages and lymphocytes. J Clin Invest. 1987;80:1550–60. 109. Chollet-Martin S, Montravers P, Gibert C, Elbim C, Desmonts JM, Fagon JY, et al. Subpopulation of hyperresponsive polymorphonuclear neutrophils in patients with adult respiratory distress syndrome. Role of cytokine production. Am Rev Respir Dis. 1992;146(4):990–6. 110. Bass DA, Olbrantz P, Szejda P, Seeds MC, McCall CE. Subpopulations of neutrophils with increased oxidative product formation in blood of patients with infection. J Immunol. 1986;136:860–6. 111. Macdonald SPJ, Stone SF, Neil CL, Van Eeden PE, Fatovich DM, Arendts G, et al. Sustained elevation of resistin, NGAL and IL-8 are associated with severe sepsis/septic shock in the emergency department. PLoS One. 2014;9:e110678. 112. Eagan TM, Damås JK, Ueland T, Voll-Aanerud M, Mollnes TE, Hardie JA, et al. Neutrophil gelatinase-associated lipocalin: A biomarker in COPD. Chest. 2010;138:888–95. 113. Al Mutairi SS, Mojiminiyi OA, Shihab-Eldeen A, Al Rammah T, Abdella N. Putative roles of circulating resistin in patients with asthma, COPD and cigarette smokers. Dis Markers. 2011;31:1–7. 114. Zhang M, Tang J, Yin J, Wang X, Feng X, Yang X, et al. The clinical implication of serum cyclophilin A in patients with chronic obstructive pulmonary disease. Int J COPD. 2018;13:357–63. 115. Saffarzadeh M, Juenemann C, Queisser MA, Lochnit G, Barreto G, Galuska SP, et al. Neutrophil extracellular traps directly induce epithelial and endothelial cell death: A predominant role of histones. PLoS One. 2012;7:e32366. 116. Nakazawa D, Kumar S V., Marschner J, Desai J, Holderied A, Rath L, et al. Histones and neutrophil extracellular traps enhance tubular necrosis and remote organ injury in ischemic AKI. J Am Soc Nephrol. 2017;28:1753–68. 117. Tohme S, Yazdani HO, Al-khafaji AB, Chidi AP, Loughran P, Mowen K, et al. Neutrophil extracellular traps promote the development and progression of liver metastases after surgical stress. Cancer Res. 2016;76:1367–80. 118. Fridlender ZG, Sun J, Kim S, Kapoor V, Cheng G, Worthen GS, et al. Polarization of tumor-associated neutrophil (TAN) phenotype by TGF-β: “N1” versus “N2” TAN. Cancer Cell. 2009;16:183–94. 119. Shamamian P, Schwartz JD, Pocock BJZ, Monea S, Whiting D, Marcus SG, et al. Activation of progelatinase A (MMP-2) by neutrophil elastase, cathepsin G, and proteinase-3: A role for inflammatory cells in tumor invasion and angiogenesis. J Cell Physiol. 2001;189:197–206. 209 120. Spencer J, Ferraro F, Roussakis E, Klein A, Wu J, Runnels J, et al. Direct measurement of local oxygen concentration in the bone marrow of live animals. Nature. 2014;508:269–73. 121. Parmar K, Mauch P, Vergilio J, Sackstein R, Down JD. Distribution of hematopoietic stem cells in the bone marrow according to regional hypoxia. PNAS. 2007;104(13):5431–6. 122. Jing D, Wobus M, Poitz DM, Bornhäuser M, Ehninger G, Ordemann R. Oxygen tension plays a critical role in the hematopoietic microenvironment in vitro. Haematologica. 2012;97(3):331–9. 123. Ward JPT. Oxygen sensors in context. Biochim Biophys Acta. 2008;1777:1–14. 124. Nolte D, Steinhauser P, Pickelmann S, Berger S, Roger H, Messmer K. Effects of diaspirin-cross-linked hemoglobin (DCLHb) on local tissue oxygen tension in striated skin muscle: An efficacy study in the hamster. J Lab Clin Med. 1997;130(3):328–38. 125. Zheng L, Kelly CJ, Colgan SP. Physiologic hypoxia and oxygen homeostasis in the healthy intestine. A Review in the Theme: Cellular responses to hypoxia. Am J Physiol Cell Physiol. 2015;309:C350-360. 126. Stewart FA, Denekamp J, Randhawa VS. Skin sensitization by misonidazole: a demonstration of uniform mild hypoxia. Br J Cancer. 1982;45:869–77. 127. Campbell EL, Bruyninckx WJ, Kelly CJ, GLover LE, Mcnamee EN, Bowers BE, et al. Transmigrating neutrophils shape the mucosal microenvironment through localized oxygen depletion to influence resolution of inflammation. Immunity. 2014;40:66–77. 128. Mani R, White JE, Barrett DF, Weaver PW. Tissue oxygenation, venous ulcers and fibrin cuffs. J R Soc Med. 1989;82:345–6. 129. Simmen H, Blaser J. Analysis of pH and p02 in abscesses, peritoneal fluid, and drainage fluid in the presence or absence of bacterial infection during and after abdominal surgery. Am J Surg. 1993;165:24–7. 130. Jiang BH, Semenza GL, Bauer C, Marti HH. Hypoxia-inducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension. Am J Physiol. 1996;271:C1172–80. 131. Polosukhin V V., Lawson WE, Milstone AP, Egunova SM, Kulipanov AG, Tchuvakin SG, et al. Association of progressive structural changes in the bronchial epithelium with subepithelial fibrous remodeling: A potential role for hypoxia. Virchows Arch. 2007;451(4):793–803. 132. Lee SH, Lee SH, Kim CH, Yang KS, Lee EJ, Min KH, et al. Increased expression of vascular endothelial growth factor and hypoxia inducible factor-1α in lung tissue of patients with chronic bronchitis. Clin Biochem. 2014;47:552–9. 133. Giatromanolaki A, Sivridis E, Maltezos E, Papazoglou D, Simopoulos C, Gatter KC, et al. Hypoxia inducible factor 1 alpha and 2 alpha overexpression in inflammatory bowel disease. J Clin Pathol. 2003;56:209–13. 134. Grahl N, Puttikamonkul S, Macdonald JM, Gamcsik MP, Ngo LY, Hohl TM, et al. In vivo hypoxia and a fungal alcohol dehydrogenase influence the pathogenesis of invasive pulmonary aspergillosis. PLoS Pathog. 2011 Jul;7:e1002145. 210 135. Mahnke A, Meier RJ, Schatz V, Hofmann J, Castiglione K, Schleicher U, et al. Hypoxia in Leishmania major skin lesions impairs the NO-dependent leishmanicidal activity of macrophages. J Invest Dermatol. 2014;134:2339–46. 136. Belton M, Brilha S, Manavaki R, Mauri F, Nijran K, Hong YT, et al. Hypoxia and tissue destruction in pulmonary TB. Thorax. 2016;71:1145–53. 137. Eckle T, Faigle M, Grenz A, Laucher S, Thompson LF, Eltzschig HK. A2B adenosine receptor dampens hypoxia-induced vascular leak. Blood. 2008;111:2024–35. 138. Hartmann G, Tscho M, Fischer R, Bidlingmaier C, Riepl R, Tscho K. High altitude increases circulating interleukin-6, interleukin-1 receptor antagonist and C-reactive protein. Cytokine. 2000;12(3):246–52. 139. Green GM, Kass EH. The influence of bacterial species on pulmonary resistance to infection in mice subjected to hypoxia, cold stress and ethanolic intoxication. Br J Exp Pathol. 1965;46(3):360–6. 140. Knighton RD, Halliday B, Hunt TK. Oxygen as an antibiotic. The effect of inspired oxygen on infection. Arch Surg. 1984;119:199–204. 141. Shaw JJ, Psoinos C, Emhoff TA, Shah SA, Santry HP. Not Just Full of Hot Air : Hyperbaric Oxygen Therapy Increases Survival in Cases of Necrotizing Soft Tissue Infections. Surg Infect (Larchmt). 2014;15(3):328–35. 142. Walmsley SR, McGovern NN, Whyte MKB, Chilvers ER. The HIF/VHL pathway: from oxygen sensing to innate immunity. Am J Respir Cell Mol Biol. 2008 Mar;38(3):251–5. 143. Cramer T, Yamanishi Y, Clausen BE, Förster I, Pawlinski R, Mackman N, et al. HIF- 1α is essential for myeloid cell-mediated inflammation. Cell. 2003;112(5):645–57. 144. Peyssonnaux C, Cejudo-Martin P, Doedens A, Zinkernagel a. S, Johnson RS, Nizet V. Cutting Edge: Essential Role of Hypoxia Inducible Factor-1 in Development of Lipopolysaccharide-Induced Sepsis. J Immunol. 2007;178(12):7516–9. 145. Walmsley SR, Print C, Farahi N, Peyssonnaux C, Johnson RS, Cramer T, et al. Hypoxia-induced neutrophil survival is mediated by HIF-1α -dependent NF-kB activity. J Exp Med. 2005 Jan 3;201(1):105–15. 146. Arsham AM, Howell JJ, Simon MC. A novel hypoxia-inducible factor-independent hypoxic response regulating mammalian target of rapamycin and its targets. J Biol Chem. 2003;278:29655–60. 147. Wottawa M, Naas S, Böttger J, van Belle GJ, Möbius W, Revelo NH, et al. Hypoxia- stimulated membrane trafficking requires T-plastin. Acta Physiol. 2017;221:59–73. 148. McGovern NN, Cowburn AS, Porter L, Walmsley SR, Summers C, Thompson AAR, et al. Hypoxia selectively inhibits respiratory burst activity and killing of Staphylococcus aureus in human neutrophils. J Immunol. 2011;186:453–63. 149. Hoenderdos K, Lodge K, Hirst R, Chen C, Palazzo S, Emerenciana A, et al. Hypoxia upregulates neutrophil degranulation and potential for tissue injury. Thorax. 2016;71:1030–8. 150. Scannell G, Waxman K, Vaziri N, Zhang J, Kaupke C, Jalali M, et al. Hypoxia-induced alterations of neutrophil membrane receptors. J Surg Res. 1995;59:141–5. 211 151. Kong T, Eltzschig HK, Karhausen J, Colgan SP, Shelley CS. Leukocyte adhesion during hypoxia is mediated by HIF-1-dependent induction of β2 integrin gene expression. PNAS. 2004;101:10440–5. 152. Baudry N, Danialou G, Boczkowski J, Vicaut E. In vivo study of the effect of systemic hypoxia on leukocyte–endothelium interactions. Am J Respir Crit Care Med. 1998;158:477–83. 153. Wang J, Liu H. Systemic hypoxia enhances bactericidal activities of human polymorphonuclear leucocytes. Clin Sci. 2009;116:805–17. 154. Tamura DY, Moore EE, Partrick DA, Johnson JL, Offner PJ, Silliman CC. Acute hypoxemia in humans enhances the neutrophil inflammatory response. Shock. 2002;17(4):269–73. 155. Antonova O, Loktionova S, Golubeva N, Romanov YA, Mazurov A V. Damage and activation of endothelial cells during in vitro hypoxia. Bull Exp Biol Med. 2007;144(10):384–6. 156. Schmitz K, Jennewein M, Pohlemann T, Seekamp A, Oberringer M. Reoxygenation attenuates the adhesion of neutrophils to microvascular endothelial cells. Angiology. 2011;62:155–62. 157. Wood JG, Mattioli LF, Gonzalez NC. Hypoxia causes leukocyte adherence to mesenteric venules in nonacclimatized, but not in acclimatized, rats. J Appl Physiol. 1999;87(3):873–81. 158. Colgan SP, Dzus AL, Parkos CA. Epithelial exposure to hypoxia modulates neutrophil transepithelial migration. J Exp Med. 1996;184:1003–15. 159. Wood JG, Johnson JS, Mattioli LF, Gonzalez NC. Systemic hypoxia increases leukocyte emigration and vascular permeability in conscious rats. J Appl Physiol. 2000;89:1561–8. 160. Eltzschig HK, Thompson LF, Karhausen J, Cotta RJ, Ibla JC, Robson SC, et al. Endogenous adenosine produced during hypoxia attenuates neutrophil accumulation: coordination by extracellular nucleotide metabolism. Blood. 2004;104:3986–92. 161. Peyssonnaux C, Datta V, Cramer T, Doedens A, Theodorakis EA, Gallo RL, et al. HIF-1α expression regulates the bactericidal capacity of phagocytes. J Clin Invest. 2005;115:1806–15. 162. Rotstein O, Fiegel V, Simmons R, Knighton D. The deleterious effect of reduced pH and hypoxia on neutrophil migration in vitro. J Surg Res. 1988;45:298–303. 163. Ma S, Cai C, Ma Y, Bai Z, Meng X, Yang X, et al. Store-operated Ca 2+ entry mediated regulation of polarization in differentiated human neutrophil-like HL-60 cells under hypoxia. Mol Med Rep. 2014;9:819–24. 164. Lin AE, Beasley FC, Olson J, Keller N, Shalwitz RA, Hannan TJ, et al. Role of hypoxia inducible factor-1α (HIF-1α) in innate defense against uropathogenic Escherichia coli infection. PLoS Pathog. 2015;11(4):e1004818. 165. Zinkernagel AS, Peyssonnaux C, Johnson RS, Nizet V. Pharmacologic augmentation of hypoxia-inducible factor-1α with mimosine boosts the bactericidal capacity of phagocytes. J Infect Dis. 2008;197:214–7. 212 166. Berger EA, McClellan SA, Vistisen KS, Hazlett LD. HIF-1α is essential for effective PMN bacterial killing, antimicrobial peptide production and apoptosis in Pseudomonas aeruginosa keratitis. PLoS Pathog. 2013;9(7):e1003457. 167. Fritzenwanger M, Jung C, Goebel B, Lauten A, Figulla HR. Impact of short-term systemic hypoxia on phagocytosis, cytokine production, and transcription factor activation in peripheral blood cells. Mediators Inflamm. 2011;2011:429501. 168. Freischlag JA, Hanna D. Neutrophil (PMN) phagocytosis and chemotaxis after 2 hr of ischemia. J Surg Res. 1991;50:648–52. 169. Lu Y, Wu C, Huang Y, Huang C, Yang C, Lee T, et al. Neutrophil priming by hypoxic preconditioning protects against epithelial barrier damage and enteric bacterial translocation in intestinal ischemia/reperfusion. Lab Investig. 2012;92:783–96. 170. Knowles R, Keeping H, Graeber T, Nguyen K, Garner C, D’Amico R, et al. Cytokine control of PMN phagocytosis: regulatory effects of hypoxemia and hypoxemia- reoxygenation. Am J Physiol Cell Physiol. 1997;272(41):C1352–64. 171. Knowles R, Keeping H, Nguyen K, Garner C, D’Amico R, Simms HH. Hypoxemia up- regulates inerleukin-8 stimulated phagocytosis of polymorphonuclear leukocytes by differential regulation of CD32w and CD35 messenger RNA expression. Surgery. 1995;118:177–84. 172. Simms HH, D’Amico RD. Hypoxia induces an opsonic mismatch on the polymorphonuclear leukocyte surface-reversal via arg-gly-asp-ser-mediated adhesion. J Surg Res. 1993;54:299–304. 173. Simms H, D’Amico RD. Regulation of polymorphonuclear neutrophil CD16 and CD11b/CD18 expression by matrix proteins during hypoxia is VLA-5, VLA-6 dependent. J Immunol. 1995;155:4979–90. 174. Walmsley SR, Cowburn AS, Clatworthy MR, Morrell NW, Roper EC, Singleton V, et al. Neutrophils from patients with heterozygous germline mutations in the von Hippel Lindau protein (pVHL) display delayed apoptosis and enhanced bacterial phagocytosis. Blood. 2006;108(9):3176–8. 175. Almzaiel AJ, Billington R, Smerdon G, Moody AJ. Effects of hyperbaric oxygen treatment on antimicrobial function and apoptosis of differentiated HL-60 (neutrophil- like) cells. Life Sci. 2013;93:125–31. 176. Schulz R, Mahmoudi S, Hattar K, Sibelius ULF, Olschewski H, Mayer K, et al. Enhanced release of superoxide from polymorphonuclear neutrophils in obstructive sleep apnea. Am J Resp Crit Care Med. 2000;162:566–70. 177. Simpson R, Alon R, Kobzik L, Valeri CR, Shepro D, Hechtman HB. Neutrophil and nonneutrophil-mediated injury in intestinal ischemia-reperfusion. Ann Surg. 1993;218:444–54. 178. Ellson CD, Davidson K, Ferguson GJ, O’Connor R, Stephens LR, Hawkins PT. Neutrophils from p40 phox −/− mice exhibit severe defects in NADPH oxidase regulation and oxidant-dependent bacterial killing. J Exp Med. 2006;203(8):1927–37. 179. Mandell GL. Bactericidal activity of aerobic and anaerobic polymorphonuclear neutrophils. Infect Immun. 1974;9(2):337–41. 180. Jonsson K, Hunt TK, Mathes SJ. Oxygen as an isolated variable influences resistance to infection. Ann Surg. 1988;208(6):783–7. 213 181. Okumura C, Hollands A, Tran D, Olson J, Dahesh S, von Köckritz-Blickwede M, et al. A new pharmacological agent (AKB-4924) stabilizes hypoxia inducible factor (HIF) and increases skin innate defenses against bacterial infection. J Mol Med. 2012;90(9):1079–89. 182. Ong CWM, Fox K, Ettorre A, Elkington PT, Friedland JS. Hypoxia increases neutrophil-driven matrix destruction after exposure to Mycobacterium tuberculosis. Sci Rep. 2018;8:11475. 183. Christoffersson G, Vågesjo E, Vandooren J, Liden M, Massena S, Reinert RB, et al. VEGF-A recruits a proangiogenic MMP-9 – delivering neutrophil subset that induces angiogenesis in transplanted hypoxic tissue. Blood. 2012;120:4653–62. 184. McInturff AM, Cody MJ, Elliott EA, Glenn JW, Rowley JW, Rondina MT, et al. Mammalian target of rapamycin regulates neutrophil extracellular trap formation via induction of hypoxia-inducible factor 1α. Blood. 2012;120:3118–25. 185. Vollger L, Akong-Moore K, Cox L, Goldmann O, Wang Y, Schafer ST, et al. Iron- chelating agent desferrioxamine stimulates formation of neutrophil extracellular traps (NETs) in human blood-derived neutrophils. Biosci Rep. 2016;36:e00333. 186. Branitzki-Heinemann K, Möllerherm H, Völlger L, Hussein D, de Buhr N, Blodkamp S, et al. Formation of neutrophil extracellular traps under low oxygen level. Front Immunol. 2016;7:518. 187. Chen Y-C, Ho C-W, Tsai H-H, Wang J-S. Interval and continuous exercise regimens suppress neutrophil-derived microparticle formation and neutrophil-promoted thrombin generation under hypoxic stress. Clin Sci. 2015;128(7):425–36. 188. Tremblay JC, Thom SR, Yang M, Ainslie PN. Oscillatory shear stress, flow-mediated dilatation, and circulating microparticles at sea level and high altitude. Atherosclerosis. 2017;256:115–22. 189. Ayers L, Stoewhas A, Ferry B, Latshang T, Lo Cascio C, Sadler R, et al. Circulating levels of cell-derived microparticles are reduced by mild hypobaric hypoxia: data from a randomised controlled trial. Eur J Appl Physiol. 2014;114:1067–73. 190. Priou P, Gagnadoux F, Tesse A, Mastronardi ML, Agouni A, Meslier N, et al. Endothelial dysfunction and circulating microparticles from patients with obstructive sleep apnea. Am J Pathol. 2010;177:974–83. 191. Elks PM, Eeden FJ Van, Dixon G, Wang X, Reyes-aldasoro CC, Ingham PW, et al. Activation of hypoxia-inducible factor-1α (Hif-1α) delays inflammation resolution by reducing neutrophil apoptosis and reverse migration in a zebrafish inflammation model. Blood. 2011;118(3):712–23. 192. Dyugovskaya L, Polyakov A, Lavie P, Lavie L. Delayed Neutrophil Apoptosis in Patients with Sleep Apnea. Am J Resp Crit Care Med. 2008;177:544–54. 193. Whitlock BB, Gardai S, Fadok V, Bratton D, Henson PM. Differential Roles for alphaMbeta2 Integrin Clustering or Activation in the Control of Apoptosis Via Regulation of Akt and ERK Survival Mechanisms. J Cell Biol. 2000;151:1305–20. 194. Walmsley SR, Chilvers ER, Thompson AA, Vaughan K, Marriott HM, Parker LC, et al. Prolyl hydroxylase 3 (PHD3) is essential for hypoxic regulation of neutrophilic inflammation in humans and mice. J Clin Invest. 2011;121(3):1053–63. 214 195. Kiss J, Mollenhauer M, Walmsley SR, Kirchberg J, Radhakrishnan P, Dudda J, et al. Loss of the oxygen sensor PHD3 enhances the innate immune response to abdominal sepsis. J Immunol. 2012;189:1955–65. 196. Lodge KM, Thompson AAR, Chilvers ER, Condliffe AM. Hypoxic regulation of neutrophil function and consequences for Staphylococcus aureus infection. Microbes Infect. 2017;19:166–76. 197. Turato G, Di Stefano A, Maestrelli P, Mapp C, Ruggieri M, Roggeri A, et al. Effect of smoking cessation on airway inflammation of rats with chronic bronchitis. Am J Resp Crit Care Med. 1995;152:1262–7. 198. Pesci A, Majori M, Cuomo A, Borciani N, Bertacco S, Cacciani G, et al. Neutrophils infiltrating bronchial epithelium in chronic obstructive pulmonary disease. Respir Med. 1998;92:863–70. 199. Restrepo MI, Mortensen EM, Pugh J a, Anzueto a. COPD is associated with increased mortality in patients with community-acquired pneumonia. Eur Respir J Off J Eur Soc Clin Respir Physiol. 2006;28(2):346–51. 200. Hiemstra PS, van Wetering S, Stolk J. Neutrophil serine proteinases and defensins in chronic obstructive pulmonary disease: effects on pulmonary epithelium. Eur Respir J. 1998;12:1200–8. 201. Hou H-H, Cheng S-L, Chung K-P, Wei S-C, Tsao P-N, Lu H-H, et al. PIGF mediates neutrophil elastase-induced airway epithelial cell apoptosis and emphysema. Respir Res. 2014;15:106. 202. Shapiro SD, Goldstein NM, Houghton AM, Kobayashi DK, Kelley D, Belaaouaj A. Neutrophil elastase contributes to cigarette smoke-induced emphysema in mice. Am J Pathol. American Society for Investigative Pathology; 2003;163:2329–35. 203. Paone G, Conti V, Vestri A, Leone A, Puglisi G, Benassi F, et al. Analysis of sputum markers in the evaluation of lung inflammation and functional impairment in symptomatic smokers and COPD patients. Dis Markers. 2011;31:91–100. 204. Carter RI, Ungurs MJ, Mumford RA, Stockley RA. Aα-Val360: A marker of neutrophil elastase and COPD disease activity. Eur Respir J. 2013;41:31–8. 205. Sinden NJ, Stockley RA. Proteinase 3 activity in sputum from subjects with alpha-1- antitrypsin deficiency and COPD. Eur Respir J. 2013;41:1042–50. 206. Zhu A, Ge D, Zhang J, Teng Y, Yuan C, Huang M, et al. Sputum myeloperoxidase in chronic obstructive pulmonary disease. Eur J Med Res. 2014;19:12. 207. Vernooy J, Lindeman J, Jacobs J, Hanemaaijer R, Wouters E. Increased activity of matrix metalloproteinase-8 and matrix metalloproteinase-9 in induced sputum from patients with COPD. Chest. 2004;126:1802–10. 208. Thatcher TH, Hsiao H-M, Pinner E, Laudon M, Pollock SJ, Sime PJ, et al. Neu-164 and Neu-107, two novel antioxidant and anti-myeloperoxidase compounds, inhibit acute cigarette smoke-induced lung inflammation. Am J Physiol lung cell Mol Physiol. 2013;305:L165-174. 209. Churg A, Marshall C V, Sin DD, Bolton S, Zhou S, Thain K, et al. Late intervention with a myeloperoxidase inhibitor stops progression of experimental chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2012 Jan 1;185(1):34–43. 215 210. Owen CA, Campbell EJ. The cell biology of leukocyte-mediated proteolysis. J Leukoc Biol. 1999;65:137–50. 211. Koo HK, Hong Y, Lim MN, Yim JJ, Kim WJ. Relationship between plasma matrix metalloproteinase levels, pulmonary function, bronchodilator response, and emphysema severity. Int J COPD. 2016;11:1129–37. 212. Vandenbroucke RE, Dejonckheere E, Libert C. A therapeutic role for matrix metalloproteinase inhibitors in lung diseases? Eur Respir J. 2011;38:1200–14. 213. Sinden NJ, Baker MJ, Smith DJ, Kreft J-U, Dafforn TR, Stockley RA. α-1-Antitrypsin variants and the proteinase/antiproteinase imbalance in chronic obstructive pulmonary disease. Am J Physiol Lung Cell Mol Physiol. 2015;308:L179-190. 214. Liou TG, Campbell EJ. Quantum proteolysis resulting from release of single granules by human neutrophils: a novel, nonoxidative mechanism of extracellular proteolytic activity. J Immunol. 1996;157:2624–31. 215. Owen CA, Campbell MA, Sannes PL, Boukedes SS, Campbell EJ. Cell surface-bound elastase and cathepsin G on human neutrophils: A novel, non-oxidative mechanism by which neutrophils focus and preserve catalytic activity of serine proteinases. J Cell Biol. 1995;131:775–89. 216. Sapey E, Stockley JA, Greenwood H, Ahmad A, Bayley D, Lord JM, et al. Behavioral and structural differences in migrating peripheral neutrophils from patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2011 May 1;183:1176–86. 217. Walton G, Stockley J, Griffiths D, Sadhra C, Purvis T, Sapey E. Repurposing treatments to enhance innate immunity. Can statins improve neutrophil functions and clinical outcomes in COPD? J Clin Med. 2016;5:89. 218. Teresi RE, Planchon SM, Waite KA, Eng C. Regulation of the PTEN promoter by statins and SREBP. Hum Mol Genet. 2008;17:919–28. 219. Shahriary A, Mehrani H, Ghanei M, Parvin S. Comparative proteome analysis of peripheral neutrophils from sulfur mustard-exposed and COPD patients. J Immunotoxicol. 2015;12:132–9. 220. Langereis JD, Schweizer RC, Lammers J-WJ, Koenderman L, Ulfman LH. A unique protein profile of peripheral neutrophils from COPD patients does not reflect cytokine- induced protein profiles of neutrophils in vitro. BMC Pulm Med. 2011;11:44. 221. Loi ALT, Hoonhorst S, van Aalst C, Langereis J, Kamp V, Sluis-Eising S, et al. Proteomic profiling of peripheral blood neutrophils identifies two inflammatory phenotypes in stable COPD patients. Respir Res. Respiratory Research; 2017;18:100. 222. Mall MA, Harkema JR, Trojanek JB, Treis D, Livraghi A, Schubert S, et al. Development of chronic bronchitis and emphysema in β-epithelial Na+ channel- overexpressing mice. Am J Respir Crit Care Med. 2008;177:730–42. 223. Nie X, Laforest R, Elvington A, Randolph GJ, Zheng J, Voller T, et al. PET/MRI of hypoxic atherosclerosis using 64Cu-ATSM in a rabbit model. J Nucl Med. 2016;57:2006–11. 216 224. Joshi FR, Manavaki R, Fryer TD, Figg NL, Sluimer JC, Aigbirhio FI, et al. Vascular imaging with 18F-Fluorodeoxyglucose Positron Emission Tomography is influenced by hypoxia. J Am Coll Cardiol. 2017;69:1873–4. 225. MacLay JD, MacNee W. Cardiovascular disease in COPD: Mechanisms. Chest. 2013;143:798–807. 226. Ye C, Younus A, Malik R, Roberson L, Shaharyar S, Veledar E, et al. Subclinical cardiovascular disease in patients with chronic obstructive pulmonary disease: a systematic review. QJM. 2017;110:341–9. 227. Feary JR, Rodrigues LC, Smith CJ, Hubbard RB, Gibson JE. Prevalence of major comorbidities in subjects with COPD and incidence of myocardial infarction and stroke: a comprehensive analysis using data from primary care. Thorax. 2010;65:956–62. 228. Curkendall SM, DeLuise C, Jones JK, Lanes S, Stang MR, Goehring E, et al. Cardiovascular disease in patients with chronic obstructive pulmonary disease, Saskatchewan Canada: Cardiovascular disease in COPD patients. Ann Epidemiol. 2006;16:63–70. 229. Enriquez JR, De Lemos JA, Parikh S V., Peng SA, Spertus JA, Holper EM, et al. Association of chronic lung disease with treatments and outcomes patients with acute myocardial infarction. Am Heart J. 2013;165:43–9. 230. Chen W, Thomas J, Sadatsafavi M, FitzGerald J. Risk of cardiovascular comorbidity in patients with chronic obstructive pulmonary disease: a systematic review and meta- analysis. Lancet Respir Med. 2015;3:631–9. 231. Morgan AD, Rothnie KJ, Bhaskaran K, Smeeth L, Quint JK. Chronic obstructive pulmonary disease and the risk of 12 cardiovascular diseases: A population-based study using UK primary care data. Thorax. 2018;73:877–9. 232. Rothnie K, Connell O, Mullerova H, Smeeth L, Pearce N, Douglas I, et al. Myocardial infarction and ischaemic stroke following exacerbations of chronic obstructive pulmonary disease. Ann Am Thorac Soc. 2018;15:935–46. 233. Morgan A, Zakeri R, Quint J. Defining the relationship between COPD and CVD: what are the implications for clinical practice? Ther Adv Respir Dis. 2018;12:1–16. 234. Eickhoff P, Valipour A, Kiss D, Schreder M, Cekici L, Geyer K, et al. Determinants of Systemic Vascular Function in Patients with Stable Chronic Obstructive Pulmonary Disease. Am J Respir Crit Care Med. 2008;178:1211–8. 235. Steyers CM, Miller FJ. Endothelial dysfunction in chronic inflammatory diseases. Int J Mol Sci. 2014;15:11324–49. 236. Deanfield JE, Halcox JP, Rabelink TJ. Endothelial function and dysfunction: Testing and clinical relevance. Circulation. 2007;115:1285–95. 237. Mozos I, Malainer C, Horbańczuk J, Gug C, Stoian D, Luca CT, et al. Inflammatory markers for arterial stiffness in cardiovascular diseases. Front Immunol. 2017;8:1058. 238. Oyama J, Nagatomo D, Yoshioka G, Yamasaki A, Kodama K, Sato M, et al. The relationship between neutrophil to lymphocyte ratio, endothelial function, and severity in patients with obstructive sleep apnea. J Cardiol. 2016;67:295–302. 217 239. Mazor R, Shurtz-Swirski R, Farah R, Kristal B, Shapiro G, Dorlechter F, et al. Primed polymorphonuclear leukocytes constitute a possible link between inflammation and oxidative stress in hyperlipidemic patients. Atherosclerosis. 2008;197:937–43. 240. Drechsler M, Megens RTA, Van Zandvoort M, Weber C, Soehnlein O. Hyperlipidemia- triggered neutrophilia promotes early atherosclerosis. Circulation. 2010;122:1837–45. 241. Lomas-Neira J, Venet F, Chung C, Thakkar R, Heffernan D, Ayala A. Neutrophil – Endothelial Interactions Mediate Angiopoietin-2 – Associated Pulmonary Endothelial Cell Dysfunction in Indirect Acute Lung Injury in Mice. Am J Respir Cell Mol Biol. 2014;50:193–200. 242. Gill SE, Taneja R, Rohan M, Wang L, Mehta S. Pulmonary microvascular albumin leak is associated with endothelial cell death in murine sepsis-induced lung injury in vivo. PLoS One. 2014;9:e88501. 243. Fox ED, Heffernan DS, Cioffi WG, Reichner JS. Neutrophils from critically ill septic patients mediate profound loss of endothelial barrier integrity. Crit Care. 2013;17:R226. 244. Jerome SN, Akimitsu T, Korthuis RJ. Leukocyte adhesion, edema, and development of postischemic capillary no-reflow. Am J Physiol. 1994;267:H1329-1336. 245. Pinsky DJ, Naka Y, Liao H, Oz MC, Wagner DD, Mayadas TN, et al. Hypoxia-induced exocytosis of endothelial cell weibel-palade bodies: A mechanism for rapid neutrophil recruitment after cardiac preservation. J Clin Invest. 1996;97:493–500. 246. Sager HB, Dutta P, Dahlman JE, Hulsmans M, Courties G, Sun Y, et al. RNAi targeting multiple cell adhesion molecules reduces immune cell recruitment and vascular inflammation after myocardial infarction. Sci Transl Med. 2016;8:342ra80. 247. Jerke U, Rolle S, Purfürst B, Luft FC, Nauseef WM, Kettritz R. Β2 integrin-mediated cell-cell contact transfers active myeloperoxidase from neutrophils to endothelial cells. J Biol Chem. 2013;288:12910–9. 248. Naruko T, Ueda M, Haze K, Van der Wal AC, Van der Loos CM, Itoh A, et al. Neutrophil infiltration of culprit lesions in acute coronary syndromes. Circulation. 2002;106:2894–900. 249. Quillard T, Araújo HA, Franck G, Shvartz E, Sukhova G, Libby P. TLR2 and neutrophils potentiate endothelial stress, apoptosis and detachment: Implications for superficial erosion. Eur Heart J. 2015;36:1394–404. 250. Ionita MG, van den Borne P, Catanzariti LM, Moll FL, de Vries JPPM, Pasterkamp G, et al. High neutrophil numbers in human carotid atherosclerotic plaques are associated with characteristics of rupture-prone lesions. Arterioscler Thromb Vasc Biol. 2010;30:1842–8. 251. Darbousset R, Thomas GM, Mezouar S, Frère C, Bonier R, Mackman N, et al. Tissue factor−positive neutrophils bind to injured endothelial wall and initiate thrombus formation. Blood. 2012;120:2133–43. 252. DiStasi MR, Ley K. Opening the flood-gates: How neutrophil-endothelial interactions regulate permeability. Trends Immunol. 2009;30:547–56. 253. Giambelluca MS, Bertheau-Mailhot G, Laflamme C, Rollet-Labelle E, Servant MJ, Pouliot M. TNF-α expression in neutrophils and its regulation by glycogen synthase kinase-3: A potentiating role for lithium. FASEB J. 2014;28:3679–90. 218 254. Meegan JE, Yang X, Coleman DC, Jannaway M, Yuan SY. Neutrophil-mediated vascular barrier injury: Role of neutrophil extracellular traps. Microcirculation. 2017;24:e12352. 255. Klebanoff SJ. Myeloperoxidase: friend and foe. J Leukoc Biol. 2005;77:598–625. 256. Patterson EK, Fraser DD, Capretta A, Potter RF, Cepinskas G. Carbon monoxide- releasing molecule 3 inhibits myeloperoxidase (MPO) and protects against MPO- induced vascular endothelial cell activation/dysfunction. Free Radic Biol Med. 2014;70:167–73. 257. Rudolph TK, Rudolph V, Baldus S. Contribution of myeloperoxidase to smoking- dependent vascular inflammation. Proc Am Thorac Soc. 2008;5:820–3. 258. Yang JJ, Kettritz R, Falk RJ, Jennette JC, Gaido ML. Apoptosis of endothelial cells induced by the neutrophil serine proteases proteinase 3 and elastase. Am J Pathol. 1996;149:1617–26. 259. Zimmerman BJ, Granger DN. Reperfusion-induced leukocyte infiltration: Role of elastase. Am J Physiol. 1990;259:H390-394. 260. Leclercq A, Houard X, Philippe M, Ollivier V, Sebbag U, Meilhac O, et al. Involvement of intraplaque hemorrhage in atherothrombosis evolution via neutrophil protease enrichment. J Leukoc Biol. 2007;82:1420–9. 261. Alfaidi M, Wilson H, Daigneault M, Burnett A, Ridger V, Chamberlain J, et al. Neutrophil elastase promotes interleukin-1β secretion from human coronary endothelium. J Biol Chem. 2015;290:24067–78. 262. Moreno JA, Ortega-Gómez A, Delbosc S, Beaufort N, Sorbets E, Louedec L, et al. In vitro and in vivo evidence for the role of elastase shedding of CD163 in human atherothrombosis. Eur Heart J. 2012;33:252–63. 263. Wen G, An W, Chen J, Maguire EM, Chen Q, Yang F, et al. Genetic and pharmacologic inhibition of the neutrophil elastase inhibits experimental atherosclerosis. J Am Heart Assoc. 2018;7:e008187. 264. Smith FB, Fowkes FGR, Rumley A, Lee AJ, Lowe GDO, Hau CM. Tissue plasminogen activator and leucocyte elastase as predictors of cardiovascular events in subjects with angina pectoris: Edinburgh artery study. Eur Heart J. 2000;21:1607– 13. 265. Florence JM, Krupa A, Booshehri LM, Allen TC, Kurdowska AK. Metalloproteinase-9 contributes to endothelial dysfunction in atherosclerosis via protease activated receptor-1. PLoS One. 2017;12:e0171427. 266. Yasmin, Wallace S, McEniery CM, Dakham Z, Pusalkar P, Maki-Petaja K, et al. Matrix metalloproteinase-9 (MMP-9), MMP-2, and serum elastase activity are associated with systolic hypertension and arterial stiffness. Arterioscler Thromb Vasc Biol. 2005;25:372–8. 267. Hashimoto M, Kuriiwa S, Kojima A, Minagawa S, Numata T, Hara H, et al. Aortic rupture involving matrix metalloproteinases 8 and 9 during Staphylococcus aureus pneumonia. Thorax. 2018;73:397–8. 268. Döring Y, Drechsler M, Wantha S, Kemmerich K, Lievens D, Vijayan S, et al. Lack of neutrophil-derived CRAMP reduces atherosclerosis in mice. Circ Res. 2012;110:1052–6. 219 269. Tanaka K, Koike Y, Shimura T, Okigami M, Ide S, Toiyama Y, et al. In vivo characterization of neutrophil extracellular traps in various organs of a murine sepsis model. PLoS One. 2014;9:e111888. 270. Borissoff JI, Joosen IA, Versteylen MO, Brill A, Fuchs TA, Savchenko AS, et al. Elevated levels of circulating DNA and chromatin are independently associated with severe coronary atherosclerosis and a prothrombotic state. Arter Thromb Vasc Biol. 2013;33:2032–40. 271. Wang H, Wang Q, Venugopal J, Wang J, Kleiman K, Guo C, et al. Obesity-induced endothelial dysfunction is prevented by neutrophil extracellular trap inhibition. Sci Rep. 2018;8:4881. 272. Knight JS, Zhao W, Luo W, Subramanian V, O’Dell AA, Yalavarthi S, et al. Peptidylarginine deiminase inhibition is immunomodulatory and vasculoprotective in murine lupus. J Clin Invest. 2013;123:2981–93. 273. Knight JS, Luo W, O’Dell AA, Yalavarthi S, Zhao W, Subramanian V, et al. Peptidylarginine deiminase inhibition reduces vascular damage and modulates innate immune responses in murine models of atherosclerosis. Circ Res. 2014;114:947–56. 274. Johnson III BL, Kuethe JW, Caldwell CC. Neutrophil derived microvesicles: Emerging role of a key mediator to the immune response. Endocr Metab Immune Disord - Drug Targets. 2014;14:210–7. 275. Pluskota E, Woody NM, Szpak D, Ballantyne CM, Soloviev D a, Simon DI, et al. Expression, activation, and function of integrin alphaMbeta2 (Mac-1) on neutrophil- derived microparticles. Blood. 2008 Sep 15;112(6):2327–35. 276. Nolan S, Dixon R, Norman K, Hellewell P, Ridger V. Nitric oxide regulates neutrophil migration through microparticle formation. Am J Pathol. 2008;172:265–73. 277. Hong Y, Eleftheriou D, Hussain A a K, Price-Kuehne FE, Savage CO, Jayne D, et al. Anti-neutrophil cytoplasmic antibodies stimulate release of neutrophil microparticles. J Am Soc Nephrol. 2012 Jan;23(1):49–62. 278. Lim K, Sumagin R, Hyun Y-M. Extravasating Neutrophil-derived Microparticles Preserve Vascular Barrier Function in Inflamed Tissue. Immune Netw. 2013 Jun;13(3):102–6. 279. Agustí A, Edwards LD, Rennard SI, MacNee W, Tal-Singer R, Miller BE, et al. Persistent systemic inflammation is associated with poor clinical outcomes in COPD: A novel phenotype. PLoS One. 2012;7:e37483. 280. Coulson JM, Rudd JHF, Duckers JM, Rees JIS, Shale DJ, Bolton CE, et al. Excessive aortic inflammation in chronic obstructive pulmonary disease: An 18 F-FDG PET pilot study. J Nucl Med. 2010;51:1357–60. 281. Kunisaki KM, Dransfield MT, Anderson JA, Brook RD, Calverley PMA, Celli BR, et al. Exacerbations of chronic obstructive pulmonary disease and cardiac events: A cohort analysis. Am J Respir Crit Care Med. 2018;198:51–7. 282. Aaron SD, Vandemheen KL, Ramsay T, Zhang C, Avnur Z, Nikolcheva T, et al. Multi analyte profiling and variability of inflammatory markers in blood and induced sputum in patients with stable COPD. Respir Res. 2010;11:41. 220 283. Oudijk E-JD, Gerritsen WBM, Nijhuis EHJ, Kanters D, Maesen BLP, Lammers J-WJ, et al. Expression of priming-associated cellular markers on neutrophils during an exacerbation of COPD. Respir Med. 2006;100:1791–9. 284. Pletz MWR, Ioanas M, de Roux A, Burkhardt O, Lode H. Reduced spontaneous apoptosis in peripheral blood neutrophils during exacerbation of COPD. Eur Respir J. 2004;23:532–7. 285. Li Y, Lu Y, Zhao Z, Wang J, Li J, Wang W, et al. Relationships of MMP-9 and TIMP-1 proteins with chronic obstructive pulmonary disease risk: A systematic review and meta-analysis. J Res Med Sci. 2016;21:12. 286. Park HY, Man SFP, Tashkin D, Wise RA, Connett JE, Anthonisen NA, et al. The relation of serum myeloperoxidase to disease progression and mortality in patients with chronic obstructive pulmonary disease (COPD). PLoS One. 2013 Jan;8:e61315. 287. Noguera A, Busquets X, Sauleda J, Villaverde J, MacNee W, Agusti A. Expression of adhesion molecules and G proteins in circulating neutrophils in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 1998;158:1664–8. 288. Yamagata T, Sugiura H, Yokoyama T, Yanagisawa S, Ichikawa T, Ueshima K, et al. Overexpression of CD-11b and CXCR1 on circulating neutrophils: its possible role in COPD. Chest. 2007;132:890–9. 289. Woolhouse IS, Bayley DL, Lalor P, Adams DH, Stockley RA. Endothelial interactions of neutrophils under flow in chronic obstructive pulmonary disease. Eur Respir J. 2005;25:612–7. 290. Mackarel AJ, Russell KJ, Ryan CM, Hislip SJ, Rendall JC, FitzGerald MX, et al. CD18 dependency of transendothelial neutrophil migration differs during acute pulmonary inflammation. J Immunol. 2001;167:2839–46. 291. MacNee W, Wiggs B, Belzberg A, Hogg J. The effect of cigarette smoking on neutrophil kinetics in human lungs. N Engl J Med. 1989;321:924–8. 292. Hueper K, Vogel-Claussen J, Parikh MA, Austin JHM, Bluemke DA, Carr J, et al. Pulmonary microvascular blood flow in mild chronic obstructive pulmonary disease and emphysema. Am J Respir Crit Care Med. 2015;192:570–80. 293. Arnould T, Michiels C, Remacle J. Increased PMN adherence on endothelial cells after hypoxia: involvement of PAF, CD18/CD11b, and ICAM-1. Am J Physiol. 1993;264:C1102-1110. 294. Zhou Y, Huang X, Zhao T, Qiao M, Zhao X, Zhao M, et al. Hypoxia augments LPS- induced inflammation and triggers high altitude cerebral edema in mice. Brain Behav Immun. 2017;64:266–75. 295. Baldi S, Pinna GD, Mombaruzzo P, Biglieri M, De Martini A, Palange P. C-reactive protein correlates with tissue oxygen availability in patients with stable COPD. Int J COPD. 2008;3:745–51. 296. Dinh-Xuan AT, Higenbottam TW, Clelland CA, Pepke-Zaba J, Cremona G, Butt Y, et al. Impairment of endothelium-dependent pulmonary-artery relaxation in chronic obstructive lung disease. N Engl J Med. 1991;324:1539–47. 297. Fein D, Zaidi A, Sulica R. Pulmonary hypertension due to common respiratory conditions: Classification, evaluation and management strategies. J Clin Med. 2016;5:75. 221 298. Oliveira M V., Abreu SC, Padilha GA, Rocha NN, Maia LA, Takiya CM, et al. Characterization of a mouse model of emphysema induced by multiple instillations of low-dose elastase. Front Physiol. 2016;7:457. 299. Maruyama K, Ye CL, Woo M, Venkatacharya H, Lines LD, Silver MM, et al. Chronic hypoxic pulmonary hypertension in rats and increased elastolytic activity. Am J Physiol. 1991;261:H1716–26. 300. Klinke A, Berghausen E, Friedrichs K, Molz S, Lau D, Remane L, et al. Myeloperoxidase aggravates pulmonary arterial hypertension by activation of vascular Rho-kinase. J Clin Investig Insight. 2018;3:e97530. 301. Kim V, Benditt JO, Wise RA, Sharafkhaneh A. Oxygen therapy in chronic obstructive pulmonary disease. Proc Am Thorac Soc. 2008;5:513–8. 302. McGovern N. The Effects of Hypoxia on Neutrophil Biology. PhD thesis. 2010; 303. Khan MA, Philip LM, Cheung G, Vadakepeedika S, Grasemann H, Sweezey N, et al. Regulating NETosis: Increasing pH promotes NADPH oxidase-dependent NETosis. Front Med. 2018;5:19. 304. Juss JK, Hayhoe RP, Owen CE, Bruce I, Walmsley SR, Cowburn AS, et al. Functional redundancy of class I phosphoinositide 3-kinase (PI3K) isoforms in signaling growth factor-mediated human neutrophil survival. PLoS One. 2012;e45933. 305. Uhlmann J, Siemens N, Kai-Larsen Y, Fiedler T, Bergman P, Johansson L, et al. Phosphoglycerate kinase - a novel streptococcal factor involved in neutrophil activation and degranulation. J Infect Dis. 2016;214:1876–83. 306. Malmström E, Davidova A, Mörgelin M, Linder A, Larsen M, Qvortrup K, et al. Targeted mass spectrometry analysis of neutrophil-derived proteins released during sepsis progression. Thromb Haemost. 2014;112:1230–43. 307. Zhang D, Shooshtarizadeh P, Laventie B-J, Colin DA, Chich J-F, Vidic J, et al. Two chromogranin A-derived peptides induce calcium entry in human neutrophils by calmodulin-regulated calcium independent phospholipase A2. PLoS One. 2009;4:e4501. 308. Aslam R, Laventie B-J, Marban C, Prevost G, Keller D, Strub J-M, et al. Activation of neutrophils by the two-component leukotoxin LukE/D from Staphylococcus aureus: Proteomic analysis of the secretions. J Proteome Res. 2013;12:3667–78. 309. Boussac M, Garin J. Calcium-dependent secretion in human neutrophils: A proteomic approach. Electrophoresis. 2000;21:665–72. 310. Thompson A, Schäfer J, Kuhn K, Kienle S, Schwarz J, Schmidt G, et al. Tandem mass tags: A novel quantification strategy for comparative analysis of complex protein mixtures by MS/MS. Anal Chem. 2003;75:1895–904. 311. Ettelaie C, Collier MEW, Maraveyas A, Ettelaie R. Characterization of physical properties of tissue factor-containing microvesicles and a comparison of ultracentrifuge-based recovery procedures. J Extracell Vesicles. 2014;3:23592. 312. Johnson PC, Vandegriff K, Tsai AG, Intaglietta M, Paul C, Vandegriff K, et al. Effect of acute hypoxia on microcirculatory and tissue oxygen levels in rat cremaster muscle. 2005;(35):1177–84. 222 313. Imtiyaz HZ, Simon MC. Hypoxia-inducible factors as essential regulators of inflammation. Curr Top Microbiol Immunol. 2010;345:105–20. 314. Kokura S, Norimasa Y, Yoshikawa T. Serial Review : Vascular Dysfunction and Free Radicals. Free Radic Biol Med. 2002;33(4):427–32. 315. Hoenderdos K. Modulation of neutrophil degranulation by hypoxia. PhD thesis. 2014; 316. Rius J, Guma M, Schachtrup C, Akassoglou K, Zinkernagel AS, Nizet V, et al. NF-κB links innate immunity to the hypoxic response through transcriptional regulation of HIF-1α. Nature. 2008;453:807–11. 317. Witko-Sarsat V, Lesavre P, Lopez S, Bessou G, Hieblot C, Prum B, et al. A large subset of neutrophils expressing membrane proteinase 3 is a risk factor for vasculitis and rheumatoid arthritis. J Am Soc Nephrol. 1999;10:1224–33. 318. Johansson L, Linner A, Sunden-Cullberg J, Haggar A, Herwald H, Lore K, et al. Neutrophil-derived hyperresistinemia in severe acute streptococcal infections. J Immunol. 2009;183:4047–54. 319. Csernok E, Ernst M, Schmitt W, Bainton DF, Gross WL. Activated neutrophils express proteinase 3 on their plasma membrane in vitro and in vivo. Clin Exp Immunol. 1994;95:244–50. 320. Fujimura N, Obara H, Suda K, Takeuchi H, Miyasho T, Kawasako K, et al. Neutrophil elastase inhibitor improves survival rate after ischemia reperfusion injury caused by supravisceral aortic clamping in rats. J Surg Res. 2013;180:e31-36. 321. Ruscitti P, Cipriani P, Masedu F, Romano S, Berardicurti O, Liakouli V, et al. Increased cardiovascular events and subclinical atherosclerosis in rheumatoid arthritis patients: 1 year prospective single centre study. PLoS One. 2017;12:e0170108. 322. Chironi G, Pagnoux C, Simon A, Pasquinelli-Balice M, Del-Pino M, Gariepy J, et al. Increased prevalence of subclinical atherosclerosis in patients with small-vessel vasculitis. Heart. 2007;93:96–9. 323. Zhang P, Zhu D, Chen X, Li Y, Li N, Gao Q, et al. Prenatal hypoxia promotes atherosclerosis via vascular inflammation in the offspring rats. Atherosclerosis. 2016;245:28–34. 324. Burke DL, Frid MG, Kunrath CL, Karoor V, Anwar A, Wagner BD, et al. Sustained hypoxia promotes the development of a pulmonary artery-specific chronic inflammatory microenvironment. Am J Physiol Lung Cell Mol Physiol. 2009;297:L238- 250. 325. Stark A, Chandra A, Chakraborty K, Alam R, Carbonaro V, Clark J, et al. PI3Kδ hyper-activation promotes development of B cells that exacerbate Streptococcus pneumoniae infection in an antibody-independent manner. Nat Commun. 2018;9:3174. 326. Okkenhaug K, Bilancio A, Farjot G, Priddle H, Sancho S, Peskett E, et al. Impaired B and T Cell Antigen Receptor Signaling in p110δ PI 3-Kinase Mutant Mice. Science (80- ). 2002;297:1031–4. 327. Hirsch E, Katanaev VL, Garlanda C, Azzolino O, Pirola L, Silengo L, et al. Central role for G Protein–coupled Phosphoinositide 3-Kinase γ in inflammation. Science (80- ). 2013;287:1049-10–53. 223 328. Condliffe AM, Davidson K, Anderson KE, Ellson CD, Crabbe T, Okkenhaug K, et al. Sequential activation of class IB and class IA PI3K is important for the primed respiratory burst of human but not murine neutrophils. Blood. 2005;106:1432–40. 329. Vercellotti G, Wickham N, Gustafson K, Yin H, Hebert M, Jacob H. Thrombin-treated endothelium primes neutrophil functions: inhibition by platelet-activating factor receptor antagonists. J Leukoc Biol. 1989;45:483–90. 330. Mateo J, Izquierdo-Garcia D, Badimon J, Fayad Z, Fuster V. Noninvasive assessment of hypoxia in rabbit advanced atherosclerosis using 18F-fluoromisonidazole PET imaging. Circ Cardiovasc Imaging. 2014;7:312–20. 331. Sluimer JC, Gasc JM, van Wanroij JL, Kisters N, Groeneweg M, Sollewijn Gelpke MD, et al. Hypoxia, hypoxia-inducible transcription factor, and macrophages in human atherosclerotic plaques are correlated with intraplaque angiogenesis. J Am Coll Cardiol. 2008;51:1258–65. 332. Baze MM, Hunter K, Hayes JP. Chronic hypoxia stimulates an enhanced response to immune challenge without evidence of an energetic tradeoff. J Exp Biol. 2011;214:3255–68. 333. Peyssonnaux C, Cejudo-martin P, Doedens A, Zinkernagel AS, Johnson RS, Nizet V. Cutting Edge: Essential role of hypoxia inducible factor-1 α in development of lipopolysaccharide-induced sepsis. J Immunol. 2007;178:7516–9. 334. Koenderman L, Kanters D, Maesen B, Raaijmakers J, Lammers J-WJ, de Kruif J, et al. Monitoring of neutrophil priming in whole blood by antibodies isolated from a synthetic phage antibody library. J Leukoc Biol. 2000;68:58–64. 335. Oudijk EJD, Nijhuis EHJ, Zwank MD, van de Graaf EA, Mager HJ, Coffer PJ, et al. Systemic inflammation in COPD visualised by gene profiling in peripheral blood neutrophils. Thorax. 2005;60:538–44. 336. Shimizu M, Hasegawa N, Nishimura T, Endo Y, Shiraishi Y, Yamasawa W, et al. Effects of TNF-α-converting enzyme inhibition on acute lung injury induced by endotoxin in the rat. Shock. 2009;32:535–40. 337. Keatings VM, Collins PD, Scott DM, Barnes PJ. Differences in interleukin-8 and tumor necrosis factor-α in induced sputum from patients with chronic obstructive pulmonary disease or asthma. Am J Respir Crit Care Med. 1996;153:530–4. 338. Di Francia M, Barbier D, Mege JL, Orehek J. Tumor necrosis factor-alpha levels and weight loss in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 1994;150:1453–5. 339. Takabatake N, Nakamura H, Abe S, Inoue S, Hino T, Saito H, et al. The relationship between chronic hypoxemia and activation of the tumor necrosis factor-α system in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2000;161:1179–84. 340. Aaron SD, Angel JB, Lunau M, Wright K, Fex C, Le Saux N, et al. Granulocyte inflammatory markers and airway infection during acute exacerbation of chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2001;163:349–55. 341. Zheng J, Shi Y, Xiong L, Zhang W, Li Y, Gibson PG, et al. The expression of IL-6, TNF-α and MCP-1 in respiratory viral infection in acute exacerbations of chronic obstructive pulmonary disease. J Immunol Res. Hindawi; 2017;8539294. 224 342. Churg A, Wang RD, Tai H, Wang X, Xie C, Wright JL. Tumor necrosis factor-α drives 70% of cigarette smoke-induced emphysema in the mouse. Am J Respir Crit Care Med. 2004;170:492–8. 343. Singh NRP, Johnson A, Peters a M, Babar J, Chilvers ER, Summers C. Acute lung injury results from failure of neutrophil de-priming: a new hypothesis. Eur J Clin Invest. 2012 Dec;42(12):1342–9. 344. Bitencourt CS, Bessi VL, Huynh DN, Menard L, Lefebvre JS, Levesque T, et al. Cooperative role of endogenous leucotrienes and platelet-activating factor in ischaemia–reperfusion-mediated tissue injury. J Cell Mol Med. 2013;17:1554–65. 345. Sandoval D, Gukovskaya A, Reavey P, Gukovsky S, Sisk A, Braquet P, et al. The role of neutrophils and platelet-activating factor in mediating experimental pancreatitis. Gastroenterology. 1996;111:1081–91. 346. Mariano F, Guida G, Donati D, Tetta C, Cavalli PL, Verzetti G, et al. Nephrology Dialysis Transplantation Production of platelet-activating factor in patients with sepsis- associated acute renal failure. 1999;1150–7. 347. Sapey E, Wood AM, Ahmad A, Stockley RA. Tumor necrosis factor-α rs361525 polymorphism is associated with increased local production and downstream inflammation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2010;182:192–9. 348. Liu L, Puri KD, Penninger JM, Kubes P. Leukocyte PI3Kγ and PI3Kδ have temporally distinct roles for leukocyte recruitment in vivo. Blood. 2007;110:1191–8. 349. Sadhu C, Masinovsky B, Dick K, Sowell CG, Staunton DE. Essential role of phosphoinositide 3-kinase in neutrophil directional movement. J Immunol. 2003;170:2647–54. 350. Madeddu P, Kraenkel N, Barcelos LS, Siragusa M, Campagnolo P, Oikawa A, et al. Phosphoinositide 3-kinase γ gene knockout impairs postischemic neovascularization and endothelial progenitor cell functions. Arterioscler Thromb Vasc Biol. 2008;28:68– 76. 351. Alvarez-Tejado M, Naranjo-sua S, Jimenez C, Carrera AC, Landazuri MO, del Peso L. Hypoxia induces the activation of the phosphatidylinositol 3-kinase/Akt cell survival pathway in PC12 cells. J Biol Chem. 2001;276:22368–74. 352. Alloatti G, Levi R, Malan D, Del Sorbo L, Bosco O, Barberis L, et al. Phosphoinositide 3-kinase γ-deficient hearts are protected from the PAF-dependent depression of cardiac contractility. Cardiovasc Res. 2003;60:242–9. 353. Vogt KL, Summers C, Chilvers ER, Condliffe AM. Priming and de-priming of neutrophil responses in vitro and in vivo. Eur J Clin Invest. 2018;e12967:Epub ahead of print. 354. Savage CO, Pottinger BE, Gaskin G, Pusey CD, Pearson JD. Autoantibodies developing to myeloperoxidase and proteinase 3 in systemic vasculitis stimulate neutrophil cytotoxicity toward cultured endothelial cells. Am J Pathol. 1992;141:335– 42. 355. Mihara K, Ramachandran R, Renaux B, Saifeddine M, Hollenberg MD. Neutrophil elastase and proteinase-3 trigger G protein-biased signaling through proteinase- activated receptor-1 (PAR1). J Biol Chem. 2013;288:32979–90. 225 356. Yoon C, Hur J, Oh I, Park K, Kim T, Shin J, et al. Intercellular adhesion molecule-1 is upregulated in ischemic muscle, which mediates trafficking of endothelial progenitor cells. Arterioscler Thromb Vasc Biol. 2006;26:1066–72. 357. Milhoan K, Lane T, Bloor C. Hypoxia induces endothelial cells to increase their adherence for neutrophils: role of PAF. Am J Physiol. 1992;263:H956-962. 358. Salas A, Panes J, Elizalde JI, Granger DN, Pique JM. Reperfusion-induced oxidative stress in diabetes: cellular and enzymatic sources. J Leukoc Biol. 1999;66:59–66. 359. Vercellotti G, Yin H, Gustafson K, Nelson R, Jacob H. Platelet-activating factor primes neutrophil responses to agonists: role in promoting neutrophil-mediated endothelial damage. Blood. 1988;71:1100–7. 360. Koshio O, Nagao T, Mabuchi A, Ono Y, Suzuki K. Apoptotic signaling in endothelial cells with neutrophil activation. Mol Cell Biochem. 2012;363:269–80. 361. Jerke U, Hernandez DP, Beaudette P, Korkmaz B, Dittmar G, Kettritz R. Neutrophil serine proteases exert proteolytic activity on endothelial cells. Kidney Int. 2015;88:764–75. 362. Maruyama K, Ye C, Woo M, Venkatacharya H, Lines L, Silver M, et al. Chronic hypoxic pulmonary hypertension in rats and increased elastolytic activity. Am J Physiol. 1991;261:H1716-1726. 363. Audain E, Uszkoreit J, Sachsenberg T, Pfeuffer J, Liang X, Hermjakob H, et al. In- depth analysis of protein inference algorithms using multiple search engines and well- defined metrics. J Proteomics. 2017;150:170–82. 364. Jethwaney D, Islam MR, Leidal KG, de Bernabe DB-V, Campbell KP, Nauseef WM, et al. Proteomic analysis of plasma membrane and secretory vesicles from human neutrophils. Proteome Sci. 2007;5:12. 365. Niemann CU, Cowland JB, Klausen P, Askaa J, Calafat J, Borregaard N. Localization of serglycin in human neutrophil granulocytes and their precursors. J Leukoc Biol. 2004;76(2):406–15. 366. Timár CI, Lorincz ÁM, Csépányi-Kömi R, Vályi-Nagy A, Nagy G, Buzás EI, et al. Antibacterial effect of microvesicles released from human neutrophilic granulocytes. Blood. 2013;121(3):510–8. 367. Aquino EN, Neves ACD, Santos KC, Uribe CE, Souza PEN, Correa JR, et al. Proteomic analysis of neutrophil priming by PAF. Protein Pept Lett. 2016;23:142–51. 368. Langereis JD, Franciosi L, Ulfman LH, Koenderman L. GM-CSF and TNFα modulate protein expression of human neutrophils visualized by fluorescence two-dimensional difference gel electrophoresis. Cytokine. 2011;56:422–9. 369. Boldt K, Rist W, Weiss SM, Weith A, Lenter MC. FPRL-1 induces modifications of migration-associated proteins in human neutrophils. Proteomics. 2006;6:4790–9. 370. Arshid S, Tahir M, Fontes B, Frasson de Souza Montero E, Castro MS, Sidoli S, et al. High performance mass spectrometry based proteomics reveals enzyme and signaling pathway regulation in neutrophils during the early stage of surgical trauma. Proteomics Clin Appl. 2017;11:1600001. 226 371. Zhou JY, Krovvidi RK, Gao Y, Gao H, Petritis BO, De AK, et al. Trauma-associated human neutrophil alterations revealed by comparative proteomics profiling. Proteomics Clin Appl. 2013;7(7):571–83. 372. Limon D, Fabrega MJ, Calpena AC, Badía J, Baldoma L, Pérez-García L. Multifunctional serine protease inhibitor-coated water-soluble gold nanoparticles as a novel targeted approach for the treatment of inflammatory skin diseases. Bioconjug Chem. 2018;29:1060–72. 373. Pascovici D, Handler D, Wu JX, Haynes PA. Multiple testing corrections in quantitative proteomics: a useful but blunt tool. Proteomics. 2016;2448–53. 374. Šenolt L, Housa D, Vernerová Z, Jirásek T, Svobodová R, Veigl D, et al. Resistin in rheumatoid arthritis synovial tissue, synovial fluid and serum. Ann Rheum Dis. 2007;66:458–63. 375. Oikonomou KA, Kapsoritakis AN, Theodoridou C, Karangelis D, Germenis A, Stefanidis I, et al. Neutrophil gelatinase-associated lipocalin (NGAL) in inflammatory bowel disease: Association with pathophysiology of inflammation, established markers, and disease activity. J Gastroenterol. 2012;47:519–30. 376. Fu C, Jiang L, Zhu F, Liu Z, Li W, Jiang H, et al. Chronic intermittent hypoxia leads to insulin resistance and impaired glucose tolerance through dysregulation of adipokines in non-obese rats. Sleep Breath. 2015;19(4):1467–73. 377. Mellor A, Boos C, Stacey M, Hooper T, Smith C, Begley J, et al. Neutrophil Gelatinase-Associated Lipocalin: Its Response to Hypoxia and Association with Acute Mountain Sickness. Dis Markers. 2013;35(5):537–42. 378. Snäll J, Linnér A, Uhlmann J, Siemens N, Ibold H, Janos M, et al. Differential neutrophil responses to bacterial stimuli: Streptococcal strains are potent inducers of heparin-binding protein and resistin-release. Sci Rep. 2016;6:21288. 379. Spencer LA, Szela CT, Perez SAC, Kirchhoffer CL, Neves JS, Radke AL, et al. Human eosinophils constitutively express multiple Th1, Th2, and immunoregulatory cytokines that are secreted rapidly and differentially. J Leukoc Biol. 2009;85:117–23. 380. Song F, Zhang X, Ren X-B, Zhu P, Xu J, Wang L, et al. Cyclophilin A (CyPA) induces chemotaxis independent of its peptidylprolyl cis-trans isomerase activity. Direct binding between CyPA and the ectodomain of CD147. J Biol Chem. 2011;286:8197– 203. 381. Wang L, Wang C-H, Jia J-F, Ma X-K, Li Y, Zhu H-B, et al. Contribution of cyclophilin A to the regulation of inflammatory processes in rheumatoid arthritis. J Clin Immunol. 2010;30:24–33. 382. Tegeder I, Schumacher A, John S, Geiger H, Geisslinger G, Bang H, et al. Elevated serum cyclophilin levels in patients with severe sepsis. J Clin Immunol. 1997;17:380– 6. 383. Piechota-Polanczyk A, Włodarczyk M, Sobolewska-Włodarczyk A, Jonakowski M, Pilarczyk A, Stec-Michalska K, et al. Serum cyclophilin A correlates with increased tissue MMP-9 in patients with ulcerative colitis, but not with crohn’s disease. Dig Dis Sci. 2017;62:1511–7. 227 384. Bhargava M, Becker TL, Viken KJ, Jagtap PD, Dey S, Steinbach MS, et al. Proteomic profiles in acute respiratory distress syndrome differentiates survivors from non- survivors. PLoS One. 2014;9:e109713. 385. Billich A, Winkler G, Aschauer H, Rot A, Peichl P. Presence of cyclophilin A in synovial fluids of patients with rheumatoid arthritis. J Exp Med. 1997;185(5):975–80. 386. Sherry B, Yarlett N, Strupp A, Cerami A. Identification of cyclophilin as a proinflammatory secretory product of lipopolysaccharide-activated macrophages. Proc Natl Acad Sci. 1992;89:3511–5. 387. Seko Y, Fujimura T, Taka H, Mineki R, Murayama K, Nagai R. Hypoxia followed by reoxygenation induces secretion of cyclophilin A from cultured rat cardiac myocytes. Biochem Biophys Res Commun. 2004;317:162–8. 388. Suzuki J, Jin ZG, Meoli DF, Matoba T, Berk BC. Cyclophilin A is secreted by a vesicular pathway in vascular smooth muscle cells. Circ Res. 2006;98:811–7. 389. Teoh NC, Ajamieh H, Wong HJ, Croft K, Mori T, Allison AC, et al. Microparticles mediate hepatic ischemia-reperfusion injury and are the targets of Diannexin (ASP8597). PLoS One. 2014;9:e104376. 390. Ferrer E, Dunmore B, Hassan D, Ormiston M, Moore S, Deighton J, et al. A potential role for exosomal TCTP export in vascular remodeling in pulmonary arterial hypertension. Am J Respir Cell Mol Biol. 2018;Epub ahead. 391. Grygorczyk R, Orlov SN. Effects of hypoxia on erythrocyte membrane properties — Implications for intravascular hemolysis and purinergic control of blood flow. Front Physiol. 2017;8:1110. 392. Rubartelli A, Bajetto A, Allavena G, Wollman E, Sitia R. Secretion of thioredoxin by normal and neoplastic cells through a leaderless secretory pathway. J Biol Chem. 1992;267:24161–4. 393. Rammes A, Roth J, Goebeler M, Klempt M, Hartmann M, Sorg C. Myeloid-related protein (MRP) 8 and MRP14, calcium-binding proteins of the S100 family, are secreted by activated monocytes via a novel, tubulin- dependent pathway. J Biol Chem. 1997;272(14):9496–502. 394. Kido J, Kido R, Kataoka M, Mk F, Calprotectin NT. Calprotectin release from human neutrophils is induced by Porphyromonas gingivalis lipopolysaccharide via the CD-14 – Toll-like receptor – nuclear factor KB pathway. J Peridont Res. 2003;38:557–63. 395. Okubo K, Kosaka S, Isowa N, Hirata T, Hitomi S, Yodoi J, et al. Amelioration of ischemia-reperfusion injury by human thioredoxin in rabbit lung. J Thorac Cardiovasc Surg. 1997;113:1–9. 396. Nakamura H, Herzenberg LA, Bai J, Araya S, Kondo N, Nishinaka Y, et al. Circulating thioredoxin suppresses lipopolysaccharide- induced neutrophil chemotaxis. PNAS. 2001;98:15143–15148. 397. Hessian PA, Edgeworth J, Hogg N. MRP-8 and MRP-14, two abundant Ca(2+)- binding proteins of neutrophils and monocytes. J Leukoc Biol. 1993;53:197–204. 398. Gomes LH, Raftery MJ, Xing W, Goyette JD, Thomas PS, Geczy CL. S100A8 and S100A9 — oxidant scavengers in inflammation. Free Radic Biol Med. 2013;58:170– 86. 228 399. Isaksen B, Fagerhol MK. Calprotectin inhibits matrix metalloproteinases by sequestration of zinc. J Clin Pathol Mol Pathol. 2001;54:289–92. 400. Hiroshima Y, Hsu K, Tedla N, Wong SW, Chow S, Kawaguchi N, et al. S100A8/A9 and S100A9 reduce acute lung injury. Immunol Cell Biol. 2017;95:461–72. 401. Achouiti A, Vogl T, Meer AJ Van Der, Stroo I, Florquin S, de Boer OJ, et al. Myeloid- related protein-14 deficiency promotes inflammation in staphylococcal pneumonia. Eur Respir J. 2015;46:464–73. 402. Ryckman C, Vandal K, Rouleau P, Talbot M, Tessier PA. Proinflammatory activities of S100: Proteins S100A8, S100A9, and S100A8/A9 induce neutrophil chemotaxis and adhesion. J Immunol. 2003 Mar 15;170:3233–42. 403. Simard J-C, Girard D, Tessier PA. Induction of neutrophil degranulation by S100A9 via a MAPK-dependent mechanism. J Leukoc Biol. 2010 May;87:905–14. 404. Simard J-C, Noël C, Tessier PA, Girard D. Human S100A9 potentiates IL-8 production in response to GM-CSF or fMLP via activation of a different set of transcription factors in neutrophils. FEBS Lett. 2014 Jun 13;588:2141–6. 405. Vogl T, Tenbrock K, Ludwig S, Leukert N, Ehrhardt C, van Zoelen MAD, et al. Mrp8 and Mrp14 are endogenous activators of Toll-like receptor 4, promoting lethal, endotoxin-induced shock. Nat Med. 2007;13:1042–9. 406. Jin ZG, Lungu AO, Xie L, Wang M, Wong C, Berk BC. Cyclophilin A is a proinflammatory cytokine that activates endothelial cells. Arterioscler Thromb Vasc Biol. 2004;24:1186–91. 407. Xue C, Sowden M, Berk BC. Extracellular cyclophilin A, especially acetylated, causes pulmonary hypertension by stimulating endothelial apoptosis, redox stress, and inflammation. Arterioscler Thromb Vasc Biol. 2017;37:1138–46. 408. Verma S, Li SH, Wang C-H, Fedak PWM, Li R-K, Weisel RD, et al. Resistin promotes endothelial cell activation: Further evidence of adipokine-endothelial interaction. Circulation. 2003;108:736–40. 409. Macdonald SPJ, Bosio E, Neil C, Arendts G, Burrows S, Smart L, et al. Resistin and NGAL are associated with inflammatory response, endothelial activation and clinical outcomes in sepsis. Inflamm Res. 2017;66:611–9. 410. Solini A, Stea F, Santini E, Bruno RM, Duranti E, Taddei S, et al. Adipocytokine levels mark endothelial function in normotensive individuals. Cardiovasc Diabetol. 2012;11:103. 411. Eilenberg W, Stojkovic S, Kaider A, Kozakowski N, Domenig CM, Burghuber C, et al. NGAL and MMP-9/NGAL as biomarkers of plaque vulnerability and targets of statins in patients with carotid atherosclerosis. Clin Chem Lab Med. 2018;56:147–56. 412. Choi KM, Lee JS, Kim EJ, Baik SH, Seo HS, Choi DS, et al. Implication of lipocalin-2 and visfatin levels in patients with coronary heart disease. Eur J Endocrinol. 2008;158:203–7. 413. Naka KK, Papathanassiou K, Bechlioulis A, Pappas K, Tigas S, Makriyiannis D, et al. Association of vascular indices with novel circulating biomarkers as prognostic factors for cardiovascular complications in patients with type 2 diabetes mellitus. Clin Biochem. 2018;53:31–7. 229 414. Ramachandran S, Venugopal A, Raman Kutty V, Vinitha A, Divya G, Chitrasree V, et al. Plasma level of cyclophilin A is increased in patients with type 2 diabetes mellitus and suggests presence of vascular disease. Cardiovasc Diabetol. Cardiovascular Diabetology; 2014;13:38. 415. Stănescu D, Sanna A, Veriter C, Kostianev S, Calcagni PG, Fabbri LM, et al. Airways obstruction, chronic expectoration, and rapid decline of FEV1 in smokers are associated with increased levels of sputum neutrophils. Thorax. 1996;51:267–71. 416. Faner R, Tal-singer R, Riley JH, Celli B, Vestbo J, MacNee W, et al. Lessons from ECLIPSE: a review of COPD biomarkers. Thorax. 2014;69:666–72. 417. Blidberg K, Palmberg L, Dahlén B, Lantz AS, Larsson K. Chemokine release by neutrophils in chronic obstructive pulmonary disease. Innate Immun. 2011;18:503–10. 418. Baines KJ, Simpson JL, Gibson PG. Innate immune responses are increased in chronic obstructive pulmonary disease. PLoS One. 2011;6:e18426. 419. Stockley RA. Neutrophils and protease/antiprotease imbalance. Am J Respir Crit Care Med. 1999;160:S49–52. 420. Keatings VM, Barnes PJ. Granulocyte activation markers in induced sputum: Comparison between chronic obstructive pulmonary disease, asthma, and normal subjects. Am J Respir Crit Care Med. 1997;155:449–53. 421. Chen Y-C, Ho C-W, Tsai H-H, Wang J-S. Interval and continuous exercise regimens suppress neutrophil-derived microparticle formation and neutrophil-promoted thrombin generation under hypoxic stress. Clin Sci. 2015;128:425–36. 422. Martínez GJJ, Barraclough JYY, Nakhla S, Kienzle V, Robertson S, Mallat Z, et al. Neutrophil-derived microparticles are released into the coronary circulation following percutaneous coronary intervention in acute coronary syndrome patients. Biosci Rep. 2017;37:BSR20160430. 423. Chironi G, Simon A, Hugel B, Del Pino M, Gariepy J, Freyssinet JM, et al. Circulating leukocyte-derived microparticles predict subclinical atherosclerosis burden in asymptomatic subjects. Arterioscler Thromb Vasc Biol. 2006;26:2775–80. 424. Tan DBA, Armitage J, Teo T-H, Ong NE, Shin H, Moodley YP. Elevated levels of circulating exosome in COPD patients are associated with systemic inflammation. Respir Med. 2017;132:261–4. 425. Lok LSC, Farahi N, Juss JK, Loutsios C, Solanki CK, Peters AM, et al. Effects of tocilizumab on neutrophil function and kinetics. Eur J Clin Invest. 2017;47:736–45. 426. Grechowa I, Horke S, Wallrath A, Vahl C-F, Dorweiler B. Human neutrophil elastase induces endothelial cell apoptosis by activating the PERK-CHOP branch of the unfolded protein response. FASEB J. 2017;31:3868–81. 427. Kumor-Kisielewska A, Kierszniewska-Stepien D, Pietras T, Kroczynska-Bednarek J, Kurmanowska Z, Antczak A, et al. Assessment of leptin and resistin levels in patients with chronic obstructive pulmonary disease. Pol Arch Med Wewn. 2013;123:215–20. 428. Chemaly ER, Hadri L, Zhang S, Kim M, Kohlbrenner E, Sheng J, et al. Long-term in vivo resistin overexpression induces myocardial dysfunction and remodeling in rats. J Mol Cell Cardiol. 2011;51:144–55. 230 429. Weikert C, Westphal S, Berger K, Dierkes J, Möhlig M, Spranger J, et al. Plasma resistin levels and risk of myocardial infarction and ischemic stroke. J Clin Endocrinol Metab. 2008;93:2647–53. 430. Prugger C, Luc G, Haas B, Morange PE, Ferrieres J, Amouyel P, et al. Multiple biomarkers for the prediction of ischemic stroke: The PRIME study. Arterioscler Thromb Vasc Biol. 2013;33:659–66. 431. Muse E, Feldman DI, Blaha MJ, Dardari Z, Blumenthal RS, Budoff MJ, et al. The association of resistin with cardiovascular disease in the multi-ethnic study of atherosclerosis. Atherosclerosis. 2015;239:101–8. 432. Cockayne DA, Cheng DT, Waschki B, Sridhar S, Ravindran P, Hilton H, et al. Systemic biomarkers of neutrophilic inflammation, tissue injury and repair in COPD patients with differing levels of disease severity. PLoS One. 2012;7:e38629. 433. Bchir S, Nasr H ben, Bouchet S, Benzarti M, Garrouch A, Tabka Z, et al. Concomitant elevations of MMP-9, NGAL, proMMP-9/NGAL and neutrophil elastase in serum of smokers with chronic obstructive pulmonary disease. J Cell Mol Med. 2017;21:1280– 91. 434. Wang Y, Jia M, Yan X, Cao L, Barnes PJ, Adcock IM, et al. Increased neutrophil gelatinase-associated lipocalin (NGAL) promotes airway remodelling in chronic obstructive pulmonary disease. Clin Sci. 2017;131:1147–59. 435. Lindberg S, Pedersen SH, Mogelvang R, Jensen JS, Flyvbjerg A, Galatius S, et al. Prognostic utility of neutrophil gelatinase-associated lipocalin in predicting mortality and cardiovascular events in patients with ST-segment elevation myocardial infarction treated with primary percutaneous coronary intervention. J Am Coll Cardiol. 2012;60:339–45. 436. Hu R, Ouyang Q, Dai A, Tan S, Xiao Z, Tang C. Heat shock protein 27 and cyclophilin A associate with the pathogenesis of COPD. Respirology. 2011;16:983–93. 437. Zhang T, Zhang J, Ge H. Functions of cyclophilin A in atherosclerosis. Exp Clin Cardiol. 2013;18:e118–24. 438. Noguera A, Batle S, Miralles C, Iglesias J, Busquets X, Macnee W, et al. Enhanced neutrophil response in chronic obstructive pulmonary disease. Thorax. 2001;56:432– 7. 439. Almansa R, Socias L, Sanchez-Garcia M, Martín-Loeches I, Del Olmo M, Andaluz- Ojeda D, et al. Critical COPD respiratory illness is linked to increased transcriptomic activity of neutrophil proteases genes. BMC Res Notes. 2012;5:401. 440. Obeidat M, Nie Y, Chen V, Shannon CP, Andiappan AK, Lee B, et al. Network-based analysis reveals novel gene signatures in peripheral blood of patients with chronic obstructive pulmonary disease. Respir Res. 2017;18:72. 441. Titz B, Sewer A, Schneider T, Elamin A, Martin F, Dijon S, et al. Alterations in the sputum proteome and transcriptome in smokers and early-stage COPD subjects. J Proteomics. 2015;128:306–20. 442. Bjerregaard MD, Jurlander J, Klausen P, Borregaard N, Cowland JB. The in vivo profile of transcription factors during neutrophil differentiation in human bone marrow. Blood. 2003;101:4322–32. 231 443. Kawayama T, Kinoshita T, Matsunaga K, Kobayashi A, Hayamizu T, Johnson M, et al. Responsiveness of blood and sputum inflammatory cells in Japanese COPD patients, non-COPD smoking controls, and non-COPD nonsmoking controls. Int J COPD. 2016;11:295–303. 444. Shindo K, Koide K, Fukumura M. Enhancement of leukotriene B4 release in stimulated asthmatic neutrophils by platelet activating factor. Thorax. 1997;52:1024– 9. 445. Lacedonia D, Carpagnano GE, Trotta T, Palladino GP, Panaro MA, Zoppo LD, et al. Microparticles in sputum of COPD patients: a potential biomarker of the disease? Int J COPD. 2016;11:527–33. 446. Park HK, Ahima RS. Resistin in rodents and humans. Diabetes Metab J. 2013;37:404–14. 447. Mårtensson J, Bellomo R. The rise and fall of NGAL in acute kidney injury. Blood Purif. 2014;37:304–10. 448. Carter RI, Mumford R a, Treonze KM, Finke PE, Davies P, Si Q, et al. The fibrinogen cleavage product Aα-Val360, a specific marker of neutrophil elastase activity in vivo. Thorax. 2011 Aug;66:686–91. 449. Dorward DA, Lucas CD, Doherty MK, Chapman GB, Scholefield EJ, Conway Morris A, et al. Novel role for endogenous mitochondrial formylated peptide-driven formyl peptide receptor 1 signalling in acute respiratory distress syndrome. Thorax. 2017;72:928–36. 450. Hauser CJ, Sursal T, Rodriguez EK, Appleton PT, Zhang Q, Itagaki K. Mitochondrial DAMPs from femoral reamings activate neutrophils via formyl peptide receptors and P44/42 MAP Kinase. J Orthop Trauma. 2010;24:534–8. 451. Kato H, Kuriyama N, Duarte S, Clavien PA, Busuttil RW, Coito AJ. MMP-9 deficiency shelters endothelial PECAM-1 expression and enhances regeneration of steatotic livers after ischemia and reperfusion injury. J Hepatol. 2014;60:1032–9. 452. Sethi S. Molecular diagnosis of respiratory tract infection in acute exacerbations of chronic obstructive pulmonary disease. Clin Infect Dis. 2011;52 Suppl 4(Suppl 4):S290–5. 453. Navaratnam V, Millett ERC, Hurst JR, Thomas SL, Smeeth L, Hubbard RB, et al. Bronchiectasis and the risk of cardiovascular disease: a population-based study. Thorax. 2017;72:161–6. 454. Carden DL, Granger DN. Pathophysiology of ischaemia-reperfusion injury. J Pathol. 2000;190(3):255–66. 455. Chen L-W, Lin M-W, Hsu C-M. Different pathways leading to activation of extracellular signal-regulated kinase and p38 MAP kinase by formyl-methionyl-leucyl- phenylalanine or platelet activating factor in human neutrophils. J Biomed Sci. 2005;12:311–9. 456. Wright HL, Thomas HB, Moots RJ, Edwards SW. RNA-Seq reveals activation of both common and cytokine-specific pathways following neutrophil priming. PLoS One. 2013;8:e58598. 457. Al-Shami A, Naccache PH. Granulocyte-macrophage colony-stimulating factor- activated signaling pathways in human neutrophils. J Biol Chem. 1999;274:5333–8. 232 458. Corey S, Eguinoa A, Puyana-Theall K, Bolen JB, Cantley L, Mollinedo F, et al. Granulocyte macrophage-colony stimulating factor stimulates both association and activation of phosphoinositide 3OH-kinase and src-related tyrosine kinase(s) in human myeloid derived cells. Embo J. 1993;12:2681–90. 459. Bienes-Martínez R, Ordóñez A, Feijoo-Cuaresma M, Corral-Escariz M, Mateo G, Stenina O, et al. Autocrine stimulation of clear-cell renal carcinoma cell migration in hypoxia via HIF-independent suppression of thrombospondin-1. Sci Rep. 2012;2:788. 460. Xu Y, Loison F, Luo HR. Neutrophil spontaneous death is mediated by down- regulation of autocrine signaling through GPCR, PI3K , ROS, and actin. Proc Natl Acad Sci. 2010;107:2950–5. 461. Jog NR, Rane MJ, Lominadze G, Luerman GC, Ward RA, McLeish KR. The actin cytoskeleton regulates exocytosis of all neutrophil granule subsets. Am J Physiol Cell Physiol. 2007;292:C1690–700. 462. Jiang L, Chughtai K, Purvine SO, Bhujwalla ZM, Raman V, Paša-Tolić L, et al. MALDI-mass spectrometric imaging revealing hypoxia-driven lipids and proteins in a breast tumor model. Anal Chem. 2015;87:5947–56. 463. Paclet M-H, Davis C, Kotsonis P, Godovac-Zimmermann J, Segal AW, Dekker L V. N- Formyl peptide receptor subtypes in human neutrophils activate L-plastin phosphorylation through different signal transduction intermediates. Biochem J. 2004;377:469–77. 464. Hirota K, Semenza GL. Rac1 activity is required for the activation of hypoxia-inducible factor 1. J Biol Chem. 2001;276:21166–72. 465. Yang MQ, Du Q, Goswami J, Varley PR, Chen B, Wang RH, et al. Interferon regulatory factor 1–Rab27a regulated extracellular vesicles promote liver ischemia/reperfusion injury. Hepatology. 2018;67:1056–70. 466. Yamaoka M, Ando T, Terabayashi T, Okamoto M, Takei M, Nishioka T, et al. PI3K regulates endocytosis after insulin secretion by mediating signaling crosstalk between Arf6 and Rab27a. J Cell Sci. 2016;129:637–49. 467. Yan B, Wei J-J, Yuan Y, Sun R, Li D, Luo J, et al. IL-6 cooperates with G-CSF to induce protumor function of neutrophils in bone marrow by enhancing STAT3 activation. J Immunol. 2013;190:5882–93. 468. van der Linden M, Westerlaken GHA, van der Vlist M, van Montfrans J, Meyaard L. Differential signalling and kinetics of neutrophil extracellular trap release revealed by quantitative live imaging. Sci Rep. Springer US; 2017;7:6529. 469. Kimura H, Matsuyama Y, Araki S, Koizumi A, Kariya Y, Takasuga S, et al. The effect and possible clinical efficacy of in vivo inhibition of neutrophil extracellular traps by blockade of PI3K-gamma on the pathogenesis of microscopic polyangiitis. Mod Rheumatol. 2018;28:530–41. 470. Owen CA, Campbell MA, Sannes PL, Boukedes SS, Campbell EJ. Cell surface-bound elastase and cathepsin G on human neutrophils: A novel, non-oxidative mechanism by which neutrophils focus and preserve catalytic activity of serine proteinases. J Cell Biol. 1995;131:775–89. 233 471. Porter LM, Cowburn AS, Farahi N, Deighton J, Farrow SN, Fiddler CA, et al. Hypoxia causes IL-8 secretion, Charcot Leyden crystal formation, and suppression of corticosteroid-induced apoptosis in human eosinophils. Clin Exp Allergy. 2016;47:770–84. 472. Björnsdottir H, Welin A, Dahlgren C, Karlsson A, Bylund J. Quantification of heterotypic granule fusion in human neutrophils by imaging flow cytometry. Data Br. 2016;6:386–93. 473. Rice WG, Ganz T, Kinkade JM, Selsted ME, Lehrer RI, Parmley RT. Defensin-rich dense granules of human neutrophils. Blood. 1987;70:757–65. 474. Jung J, Bohn G, Allroth A, Boztug K, Brandes G, Sandrock I, et al. Identification of a homozygous deletion in the AP3B1 gene causing Hermansky-Pudlak syndrome, type 2. Blood. 2006;108:362–9. 475. Niemann CU, Abrink M, Pejler G, Fischer RL, Christensen EI, Knight SD, et al. Neutrophil elastase depends on serglycin proteoglycan for localization in granules. Blood. 2007;109:4478–86. 476. Vogelmeier C, Aquino T, O’Brien C, Perrett J, Gunawardena K. A randomised, placebo-controlled, dose-finding study of AZD9668, an oral inhibitor of neutrophil elastase, in patients with chronic obstructive pulmonary disease treated with tiotropium. COPD. 2012;9:111–20. 477. Kuna P, Jenkins M, O’Brien CD, Fahy WA. AZD9668, a neutrophil elastase inhibitor, plus ongoing budesonide/ formoterol in patients with COPD. Respir Med. 2012;106:531–9. 478. von Nussbaum F, Li VMJ. Neutrophil elastase inhibitors for the treatment of (cardio)pulmonary diseases: Into clinical testing with pre-adaptive pharmacophores. Bioorganic Med Chem Lett. 2015;25:4370–81. 479. Nymo SH, Hartford M, Ueland T, Yndestad A, Lorentzen E, Truvé K, et al. Serum neutrophil gelatinase-associated lipocalin (NGAL) concentration is independently associated with mortality in patients with acute coronary syndrome. Int J Cardiol. 2018;262:79–84. 480. Fontana A, Spadaro S, Copetti M, Spoto B, Salvemini L, Pizzini P, et al. Association between resistin levels and all-cause and cardiovascular mortality: A new study and a systematic review and meta-analysis. PLoS One. 2015;10:e0120419. 481. Saleh AD, Kwok B, Brown JS, Hurst JR. Correlates and assessment of excess cardiovascular risk in bronchiectasis. Eur Respir J. 2017;50:1701127. 482. Yanagisawa S, Baker JR, Vuppusetty C, Fenwick P, Donnelly LE, Ito K, et al. Decreased phosphatase PTEN amplifies PI3K signaling and enhances pro- inflammatory cytokine release in COPD. Am J Physiol Lung Cell Mol Physiol. 2017;313:L230–9. 483. Yu J, Liu R, Huang J, Wang L, Wang W. Inhibition of Phosphatidylinositol 3-kinease suppresses formation and progression of experimental abdominal aortic aneurysms. Sci Rep. 2017;7:15208. 484. Takeda M, Ito W, Tanabe M, Ueki S, Kato H, Kihara J, et al. Allergic airway hyperresponsiveness, inflammation, and remodeling do not develop in phosphoinositide 3-kinase γ-deficient mice. J Allergy Clin Immunol. 2009;123:805–12. 234 485. Galluzzo M, Ciraolo E, Lucattelli M, Hoxha E, Ulrich M, Campa CC, et al. Genetic deletion and pharmacological inhibition of PI3Kγ reduces neutrophilic airway inflammation and lung damage in mice with cystic fibrosis-like lung disease. Mediators Inflamm. 2015;2015:545417. 486. Green CE, Turner AM. The role of the endothelium in asthma and chronic obstructive pulmonary disease (COPD). Respir Res. Respiratory Research; 2017;18:20. 487. Amsellem V, Gary-Bobo G, Marcos E, Maitre B, Chaar V, Validire P, et al. Telomere dysfunction causes sustained inflammation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med. 2011;184:1358–66. 488. Page A V, Liles WC. Biomarkers of endothelial activation/dysfunction in infectious diseases. Virulence. 2013;4:507–16. 489. Stockley RA, Turner AM. α-1-Antitrypsin deficiency: Clinical variability, assessment, and treatment. Trends Mol Med. 2014;20:105–15. 490. Valdez S, Patterson S, Ezquer M, Torrecilla M, Lama MC, Seltzer A. Acute sublethal global hypoxia induces transient increase of GAP-43 immunoreactivity in the striatum of neonatal rats. Synapse. 2007;61:124–37.