Investigating the control of epithelial polarisation in human kidney organoids

Abstract

Polarised epithelial tubes make up the bulk of the functional domains of adult organs, and failures in either the development or maintenance of these structures is the basis of many adult pathologies, such as polycystic kidney disease. One of the key mechanisms of tube formation during development is so-called de novo acquisition of polarity, where initially unpolarised cells polarise and orient themselves with respect to each other, swapping out a mesenchymal front-back polarity for an epithelial apical-basal polarity, concomitant with the establishment of intercellular junctions and adherence to an underlying basement membrane. Currently, this process is poorly understood in many in vivo situations. During mammalian kidney development, intermediate mesoderm-derived nephron progenitor cells of unpolarised mesenchymal character undergo a mesenchymal-to-epithelial transition (MET) to yield the renal vesicle (RV), the first epithelial structure from which the main tubular component of the kidney arises. Despite extensive research into the molecular mechanisms controlling several aspects of kidney development over several decades, this transition remains poorly understood. WNT signalling from an adjacent epithelial tissue, the ureteric bud, is required to induce condensation of nephron progenitors into a pre-tubular aggregate (PTA) that subsequently undergoes MET, but the behaviours of individual cells within the aggregate leading to the acquisition of apical-basal polarity has not been addressed, although it is known that failures at this stage can lead to a suite of congenital abnormalities that may present as renal dysplasia, hypoplasia or, in extreme cases, complete renal agenesis. In this thesis, I have investigated the control of the PTA-RV transition as it relates to human kidney development, using human kidney organoids as a model. Grown from induced pluripotent stem cells, intermediate mesoderm-like cells are aggregated into spheres of 1000 cells, and over the course of four days spontaneously polarise and organise into tube structures in the absence of external growth factors or supplied matrix. In chapter 1, I characterised the process of MET using time-staged immunofluorescence imaging and reanalysed an existing RNA-sequencing dataset to show that kidney organoids faithfully and accurately modelled the PTA-RV transition. In chapter 2, drawing on published examples, as well as principles of developmental self-organisation, I developed a theoretical framework to determine the symmetry breaking cue that leads to the acquisition of epithelial polarity in kidney organoids, identifying several potential cell-cell and cell-extracellular matrix (ECM) interactions from a differential gene expression analysis. In particular, I showed that cadherins, specifically cadherin-6 (CDH6), were expressed early during the development of kidney organoids, and prefigured the site of initiation of the apical domain. I also analysed the distribution of the cell surface receptor family integrins and their ECM ligands, including laminins, another known polarising interaction, and further characterised the ECM niche within which nephron progenitors and their surrounding stroma develop. Finally, in chapter 3, I used a CRISPR/Cas9-based genetic intervention system to reversibly knock down the candidates identified in chapter 2 specifically across this window of MET, and also adapted the organoid culture system to observe isolated single cell adhesion events in 2D, concluding that, contrary to prevalent hypotheses regarding de novo acquisition of polarity during kidney development, cell-cell interactions appear to dominate over cell-ECM interactions. Kidney disease affects an estimated 10% of people worldwide, and congenital abnormalities of the kidney and urinary tract are some of the most common of all congenital defects observed clinically. Nephron number at birth is a strong predictor of the likelihood of developing kidney disease in later life and the requirement for a transplant, which are, at present, severely limiting. Yet still, accurate models of human kidney development are lacking. Stem cell-derived organoids may present a potential solution to this issue. Understanding how to more precisely direct nephrogenesis in vitro may help to overcome the reproducibility bottleneck experienced by all organoid fields, and possibly guide novel regenerative therapies in the future.