Microglia are macrophages resident in the central nervous system (CNS). Microglia contribute to CNS development, homeostasis and pathology, and microglial dysfunction has been associated with several neurodegenerative diseases. CD33 and triggering receptor expressed on myeloid cells 2 (TREM2) are microglial receptors, variants of which affect Alzheimer’s disease (AD) risk, and therefore are potential targets for therapeutic intervention in AD.

In this work, I show a conserved inhibitory role of CD33, a sialic acid-binding immunoglobulin-type lectin (Siglec) in both mice and human microglial cells. Knock down of CD33 in a mouse microglial cell line, BV-2, resulted in increased phagocytosis of several targets and increased migration towards ATP.

Human CD33 has two distinct variants: CD33M, associated with increased risk of developing AD, and CD33m, associated with decreased risk of developing AD by genome wide association studies (GWAS). However, how these variants confer differential risk of AD was unclear. By overexpression of the receptors in CHME-3 cells, I show that CD33M inhibits phagocytosis, migration and proliferation, but increases adhesion to components of the extracellular matrix when highly overexpressed, similar to mouse CD33. In contrast, CD33m, increases phagocytosis, migration and proliferation but decreases adhesion to components of the extracellular matrix. This suggests that CD33m has gained (rather than lost) function relative to CD33M, and that CD33m may be protective by providing efficient removal of toxic substrates such as neuronal debris.

CD33 may oppose the signalling function of the phagocytic receptor TREM2. Humans with rare TREM2 variants (particularly R47H) have increased risk of developing AD. In this work I show a potential novel gain of function role of the R47H variant of TREM2, which results in increased loss of neurons, inflammation and phagocytosis of isolated synapses. This gain of function could explain why individuals heterozygous for R47H have a several-fold increased risk of developing AD. Interestingly, I found that the presence of wild-type TREM2 inhibited phagocytosis of phosphatidylserine-coated beads, isolated synapses, neurons and neuronal debris. This inhibition could be mediated by TREM2-dependent expression of cystatin F, a lysosomal protein upregulated during neurodegeneration. Such inhibition of phagocytosis could be detrimental in early AD by reducing clearance of Aβ plaques, but beneficial in later stages of AD by reducing phagocytosis of synapses and neurons. Although, it should be taken into account these findings need to be confirmed in other more physiological cell types or systems to ensure this effect is not artefactual.

Finally, I found that CD33M inhibits TREM2 mediated activation of the tyrosine kinase, SYK, whereas CD33m increases SYK activation, potentially contributing to TREM2’s protective effects and showing a direct link between CD33 and TREM2 signalling.

Together, this work elucidates key roles of CD33, TREM2 and their variants in microglial function and dysfunction, which helps to clarify their potential roles in neurodegenerative disease.