Glia crosstalk in modulating synaptic dysfunctions in Alzheimer's disease (AD).

Region-specific excitatory synapse loss correlates strongly to Alzheimer's disease (AD) cognitive decline (1,2). In mouse models of amyloidosis, excitatory synapses externalize phosphatidylserine, which microglia recognize as an eat-me signal via specific receptors, mainly TREM2 (3-7). Another mechanism underlying microglia engulfment requires cooperation with astrocytes.

Astrocytes are the main responders to synaptic information in tripartite synapses and develop an intimate interaction with microglia upon injury (8,9,6). Although astrocytes engulf synapses during development and express immune clearance-related genes during injury, interestingly, in models of amyloidosis, we do not see astrocytes engulfing excitatory synapses (Hong lab unpublished data).

Instead, we see astrocytes in brain regions where synapses are affected secreting a molecule called MFGE8, which has been shown to act as a bifunctional linker of phagocytes and externalized phosphatidylserine (Hong lab unpublished data,6). Blocking astrocytic MFGE8 in mouse models of amyloidosis using CRISPR-Cas9 technology prevents microglia-synapse engulfment and ameliorates synapse loss in these mice.

Altogether, these data suggest glia-immune-synapse crosstalk that determine the specificity of synapse elimination in early stages of AD, at least in context of excitatory synapses.

Few studies have investigated the inhibitory synapse loss in AD. In mouse models of amyloidosis show decreased amplitude/frequency of inhibitory post-synaptic currents on principal cells at pre-plaque stages and interneurons loss with Nav1.1 defects in parvalbumin cells in later stages (10,11,12,13). Hong lab showed a loss of post-synaptic gephyrin in amyloidosis mice hippocampi, suggesting that interneurons are important effectors of early changes (Hong lab unpublished data).

Interestingly, pilot studies suggest that TREM2-mediated synapse elimination by microglia in hippocampus of amyloid mouse models may affect only excitatory but not inhibitory synapses (Hong lab unpublished data,3,7). This raises an intriguing question of what mediates the loss of inhibitory synapse loss, and the roles of astrocytes and microglia in this process.

Overall, this can introduce a new paradigm in glial-synapse crosstalk, where there may be distinct mechanisms to how inhibitory synapses are lost in the AD brain versus excitatory synapses. First, I will elucidate whether various 'eat-me' signals are expressed differentially on excitatory vs. inhibitory synapses, for e.g., if phosphatidylserine is externalized on excitatory synapses but not on inhibitory synapses.

During development, specific GABA-receptive microglia selectively engulf inhibitory synapses around the CA1 pyramidal cells soma - it would be interesting to test this in AD (14). Another option is the contribution of microglia and synaptic non-intrinsic different signals from different hubs. Astrocytes' expression of specific signals may determine the microglia's engulfment specificity.

A final route might involve astrocytes' direct engulfment. Astrocytes can control inhibitory synapse formation via secreted proteins (15-17). In amyloidosis mice, a new bulbous astrocyte with increased p62 level was found (Hong's lab unpublished data). p62 indicate autophagic degradative activity (18).

This astrocyte type may release specific signals affecting the selective microglia's engulfment, or they may be responsible for engulfment.

Besides the above suggestions, I will characterize the bulbous astrocytes to understand if p62 has an upstream role in pathogenesis, controlling astrocyte morphogenesis and affecting synapse loss. I will assess the effects of p62 loss in primary astrocyte and neuron cultures to then investigate in vivo the temporal relationship between synapse degeneration and p62+ astrocytes' appearance (19).

Moreover, to understand the glia crosstalk effect on synapsis engulfment specificity, the in vivo chemico-genetic approach Split-TurboID tha