Investigating direct vs. indirect effects of Amyloid-β on neurons and astroglia in human and mouse tissue

Abstract

Alzheimer’s Disease (AD) has, since its original characterisation over a century ago, been defined by the presence of two neuropathological hallmarks in the brain— extracellular plaques comprised of Amyloid-β (Aβ) and intracellular hyperphosphorylated tau protein (p-Tau). Of these hallmarks, Aβ has, for many decades, adopted a central role in the study of AD pathogenesis and progression; captured by Hardy and Higgins’ Amyloid Cascade Hypothesis (1992). This hypothesis, in its original form, posited that the deposition of Aβ in the brain represented a key upstream mediator of disease onset and progression—leading to the supposition in the field that, by therapeutically clearing Aβ, AD could be effectively halted in its tracks.

However, today it is well-understood that the degree of plaque deposition in the brain is in fact a poor correlate of cognitive symptoms observed in the clinic. Furthermore, a historic failure across human clinical trials—until recently—of Aβ-targeting immunotherapies had led to a shift away from the prevailing view of Aβ as a key upstream agent in AD. Despite this, a case can indeed be made to propose that Aβ does play an important role in disease pathogenesis and progression. While insoluble plaques themselves do not correlate with cognitive decline, the insoluble pool of Aβ exhibits a closer relationship. This soluble pool contains a heterogeneous population of Aβ oligomers that have, across several studies, been demonstrated to be directly cyto- and synaptotoxic in vitro. Additionally, Aβ oligomers can be observed in vivo to accumulate within synapses of AD postmortem brain—thus indirectly linking them to the pathologically-relevant process of synapse loss, which in turn, is known to be the strongest correlate of cognitive decline seen in patients. Furthermore, with a modest slowing of progressive cognitive decline reported from clinical trials involving recent passive immunotherapies, lecanemab and donanemab, it would appear that a targeting of particular soluble Aβ species may in fact represent an effective therapeutic strategy. Despite this, the clearance of Aβ alone across these trials exhibited only a modest clinical benefit and failed to fully resolve cognitive decline. Overall, this therefore emphasises the need to better understand other factors in the brain that may interact with Aβ to influence AD pathogenesis and progression.

Reactive gliosis also represents a cardinal feature of AD, with the accumulation of both activated microglia and reactive astrocytes around extracellular plaques being documented as far back as the original work of Alzheimer. Of these, microglia are today understood to congregate in the vicinity of plaques, wherein they adopt distinct transcriptomic signatures and interact directly with Aβ deposits. Furthermore, human Genome-Wide Association Study (GWAS) has directly implicated microglia-specific TREM2 as a prominent risk factor for the development of AD, in addition to a number of other microglia-specific genes. Activated microglia are hypothesised to ingest extracellular Aβ material via numerous mechanisms and, in the literature, have been touted as being both protective and potentially harmful in disease. Microglia may act to clear toxic Aβ oligomers from the brain to prevent neurodegeneration and synapse loss; however, evidence from emerging in vivo microglial knockout models has suggested that, in mice, they may alternatively act to exacerbate Aβ pathology. Meanwhile, reactive astrocytes, which also accumulate spatially around extracellular plaques, remain less well understood but also appear to adopt unique transcriptomic profiles and have been associated with both trophic support of vulnerable neurons, in addition to potential neurodegenerative effects.

With newly-developed in vivo mouse models allowing for full knockout of microglia from the brain without early lethality or major health deficits, we investigated whether microglia could act as an indirect mediator for Aβ pathology: posing the question as to what extent Aβ may influence disease-relevant phenotypes directly vs. indirectly through microglia. Employing the novel Csf1rΔFIRE/ΔFIRE (“FIRE”) mouse, which completely lacks brain-resident microglia, we generated a cross with APPswe/PS1dE9 (“APP/PS1”) mice that exhibit Aβ pathology with no microglial influence. With this cross (“APP/PS1 x FIRE”), we investigated the effect of microglial absence on disease-relevant phenotypes in the neocortex, including plaque load, plaque- associated reactive astrogliosis and plaque-associated synapse loss. The findings of this thesis demonstrate that, in the absence of microglia, both cortical plaque load and plaque-associated reactive astrogliosis are significantly reduced; suggesting that microglia may play a role in exacerbating both plaque deposition and the downstream astrocytic response to Aβ. Regarding, plaque-associated excitatory synapse loss, no significant effect of microglia was observed in the cortex; implying that, contrary to expectations from existing literature, microglia do not appear to play a prominent role regarding synaptopathy in response to Aβ. Furthermore, by also investigating whether plaque-associated reactive astrocytes were seen to internalise synapses in the plaque microenvironment, little- to no internalisation could be seen. Taken together, this thesis therefore proposes that Aβ more likely acts as a direct driver of synapse loss in the cortex. Finally, using postmortem human AD tissue, we investigated whether therapeutic antibody lecanemab may bind to synaptic Aβ oligomers, thus preventing these putative direct toxic effects of Aβ and rescuing synapse loss; potentially providing a biological mechanism by which lecanemab may slow cognitive decline in human patients. This thesis reports that, at the level of light microscopy, colocalisation of lecanemab can be seen at surviving excitatory synapses found in close proximity to dense core extracellular plaques—a region known to exhibit profound synapse loss. Preliminary immunogold electron microscopy suggests that this colocalisation may represent a physical binding of lecanemab to Aβ oligomers found inside synapses. Altogether, by identifying specific processes by which Aβ pathology may be modified either directly or indirectly, this thesis provides evidence going towards a further understanding of how said processes, including plaque deposition, reactive astrogliosis and synapse loss could more effectively be therapeutically targeted going into the future.