TREM2 as an evolving therapeutic target in Alzheimer's disease.

Abstract

Triggering receptor expressed on myeloid cells 2 (TREM2) is type I transmembrane protein receptor found in osteoclasts, dendritic cells, macrophages, and microglia (Ulland and Colonna, 2018). In the central nervous system, TREM2 expression is confined exclusively to microglia. TREM2 comprises an extracellular V-type immunoglobulin (Ig) ectodomain, a connecting stalk followed by a transmembrane region and a short C-terminal tail. TREM2 signaling involves its association with the adaptor proteins DAP10 and DAP12 with the recruitment of PI3K and SYK, respectively that drive multiple downstream events resulting in Ca2+ mobilization and activation of mitogen-activated protein kinase-mediated cascades as well as other pathways. The downstream signals result in the modulation of key functional properties of the macrophage/microglia including proliferation, phagocytosis, survival, actin cytoskeleton remodeling, and cell metabolism (Ulland and Colonna, 2018; Deczkowska et al., 2020; Box 1).Box 1: Effects of microglial triggering receptor expressed on myeloid cells 2 (TREM2) activation in the central nervous systemSoluble TREM2 (sTREM2) can be produced from both proteolytic cleavage of the extracellular domain by ADAM 10/17 and/or alternative splicing (Filipello et al., 2022). sTREM2 has separate and (receptor) independent activities that mirror and at the same time promote inflammation thus establishing it as an interesting biomarker but also as a potential pathogenetic factor in neurodegeneration (Filipello et al., 2022). This is evident by studies showing that soluble TREM2 drives neuroinflammation by direct intracerebral injection but also inhibits beta amyloid oligomerization. Several studies have shown that cerebrospinal fluid sTREM2 levels rise in the initial symptomatic stages of Alzheimer’s disease (AD) thus attesting for the active neuroinflammatory process present at this time (soluble TREM2 shedding from proliferating, active microglia prior to the burned-out state). It also may provide a window for therapeutic intervention based on the assumption that a more robust expression of TREM2 in microglia within the central nervous system is apparent in the early stages of the disease. Interest in TREM2 in the context of neurodegenerative diseases sparked in 2013 following initial studies showing that variants of TREM2 are associated with increased prevalence of AD (Guerreiro et al., 2013), but also with several other neurodegenerative diseases. This observation drove huge interest in this hemostatic protein, as it stands as it supports the role of innate immunity as an important factor influencing the pathogenesis of neurodegeneration. This theme somehow challenged the previously popular amyloid cascade hypothesis as the sole driver of AD, highlighting the pathogenetic role of beta amyloid and Tau. The role of TREM2 in AD has been complex to decipher, as modeling of the disease in animals suffers several caveats including differential timing of amyloid deposition, lack of appropriate mirroring of the sequence of pathology in humans where amyloid plaque accumulation precedes Tau pathology and cognitive decline (Ulland and Colonna, 2018; Deczkowska et al., 2020). Moreover, TREM2 deletion by gene manipulation does not necessarily result in an opposing effect produced by TREM2 agonism. What can be concluded at present with regard to the role of TREM2 in experimental amyloid-based models is that the protein has an important role in plaque entrainment by microglia, where it appears to be overexpressed and drive transit of these innate immune cells to a disease-associated microglial (DAM) state. This transition is marked by microglial priming that is evident by a set of overexpressed genes and an improved phagocytic capacity. The impact of TREM2, however, appears to be more complicated in Tau-based AD models where inconsistent results have been reported although different models and readouts have been used. The strong genetic-based data and corresponding animal data led to an interest in the concept of agonizing TREM2 on microglia in an attempt to functionalize/activate microglia with the hope of achieving a more efficient attenuation of amyloid and Tau pathologies. In this sense, monoclonal antibodies (mAb) have a potential advantage over small molecules as they can engage specific epitopes within TREM2 without penetrating the cells thus avoiding unpredictable side effects. One attractive method of agonizing TREM2 was to develop a mAb that binds TREM2 and blocks its cleavage and subsequent shedding in the stalk region, thus stabilizing the protein on the membrane and allowing constant activation (Schlepckow et al., 2020). Indeed, it has been reported that this strategy led to reduced shedding, enhanced microglial capacity to engulf beta amyloid and neuronal debris and high doses of the mAb led to a reduction in amyloid deposition over a short period in an experimental model and using very high doses of the agent (Schlepckow et al., 2020). No data were reported with regard to the effects of this antibody on Tau burden in models of Tauopathies. This strategy of blocking cleavage site is co-developed by Denali/Takeda and is reported to be in initial clinical trials where the antibody is combined with a cargo using the transferrin receptor to facilitate blood-brain barrier crossing. An additional approach is to engage TREM2 at the extracellular domain driving activation due to the potential sequestration of the protein on the cellular membrane. In experimental studies using the murine and human surrogates of AL002, an extracellular domain targeting mAb currently developed by Alector, beneficial effects on plaque size and composition were demonstrated (Price et al., 2020; Wang et al., 2020). These findings were corroborated by in vitro studies showing that AL002 activates TREM2 in microglia and resulted in an increase in their proliferative state. Importantly, the phase I study clearly shows that the antibody achieves target engagement evident by a dose-dependent saturation of soluble TREM2 levels. Yet, a very recent study that tested this antibody in the 5xFAD amyloid-based model where human Tau is delivered intracerebrally, when amyloid pathology is already present, showed exacerbation of neuritic plaques and Tau pathology (Jain et al., 2023). This finding further compounds the interpretation of the preclinical studies that show opposing effects of TREM2 on Tau pathology. This finding of increased tau pathology with TREM2 agonism is of particular importance as AL002 is currently tested in a phase II study in patients with AD. We recently reported the results on our anti-TREM2 mAb, CGX101 in models of amyloidopathies (Fassler et al., 2021). Several principles were implemented in the design of the antibody: (1) We targeted the extracellular domain of TREM2 that would also engage the soluble shed form so as to achieve effective target engagement and neutralize the proinflammatory effects of sTREM2. This has the advantage of constraining long-term chronic neuroinflammation due to the activation of TREM2 by the agonizing antibody. (2) The antibody was developed to recognize the human and murine TREM2 with significant affinity so as to improve the potential translatability of the preclinical data to humans without having to develop two different murine and human antibodies. (3) Unlike the current mAbs in clinical trials, the isotype of CGX101 is IgG4 rather than IgG1 with the idea being to attenuate non-specific proinflammatory effects that could be associated with amyloid-related imaging abnormalities evidenced in clinical trials with anti-amyloid mAbs. We have tested CGX101 in a comprehensive set of in vivo, ex vivo, and in vitro studies. The antibody was shown efficacious in reducing amyloid burden in young and aged 5xFAD mice upon chronic treatment and when delivered acutely into the brains. Importantly, CGX101 achieved good blood-brain barrier penetration, improved activated microglial coverage of the plaques, reduced cognitive decline, and achieved effective target engagement in the brain. On a chronic basis, treatment was associated with reduced neuroinflammatory markers along with an apparent shift towards a DAM signature and attenuated neurodegeneration evident by a reduction of dystrophic neurites and preservation of synaptic markers (unpublished). There is a major question as to the established effect of mAbs on plaque removal and its association with improvement in the decline of cognitive function in clinical settings. This apparent disconnect has been highlighted with regard to the beta amyloid antibody Adacunumab that has been approved by the FDA. Our preclinical models with CGX101 clearly suggest that plaque removal is associated with the preservation of cognitive functions that may be related to the suppression of chronic neuroinflammation (Fassler et al., 2021). We have also tested CGX101 separately, in a Tau-based genetic model where human mutations are introduced to the mice (unpublished). Chronic treatment with CGX101 was associated with a significant reduction in phopho-Tau burden, improved cognitive performance, and attenuated markers of neurodegeneration. In the in vitro studies, microglial priming with CGX101 led to amelioration of p-Tau-induced seeding of neurons and this effect was corroborated by ex vivo and in vivo studies showing that treatment with the antibody reduced spreading and propagation of Tau pathology. Importantly and similar to the results in amyloid-based models, the effects on Tau pathology were corroborated by the preservation of cognitive functions in the animals. These findings are in apparent contrast with a very recent study with AL002 showing that chronic treatment was associated with exacerbation of Tau pathology in a preclinical model where human Tau was delivered to mice with preexisting beta amyloid pathology. The apparently contrasting data from our CGX101 and AL002 may result from the use of different models but can also mirror the different epitope recognition by the different antibodies and also the unique attribute of CGX101 to attenuate chronic neuroinflammation as outlined above. The potential interest in TREM2 in neurodegenerative diseases also extends to the rare disease ALSP, a rapidly progressive, fatal autosomal dominant disease caused by CSF1R gene mutations that results in microglial dysfunction (Ferrer, 2022). Vigil Neuroscience licensed a fully human mAb to TREM2 from Amgen and is advancing it to phase II in this rare disease further exemplifying the growing interest in targeting this innate immune receptor in neurodegenerative diseases. Initial phase I data suggest that the antibody achieved good target engagement evident by a dose-dependent reduction in cerebrospinal fluid sTREM2 levels. In conclusion, TREM2 raises growing interest as a potential dominant player in shaping the inflammatory response in the central nervous system, which could influence the initiation and progression of neurodegenerative diseases. The most advanced therapeutic approach to agonizing this microglial receptor involves the use of mAbs due to their ability to confer specific targeting of TREM2 devoid of unwanted activation of internal signaling pathways typically conferred by small molecules. However, gaps in understanding of the exact role of TREM2 as well as important differences in epitope recognition between the different studied mAbs are still to be reconciled and a class effect of the different agents does not seem probable. Regardless of the ambiguities, TREM2 engagement and targeting has the advantage of acting irrespective of the triggering aggregated protein, whether being beta amyloid, Tau, alpha synuclein, or any other incriminated misfolded trigger. The first proof of concept studies to inform us as to the clinical utility of targeting TREM2 will be available in the next couple of years. C-Editors: Zhao M, Liu WJ, Qiu Y; T-Editor: Jia Y