TREM2-DAP12 Signalosome Enhancement — Boosting PI3K-AKT-mTOR Axis for Microglial Metabolic Fitness

Core Hypothesis and Rationale

The central hypothesis posits that selective enhancement of the TREM2-DAP12 signalosome—specifically by augmenting downstream PI3K-AKT-mTOR axis signaling—will restore and sustain the metabolic fitness required for disease-associated microglia (DAM) to execute their neuroprotective amyloid surveillance functions during the early-to-mid stages of Alzheimer's disease pathology.

Core Hypothesis and Rationale

The central hypothesis posits that selective enhancement of the TREM2-DAP12 signalosome—specifically by augmenting downstream PI3K-AKT-mTOR axis signaling—will restore and sustain the metabolic fitness required for disease-associated microglia (DAM) to execute their neuroprotective amyloid surveillance functions during the early-to-mid stages of Alzheimer's disease pathology. This hypothesis explicitly sides with the agonist camp of the TREM2 debate, but with a critical mechanistic refinement: rather than simply increasing TREM2 ligand engagement at the ectodomain level, the proposed intervention targets the intracellular signal transduction machinery that couples DAP12 (encoded by TYROBP) immunoreceptor tyrosine-based activation motif (ITAM) phosphorylation to anabolic metabolic reprogramming via SYK kinase and the PI3K p110δ isoform.

The novelty lies in the observation that in late-stage or chronically stimulated microglia, TREM2 ectodomain shedding by ADAM10/ADAM17 produces soluble TREM2 (sTREM2) that decouples ligand sensing from intracellular signaling, creating a state in which surface TREM2 engagement is partially preserved but downstream mTORC1-driven anabolism collapses. This metabolic collapse—manifesting as impaired mitochondrial oxidative phosphorylation, reduced cholesterol processing capacity, and failure to sustain lysosomal biogenesis—is proposed to represent the primary reason that DAM transition from a Stage 1 homeostatic-exit state to a dysfunctional, inflammatory Stage 2 state incapable of effective plaque compaction. Critically, this hypothesis reframes the antagonism argument: those observations favoring TREM2 suppression in late-stage disease likely reflect attempts to dampen an already metabolically failed, inflammatory DAM phenotype, rather than addressing the upstream metabolic deficiency that caused that failure. Boosting the PI3K-AKT-mTOR axis specifically during the window when DAM retain mitochondrial competence would prevent this collapse entirely.

Mechanistic Evidence

The molecular cascade begins at the DAP12 ITAM, which upon TREM2 clustering by phosphatidylserine, sulfatide, or APOE-lipid complexes undergoes dual tyrosine phosphorylation (pY62 and pY72 in the canonical human TYROBP sequence), creating docking sites for the tandem SH2 domains of SYK. SYK autophosphorylation at Y352 and Y525/526 then recruits the p85α regulatory subunit of PI3Kδ, generating PIP3 at the inner leaflet of the plasma membrane. PIP3 recruits AKT via its pleckstrin homology domain; subsequent phosphorylation at T308 by PDK1 and S473 by mTORC2 produces fully active AKT. Active AKT phosphorylates TSC2 (inhibiting the TSC1/TSC2 complex), thereby disinhibiting Rheb GTPase and activating mTORC1. mTORC1 then drives S6K1 phosphorylation and 4E-BP1 inactivation, coordinating ribosome biogenesis, lysosomal expansion via TFEB nuclear translocation suppression/activation dynamics, and mitochondrial anabolism through PGC-1α-independent mechanisms involving SREBP1 for lipid synthesis critical to myelin debris processing.

Supporting evidence comes from multiple converging lines. First, Trem2-knockout mice in 5xFAD and APP/PS1 backgrounds show markedly reduced mTORC1 activity in plaque-associated microglia (measured by phospho-S6 immunofluorescence), concurrent with failure of DAM Stage 2 expansion—a finding replicated in Tyrobp-null animals. Second, human iPSC-derived microglia expressing the TREM2 R47H AD-risk variant show attenuated SYK phosphorylation upon lipid ligand stimulation, with consequent reduction in AKT-S473 phosphorylation and impaired phagocytic cup formation. Third, single-cell proteomics of human post-mortem AD tissue (Mathys et al., 2019 framework; subsequent proteomics validation cohorts) identifies DAM subpopulations with elevated phospho-S6K1 that co-express LPL, CST7, and APOE at high levels and are spatially enriched at compact plaque cores versus diffuse plaques—suggesting that mTORC1-competent DAM are the ones actually compacting amyloid. Fourth, pharmacological mTORC1 activation via PTEN-null microglia-specific knockouts in APP mice demonstrates enhanced plaque compaction and reduced neuritic dystrophy, providing direct causal evidence that mTOR activity in microglia is neuroprotective in the amyloid context.

Disease Stage Specificity

This intervention has maximal therapeutic relevance during the prodromal-to-early symptomatic window, corresponding to Braak tangle stages II-IV and amyloid PET positivity (Centiloid score 20-70), before widespread neurofibrillary pathology co-opts microglial resources toward tau-associated neuroinflammation. Biomarker operationalization: patients with elevated CSF sTREM2 (indicating active TREM2 ectodomain shedding and thus a window where intracellular signaling enhancement would be relevant), positive amyloid PET, negative or mildly positive tau PET, and preserved hippocampal volume on MRI represent the ideal treatment population.

The stage specificity argument is mechanistically grounded: mTORC1-driven anabolism requires a functional mitochondrial electron transport chain as substrate. In late-stage AD, mitochondrial membrane potential collapse driven by hyperphosphorylated tau's interaction with complex I subunits renders mTOR activation futile or even harmful (by inducing mitophagy arrest). Furthermore, in fully transitioned Stage 2 DAM that have adopted an NF-κB-dominant inflammatory phenotype, PI3K-AKT signaling may paradoxically amplify IL-1β and TNF production via AKT-mediated IKKα phosphorylation. This is the critical reconciliation with antagonism data: late-stage TREM2 suppression experiments that show benefit are likely operating in this mitochondrially compromised, NF-κB-hyperactivated DAM context where the PI3K axis has been rewired toward pro-inflammatory rather than anabolic outputs.

Therapeutic Strategy

The preferred modality is a bispecific intrabody or cell-type-restricted gene therapy delivering a constitutively active SYK(Y352E/Y525E/Y526E) variant under a microglia-specific promoter (Cx3cr1 or P2ry12 regulatory elements) delivered via AAV-PHP.eB or next-generation capsids with enhanced CNS tropism following a single intrathecal or intravenous administration. This approach bypasses the ADAM10/ADAM17 shedding problem entirely by acting downstream of the surface receptor. Alternatively, a small molecule PI3Kδ-selective partial agonist (building on the idelalisib/duvelisib scaffold but with reduced catalytic site occupancy to avoid full immunosuppressive PI3Kδ inhibition paradox) could be deployed. Critically, dosing must achieve mTORC1 activation within a Goldilocks window: sufficient S6K1 phosphorylation to drive lysosomal biogenesis and lipid catabolism, but below the threshold that triggers S6K1-mediated IRS-1 serine phosphorylation feedback, which would paradoxically suppress AKT and undermine the intervention. For BBB penetration of small molecules, the molecular weight must remain below 450 Da with calculated logP between 1-3; the SYK inhibitor scaffold offers medicinal chemistry opportunities to engineer CNS-penetrant partial agonists rather than full inhibitors.

Key Uncertainties and Risks

The most serious mechanistic uncertainty is the precise identity of TREM2's lipid ligands in vivo and whether their availability fluctuates in ways that would compete with or synergize with intracellular signal enhancement. If DAM in the amyloid microenvironment are already maximally ligand-stimulated, adding downstream signal amplification may produce mTORC1 hyperactivation driving excessive microglial proliferation and displacement of homeostatic microglia from non-plaque regions. Safety concerns include: (1) off-target PI3Kδ activation in CNS-resident lymphocytes potentially impairing meningeal immune surveillance; (2) mTOR hyperactivation promoting mTORC1-driven suppression of autophagy, worsening tau aggregate clearance despite improving amyloid compaction—creating a pathological trade-off; (3) constitutively active SYK constructs causing microglial hyper-reactivity to sterile inflammation stimuli, worsening secondary injury responses. A fundamental risk is that metabolic rescue of DAM may simply delay rather than prevent the Stage 2 inflammatory transition, providing a narrow therapeutic window requiring precise timing that is clinically impractical with current biomarker resolution.

Experimental Roadmap

Phase 1 — In vitro mechanistic validation: Generate human iPSC-microglia expressing R47H, common variant, and sTREM2-overexpressing lines. Transduce with AAV-SYK(constitutively active) or treat with PI3Kδ partial agonists. Measure: phospho-proteomics (S6K1, 4E-BP1, AKT-S473), Seahorse metabolic flux (OCR/ECAR), cholesterol efflux capacity (BODIPY-cholesterol), lysosomal pH (LysoSensor), and phagocytic capacity using pHrodo-conjugated Aβ42 fibrils. Success criterion: ≥40% restoration of OCR and ≥2-fold increase in Aβ42 phagocytic index in R47H cells to common-variant levels.

Phase 2 — Mouse model validation: Cross AAV-PHP.eB-Cx3cr1-SYK(CA) into 5xFAD mice at 2 months (pre-plaque) and 4 months (established plaques). Primary endpoints: plaque compaction index (ThioS-positive core area/total MOAB2-positive area), neuritic dystrophy (APP-immunoreactive dystrophic neurite density per plaque), microglial metabolic phenotype (phospho-S6 IHC), and cognitive outcomes (Barnes maze, novel object recognition). Secondary: bulk and single-nucleus RNA-seq of CD11b-sorted microglia to map DAM trajectory shifts.

Phase 3 — Human biomarker correlation: Mine existing ADNI and Knight ADRC cohort CSF datasets to test whether high sTREM2 combined with low phospho-tau181 (proxy for early stage) identifies individuals with preserved microglial metabolic competence, validating the proposed treatment window. Correlate with longitudinal amyloid PET centiloid change rates as a proxy for plaque compaction efficiency.

Success criteria for the program overall: Phase 2 must demonstrate ≥25% reduction in neuritic dystrophy density without worsening tau pathology (AT8 immunoreactivity), and Phase 3 must confirm the sTREM2-high/p-tau-low biomarker signature enriches for slow amyloid accumulation in untreated individuals, validating biological plausibility of the target window before clinical translation.