assess

# Feasibility Assessment: SPP1 Therapeutic Hypotheses

## Scope and Framework

Seven hypotheses survived initial scrutiny at varying credibility levels. Two hypotheses (H1, H6) present fatal or near-fatal conceptual weaknesses—the unproven amyloid clearance premise in H1, and the absent human splice-variant evidence in H6—and are treated as secondary. The remaining five hypotheses are evaluated across druggability, biomarkers and model systems, clinical development constraints, safety, and realistic timeline/cost. A synthesis framework is provided at the end.

---

## Surviving Hypotheses Overview

| Hypothesis | Mechanism | Revised Confidence | Strategic Tier |
|---|---|---|---|
| **H4** | TREM2 agonism redirects SPP1 signaling | 0.58 | **Near-term feasible** |
| **H5** | C3/C3aR complement inhibition | 0.45 | **Reformulation required** |
| **H3** | ITGAX (αXβ2) integrin blockade | 0.42 | **Long-term investment** |
| **H7** | Transient SPP1 blockade during synaptic vulnerability window | 0.48 | **Biomarker-gated** |
| **H2** | Perivascular cell–specific SPP1 knockout | 0.40 | **Technically constrained** |

---

## H4: TREM2-Dependent Switch Hypothesis

**Highest surviving confidence; existing therapeutic platform.**

### Druggability: Strong (Tier 1)

TREM2 is a high-value target with active clinical programs. Multiple TREM2 agonist antibodies exist (AL002 [Alector/AbbVie], HFF3760 [TrueBinding], analogs from Denali). These are humanized monoclonal antibodies with established manufacturing pathways, PK/PD characterized in phase I. The addition of SPP1 modulation does not require de novo target discovery—it requires mechanistic pairing with an existing agent.

The primary feasibility question is not whether TREM2 is druggable, but whether combining TREM2 agonism with SPP1 modulation (enhancement or inhibition depending on disease stage) produces a therapeutic window not captured by either approach alone. This is a combination-strategy feasibility question, not a target-discovery feasibility question.

**Critical gap:** No existing TREM2 agonist is combined with SPP1-targeting in any current pipeline. The combination would require a new IND application, not a label expansion.

### Biomarkers and Model Systems: Adequate

**Translational biomarkers:**
- CSF sTREM2 is an established pharmacodynamic marker of TREM2 engagement (increases with agonist dosing)
- CSF NfL and neurogranin as downstream synaptic integrity markers
- Plasma p-tau217 is emerging as a high-specificity AD progression marker
- TSPO-PET for microglial activation state (though TSPO has limitations in human population genetics)

**Model systems:**
- 5xFAD/Trem2−/− mice provide the key experiment: does SPP1 overexpression in Trem2−/− background produce the same synapse loss phenotype as SPP1 overexpression in wild-type? If yes, the "switch" model is falsified. If no, TREM2 is upstream of SPP1's pathological effect.
- Human iPSC-derived microglia from R47H carriers enable mechanistic studies with human-relevant genetics
- scRNA-seq of human AD brain already shows TREM2-dependent microglial states; cross-referencing with SPP1 expression in those states is feasible using existing datasets

**Key experiment in pipeline:** The proposed RNA-seq comparison (SPP1-treated Trem2−/− vs. WT microglia) would cost approximately $30,000–60,000 (10x Genomics Chromium) and could be completed within 6 months. This is the single most decisive experiment for this hypothesis.

### Clinical Development Constraints: Moderate

**Regulatory pathway:** Combination approach with two novel mechanisms would require a dedicated phase II program. However, if TREM2 agonism is approved first (AL002 is in phase II for AD), adding a SPP1-targeting component to an existing IND is more tractable than developing both simultaneously.

**Patient stratification:** The strongest rationale is for TREM2 R47H carriers, where the switch failure is most mechanistically plausible. However, R47H represents only ~3–5% of LOAD cases. The broader population would require biomarker-based stratification (elevated CSF SPP1, CD11c+ microglia by PET, or microglial activation signature on TSPO-PET).

**Primary endpoint challenge:** Synaptic density cannot be directly measured in living humans. Surrogate endpoints would include:
- Cognitive composite (LAND, CIBIC+)
- CSF neurogranin as a synaptic proxy
- Volumetric MRI (hippocampal atrophy rate)
- Tau-PET as a downstream correlate

A trial would likely require 18–24 months of treatment with ≥500 participants per arm for powered cognitive outcomes, at substantial cost.

**Phase III planning:** Trial design must pre-specify whether the combo strategy is:
1. Chronic TREM2 agonism (already in AD) combined with episodic SPP1 blockade during defined windows, or
2. Continuous dual targeting throughout

This distinction determines dosing schedules, safety monitoring, and regulatory submission strategy.

### Safety: Manageable but Monitored

**TREM2 agonist safety profile:** Currently acceptable in phase I/II, with immune-related adverse events as primary concern. TSPO-PET studies show increased microglial activation, which is mechanistically intended but requires careful monitoring for cytokine release or ARIA (Amyloid-Related Imaging Abnormalities).

**Combination concern:** If SPP1 is truly downstream of TREM2 in pathological signaling, TREM2 agonism alone may suppress SPP1 pathology entirely, making the combination unnecessary. Conversely, if SPP1 acts through a parallel pathway, adding SPP1 blockade to TREM2 agonism risks over-suppressing microglial surveillance, potentially increasing infection risk or impairing beneficial amyloid clearance.

**Key safety study:** Microglial state profiling in treated animals—distinguishing homeostatic, DAM, and maladaptive states via RNA-seq or flow cytometry—must be incorporated into preclinical GLP toxicology.

### Realistic Timeline and Cost

| Phase | Timeline | Cost Estimate |
|---|---|---|
| Key preclinical experiments (RNA-seq, scRNA-seq validation) | 6–9 months | $800K–1.2M |
| GLP toxicology (TREM2 agonist + SPP1 combination) | 12 months | $2–3M |
| Phase I (safety, biomarker readouts) | 18–24 months | $8–15M |
| Phase II (proof-of-concept, biomarker-driven) | 30–36 months | $30–50M |
| Phase III | 48–60 months | $150–250M |
| **Total to approval (optimistic scenario)** | **10–12 years** | **$200–350M** |

**Critical path item:** The 6-month RNA-seq experiment in Trem2−/− mice determines whether the entire development program proceeds. If SPP1 effects are TREM2-independent, H4 collapses. If the switch is confirmed, this becomes the leading SPP1-related therapeutic strategy.

**Recommendation:** Fund the decisive Trem2−/− + SPP1 RNA-seq experiment immediately. If positive, this hypothesis can proceed to IND-enabling studies within 18 months using existing TREM2 agonist programs as the anchor.

---

## H5: Complement Cascade Specificity (C3/C3aR Inhibition)

**Survives primarily because complement inhibitors exist clinically; challenged by specificity.**

### Druggability: Established platform, wrong target relationship

C3 inhibition is a solved drug development problem. Eculizumab (Alexion/AstraZeneca), ravulizumab, and pegcetacoplan demonstrate the platform. C3aR antagonists (e.g., avacopan, approved for ANCA vasculitis) demonstrate the receptor-level approach.

The druggability challenge is not the mechanism but the **specificity problem**: C3 is not specific to SPP1 signaling. The hypothesis asserts that SPP1 drives C3 expression, but if C3 inhibition protects synapses via general complement blockade, the therapeutic rationale is "complement inhibition in AD" rather than "SPP1 pathway targeting." This is a strategic distinction, not a failure.

**Potential reformulation:** Rather than C3 as the target, C3aR (microglial) may offer a more localized approach—preserving systemic complement while blocking the microglial inflammatory arm. This reframes the hypothesis as "microglial C3aR blockade downstream of SPP1" rather than systemic C3 inhibition.

### Biomarkers and Model Systems: Excellent

**Translational biomarkers:**
- Plasma C3 and C3a are measurable with standard ELISA
- CSF complement activity assays exist and are used in clinical trials
- Synaptic biomarkers: CSF neurogranin, VILIP-1, and emerging plasma p-tau217/NfL ratios

**Model systems:**
- C3−/− 5xFAD mice are available; synaptic protection is documented (PMID: 28973389)
- The key missing experiment is whether C3 knockdown in SPP1−/− mice provides additive protection or converges on the same endpoint. This distinguishes the pathways.

**Human validation:**
- Complement gene expression in AD brain is already characterized (multiple RNA-seq datasets)
- CSF complement levels in AD vs. control show elevated C3 and C3a; correlation with cognitive decline is established

### Clinical Development Constraints: Moderate-low barrier, but specificity challenge

**Existing regulatory precedent:** Pegcetacoplan is in phase III for Alzheimer's (TOPAZ trial, Roche/UCB) targeting complement component C3. This establishes the regulatory pathway and FDA familiarity with complement inhibition in AD.

**Patient population:** Broad—any amyloid-positive early AD patient. No genetic stratification required (unlike H4 with TREM2 carriers). This is a larger market but also requires larger trials.

**Endpoint challenge:** Same as H4—cognitive composites are primary, synaptic biomarkers are exploratory.

**Critical trial design question:** If C3 inhibition is effective in AD, is it through SPP1 pathway modulation or general complement-mediated synapse loss? Without a mechanistic link, this is indistinguishable. A mechanistic biomarker (e.g., microglial C3aR occupancy correlating with CSF neurogranin improvement) would partially address this.

### Safety: Major concern requiring mitigation

**Systemic complement inhibition:** Eculizumab and ravulizumab carry black box warnings for serious meningococcal infections. This risk is known, manageable (vaccination, monitoring), and accepted in diseases like PNH and aHUS. However, applying this risk to early AD—where patients are less acutely ill—creates a different risk/benefit calculus.

**Potential mitigation strategies:**
1. **Microglial-specific C3aR antagonism** (avacopan-like) rather than systemic C3 inhibition—local CNS complement blockade with preserved systemic immunity. This is the most feasible safety strategy.
2. **Intermittent dosing** (monthly subcutaneous administration) rather than continuous complement blockade
3. **Biomarker-gated dosing** (stop treatment when synaptic biomarkers stabilize) to minimize cumulative risk

**Infection risk modeling:** In a 5-year AD prevention trial, the meningococcal infection risk (~0.5% per year with vaccination) may be acceptable if cognitive benefit is demonstrated. In a symptomatic treatment trial, the calculus is more favorable. The risk becomes problematic for preclinical AD populations where intervention is long-term.

### Realistic Timeline and Cost

| Phase | Timeline | Cost Estimate |
|---|---|---|
| C3aR antagonist development (if reformulated) | 18–24 months (new IND) | $20–30M |
| Phase I (safety in AD population) | 12–18 months | $10–15M |
| Phase II (synaptic biomarker readout) | 24 months | $25–40M |
| Phase III | 42–54 months | $120–180M |
| **Total to approval (reformulated as C3aR antagonist)** | **8–10 years** | **$180–250M** |

**Alternative (leverage existing):** If pegcetacoplan data from the TOPAZ trial is positive, a C3aR antagonist could be developed as a follow-on with 5–7 years to approval (existing C3 data reduces required safety package).

**Recommendation:** H5 is the most near-term feasible hypothesis only if reformulated as microglial C3aR antagonism rather than systemic C3 inhibition. The existing pegcetacoplan data (positive or negative) in AD will be highly informative—if systemic C3 inhibition shows benefit, C3aR blockade is a lower-risk follow-on. If negative, the complement pathway in AD requires re-evaluation.

---

## H3: ITGAX (αXβ2) Integrin Blockade

**Mechanistically sound receptor selectivity; lacks chemical starting point.**

### Druggability: Chemical feasibility unestablished

The conceptual framework is compelling: SPP1 binds ITGAX (CD11c/CD18) with high affinity, CD11c+ microglia correlate with pathology, and selective blockade might preserve amyloid phagocytosis while reducing synapse targeting. However, there is no identified small molecule or antibody antagonist with demonstrated CNS penetration and AD-relevant pharmacology.

**What exists:**
- Anti-CD11c antibodies exist (clone 3.9 for mouse, clone Bu15 for human), but these are research tools, not therapeutic candidates
- No CNS-penetrant small molecule ITGAX antagonist is reported in the literature
- The "ASAPI peptide" referenced in the hypothesis is hypothetical; no reference, no chemistry, no PK data

**Gap analysis:** This hypothesis requires a drug discovery program before any clinical development can begin. That adds 2–4 years and $30–60M in discovery-phase investment before an IND exists.

**Alternative approach:** An anti-CD11c antibody (like the Bu15 clone humanized) could be developed as a biologic