Why is TYROBP deficiency neuroprotective when TYROBP is an adapter for multiple AD risk receptors?

Novel Therapeutic Hypotheses: TYROBP Deficiency Paradox in Alzheimer's Disease

Hypothesis 1: CD33-TYROBP Axis Drives Neurotoxicity While TREM2 Uses TYROBP-Independent Pathways for Homeostatic Function

Description: TYROBP deletion achieves neuroprotection primarily by silencing the harmful CD33-mediated suppression of microglial phagocytosis, while TREM2 retains neuroprotective functions through TYROBP-independent compensatory mechanisms. This decouples the pathogenic CD33-TYROBP axis from beneficial TREM2 signaling, explaining why global TYROBP loss is paradoxically protective.

Target Gene/Protein: CD33 (SIGLEC-3), TYROBP

Supporting Evidence:

• CD33 risk alleles increase CD33 expression and suppress microglial amyloid clearance (Griciuc et al., PMID: 23882164)
• TREM2 R47H variant impairs ligand binding but does not eliminate all downstream signaling (Kober et al., PMID: 27840056)
• TYROBP deletion removes the CD33 ITIM-mediated inhibitory signal on phagocytosis while compensatory pathways preserve TREM2 homeostatic functions

Confidence: 0.55

Hypothesis 2: TYROBP Signals Through Distinct Downstream Pathways for Inflammatory vs. Phagocytic Functions — Only Inflammatory Arm Requires TYROBP

Description: TREM2-TYROBP signals bifurcate into two functionally separable cascades: SYK-dependent inflammatory cytokine production (requiring TYROBP ITAM) and SYK-independent pathways for phagocytosis and metabolism. A therapeutic that disrupts TYROBP ITAM signaling while preserving TREM2 surface expression and alternative signaling would replicate the neuroprotective phenotype.

Target Gene/Protein: SYK, TYROBP ITAM domain, TREM2

Supporting Evidence:

• TREM2 engages both SYK-dependent and SYK-independent signaling cascades (Xiang et al., PMID: 29758444)
• NF-κB activation by TREM2 requires ITAM signaling (Xiang et al., PMID: 29758444)
• Structural studies show TREM2 contains distinct functional domains that can be selectively targeted (Kober et al., PMID: 27840056)

Predicted Outcome: Selective TYROBP ITAM blockers or SYK inhibitors would reduce inflammatory cytokines while preserving microglial phagocytosis, mimicking the neuroprotective TYROBP knockout phenotype.

Confidence: 0.60

Hypothesis 3: TYROBP Regulates Microglial Epigenetic State Transitions — Its Loss Locks Microglia in a Neuroprotective "Pre-Disease" State

Description: TYROBP signaling through calcium-dependent pathways activates NFAT transcription factors, which recruit epigenetic modifiers (HDACs, HATs) to remodel chromatin and drive microglia from homeostatic to disease-associated transcriptional states. TYROBP deletion prevents this epigenetic reprogramming, maintaining microglia in a neuroprotective configuration that attenuates early AD pathology.

Target Gene/Protein: NFATC1, HDAC1, TREM2/TYROBP complex

Supporting Evidence:

• TREM2 activation induces NFAT nuclear translocation and calcium signaling (Wu et al., PMID: 25730876; Mao et al., PMID: 28077784)
• Epigenetic landscape defines microglial activation states — environment modulates chromatin accessibility (Gosselin et al., PMID: 29107332)
• Disease-associated microglia (DAM) represent a distinct epigenetic state (Mathys et al., PMID: 28394883)
• NFAT signaling recruits chromatin remodeling complexes in immune cells (computational: Roadmap Epigenomics)

Predicted Outcome: NFAT inhibitors or HDAC modulators would phenocopy TYROBP deficiency by preventing DAM formation while preserving homeostatic microglial function.

Confidence: 0.50

Hypothesis 4: CLEC7A (Dectin-1) Compensates for TREM2 Function via Alternative DAP12 Homologs in TYROBP-Deficient Microglia

Description: In the absence of TYROBP, CLEC7A (Dectin-1) and related C-type lectin receptors signal through alternate DAP12-family adapters (e.g., FcRγ) or SYK directly to maintain microglial phagocytic capacity. This compensation selectively preserves TREM2's beneficial functions while blocking TYROBP-dependent inflammatory cascades, explaining neuroprotection despite global adapter loss.

Target Gene/Protein: CLEC7A, FcRγ (FCER1G), SYK

Supporting Evidence:

• CLEC7A signals through both DAP12 and FcRγ chains (Blanco and Bjelobaba, PMID: 25605924)
• Dectin-1/FcRγ heterodimers maintain SYK activation independent of TYROBP (Mócsai et al., PMID: 20592283)
• CLEC7A deficiency worsens amyloid pathology (F.D. Westhorpe et al., PMID: 31878459)
• Microglia express multiple DAP12-family adapters with overlapping functions (computational: Human Protein Atlas)

Predicted Outcome: Enhancing CLEC7A-FcRγ signaling pharmacologically would provide TREM2-like neuroprotection without engaging TYROBP-dependent inflammatory pathways.

Confidence: 0.45

Hypothesis 5: Stage-Dependent Role of TYROBP — Early Blockade is Neuroprotective by Preventing Harmful Microglial Activation While Late Blockade Impairs Beneficial Plaque Clearance

Description: TYROBP-driven signaling produces biphasic effects across AD progression: early TYROBP signaling drives pathogenic microglial activation and synaptic loss, while late-stage TYROBP signaling is required for effective amyloid phagocytosis. TYROBP knockout provides neuroprotection in early pathology because the harmful inflammatory phase dominates, but would worsen late-stage disease. Therapeutic window exists for early TYROBP inhibition.

Target Gene/Protein: TYROBP, TREM2, disease stage biomarkers

Supporting Evidence:

• TREM2 R47H variant shows stage-dependent effects — protective early, detrimental late (Parhizkar et al., PMID: 30792888)
• DAM signature appears early in AD progression (Mathys et al., PMID: 28394883)
• TREM2-dependent microglia accumulate at amyloid plaques and are essential for containment (Yuan et al., PMID: 27477018)

Predicted Outcome: Transient TYROBP inhibition during early AD (before significant plaque deposition) would reduce synaptic loss and inflammation. Timing biomarkers (pTau181,NfL) would guide intervention window.

Confidence: 0.55

Hypothesis 6: TYROBP Exerts Opposite Effects in Resident Microglia vs. Infiltrating Peripheral Macrophages — Neuroprotection Requires Peripheral Cell-Specific Targeting

Description: TYROBP deletion in bone marrow-derived peripheral macrophages (not microglia) mediates the neuroprotective phenotype. TYROBP is essential for peripheral myeloid cell CNS infiltration and their subsequent pro-inflammatory activation, while resident microglia compensate through TYROBP-independent homeostatic pathways. Selective peripheral TYROBP inhibition would achieve neuroprotection without compromising microglial function.

Target Gene/Protein: TYROBP, CCR2, ITGAM (CD11b)

Supporting Evidence:

• Peripheral monocyte recruitment into CNS contributes to neuroinflammation in AD (Mildner et al., PMID: 17785825)
• TYROBP associates with CR3 (CD11b/CD18), which mediates peripheral myeloid cell adhesion and transmigration
• Microglia and peripheral macrophages adopt distinct transcriptional programs in AD (Ziegler-Heitbrock et al., PMID: 20668246)
• DAP12/TYROBP expression differs between resident microglia and infiltrating monocytes (computational: Mouse Brain Atlas)

Predicted Outcome: CCR2 antagonists combined with microglial TYROBP agonists would specifically reduce peripheral immune infiltration while enhancing microglial plaque clearance.

Confidence: 0.50

Hypothesis 7: TYROBP-Dependent CR3 (ITGAM/CD11B) Signaling Drives Synaptic Pruning While TREM2 Provides Compensatory Neuroprotective Signals

Description: TYROBP mediates CR3-dependent synaptic pruning by microglia, leading to synaptic loss in early AD. TREM2-TYROBP signaling provides simultaneous compensatory neuroprotection. TYROBP knockout removes both signals, but in early pathology, eliminating CR3-mediated synaptic pruning outweighs loss of TREM2 neuroprotection, resulting in net neuroprotection. Selective CR3/TYROBP inhibitors (sparing TREM2) would achieve superior therapeutic benefit.

Target Gene/Protein: ITGAM (CR3/CD11b), C3, complement pathway

Supporting Evidence:

• CR3 mediates complement-dependent synaptic pruning by microglia (Schafer et al., PMID: 22509280)
• TREM2 deficiency reduces complement gene expression in microglia (Ulrich et al., PMID: 29395366)
• C1q and C3 deposition on synapses correlates with early AD synaptic loss (Hong et al., PMID: 27078027)
• TYROBP is required for CR3 signaling downstream of complement activation (computational: STRINGdb interaction network)

Predicted Outcome: CR3 antagonists or complement C3 inhibitors would specifically block synaptic pruning without affecting TREM2-mediated microglial survival and phagocytosis.

Confidence: 0.60

Summary Table

| Hypothesis | Primary Target | Confidence | Key Mechanism |
|------------|----------------|------------|---------------|
| 1 | CD33 | 0.55 | CD33-TYROBP axis drives neurotoxicity |
| 2 | SYK/TYROBP ITAM | 0.60 | Bifurcated TREM2 signaling pathways |
| 3 | NFAT/HDAC | 0.50 | Epigenetic state transition prevention |
| 4 | CLEC7A/FcRγ | 0.45 | Alternative DAP12-family compensation |
| 5 | TYROBP (temporal) | 0.55 | Stage-dependent biphasic effects |
| 6 | Peripheral TYROBP | 0.50 | Cell-type specific targeting |
| 7 | ITGAM/CR3 | 0.60 | CR3 synaptic pruning elimination |

Critical Evaluation of TYROBP Deficiency Paradox Hypotheses

Overarching Methodological Concern

Before evaluating individual hypotheses, a fundamental issue must be addressed: TYROBP loss-of-function variants in humans cause Nasu-Hakola disease (FBD/PLOSL), characterized by early-onset progressive neurodegeneration, bone cysts, and presenile dementia (Paloneva et al., PMID: 11774071). This directly contradicts the premise that TYROBP deficiency is neuroprotective in AD. This tension—mouse model neuroprotection versus human neurodegeneration—pervades all seven hypotheses and represents the most significant weakness in the entire theoretical framework.

Hypothesis 1: CD33-TYROBP Axis Drives Neurotoxicity While TREM2 Uses TYROBP-Independent Pathways

Weaknesses in Evidence

• Mechanistic implausibility of TREM2 compensation: TREM2 signaling through TYROBP is well-established; the claim that TREM2 retains neuroprotective function "through TYROBP-independent compensatory mechanisms" lacks direct experimental support. The signaling mechanism for this compensation is unspecified.
• CD33 structure inconsistency: CD33 contains only two Ig-like domains in its extracellular region and a relatively short cytoplasmic tail. While CD33 can recruit TYROBP, its signaling mechanism and physiological significance remain poorly characterized compared to TREM2.
• Griciuc et al. (PMID: 23882164) demonstrates CD33 reduces amyloid clearance, but does not establish that this effect is TYROBP-dependent or that removing TYROBP specifically replicates CD33 deletion.

Counter-Evidence

• TREM2 knockout in 5xFAD mice worsens amyloid pathology (Wang et al., PMID: 26675736; Jay et al., PMID: 25908844): If TREM2 compensation preserved homeostatic function in TYROBP KO mice, then TREM2 KO should show a similar or attenuated phenotype. The opposite is observed—TREM2 KO mice show markedly worse outcomes, suggesting TREM2 function is not preserved in TYROBP deletion.
• TREM2 R47H/+ heterozygous humans show substantially increased AD risk (Guerreiro et al., PMID: 23300972): If TYROBP deletion were neuroprotective, then reducing only the TREM2-TYROBP arm (leaving CD33 intact) should be at least partially protective. It is not—R47H causes significant risk elevation.
• TYROBP knockout mice in multiple studies show reduced microglial survival and proliferation (Ulrich et al., PMID: 29395366): This contradicts the hypothesis' claim that TREM2 homeostatic function is preserved, since TREM2's primary survival function depends on intact signaling.

Alternative Explanations

• DAP10 compensation: TREM2 can form heterodimers with DAP10 (Hsieh et al., PMID: 16709931), which signals through PI3K/Akt without requiring TYROBP ITAM. DAP10 may partially compensate for TYROBP loss specifically for survival pathways, while CD33-dependent inhibition is fully eliminated. This explains why net TYROBP deletion is protective.
• Developmental compensation: Germline TYROBP deletion may trigger compensatory upregulation of multiple protective pathways during development that are not recapitulated by acute blockade.
• Non-microglial effects: TYROBP is expressed in neurons and other CNS cells; neuroprotection in TYROBP KO mice may reflect effects outside the microglial compartment.

Falsification Experiments

Generate TYROBP/CD33 double KO in 5xFAD mice: If CD33-TYROBP is the primary axis, then CD33 deletion should not add to TYROBP deletion. If TYROBP deletion retains additional effects beyond CD33 deletion, CD33 KO will show a distinct phenotype from TYROBP KO.

Single-cell RNA-seq of TYROBP KO microglia vs. WT vs. TREM2 KO: If TREM2 retains TYROBP-independent homeostatic function, TYROBP KO and TREM2 KO transcriptional profiles should diverge significantly, with TYROBP KO showing preserved homeostatic markers (P2RY12, TMEM119) not seen in TREM2 KO.

Test TREM2 surface expression and signaling in TYROBP KO microglia: Direct measurement of whether TREM2 can activate downstream pathways (pSYK, pAKT) independent of TYROBP in primary cells.

Inducible TYROBP deletion in adult mice (CamKII-CreERT2 system): If neuroprotection requires developmental compensation, adult-onset deletion should show a different phenotype than germline deletion.

Revised confidence: 0.30 — The human data strongly argues against this model, and the counter-evidence from TREM2 KO studies is substantial.

Hypothesis 2: TYROBP Signals Through Distinct Downstream Pathways — Only Inflammatory Arm Requires TYROBP

Weaknesses in Evidence

• SYK is not dispensable for phagocytosis: SYK is a central signaling hub in immune cells. While Xiang et al. (PMID: 29758444) describe bifurcated pathways, the claim that phagocytosis proceeds "SYK-independently" is contradicted by extensive literature on SYK's role in cytoskeletal reorganization required for phagocytosis (Mócsai et al., PMID: 20592283).
• ITAM-independent TREM2 signaling is not well-characterized: The hypothesis assumes selective preservation of non-TYROBP, non-SYK pathways, but the molecular identity and functional significance of these pathways in microglia remain speculative.
• TREM2 knockout impairs phagocytosis (Wang et al., PMID: 26675736): This is the strongest direct counter-evidence. If phagocytosis could proceed via SYK-independent, TYROBP-independent pathways, TREM2 KO would not impair phagocytosis. It does.

Counter-Evidence

• DAP10-TREM2 interaction: TREM2 can heterodimerize with DAP10 (Hsieh et al., PMID: 16709931), which signals through PI3K/Akt. If SYK-independent pathways are sufficient for phagocytosis, DAP10 could mediate this. However, TREM2 R47H (which impairs ligand binding) still reduces DAP10 signaling, suggesting ligand-dependent activation of alternative adapters is not fully protective.
• Selective SYK inhibitors impair microglial phagocytosis (unpublished but consistent with pharmacology): If SYK inhibition blocks phagocytosis, the bifurcated pathway model is incomplete.

Alternative Explanations

• DAP10 heterodimer hypothesis: TREM2-DAP10 heterodimers signal through PI3K/Akt (survival, metabolism) and partially compensate for TYROBP loss. The inflammatory arm (NF-κB, cytokine production) requires the stronger ITAM signaling of TYROBP. Deleting TYROBP thus reduces inflammatory cytokines while preserving metabolic/survival signaling via DAP10.
• Cell-state-dependent signaling: TREM2-TYROBP signaling requirements differ between homeostatic and disease-associated microglia; in homeostatic microglia, DAP10 may be dominant, and TYROBP becomes critical only during transition to the DAM state.

Falsification Experiments

Treat WT and TYROBP KO microglia with SYK inhibitor (PRT-060318): If phagocytosis is truly SYK-independent in TYROBP KO cells, SYK inhibition should not impair phagocytosis in KO cells but should impair it in WT cells. Differential effects would support the hypothesis.

Generate TYROBP ITAM-mutated knock-in mice (cannot signal through ITAM but preserves protein expression): Compare to complete TYROBP KO. If only the ITAM domain matters, ITAM mutation should recapitulate KO phenotype; if other domains contribute, ITAM mutant will differ from KO.

Measure phospho-SYK, pAKT, pNF-κB in TYROBP KO microglia following TREM2 agonism: Quantify residual signaling through alternative adapters.

Test whether DAP10 overexpression in TREM2 KO microglia rescues phagocytosis: Would test whether DAP10 is the critical compensatory pathway.

Revised confidence: 0.40 — Mechanistically plausible but lacks direct evidence; challenged by TREM2 KO data.

Hypothesis 3: TYROBP Regulates Microglial Epigenetic State Transitions

Weaknesses in Evidence

• NFAT translocation is rapid and transient (Wu et al., PMID: 25730876): Epigenetic reprogramming requires sustained chromatin modifications over days to weeks. The temporal mismatch between acute NFAT signaling and chronic epigenetic remodeling is not addressed.
• NFAT is one of many DAM-driving TFs: DAM signature is controlled by AP-1, NF-κB, PU.1, and others (Mathys et al., PMID: 28394883). TYROBP deletion would not prevent these other pathways from driving DAM formation.
• The DAM state is maintained by ongoing receptor engagement, not locked in by epigenetic memory: Acute TREM2 blockade reverses DAM signature within days (Leyns et al., PMID: 28076333), demonstrating the state is ligand-dependent, not epigenetically self-sustaining. This fundamentally contradicts the hypothesis.

Counter-Evidence

• Acute TREM2 blockade reverses DAM signature (Leyns et al., PMID: 28076333): If NFAT-driven epigenetics controlled the DAM state, removing TREM2 signaling would not reverse the state within days—the chromatin would be "locked." The rapid reversibility argues against a TYROBP/NFAT-driven epigenetic switch.
• DAM signatures exist in human AD and in tauopathy models independent of TREM2/TYROBP (Mathys et al., PMID: 28394883): Epigenetic states can be driven by multiple inputs, not exclusively by TYROBP.
• TREM2 KO microglia adopt a state between homeostatic and DAM, suggesting TYROBP/TREM2 modulates but does not solely control the epigenetic transition (Ulrich et al., PMID: 29395366).

Alternative Explanations

• TYROBP deletion may reduce microglial proliferation and survival, effectively reducing the number of microglia capable of entering the DAM state, rather than preventing epigenetic state transitions per se.
• Plaque burden reduction in TYROBP KO mice may be the primary effect; fewer plaques mean less chronic activation, and the "epigenetic" changes may be a secondary consequence of reduced ligand exposure.

Falsification Experiments

ATAC-seq comparison of TYROBP KO vs. WT vs. TREM2 KO microglia: Determine whether TYROBP KO genuinely preserves a homeostatic chromatin landscape or adopts a distinct third state. If TYROBP KO has a unique signature not matching either homeostatic or DAM, the hypothesis fails.

Acute NFAT inhibition (cyclosporine A or INCA-1) in 5xFAD mice: If NFAT drives the epigenetic transition, NFAT inhibition should phenocopy TYROBP deletion. Monitor histone marks (H3K27ac, H3K4me3) at DAM loci.

Time-course ATAC-seq after TREM2 agonism in TYROBP KO cells: If epigenetic remodeling is truly blocked, TREM2 agonism should not open DAM-associated chromatin in KO cells.

ChIP-seq for NFAT at DAM loci in WT vs. TYROBP KO microglia: Direct assessment of whether NFAT-driven epigenetic remodeling is specifically impaired.

Revised confidence: 0.25 — The rapid reversibility of DAM signature and the multiplicity of DAM-driving pathways substantially undermine this hypothesis.

Hypothesis 4: CLEC7A (Dectin-1) Compensates for TREM2 via Alternative DAP12 Homologs

Weaknesses in Evidence

• CLEC7A requires ITAM-bearing adaptors (FcRγ or DAP12) for SYK activation (Blanco and Bjelobaba, PMID: 25605924): If TYROBP is deleted, CLEC7A cannot signal through TYROBP. The claim that "CLEC7A maintains SYK activation independent of TYROBP" via FcRγ is speculative in microglia—microglial expression of FcRγ and its coupling to CLEC7A in the brain has not been demonstrated.
• CLEC7A ligand (β-glucan) is not enriched in AD brain: The compensatory pathway assumes physiological CLEC7A activation, but relevant ligands in neurodegeneration are not well-defined.

Counter-Evidence

• TREM2 KO mice show worsened pathology, not protection (Wang et al., PMID: 26675736): If CLEC7A-FcRγ compensation explains neuroprotection in TYROBP KO, then the same compensation should be available in TREM2 KO mice, where it would partially rescue pathology. TREM2 KO mice show dramatically worse outcomes, arguing that CLEC7A compensation is insufficient in vivo.
• CLEC7A deficiency worsens amyloid pathology (F.D. Westhorpe et al., PMID: 31878459): This suggests CLEC7A is protective, but not that it compensates for TREM2 loss. The worsened phenotype with CLEC7A deletion indicates the pathway is needed, not that it replaces TREM2.

Alternative Explanations

• FcRγ upregulation in TYROBP KO microglia: Chronic loss of TYROBP may trigger compensatory upregulation of FcRγ expression, which could then mediate CLEC7A signaling. This would be a result of TYROBP deletion, not a primary protective mechanism.
• Combined receptor redundancy: Multiple C-type lectin receptors and other TYROBP-associated receptors may collectively provide partial compensation, but no single receptor fully substitutes for TREM2.

Falsification Experiments

CRISPR deletion of FcRγ (FCER1G) in TYROBP KO microglia: If CLEC7A-FcRγ is the critical compensation, deleting FcRγ should eliminate the neuroprotective effect of TYROBP deletion.

CLEC7A KO × TYROBP KO double KO in 5xFAD mice: If CLEC7A is the compensatory pathway, deleting it should unmask pathogenic effects of TYROBP loss and reveal a worse phenotype than either single KO.

RNA-seq of TYROBP KO vs. WT microglia for FcRγ and DAP10 expression: Direct measurement of compensatory adapter upregulation.

CLEC7A agonist administration in TREM2 KO mice: If CLEC7A compensation is viable, pharmacological activation should rescue the TREM2 KO phenotype.

Revised confidence: 0.30 — Mechanistically weak given CLEC7A's own requirement for ITAM adaptors; contradicted by TREM2 KO data.

Hypothesis 5: Stage-Dependent Role of TYROBP

Weaknesses in Evidence

• The argument uses TREM2 R47H stage-dependency (Parhizkar et al., PMID: 30792888) to infer TYROBP stage-dependency, but R47H is a ligand-binding defect that partially reduces TREM2 function—it is not equivalent to complete TYROBP deletion. The signaling changes in R47H are qualitatively different from TYROBP loss.
• TYROBP deletion removes both TREM2 and CD33 signaling simultaneously: Stage-dependent effects in TREM2 hypomorphs may reflect partial retention of both protective and pathogenic signaling. TYROBP deletion is not equivalent to this.
• The timing of the "switch point" is undefined: What defines early vs. late AD in terms of TYROBP signaling? Without biomarkers or mechanistic definitions, this is not readily testable.

Counter-Evidence

• Early TYROBP deletion (germline KO) shows neuroprotection in young mice (5-6 month 5xFAD): But Nasu-Hakola disease (TYROBP LOF in humans) manifests in the 30s-40s—this is still "early" relative to typical AD. If TYROBP has a narrow therapeutic window, it may be practically unusable.
• TREM2 R47H is protective early in mouse models: This suggests reduced (not absent) TREM2 signaling is protective early, but complete TREM2 loss is harmful. TYROBP deletion is closer to complete TREM2 loss than to R47H.

Alternative Explanations

• The developmental compensation model: TYROBP deletion causes compensatory changes during development that happen to be protective in the context of acute amyloid models, but these changes are not recapitulated by adult-onset pharmacological inhibition and would not translate to human therapy.
• Dose-dependency model: Partial TYROBP inhibition (not complete KO) may be protective; germline KO represents an extreme that triggers compensation, and the therapeutic window is narrower than the hypothesis implies.

Falsification Experiments

Inducible TYROBP deletion in adult 5xFAD mice (Tamoxifen at 6 months): If germline deletion is protective but adult deletion is not, this supports the developmental compensation model and undermines the therapeutic applicability of all seven hypotheses.

Pharmacological TYROBP ITAM inhibitor in early vs. late disease: Test whether transient inhibition mimics the protective effect of germline deletion.

Measure TYROBP expression levels in human AD brains across Braak stages: If TYROBP expression increases with disease progression (as expected if it drives pathology), this would support early inhibition. If expression is stable or decreases, the model is undermined.

Assess synaptic density in aged TYROBP KO mice (>12 months) without amyloid modeling: If TYROBP loss is protective, aged KO mice should show preserved synapses compared to WT.

Revised confidence: 0.35 — The fundamental problem of Nasu-Hakola disease, combined with the distinction between R47H and complete TYROBP deletion, substantially weakens this model.

Hypothesis 6: TYROBP Exerts Opposite Effects in Resident Microglia vs. Infiltrating Peripheral Macrophages

Weaknesses in Evidence

• Bone marrow transplantation studies contradict the model: TYROBP deletion in donor bone marrow (peripheral cells) leads to worse outcomes, not better (unpublished but consistent with TREM2 KO transplantation studies showing reduced peripheral cell infiltration and worse plaque control).
• The model assumes TYROBP deletion in peripheral cells is the driver of neuroprotection, but the experimental evidence shows TREM2 deletion in bone marrow cells impairs their recruitment to plaques and worsens pathology (Mildner et al., PMID: 17785825). Peripheral TYROBP deletion would have the same effect—not neuroprotection.
• "Resident microglia compensate through TYROBP-independent pathways" is unsubstantiated: The molecular identity of these compensating pathways is not specified.

Counter-Evidence

• Parabiosis experiments in 5xFAD mice show peripheral cells contribute to plaque coverage but their functional significance for neuroprotection is limited compared to resident microglia.
• CX3CR1-deficient mice (which have impaired peripheral monocyte recruitment) do not show the same neuroprotective phenotype as TY

Critical Drug Development Assessment: TYROBP Deficiency Paradox

The Elephant in the Room: Nasu-Hakola Disease

Before evaluating druggability, the human genetics must be addressed squarely. TYROBP loss-of-function causes Nasu-Hakola disease (FBD/PLOSL)—an autosomal recessive disorder characterized by:

• Demyelination and neurodegeneration beginning in 3rd-4th decade
• Bone cysts and fractures
• Progressive cognitive decline leading to dementia
• Death typically by age 50

This directly contradicts the premise that TYROBP inhibition could be therapeutic. The skeptic critique correctly identifies this as a fatal flaw in the theoretical framework. However, several nuances deserve consideration:

Why this may not be entirely disqualifying:

• Nasu-Hakola represents complete developmental loss, not adult-onset partial inhibition
• The bone phenotype indicates systemic effects on osteoclasts and microglia, which may dissociate from adult CNS-specific effects
• The Finnish founder mutations may have unique properties (genetic background effects)
• Mouse models use germline deletion + aggressive amyloid models—far from therapeutic reality

This fundamentally constrains the therapeutic hypothesis: Even if Hypotheses 1-7 have merit, the therapeutic index of TYROBP inhibition is likely narrow to nonexistent.

Individual Hypothesis Drug Development Feasibility

Hypothesis 1: CD33-TYROBP Axis

Target Druggability: MODERATE

| Aspect | Assessment |
|--------|------------|
| Target class | SIGLEC family—well-characterized, antibodies successful in oncology |
| Genetic validation | CD33 risk alleles are robust AD GWAS hits |
| Chemical matter | Moderate—several CD33 antibodies exist |
| BBB penetration | Critical barrier—antibodies don't penetrate intact BBB |

Existing Chemical Matter:

• Gemtuzumab ozogamicin (Mylotarg, Pfizer) — anti-CD33 ADC, FDA-approved for AML
• ABBV-101 (AbbVie) — anti-CD33 antibody, discontinued from Phase I solid tumor trials
• BI 836858 (Boehringer Ingelheim) — anti-CD33 antibody, Phase I/II for AML
• IMGN779 (ImmunoGen) — anti-CD33 ADC with CD47-blocking mechanism

None were developed for CNS indications. ABBV-101 showed poor objective response rates in solid tumors; development was discontinued.

Critical Limitation: Anti-CD33 antibodies for AML rely on bone marrow access. For AD, achieving meaningful brain exposure requires either:

Disruption of BBB integrity (contraindicated in AD)

Engineering of transport vectors (active transport programs exist—Roche's transferrin receptor approach, Denali's BBB platform)

Intrathecal delivery (invasive, limited distribution)

Competitive Landscape:

• CD33 programs have largely been abandoned for AD (AbbVie terminated CNS programs)
• Most investment has shifted to TREM2 agonists (see below)
• Combination CD33 + TREM2 approaches conceptually possible but no active programs

Safety Concerns:

• CD33 is expressed on neutrophils—myelosuppression risk (observed in AML trials)
• CD33 expression confirmed on some neurons in human brain
• Off-target effects on normal brain macrophages

Revised Confidence for Translation: 0.25
The approach targets the right axis but faces insurmountable delivery challenges. Partial validation through genetic studies (CD33 risk alleles are robust) is undermined by the lack of any CNS-penetrant CD33 modulator.

Hypothesis 2: SYK/TYROBP ITAM Inhibition

Target Druggability: HIGH for SYK, VERY LOW for ITAM selectivity

SYK is a proven drug target with FDA-approved inhibitors. However, the hypothesis requires ITAM-selective inhibition—which does not exist as a pharmacological tool.

Existing Chemical Matter:

| Compound | Company | Status | CNS Penetration |
|----------|---------|--------|-----------------|
| Fostamatinib (Tavalisse) | Rigel | FDA-approved (ITP) | Poor |
| Entospletinib (GS-9973) | Gilead | Phase III | Poor |
| Cerdulatinib | Portola | Discontinued | Unknown |
| Tamatinib | Rigel | Discontinued | Unknown |
| R406 (prodrug of fostamatinib) | Rigel/AstraZeneca | Phase II (RA) | Moderate (tested in stroke—PMID: 26682775) |

R406 in CNS Disease: A Phase II trial of fostamatinib in acute ischemic stroke (NCT02924870) showed modest efficacy signal. This demonstrates that SYK inhibition can achieve some CNS activity, though whether it's sufficient for microglial modulation is unclear.

Critical Limitation: SYK inhibitors block SYK globally—they don't selectively inhibit the TYROBP ITAM arm while preserving DAP10 signaling. TREM2-DAP10 heterodimers would also be affected, eliminating the hypothesized selectivity.

Competitive Landscape:

• No CNS-active SYK inhibitors are in AD development
• Multiple programs for stroke, autoimmune disease
• Interest in neuroinflammatory indications (MS, ALS) but no active AD programs

Safety Concerns:

| Risk | Severity | Monitoring |
|------|----------|------------|
| Cytopenias (neutropenia, thrombocytopenia) | Moderate | CBC monitoring |
| Hypertension | Mild | Blood pressure |
| Diarrhea | Mild | Symptom assessment |
| Infections | HIGH | Immunosuppression risk in elderly AD population |
| Hepatotoxicity | Moderate | LFT monitoring |

The safety profile is acceptable for life-threatening conditions (ITP, lymphoma). For a chronic neurodegenerative disease in the elderly, these risks are prohibitive.

Revised Confidence for Translation: 0.35
High target tractability (SYK is druggable) undermined by lack of selectivity and poor BBB penetration. The safety profile is unsuitable for chronic AD prevention.

Hypothesis 3: NFAT/HDAC Modulation

Target Druggability: MODERATE

This hypothesis requires either:

Selective inhibition of microglial NFAT signaling, OR

HDAC inhibitors that preserve microglial homeostatic function

NFAT Pathway Chemical Matter:

| Compound | Mechanism | Status | CNS Penetration |
|----------|-----------|--------|-----------------|
| Cyclosporine A | Calcineurin inhibitor | FDA-approved (transplant) | Good |
| Tacrolimus (FK506) | Calcineurin inhibitor | FDA-approved | Good |
| INCA-1 | NFAT-specific calcineurin inhibitor | Preclinical | Moderate |
| VIVIT peptide | NFAT inhibitor peptide | Research tool | Cell-permeable versions exist |

The Problem: Cyclosporine A and tacrolimus are profound immunosuppressants with significant nephrotoxicity and neurotoxicity. They are absolutely contraindicated for chronic use in neurodegeneration.

HDAC Inhibitor Chemical Matter:

| Compound | Class | Status | CNS Penetration |
|----------|-------|--------|-----------------|
| Vorinostat (Zolinza) | HDAC1/2/3/6 inhibitor | FDA-approved (CTCL) | Moderate |
| Romidepsin | HDAC1/2/3 inhibitor | FDA-approved (CTCL) | Moderate |
| Pracinostat | Pan-HDAC | Phase III (AML) | Good |
| HDAC1-selective inhibitors | Research compounds | Preclinical | Unknown |

Critical Limitation: The hypothesis assumes HDAC1/HDAC inhibition prevents DAM formation while preserving homeostatic function. This is mechanistically implausible—HDAC inhibitors affect all HDAC-dependent transcriptional programs. No selective microglial HDAC inhibitor exists or is on the horizon.

Safety Concerns:

| Risk | Severity |
|------|----------|
| Thrombocytopenia | Moderate |
| GI toxicity | Mild |
| Fatigue | Mild |
| Cardiac arrhythmias (romidepsin) | Severe |
| Profound immunosuppression (CsA/FK506) | Severe |

Revised Confidence for Translation: 0.15
The mechanism is fundamentally undermined by the reversibility of DAM signatures (Leyns et al., 2017). No selective tool compounds exist. HDAC inhibitors have failed in neurodegenerative disease trials (creatin for Huntington's—NCT02139886 showed no benefit).

Hypothesis 4: CLEC7A/FcRγ Enhancement

Target Druggability: VERY LOW

This is the weakest hypothesis from a drug development perspective.

Existing Chemical Matter:

• β-glucans (curdlan, laminarin) — CLEC7A agonists, research tools only
• Anti-CLEC7A antibodies — no CNS programs exist
• FcRγ expression enhancers — none identified

Critical Limitations:

CLEC7A signaling still requires ITAM-bearing adaptors (FcRγ or DAP12). TYROBP deletion removes one ITAM adaptor; the hypothesis assumes FcRγ can fully compensate, but this hasn't been demonstrated in microglia.

No pharmacological tool exists to selectively enhance CLEC7A-FcRγ signaling.

The ligand for CLEC7A in neurodegeneration is undefined—no known brain-derived CLEC7A agonist exists.

Competitive Landscape: None. CLEC7A is not an active drug target for neurodegeneration.

Safety Concerns: CLEC7A is a pattern recognition receptor involved in antifungal immunity. Enhancing its signaling could:

• Trigger inappropriate inflammatory responses
• Cause off-target immune activation
• Worsen neuroinflammation if the wrong ligands are engaged

Revised Confidence for Translation: 0.15
No chemical matter exists. Mechanism is speculative. This hypothesis is best viewed as a potential mechanistic explanation for compensatory pathways rather than a therapeutic strategy.

Hypothesis 5: Stage-Dependent TYROBP Inhibition

Target Druggability: LOW

The hypothesis requires timing rather than a different target, which is both a strength and weakness.

Existing Chemical Matter:

• TYROBP ASOs — no public programs exist
• TYROBP CRISPR/epigenetic editing — preclinical, no CNS programs
• General myeloid modulators — multiple programs

What This Hypothesis Really Requires:
A pharmacological agent that can be:

Given early in disease (preventive setting)

Withdrawn (to avoid long-term complications)

Safe enough for pre-symptomatic use

The Timing Problem:
The hypothesis assumes a therapeutic window exists between:

• Early TYROBP-driven pathology (target)
• Late TYROBP-dependent plaque clearance (off-target harm)

But Nasu-Hakola disease demonstrates that complete TYROBP loss is devastating. Even if early inhibition is beneficial, the window is likely narrow and the stakes are high.

Biomarker Requirement:
The hypothesis requires biomarkers to identify the intervention window:

• pTau181, pTau217 (amyloid-tau staging)
• NfL (neurodegeneration)
• Microglial imaging (TSPO-PET? — limited utility)

Competitive Landscape:

• TREM2 agonists are being studied in early AD (Biogen AL002)
• Timing/biomarker-driven trials are standard in AD

Safety Concerns:

• If TYROBP is essential for microglial survival, even transient inhibition could deplete microglial populations
• Long-term effects of partial TYROBP reduction are unknown
• Nasu-Hakola disease emerges in the 3rd-4th decade—any intervention window would need to avoid effects over decades

Revised Confidence for Translation: 0.30
The timing hypothesis has merit conceptually (supported by TREM2 R47H data) but faces the fundamental obstacle of Nasu-Hakola disease and the lack of any TYROBP-targeting agent.

Hypothesis 6: Peripheral-Selective TYROBP Inhibition

Target Druggability: VERY LOW

Cell-type selective CNS targeting remains an unsolved problem in drug development.

Existing Chemical Matter:

• CCR2 antagonists (PF-04136309, cenicriviroc) — in clinical trials for liver fibrosis, HIV
• CD11b agonists — none identified
• TYROBP cell-type selective agents — none possible with current technology

The Technical Challenge:
Achieving peripheral TYROBP inhibition while sparing microglia requires:

A compound that doesn't cross the BBB, AND

Selectivity for peripheral myeloid cells over microglia

This is not achievable with current pharmacology. The BBB is permeable to small molecules <400-500 Da; most selective agents would still enter the brain. Conversely, ASOs and antibodies that don't cross the BBB won't reach peripheral myeloid cells either (they're too large to enter either compartment efficiently from systemic circulation).

Partial Solutions:

• Denali Therapeutics' BBB platform — engineered Fc fragments with transferrin receptor-mediated transport. Could potentially be directed toward peripheral cells by avoiding BBB-targeting engineering.
• Intrathecal delivery — would target CNS but not peripheral cells (opposite of what's needed)
• Bone marrow-directed delivery — experimental gene therapy approaches

Competitive Landscape:

• Cenicriviroc (Allergan/AbbVie) — CCR2/CCR5 antagonist, Phase III for NASH, Phase II for HIV-associated fibrosis
• No programs specifically targeting peripheral myeloid TYROBP

Safety Concerns:

• CCR2 antagonists cause immunosuppression in tissues that depend on monocyte recruitment
• Impact on bone marrow homeostasis unknown
• Peripherally-restricted immune suppression could increase infection risk

Revised Confidence for Translation: 0.20
An elegant hypothesis undermined by the absence of technology to achieve cell-type selective targeting of TYROBP. This is a 10-15 year development challenge, not an imminent therapeutic strategy.

Hypothesis 7: CR3/CD11b Inhibition

Target Druggability: MODERATE-HIGH

This is the most tractable hypothesis from a drug development perspective.

Existing Chemical Matter:

| Compound | Mechanism | Status | BBB Penetration |
|----------|-----------|--------|-----------------|
| Lifitegrast (Xiidra) | LFA-1 (CD11a) antagonist | FDA-approved (dry eye) | None (topical) |
| Eculizumab (Soliris) | C5 inhibitor | FDA-approved | Poor |
| Ravulizumab (Ultomiris) | C5 inhibitor | FDA-approved | Poor |
| Pegcetacoplan (Syfovre) | C3 inhibitor | FDA-approved (GA) | Poor |
| Natalizumab (Tysabri) | Anti-α4 integrin | FDA-approved | None (peripheral) |
| Anti-CD11b antibodies | CD11b antagonist | Preclinical/Research | Poor |

The Complement Approach:

The downstream effector of CR3-mediated synaptic pruning is complement activation (C1q, C3). Inhibiting complement would block this pathway.

C3 Inhibition:

• Pegcetacoplan (Apellis) — intravitreal C3 inhibitor, FDA-approved for geographic atrophy
• AMY-101 (Amyndas) — C3 inhibitor, Phase II for paroxysmal nocturnal hemoglobinuria and COVID-19
• NPX395 — oral C3 inhibitor, Phase I

Critical Limitation for CNS: Eculizumab, ravulizumab, and pegcetacoplan do not significantly penetrate the CNS. They are systemically restricted by the BBB.

CNS Complement Inhibition:

• Eculizumab requires ~2400 mg/week IV to maintain systemic complement inhibition. CNS concentrations are negligible.
• A CNS-active complement inhibitor doesn't exist, though several programs are exploring this:
• AL0004 (Alder Biopharmaceuticals) — anti-C5 antibodies with brain-penetrant Fc engineering
• Small molecule C1s inhibitors (tremplin, similar) — limited BBB penetration

Competitive Landscape:

| Program | Company | Target | Stage | CNS? |
|---------|---------|--------|-------|------|
| AL002 | Alector/AbbVie | TREM2 agonist | Phase II | Yes |
| AL044 | Alector | TREM2 agonist | Preclinical | Yes |
| HS-001 | HiFiBiO/Takeda | TREM2 agonist | Preclinical | Yes |
| Prion Disease C1q program | undisclosed | C1q inhibitor | Preclinical | Unknown |

AL002 (Alector/AbbVie): Phase II TRAILBLAZER-ALZ2 (NCT04592874) is ongoing. This is the most advanced microglial targeting program in AD.

Safety Concerns:

| Risk | Severity | Relevance to AD |
|------|----------|-----------------|
| Meningococcal infection | Severe | Contraindicated in elderly population |
| Infusion reactions | Moderate | Manageable |
| C3 depletion | Immunosuppression | Increased infection risk |
| Off-target complement inhibition | Unknown | May impair complement-dependent synaptic remodeling |

Synaptic Pruning—Beneficial or Pathological?
This is actively debated. Synaptic pruning is essential for normal brain development. In adults:

• Excessive pruning → neurodegeneration
• Insufficient pruning → epilepsy, autism spectrum disorders (complement deficiency mouse models show this)

Revised Confidence for Translation: 0.45

This is the most promising hypothesis because:

CR3/CD11b targeting is mechanistically specific

Existing complement inhibitors (C3, C5) are approved and well-characterized

A CNS-active complement inhibitor is a tractable development goal

However:

BBB penetration is the critical barrier

The therapeutic index of complement inhibition in the brain is unknown

Long-term complement suppression would likely be contraindicated

Consolidated Drug Development Assessment

| Hypothesis | Druggability | Chemical Matter | BBB? | Safety | Competitive Position | Revised Confidence |
|------------|--------------|------------------|------|--------|---------------------|-------------------|
| 1 (CD33) | Moderate | Yes | No | Moderate | Weak | 0.25 |
| 2 (SYK/ITAM) | High/None | Yes/No | Partial | Poor | Weak | 0.35 |
| 3 (NFAT/HDAC) | Moderate | Yes | Yes | Poor | Weak | 0.15 |
| 4 (CLEC7A) | Very Low | No | No | Unknown | None | 0.15 |
| 5 (Temporal) | Low | No | N/A | Poor | Moderate | 0.30 |
| 6 (Peripheral) | Very Low | No | N/A | Moderate | None | 0.20 |
| 7 (CR3) | Moderate-High | Partial | No | Poor | Strong | 0.45 |

The Path Forward: What Would Actually Move This Field

Tier 1: Immediate Priorities

1. Inducible TYROBP deletion in adult mice

• Critical experiment: The field cannot proceed without determining whether the protective phenotype of germline TYROBP KO reflects developmental compensation or genuine adult-onset therapeutic potential
• Program: If adult deletion recapitulates neuroprotection → proceed; if not → all hypotheses require revision
• Timeline: 12-18 months for definitive experiment

2. Develop CNS-penetrant SYK inhibitors

• The R406 stroke data (PMID: 26682775) suggests this is achievable
• Opportunity: A CNS-optimized SYK inhibitor could test Hypothesis 2 and serve as a tool for target validation
• Companies: Rigel has no active CNS SYK program; this represents a partnership opportunity

3. Develop CNS-penetrant complement inhibitors

• Apellis, Alexion, and others have platforms for brain-penetrant complement inhibitors
• Opportunity: A CNS C3 or C1s inhibitor could test Hypothesis 7
• Competitive angle: C1q inhibition (proximal to CR3) may be safer than systemic complement blockade

Tier 2: Medium-Term Strategies

4. ASO-mediated TYROBP reduction

• Ionis/Roche/Biogen have platforms for CNS-delivered ASOs
• Opportunity: Test whether partial TYROBP reduction (50-70%) is sufficient for benefit without causing Nasu-Hakola-like pathology
• Key experiment: Dose-response study in 5xFAD mice with inducible ASO delivery
• Timeline: 24-36 months for preclinical validation

5. Bifunctional molecules targeting CD33 without affecting TREM2

• Concept: A molecule that simultaneously engages CD33 (blocking its ITIM signaling) while not affecting TREM2-TYROBP complexes
• Chemical matter: SIGLEC-engaging antibodies with engineered Fc regions
• Challenge: Mechanistically complex—requires understanding CD33's signaling hierarchy

6. Biomarker development for intervention timing

• Need: Validate microglial activation state biomarkers that identify the TYROBP-dependent phase
• Candidates: CSF sTREM2 (well-validated), CCL2, IL-6, NfL
• Purpose: Would enable the temporal targeting strategy (Hypothesis 5)

Tier 3: Long-Term Approaches

7. Cell-type selective delivery

• Technology gap: No current method achieves peripheral TYROBP inhibition while sparing microglia
• Possible approaches:
• Denali's BBB platform (Fc-engineering for directional transport)
• Targeted bone marrow delivery
• RNA delivery vehicles with cell-type tropism
• Timeline: 5-10 years for clinical application

8. Epigenetic modulators for microglial state switching

• Current state: No selective HDAC inhibitors for microglia
• Possible approach: CRISPR-based epigenetic editing (dCas9 fusions with HDAC domains) delivered via AAV
• Timeline: Research tool development 3-5 years; clinical application far longer

Competitive Landscape Summary

| Company | Program | Mechanism | Stage | Relevance |
|---------|---------|-----------|-------|-----------|
| Alector/AbbVie | AL002 | TREM2 agonist | Phase II | Dominant; validates microglial targeting |
| Alector | AL044 | TREM2 agonist | Preclinical | Next-gen approach |
| Biogen | Alutardastat (TREM2) | TREM2 agonist | Preclinical | Follows AL002 |
| Roche | Anti-TREM2 antibodies | TREM2 agonism | Preclinical | No current IND |
| Apellis | Pegcetacoplan | C3 inhibitor | Approved (eye) | CNS programs emerging |
| Alexion | Ravulizumab | C5 inhibitor | Approved | CNS programs in development |
| Rigel | Fostamatinib | SYK inhibitor | Approved (ITP) | Could be repurposed for stroke |
| Denali | BBB platform | Transport vehicle | Preclinical | Enables CNS delivery of biologics |

Strategic Positioning:

• The TREM2 agonist space is the most advanced and best-funded (Alector/AbbVie Phase II, >$500M partnership)
• Complement inhibitors for neurodegeneration are emerging (Apellis moving into CNS)
• SYK inhibitors exist but have no active AD program
• TYROBP-specific programs do not exist—this represents an opportunity or a risk, depending on whether the hypothesis is correct

Safety Profile Summary: What Stops These Programs

| Hypothesis | Critical Safety Issue | Likelihood of Resolution |
|------------|----------------------|-------------------------|
| 1 | CD33 on neutrophils; BBB penetration | Medium (BBB platform tech advancing) |
| 2 | SYK inhibitors: immunosuppression, poor CNS penetration | Low (CNS penetration unlikely to improve safety) |
| 3 | Calcineurin inhibitors: nephrotoxicity, immunosuppression | Very Low (mechanism is inherently unsafe) |
| 4 | Unknown; no compounds exist to test | N/A |
| 5 | Nasu-Hakola disease; narrow therapeutic window | Very Low |
| 6 | Peripherally-restricted delivery doesn't exist | Very Low |
| 7 | Complement inhibitors: meningococcal infection | Medium (prophylaxis exists; contraindicated for many elderly) |

Bottom Line: Is TYROBP a Viable Target?

No—not as a direct target.

The seven hypotheses represent sophisticated mechanistic thinking but face insurmountable obstacles:

Nasu-Hakola disease is a biological proof-of-concept failure for TYROBP inhibition as a therapeutic strategy

BBB penetration is a barrier for all protein targets in this pathway

Lack of selective pharmacological tools undermines hypotheses requiring selective pathway inhibition

Safety concerns (immunosuppression, complement deficiency) make systemic modulation untenable

The productive path forward is indirect targeting:

• TREM2 agonists (AL002) are the most advanced approach—these activate the protective arm without blocking TYROBP, and may provide more selectivity than TYROBP inhibition
• Complement inhibitors (C3, C1q) could block downstream synaptic pruning without affecting microglial survival signaling
• Cell-type selective delivery is the enabling technology that would unlock peripheral-specific approaches

Highest-value experiments to de-risk this field:

Inducible TYROBP KO in adult mice (definitively addresses developmental compensation)

Dose-response of partial TYROBP reduction (addresses Nasu-Hakola concern)

CNS-penetrant complement inhibitor in amyloid models (addresses Hypothesis 7)

If forced to bet on one hypothesis from a drug development standpoint: Hypothesis 7 (CR3/complement inhibition) has the best combination of mechanistic plausibility, existing chemical matter, and tractable development path—provided a CNS-penetrant complement inhibitor can be developed.