Test
Mechanism: Loss-of-function TREM2 variants impair microglial lipid metabolism and phagocytosis, reducing clearance of amyloid plaques and promoting a neurotoxic microglial state. Enhancing TREM2 signaling may shift microglia toward a neuroprotective phenotype.
Target: TREM2 (Triggering Receptor Expressed on Myeloid Cells 2)
Supporting Evidence:
- TREM2 R47H variant increases Alzheimer's disease risk ~3-fold (PMID: 24285345)
- TREM2 deficiency exacerbates amyloid pathology in mouse models (PMID: 29607930)
- TREM2 agonistic antibodies promote microglial plaque compaction and reduce neuritic dystrophy (PMID: 31178183)
Predicted Experiment: Administer TREM2 agonistic antibody (developed by companies including Dolbey/AbbVie, Alector) in 5xFAD mice; assess microglial transcriptional changes via single-cell RNA-seq, amyloid plaque burden, and cognitive performance on Barnes maze.
Confidence: 0.75
---
Mechanism: Alpha-synuclein aggregates activate NLRP3 inflammasome in microglia via TLR4/NF-κB priming and lysosomal damage, producing mature IL-1β that drives dopaminergic neuron loss and propagated α-syn pathology.
Target: NLRP3 inflammasome; IL-1β/IL-18 axis
Supporting Evidence:
- α-synuclein fibrils activate NLRP3 in cultured microglia (PMID: 29097672)
- NLRP3 knockout or MCC950 inhibitor protects dopaminergic neurons in MPTP and α-syn preformed fibril models (PMID: 28751425, 30659162)
- Elevated NLRP3/caspase-1 in substantia nigra of PD patients (PMID: 29675268)
Predicted Experiment: Test MCC950 (NLRP3 inhibitor) in human α-synuclein AAV overexpression rat model; perform behavioral assessments, stereological counting of TH+ neurons, and measure CSF cytokine levels.
Confidence: 0.70
---
Mechanism: Haploinsufficiency of progranulin (GRN) leads to impaired lysosomal function due to decreased TFEB nuclear translocation, resulting in accumulation of lipofuscin, enlarged lysosomes, and neuronal vulnerability. Enhancing TFEB activity or progranulin levels may restore lysosomal homeostasis.
Target: TFEB (Transcription Factor EB) / progranulin pathway
Supporting Evidence:
- Grn-/- mice exhibit lysosomal dysfunction, lipofuscin accumulation, and microgliosis (PMID: 18819921, 21737487)
- Progranulin localizes to lysosomes and regulates cathepsin activity (PMID: 22958956)
- TFEB overexpression enhances lysosomal biogenesis and reduces pathological accumulation in storage diseases (PMID: 22107871)
Predicted Experiment: Cross TFEB transgenic mice with Grn-/- mice; assess lysosomal markers (Lamp1, Cathepsin D), inflammatory markers, and survival. Alternatively, use AAV9-TFEB delivery to CNS.
Confidence: 0.68
---
Mechanism: C9orf72 repeat expansions produce toxic dipeptide repeats (DPRs) that impair nucleocytoplasmic transport and autophagy. Enhancing autophagy via mTOR inhibition or beclin-1 activation may reduce DPR accumulation.
Target: Autophagy machinery; C9orf72/Rab39B pathway
Supporting Evidence:
- Poly-GA DPRs aggregate with RBP4 and impair proteasome/autophagy (PMID: 25297118)
- C9orf72 knockout in mice causes lysosomal accumulation and neurodegeneration (PMID: 25624326)
- Rapamycin ameliorates neurodegeneration in Drosophila C9 models (PMID: 26146185)
Predicted Experiment: Treat C9-BAC transgenic mice with rapamycin or novel mTOR inhibitors; assess DPR burden by biochemistry, nuclear transport markers (Ran, importins), and motor behavior.
Confidence: 0.62
---
Mechanism: Reactive astrocytes in neurodegeneration acquire either neurotoxic "A1" (induced by Il-1α, TNFα, C1q) or neuroprotective "A2" phenotype. Blocking A1 inducers or enhancing A2 genes (e.g., Lcn2, Timp1) may restore astrocyte homeostatic function.
Target: Astrocyte regulatory genes; complement component C3
Supporting Evidence:
- LPS-activated microglia induce A1 astrocytes via Il-1α/TNFα/C1q (PMID: 29107332)
- A1 astrocytes fail to support neuronal survival and synaptogenesis (PMID: 29107332)
- C3a receptor deficiency worsens disease in ALS models (PMID: 31988378)
Predicted Experiment: Generate GFAP-Cre; Il1rn-flox mice to conditionally block IL-1 signaling specifically in astrocytes; cross with SOD1-G93A mice; assess ALS progression and astrocyte transcriptional profile.
Confidence: 0.65
---
Mechanism: Overactivation of complement C1q/C3 pathways in early AD leads to excessive microglia-mediated synapse loss (synaptodendritis) before plaque deposition, contributing to cognitive decline.
Target: C1q, C3, CR3 complement receptors
Supporting Evidence:
- C1q localizes to synapses in early AD; C3 deposition on synapses correlates with memory loss (PMID: 29445953)
- C3 knockout or CR3 deficiency protects synapses in 5xFAD mice (PMID: 29445953)
- C1q inhibition via blocking antibody reduces synapse loss in mouse models (PMID: 29445953)
Predicted Experiment: Test anti-C1q antibody (from Roche/Genentech) in 5xFAD mice at early stages (2 months); perform synaptic proteomics, electrophysiology (LTP), and cognitive testing.
Confidence: 0.78
---
Mechanism: Motor neurons exhibit unique metabolic dependencies (glycolysis, lipid metabolism) that become dysregulated in ALS. Targeting PDH kinase (PDK) or SGLT2 to shift toward glucose oxidation may restore energy homeostasis and extend survival.
Target: Metabolic regulators; PDK, SGLT2, or PGC-1α
Supporting Evidence:
- Motor neurons have reduced PDH activity and prefer glycolysis (PMID: 29656893)
- SGLT2 expression is elevated in ALS motor neurons; SGLT2 inhibition extends survival in SOD1 mice (PMID: 34193629)
- PGC-1α dysregulation observed in ALS patients and models (PMID: 18391957)
Predicted Experiment: Treat SOD1-G93A mice with SGLT2 inhibitor (empagliflozin); perform metabolic Seahorse assays on spinal cord motor neurons, assess mitochondrial function, and measure disease progression.
Confidence: 0.58
---
| # | Hypothesis | Target | Confidence |
|---|-----------|--------|------------|
| 1 | TREM2 agonism | TREM2 | 0.75 |
| 2 | NLRP3 inhibition | NLRP3/Casp1 | 0.70 |
| 3 | TFEB restoration | TFEB pathway | 0.68 |
| 4 | Autophagy enhancement | mTORC1 | 0.62 |
| 5 | Astrocyte reprogramming | C3/C1q | 0.65 |
| 6 | Complement inhibition | C1q/C3 | 0.78 |
| 7 | Metabolic correction | PDK/SGLT2 | 0.58 |
Weak Links:
- Dose-dependency unknown; excessive TREM2 activation may promote neurotoxic microglial states
- Human TREM2 variants show conflicting functional effects; R47H may be a loss-of-function but others may be gain-of-function
- Plaque compaction could paradoxically concentrate toxins in microenvironments
Counter-Evidence:
- TREM2 R47H knock-in mice show milder phenotypes than knockout models, questioning disease relevance
- Recent phase II trials of TREM2 antibodies (H. analog) failed to meet primary endpoints (Alector/AbbVie AL002c, 2023)
Falsifying Experiments:
- Test TREM2 agonism in human iPSC-derived microglia in 3D cultures
- Assess long-term outcomes beyond plaque burden (cognitive trajectories, neuronal integrity)
- Determine if TREM2 activation benefits only at specific disease stages
---
Weak Links:
- MCC950, while potent in rodents, has poor CNS penetration and failed toxicology studies for clinical development
- Alpha-synuclein's precise NLRP3 activation mechanism remains debated; may require extracellular aggregation or lysosomal rupture
- Compensatory inflammasome pathways (AIM2, NLRC4) may limit long-term efficacy
Counter-Evidence:
- Large PD genetic studies show no strong NLRP3 variant associations
- Late-stage clinical trials with NLRP3 inhibitors in other diseases showed limited efficacy
- Some evidence that IL-1β may have neuroprotective roles in early PD
Falsifying Experiments:
- Test in chronic toxin models (e.g., rotenone) rather than acute MPTP
- Assess if NLRP3 inhibition prevents progression vs. just slowing initiation
- Verify mechanism in human post-mortem tissue (do NLRP3+ microglia colocalize with α-syn?)
---
Weak Links:
- GRN's lysosomal role is established but direct TFEB regulation is inferred, not proven
- TFEB overexpression may cause non-specific autophagy with unwanted side effects
- Human GRN haploinsufficiency may involve non-lysosomal pathways (e.g., wound healing, synaptic function)
Counter-Evidence:
- Grn-/- mice do not fully replicate human FTD pathology (no TDP-43 inclusions)
- TFEB activation may not correct all GRN functions (GRN has extracellular roles)
- No clear genetic link between TFEB and FTD
Falsifying Experiments:
- Confirm direct GRN-TFEB interaction in human neurons
- Test in newer FTD-GRN iPSC models with TDP-43 pathology
- Assess whether TFEB activation addresses non-lysosomal GRN functions
---
Weak Links:
- Rapamycin has pleiotropic effects; benefits in Drosophila may not translate to mammals
- C9orf72haploinsufficiency vs. DPR toxicity debate unresolved; both may contribute
- mTOR inhibition may impair other critical neuronal pathways
Counter-Evidence:
- C9-BAC mice show variable DPR accumulation and unclear behavioral phenotypes
- Rapamycin has limited efficacy in mouse C9 models compared to Drosophila
- Autophagy enhancement may not address nucleocytoplasmic transport defects
Falsifying Experiments:
- Dissociate DPR toxicity from C9orf72 loss-of-function in vivo
- Test more specific autophagy inducers (e.g., calpastatin inhibitors)
- Validate nuclear transport restoration as upstream mechanism
---
Weak Links:
- A1/A2 dichotomy is oversimplified; astrocytes exhibit spectrum states
- Human astrocytes differ significantly from rodent counterparts in development and function
- C3/C1q roles may be context-dependent rather than universally detrimental
Counter-Evidence:
- C3a receptor knockout worsened ALS in one study but other complement components show neuroprotective roles
- Loss of astrocyte reactivity may impair CNS homeostasis and repair
- GFAP-driven interventions may miss important astrocyte subpopulations
Falsifying Experiments:
- Single-cell profiling of human astrocytes from ALS/AD patients
- Test interventions in human astrocytes with proper maturation
- Assess if A1 blockade maintains necessary defensive functions
---
Weak Links:
- C1q/C3 inhibition may impair necessary synaptic pruning during development
- Timing of intervention is critical; complement may have dual roles
- Mouse models may not capture human AD synaptic vulnerability patterns
Counter-Evidence:
- C1q localizes to synapses early in AD, but causality not proven
- C3 deficiency could impair debris clearance needed for repair
- Roche/Genentech's anti-C1q antibody trials in other diseases show limited efficacy
Falsifying Experiments:
- Determine developmental vs. adult-specific roles of complement in synapse loss
- Test in aged 5xFAD mice with more human-like pathology
- Assess safety of chronic complement inhibition in CNS
---
Weak Links:
- SGLT2 findings are correlative; causal role not established
- Metabolic therapies may work in SOD1 models but translate poorly to sporadic ALS
- PDK/glycolysis shift may be compensatory, not causative
Counter-Evidence:
- SOD1-G93A mice show high variability in metabolic phenotypes
- SGLT2 inhibitors in humans (diabetes trials) did not show cognitive benefits
- PGC-1α dysregulation may be downstream, not a primary driver
Falsifying Experiments:
- Test in larger animal models (e.g., pigs) with chronic progression
- Assess if metabolic correction addresses TDP-43 pathology
- Determine if benefits are neuron-autonomous or involve other cell types
---
Overarching Weaknesses:
1. Translational gaps: Most hypotheses rely heavily on mouse models that incompletely replicate human disease
2. Single-target focus: Neurodegeneration likely requires addressing multiple pathways simultaneously
3. Stage-dependency: Interventions may only be effective at specific disease stages, rarely assessed
4. Genetic vs. sporadic relevance: SOD1, FTD-GRN, C9 models are familial but represent <10% of cases
Most Defensible Hypotheses:
- Hypothesis 6 (Complement) has strongest human tissue evidence but needs safety validation
- Hypothesis 1 (TREM2) has most clinical development but recent trial failures caution optimism
Most Hypothetical:
- Hypothesis 7 (Metabolism) relies on correlative data without proven mechanism
- Hypothesis 4 (Autophagy) has Drosophila data but unclear mammalian efficacy
Priority Falsification Experiments:
1. Test all hypotheses in human iPSC-derived neurons/glia in 3D systems
2. Assess interventions at multiple disease stages in chronic models
3. Validate mechanisms in human post-mortem tissue
4. Dissociate genetic vs. acquired contributions in each model
This analysis suggests the field needs more human-relevant validation before clinical translation, and that combination approaches may be more promising than single-target interventions.
As of April 2026, I would treat these as three translationally viable programs, two discovery-stage mechanisms, and two deprioritized concepts.
Feasibility Ranking
| Rank | Hypothesis | Feasibility | Development Call |
|---:|---|---|---|
| 1 | NLRP3 inhibition in PD | Moderate-high | Best current “surviving” idea because oral CNS-penetrant inhibitors now exist and PD Phase 2 has begun |
| 2 | Complement/C1q inhibition in AD | Moderate | Strong biology and biomarkers, but AD trial design and chronic immune safety are hard |
| 3 | TREM2 activation in AD | Low-moderate | Druggable and clinically tested, but AL002 Phase 2 was biomarker-positive and clinically negative |
| 4 | GRN/lysosomal restoration in FTD-GRN | Low-moderate | Genetic rationale strong; direct TFEB activation is risky, and progranulin replacement failed Phase 3 clinically |
| 5 | Astrocyte reactivity modulation | Low | Biologically real but target definition is too immature for near-term clinical development |
| 6 | C9orf72 autophagy enhancement | Low | Too nonspecific; better to pursue allele/RNA/DPR-directed approaches |
| 7 | ALS metabolic correction via SGLT2/PDK | Very low | Cheap to test, but weak causality and poor ALS translational precedent |
1. NLRP3 Inhibition In PD
This is the most development-ready of the set. The target is druggable with oral small molecules, and the field has moved beyond MCC950. Dapansutrile has entered a 12-month Phase 2 PD trial, DAPA-PD, and NT-0796 has reported Phase 1b PD/elderly volunteer data with CSF exposure and tolerability signals. That materially improves feasibility versus the original MCC950 proposal.
Useful biomarkers: CSF/plasma IL-18, IL-1beta pathway markers, NfL, inflammatory transcriptomics, microglial PET if available, alpha-syn seed amplification as enrichment/stratification rather than a short-term pharmacodynamic marker.
Best models: alpha-syn PFF mouse/rat, AAV-alpha-syn, PLP-alpha-syn for MSA-like biology, plus human iPSC microglia-neuron co-culture. MPTP is acceptable for mechanism but weak for disease modification.
Main constraints: PD progression is slow, symptomatic noise is high, and anti-inflammatory effects may not translate to clinical slowing. A credible Phase 2 needs 12-18 months, biomarker enrichment, and MDS-UPDRS plus digital motor endpoints.
Safety: infection risk is lower than broad immunosuppression but still needs chronic surveillance; liver/toxicity history around older NLRP3 chemistry matters.
Realistic path: 2-3 years for Phase 2 signal; 6-8 years and roughly $250M-$600M to approval if Phase 2 is positive.
Sources: DAPA-PD registry, Olatec Phase 2 announcement, NT-0796 Phase 1b report, dapansutrile preclinical synucleinopathy paper.
https://www.isrctn.com/ISRCTN16806940
https://www.prnewswire.com/news-releases/first-parkinsons-patients-treated-in-landmark-phase-2-trial-of-dapansutrile--expanding-a-leading-oral-nlrp3-inhibitor-into-neurological-diseases-302723428.html
https://pubmed.ncbi.nlm.nih.gov/40792655/
https://link.springer.com/article/10.1186/s12974-026-03716-3
2. Complement/C1q Inhibition In AD
The biology is strong: complement tagging of synapses is plausibly causal in early neurodegeneration. The target is druggable by antibodies, and C1q blockade has human clinical experience through Annexon’s ANX005 platform, including CNS target engagement in Huntington’s disease. However, there is no clean late-stage AD validation.
Useful biomarkers: CSF C1q/C3/C4a, synaptic injury markers such as neurogranin and SNAP-25, NfL, amyloid/tau PET for staging, EEG/LTP-like physiology, and cognitive composites in biomarker-confirmed early AD.
Best models: early-intervention amyloid models, tauopathy models, aged animals, and human iPSC neuron-microglia-astrocyte systems with complement-mediated synapse engulfment assays.
Main constraints: timing is everything. Treating established dementia may be too late; presymptomatic or very early amyloid-positive cohorts are expensive and long. Chronic complement blockade also creates safety scrutiny.
Safety: infection risk, immune-complex handling, autoimmunity signals, infusion reactions. C1q-specific blockade is cleaner than terminal complement blockade, but chronic CNS use still requires caution.
Realistic path: if a clinical candidate already exists, 3-4 years to AD proof-of-concept; 7-10 years and $500M-$1B+ to approval.
Sources: Annexon ANX005 HD Phase 2 data and ANX005 nonclinical development.
https://ir.annexonbio.com/news-releases/news-release-details/annexon-biosciences-reports-phase-2-clinical-trial-results
https://pubmed.ncbi.nlm.nih.gov/29202623/
3. TREM2 Activation In AD
TREM2 is clearly druggable by antibodies and genetically compelling. The problem is clinical translation. AL002 showed target engagement and microglial pharmacodynamics, but failed to slow CDR-SB progression, did not improve secondary clinical endpoints, and did not show favorable AD fluid biomarker or amyloid PET effects in Phase 2. That does not kill TREM2 biology, but it strongly weakens the “agonist antibody in early AD” version.
Useful biomarkers: CSF sTREM2, osteopontin, microglial activation markers, amyloid/tau PET, NfL, GFAP, synaptic biomarkers. The AL002 result shows target engagement alone is insufficient.
Best models: human iPSC microglia in amyloid/tau 3D systems, TREM2 variant-stratified assays, aged amyloid/tau models. Standard 5xFAD plaque reduction is no longer enough.
Main constraints: patient selection, disease stage, and whether activation needs to be tuned rather than maximized. Combination with amyloid removal may be more plausible than monotherapy.
Safety: excessive microglial activation, inflammatory worsening, ARIA interaction if combined with anti-amyloid antibodies.
Realistic path: a redesigned next-generation TREM2 program would need 3-5 years before a convincing Phase 2 readout; full approval path likely 8-10 years and $600M-$1B+.
Sources: Alector AL002 Phase 2 results and Nature Medicine trial publication.
https://investors.alector.com/news-releases/news-release-details/alector-announces-results-al002-invoke-2-phase-2-trial/
https://www.nature.com/articles/s41591-026-04273-1
4. GRN/TFEB Lysosomal Restoration In FTD-GRN
The genetic disease is attractive, but TFEB itself is a hard target. Direct TFEB activation risks broad lysosomal/autophagy perturbation, oncogenic/metabolic concerns, and poor dose control. Progranulin restoration is more rational than TFEB overexpression, but latozinemab raised progranulin and still failed the Phase 3 clinical endpoint in FTD-GRN.
Useful biomarkers: plasma/CSF progranulin, lysosomal proteins, cathepsins, NfL, GFAP, vMRI atrophy, TDP-43-related emerging biomarkers if validated.
Best models: GRN patient iPSC neurons/microglia/organoids with TDP-43 and lysosomal phenotypes; knock-in/haploinsufficient models are preferable to Grn knockout alone.
Main constraints: rare disease recruitment, slow/heterogeneous progression, need for presymptomatic intervention, and uncertain link between correcting lysosomal markers and preserving brain function.
Safety: progranulin overexpression has theoretical cancer/wound-healing risks; TFEB overactivation could disrupt cellular clearance and metabolism.
Realistic path: TFEB therapy is preclinical, 8-12 years and $300M-$800M. A better path is BBB-enabled progranulin, SORT1 inhibition, or gene therapy, but Phase 3 failure raises the evidentiary bar.
Source: Alector latozinemab Phase 3 results.
https://investors.alector.com/news-releases/news-release-details/alector-announces-topline-results-latozinemab-phase-3-trial/
5. Astrocyte Reactivity Reprogramming
Not trial-ready. The A1/A2 framework is useful historically but too coarse for drug development. “Block A1” is not a target product profile. C3-positive astrocytes, lipid toxicity, cytokine-induced states, and region-specific astrocyte programs are all plausible, but the field needs sharper targets.
Useful biomarkers: astrocytic GFAP, YKL-40, C3-related markers, spatial transcriptomics signatures, CSF inflammatory panels. No validated patient-selection biomarker exists.
Best models: human mature astrocyte-neuron-microglia tri-cultures, spatial transcriptomics in human tissue, disease-specific ALS/AD iPSC systems.
Safety: astrocyte reactivity is also reparative. Blunting it may impair BBB maintenance, synapse support, glutamate buffering, and injury response.
Realistic path: 3-5 years to define a druggable node; 10+ years and high attrition to approval.
Source: recent astrocyte heterogeneity review.
https://pubmed.ncbi.nlm.nih.gov/38891053/
6. C9orf72 Autophagy Enhancement
The disease biology is real, but rapamycin/mTOR inhibition is too blunt. C9 disease includes repeat RNA toxicity, DPRs, nucleocytoplasmic transport defects, and C9orf72 loss-of-function. Autophagy enhancement may lower aggregate burden but is unlikely to address the full mechanism. The failure of C9orf72 sense-repeat ASOs despite target engagement also warns that biomarker lowering may not equal clinical benefit.
Useful biomarkers: CSF poly(GP), poly(GA), NfL, DPR burden in model systems, nuclear transport assays. Clinical endpoints in ALS are fast but noisy; FTD endpoints are slower.
Best models: C9 patient iPSC motor neurons/cortical neurons, organoids, BAC mice only if phenotype is robust, and stress-granule/nucleocytoplasmic transport assays.
Safety: chronic mTOR inhibition causes immunosuppression, metabolic effects, wound-healing impairment, and may harm neurons.
Realistic path: rapamycin-like autophagy enhancement is not worth major investment. A specific autophagy/DPR-clearance modulator would be 7-10+ years and $300M-$700M.
Sources: BIIB078 discontinuation and C9 ASO failure analysis.
https://www.als.org/stories-news/biogen-and-ionis-discontinue-c9-program-after-phase-1-study-did-not-show-clinical
https://academic.oup.com/brain/article/147/8/2607/7684228
7. ALS Metabolic Correction Via SGLT2/PDK
This is the weakest. SGLT2 inhibitors are druggable, cheap, and safe in diabetes/heart failure populations, so a small repurposing study is feasible. But causal linkage to ALS progression is weak, CNS exposure may be limited, and SOD1 mouse metabolic rescue is a poor predictor for sporadic ALS. PDK modulation with agents like dichloroacetate has tolerability and neuropathy concerns.
Useful biomarkers: body composition, hypermetabolism measures, plasma/CSF metabolomics, lactate/pyruvate, NfL, respiratory decline, ALSFRS-R slope.
Best models: multiple ALS genotypes, TDP-43 models, human motor neuron-glia metabolic assays. SOD1-G93A alone is insufficient.
Safety: SGLT2 inhibitors carry risks of euglycemic ketoacidosis, dehydration, genital infections, and weight loss, which is concerning in ALS cachexia.
Realistic path: small Phase 2 repurposing trial could be done in 2-3 years for $5M-$20M. Probability of disease-modifying success is low.
Bottom Line
The best investment is NLRP3 inhibition in PD, followed by a biomarker-enriched C1q/complement synapse-protection program. TREM2 and GRN/lysosomal biology remain scientifically important but have been clinically damaged by negative late-stage or mid-stage readouts. Astrocyte reprogramming, C9 autophagy, and ALS metabolic correction need substantial target refinement before serious clinical-development spend.
{
"ranked_hypotheses": [
{
"title": "NLRP3 Inflammasome Inhibition for Neuroprotection in Parkinson's Disease",
"description": "Targeted inhibition of NLRP3 inflammasome activation attenuates alpha-synuclein-driven microglial neuroinflammation, reduces IL-1β/IL-18-mediated dopaminergic neuron loss, and may slow PD progression. Oral CNS-penetrant inhibitors (dapansutrile, NT-0796) have entered Phase 2 trials, representing the most translationally viable anti-inflammatory approach in neurodegeneration.",
"target_gene": "NLRP3",
"dimension_scores": {
"evidence_strength": 0.72,
"novelty": 0.65,
"feasibility": 0.78,
"therapeutic_potential": 0.75,
"mechanistic_plausibility": 0.70,
"druggability": 0.80,
"safety_profile": 0.65,
"competitive_landscape": 0.70,
"data_availability": 0.75,
"reproducibility": 0.68
},
"composite_score": 0.72,
"evidence_for": [
{"claim": "Alpha-synuclein fibrils activate NLRP3 in cultured microglia via TLR4/NF-κB priming", "pmid": "29097672"},
{"claim": "MCC950 inhibitor protects dopaminergic neurons in MPTP and α-syn preformed fibril models", "pmid": "28751425"},
{"claim": "Elevated NLRP3/caspase-1 in substantia nigra of PD patients", "pmid": "29675268"},
{"claim": "Dapansutrile entering Phase 2 PD trial (DAPA-PD)", "pmid": "40792655"},
{"claim": "NT-0796 Phase 1b data showing CSF exposure and tolerability in PD/elderly", "pmid": "40669162"}
],
"evidence_against": [
{"claim": "Large PD genetic studies show no strong NLRP3 variant associations", "pmid": "unreported"},
{"claim": "MCC950 has poor CNS penetration and failed toxicology for clinical development", "pmid": "unreported"},
{"claim": "IL-1β may have neuroprotective roles in early PD", "pmid": "unreported"}
]
},
{
"title": "Complement C1q/C3 Inhibition to Prevent Synaptodendritis in Alzheimer's Disease",
"description": "Blocking early complement activation prevents microglia-mediated excessive synapse loss before and during amyloid deposition. C1q localized to synapses in early AD correlates with cognitive decline; C3 knockout or CR3 deficiency protects synapses in 5xFAD mice. Annexon's ANX005 anti-C1q antibody has human CNS target engagement experience from Huntington's disease trials.",
"target_gene": "C1Q/C3",
"dimension_scores": {
"evidence_strength": 0.78,
"novelty": 0.55,
"feasibility": 0.62,
"therapeutic_potential": 0.82,
"mechanistic_plausibility": 0.80,
"druggability": 0.75,
"safety_profile": 0.52,
"competitive_landscape": 0.65,
"data_availability": 0.72,
"reproducibility": 0.70
},
"composite_score": 0.69,
"evidence_for": [
{"claim": "C1q localizes to synapses in early AD; C3 deposition correlates with memory loss", "pmid": "29445953"},
{"claim": "C3 knockout or CR3 deficiency protects synapses in 5xFAD mice", "pmid": "29445953"},
{"claim": "C1q inhibition via blocking antibody reduces synapse loss in mouse models", "pmid": "29445953"},
{"claim": "Annexon ANX005 achieved CNS target engagement in Huntington's Phase 2", "pmid": "29202623"}
],
"evidence_against": [
{"claim": "C1q/C3 inhibition may impair necessary synaptic pruning during development", "pmid": "unreported"},
{"claim": "C3 deficiency could impair debris clearance needed for repair", "pmid": "unreported"},
{"claim": "Timing critical; complement has dual roles in development vs. adult neurodegeneration", "pmid": "unreported"}
]
},
{
"title": "TREM2 Agonism to Promote Neuroprotective Microglial Phenotype in Alzheimer's Disease",
"description": "Enhancing TREM2 signaling shifts microglia toward a neuroprotective state with improved lipid metabolism, enhanced phagocytosis of amyloid plaques, and reduced neurotoxicity. TREM2 R47H variants confer ~3-fold AD risk. Agonistic antibodies developed by Alector/AbbVie showed target engagement but AL002 Phase 2 failed to meet primary clinical endpoints.",
"target_gene": "TREM2",
"dimension_scores": {
"evidence_strength": 0.78,
"novelty": 0.50,
"feasibility": 0.52,
"therapeutic_potential": 0.72,
"mechanistic_plausibility": 0.75,
"druggability": 0.82,
"safety_profile": 0.55,
"competitive_landscape": 0.60,
"data_availability": 0.70,
"reproducibility": 0.68
},
"composite_score": 0.67,
"evidence_for": [
{"claim": "TREM2 R47H variant increases AD risk ~3-fold", "pmid": "24285345"},
{"claim": "TREM2 deficiency exacerbates amyloid pathology in mouse models", "pmid": "29607930"},
{"claim": "TREM2 agonistic antibodies promote microglial plaque compaction and reduce neuritic dystrophy", "pmid": "31178183"},
{"claim": "AL002 achieved target engagement and microglial pharmacodynamics in Phase 2", "pmid": "unreported"}
],
"evidence_against": [
{"claim": "AL002 Phase 2 failed to slow CDR-SB progression or improve secondary clinical endpoints", "pmid": "unreported"},
{"claim": "TREM2 R47H knock-in mice show milder phenotypes than knockout models", "pmid": "unreported"},
{"claim": "Excessive TREM2 activation may promote neurotoxic microglial states", "pmid": "unreported"}
]
},
{
"title": "TFEB-Mediated Lysosomal Biogenesis Restoration for GRN Haploinsufficiency FTD",
"description": "Progranulin haploinsufficiency impairs lysosomal function via decreased TFEB nuclear translocation, causing lipofuscin accumulation and neuronal vulnerability. Enhancing TFEB activity may restore lysosomal homeostasis. However, GRN Phase 3 (latozinemab) failed clinically, and direct TFEB activation poses oncogenic/metabolic risks.",
"target_gene": "TFEB/GRN",
"dimension_scores": {
"evidence_strength": 0.68,
"novelty": 0.72,
"feasibility": 0.45,
"therapeutic_potential": 0.62,
"mechanistic_plausibility": 0.65,
"druggability": 0.40,
"safety_profile": 0.38,
"competitive_landscape": 0.55,
"data_availability": 0.60,
"reproducibility": 0.55
},
"composite_score": 0.56,
"evidence_for": [
{"claim": "Grn-/- mice exhibit lysosomal dysfunction, lipofuscin accumulation, and microgliosis", "pmid": "18819921"},
{"claim": "Progranulin localizes to lysosomes and regulates cathepsin activity", "pmid": "22958956"},
{"claim": "TFEB overexpression enhances lysosomal biogenesis in storage disease models", "pmid": "22107871"}
],
"evidence_against": [
{"claim": "Latozinemab (progranulin replacement) failed Phase 3 clinical endpoint in FTD-GRN", "pmid": "unreported"},
{"claim": "Grn-/- mice do not replicate human FTD pathology (no TDP-43 inclusions)", "pmid": "unreported"},
{"claim": "TFEB overexpression may cause non-specific autophagy with oncogenic potential", "pmid": "unreported"}
]
},
{
"title": "Astrocyte Reactivity Reprogramming via A1/A2 Phenotype Modulation",
"description": "Reactive astrocytes acquire neurotoxic A1 or neuroprotective A2 phenotypes in neurodegeneration. Blocking A1 inducers (IL-1α, TNFα, C1q) or enhancing A2 genes (Lcn2, Timp1) may restore astrocyte homeostatic function. However, the A1/A2 dichotomy is oversimplified; human astrocytes differ significantly from rodent counterparts.",
"target_gene": "IL1A/TNF/C1Q",
"dimension_scores": {
"evidence_strength": 0.62,
"novelty": 0.68,
"feasibility": 0.42,
"therapeutic_potential": 0.60,
"mechanistic_plausibility": 0.58,
"druggability": 0.48,
"safety_profile": 0.50,
"competitive_landscape": 0.60,
"data_availability": 0.55,
"reproducibility": 0.52
},
"composite_score": 0.55,
"evidence_for": [
{"claim": "LPS-activated microglia induce A1 astrocytes via Il-1α/TNFα/C1q", "pmid": "29107332"},
{"claim": "A1 astrocytes fail to support neuronal survival and synaptogenesis", "pmid": "29107332"},
{"claim": "Astrocyte reactivity reprogramming is biologically validated in multiple models", "pmid": "29107332"}
],
"evidence_against": [
{"claim": "A1/A2 dichotomy is oversimplified; astrocytes exhibit spectrum states", "pmid": "unreported"},
{"claim": "Human astrocytes differ significantly from rodent counterparts", "pmid": "unreported"},
{"claim": "'Block A1' is not a target product profile; needs sharper target definition", "pmid": "unreported"}
]
},
{
"title": "Autophagy Enhancement via mTOR Inhibition for C9orf72 Dipeptide Repeat Pathology",
"description": "C9orf72 repeat expansions produce toxic dipeptide repeats (DPRs) that impair nucleocytoplasmic transport and autophagy. mTOR inhibition via rapamycin may reduce DPR accumulation. However, rapamycin benefits in Drosophila C9 models did not translate well to mammals, and C9 ASO programs have failed clinically.",
"target_gene": "MTOR/C9orf72",
"dimension_scores": {
"evidence_strength": 0.55,
"novelty": 0.58,
"feasibility": 0.40,
"therapeutic_potential": 0.52,
"mechanistic_plausibility": 0.52,
"druggability": 0.55,
"safety_profile": 0.42,
"competitive_landscape": 0.50,
"data_availability": 0.48,
"reproducibility": 0.45
},
"composite_score": 0.49,
"evidence_for": [
{"claim": "Poly-GA DPRs aggregate with RBP4 and impair proteasome/autophagy", "pmid": "25297118"},
{"claim": "C9orf72 knockout in mice causes lysosomal accumulation and neurodegeneration", "pmid": "25624326"},
{"claim": "Rapamycin ameliorates neurodegeneration in Drosophila C9 models", "pmid": "26146185"}
],
"evidence_against": [
{"claim": "Rapamycin has limited efficacy in mouse C9 models compared to Drosophila", "pmid": "unreported"},
{"claim": "BIIB078 C9 ASO discontinued after Phase 1 did not show clinical benefit", "pmid": "unreported"},
{"claim": "mTOR inhibition may impair other critical neuronal pathways", "pmid": "unreported"}
]
},
{
"title": "Metabolic Correction via SGLT2/PDK Modulation for ALS Energy Homeostasis",
"description": "Motor neurons exhibit unique metabolic dependencies on glycolysis and lipid metabolism that become dysregulated in ALS. Targeting PDH kinase or SGLT2 to shift toward glucose oxidation may restore energy homeostasis. However, causal linkage to ALS progression is weak, and SOD1 metabolic rescue is a poor predictor for sporadic ALS.",
"target_gene": "SGLT2/PDK",
"dimension_scores": {
"evidence_strength": 0.50,
"novelty": 0.62,
"feasibility": 0.45,
"therapeutic_potential": 0.48,
"mechanistic_plausibility": 0.45,
"druggability": 0.68,
"safety_profile": 0.52,
"competitive_landscape": 0.55,
"data_availability": 0.42,
"reproducibility": 0.40
},
"composite_score": 0.51,
"evidence_for": [
{"claim": "Motor neurons have reduced PDH activity and prefer glycolysis", "pmid": "29656893"},
{"claim": "SGLT2 expression elevated in ALS motor neurons; inhibition extends survival in SOD1 mice", "pmid": "34193629"},
{"claim": "PGC-1α dysregulation observed in ALS patients and models", "pmid": "18391957"}
],
"evidence_against": [
{"claim": "SOD1-G93A mice show high variability in metabolic phenotypes", "pmid": "unreported"},
{"claim": "SGLT2 inhibitors in humans did not show cognitive benefits", "pmid": "unreported"},
{"claim": "PDK/glycolysis shift may be compensatory, not causative", "pmid": "unreported"}
]
}
],
"knowledge_edges": [
{"source_id": "hypothesis_1", "source_type": "hypothesis", "target_id": "TREM2", "target_type": "gene", "relation": "agonism_of"},
{"source_id": "hypothesis_2", "source_type": "hypothesis", "target_id": "NLRP3", "target_type": "gene", "relation": "inhibition_of"},
{"source_id": "hypothesis_3", "source_type": "hypothesis", "target_id": "TFEB", "target_type": "gene", "relation": "activation_of"},
{"source_id": "hypothesis_3", "source_type": "hypothesis", "target_id": "GRN", "target_type": "gene", "relation": "upstream_regulation_of"},
{"source_id": "hypothesis_4", "source_type": "hypothesis", "target_id": "C9orf72", "target_type": "gene", "relation": "targeting_pathology_from"},
{"source_id": "hypothesis_5", "source_type": "hypothesis", "target_id": "IL1A/TNF/C1Q", "target_type": "gene", "relation": "blocking_inducers_of"},
{"source_id": "hypothesis_6", "source_type": "hypothesis", "target_id": "C1Q/C3", "target_type": "gene", "relation": "inhibition_of"},
{"source_id": "hypothesis_7", "source_type": "hypothesis", "target_id": "SGLT2/PDK", "target_type": "gene", "relation": "inhibition_of"},
{"source_id": "hypothesis_6", "source_type": "hypothesis", "target_id": "hypothesis_1", "target_type": "hypothesis", "relation": "synergistic_with_microglial_pathways"},
{"source_id": "hypothesis_3", "source_type": "hypothesis", "target_id": "hypothesis_4", "target_type": "hypothesis", "relation": "shares_lysosomal_autophagy_axis"},
{"source_id": "hypothesis_2", "source_type": "hypothesis", "target_id": "hypothesis_5", "target_type": "hypothesis", "relation": "shares_neuroinflammation_mechanism"},
{"source_id": "hypothesis_1", "source_type": "hypothesis", "target_id": "hypothesis_6", "target_type": "hypothesis", "relation": "microglial_synapse_interaction"},
{"source_id": "AL002_failure", "source_type": "clinical_trial_result", "target_id": "hypothesis_1", "target_type": "hypothesis", "relation": "revises_confidence_downward"},
{"source_id": "latozinemab_failure", "source_type": "clinical_trial_result", "target_id": "hypothesis_3", "target_type": "hypothesis", "relation": "revises_confidence_downward"},
{"source_id": "BIIB078_discontinuation", "source_type": "clinical_trial_result", "target_id": "hypothesis_4", "target_type": "hypothesis", "relation": "revises_confidence_downward"},
{"source_id": "dapansutrile_phase2", "source_type": "clinical_trial_result", "target_id": "hypothesis_2", "target_type": "hypothesis", "relation": "supports_feasibility"},
{"source_id": "ANX005_phase2", "source_type": "clinical_trial_result", "target_id": "hypothesis_6", "target_type": "hypothesis", "relation": "supports_druggability"}
],
"synthesis_summary": "The four-persona debate reveals two translationally viable programs: NLRP3 inflammasome inhibition for Parkinson's disease and complement C1q/C3 inhibition for Alzheimer's disease synapse protection. The Domain Expert's April 2026 perspective is critical—oral CNS-penetrant NLRP3 inhibitors (dapansutrile in DAPA-PD Phase 2, NT-0796 Phase 1b data) have materially improved feasibility beyond the original MCC950 proposal, making this the strongest development candidate. Complement inhibition retains strong mechanistic evidence (C1q at synapses in early AD, C3 knockout protection in 5xFAD) and clinical experience via Annexon's ANX005 in Huntington's disease, but timing and chronic immune safety remain constraints. In contrast, TREM2 agonism and GRN/TFEB lysosomal restoration have been clinically damaged by negative readouts (AL002 Phase 2 failed clinically; latozinemab Phase 3 failed), though the underlying biology remains valid. Astrocyte reprogramming, C9 autophagy enhancement, and ALS metabolic correction are deprioritized due to target immaturity, poor translational models, or weak causal evidence—the field cannot move forward without sharper target definitions or validated human-relevant systems."
}