Test

Therapeutic Hypotheses in Neurodegeneration

Hypothesis 1: TDP-43 Liquid-Liquid Phase Separation Dysregulation as a Central Mechanism in ALS/FTD

Mechanism: TDP-43 undergoes pathological liquid-liquid phase separation (LLPS) under stress conditions. In ALS/FTD, TDP-43 transitions from reversible liquid droplets to irreversible pathological aggregates due to impaired chaperone-mediated disassembly and altered post-translational modifications (hyperphosphorylation, ubiquitination). Therapeutic restoration of LLPS dynamics could prevent aggregate formation.

Target Gene/Protein/Pathway: TDP-43 (TARDBP); Hsp104 orthologs in mammalian neurons (Hsp70/Hsp40 system)

Supporting Evidence:

• TDP-43 forms stress granules via LLPS in physiological conditions (PMID: 24670997)
• Pathological TDP-43 aggregates colocalize with stress granule markers in ALS patient tissue (PMID: 28661562)
• Mutations in TDP-43 cause familial ALS, confirming its centrality (PMID: 19479373)
• TDP-43 frameshift mutations alter LLPS behavior (PMID: 31853077)

Confidence: 0.72

Hypothesis 2: GBA1 Loss-of-Function Exacerbates α-Synuclein Pathology Through Lysosomal Impairment

Mechanism: Heterozygous GBA1 mutations (causing Gaucher disease) are the strongest genetic risk factor for Parkinson's disease. GBA1 encodes β-glucocerebrosidase (GCase), and its loss-of-function leads to accumulated glucocerebroside substrates, which impair lysosomal function and create a feed-forward loop where reduced GCase activity increases α-synuclein levels, and elevated α-synuclein further inhibits GCase trafficking.

Target Gene/Protein/Pathway: GBA1/GCase; TFEB-mediated lysosomal biogenesis pathway

Supporting Evidence:

• GBA1 mutations increase PD risk 5-20 fold (PMID: 19690987)
• GCase activity is reduced in PD substantia nigra regardless of GBA1 status (PMID: 23685549)
• α-Synuclein binds to and inhibits GCase (PMID: 21799912)
• TFEB activation clears α-synuclein in cellular models (PMID: 25801896)

Confidence: 0.81

Hypothesis 3: TREM2-SYK Signaling Axis in Microglial State Transitions

Mechanism: TREM2 (Triggering Receptor Expressed on Myeloid Cells 2) is a microglial receptor critical for neurodegenerative disease progression. TREM2 activates SYK kinase signaling, promoting microglial survival, proliferation, and transition to a disease-associated microglia (DAM) state. TREM2 deficiency in Alzheimer's disease reduces plaque-associated microglial clustering and increases amyloid plaque burden. Conversely, TREM2 activation may be protective by enhancing phagocytosis of pathological seeds.

Target Gene/Protein/Pathway: TREM2-SYK axis; PI3K/AKT signaling downstream

Supporting Evidence:

• TREM2 R47H variant increases AD risk ~3-fold (PMID: 22577221)
• Trem2 knockout in 5xFAD mice worsens plaque burden and microglial dysfunction (PMID: 26621723)
• TREM2 agonistic antibodies enhance microglial response to plaques (PMID: 30509931)
• SYK is required for TREM2 downstream signaling (PMID: 26178167)

Confidence: 0.85

Hypothesis 4: C9orf72 Repeat Expansion Creates Toxic Gain-of-Function Through r(G4C2) Dipeptide Repeat Proteins

Mechanism: C9orf72 hexanucleotide repeat expansion (G4C2) is the most common genetic cause of ALS and FTD. Bidirectional transcription produces expanded repeat RNAs that undergo non-ATG translation, generating five different dipeptide repeat proteins (DPRs: poly-GA, -GR, -PA, -PR, -GP). Poly-GR and poly-PR are particularly toxic, disrupting nucleocytoplasmic transport, translational machinery, and stress granule dynamics.

Target Gene/Protein/Pathway: C9orf72 repeat RNA; RAN translation machinery; Nucleocytoplasmic transport (karyopherins, RanGAP)

Supporting Evidence:

• C9orf72 is the most common ALS/FTD mutation (PMID: 21944778)
• DPR proteins accumulate in patient neurons and correlate with pathology (PMID: 25437307)
• Poly-GR/PR disrupt nuclear import in Drosophila and mammalian models (PMID: 26138283)
• Antisense oligonucleotides reduce DPR production and improve phenotypes (PMID: 25907378)

Confidence: 0.88

Hypothesis 5: Astrocyte Senescence Drives Neuroinflammation in ALS

Mechanism: Astrocytes in ALS undergo cellular senescence, characterized by SASP (senescence-associated secretory phenotype) including IL-6, CXCL1, and GM-CSF release. This creates a chronic pro-inflammatory environment that activates microglia and accelerates motor neuron death. Senolytic intervention (removing senescent astrocytes) could restore astrocyte support functions and slow disease progression.

Target Gene/Protein/Pathway: p53/p21 and p16INK4a/RB senescence pathways; BCL-2 family dependency of senescent cells; IL-6/STAT3 signaling

Supporting Evidence:

• Astrocytes in SOD1G93A mice and ALS patients show senescence markers (PMID: 29937267)
• SASP factors are elevated in ALS CSF and post-mortem tissue (PMID: 32572062)
• Senolytics (ABT-263/Navitoclax) reduce inflammation in other neurodegenerative models (PMID: 29100065)
• Young astrocytes rescue motor neuron survival when co-cultured (PMID: 25437563)

Confidence: 0.68

Hypothesis 6: Mitophagy Induction as a Therapeutic Strategy in Sporadic Parkinson's Disease

Mechanism: Mitochondrial dysfunction is central to PD pathogenesis. PINK1 and PRKN/Parkin mediate mitophagy of damaged mitochondria. While rare mutations cause familial PD, sporadic PD involves impaired mitophagy due to reduced PINK1/Parkin expression or activity, increased mitochondrial stress, and altered mtDNA maintenance. Pharmacological activation of mitophagy could compensate for these deficits.

Target Gene/Protein/Pathway: PINK1/Parkin pathway; NAD+-dependent deacetylases (SIRT3); mitochondrial fission (DRP1/FIS1)

Supporting Evidence:

• PINK1 and PRKN mutations cause familial PD (PMID: 15133518, PMID: 14695260)
• Mitophagy is impaired in sporadic PD patient fibroblasts (PMID: 25019414)
• NAD+ precursors (nicotinamide riboside) enhance mitophagy and protect dopaminergic neurons (PMID: 29420476)
• Urolithin A (a mitophagy inducer) reduces α-synuclein pathology (PMID: 31377141)

Confidence: 0.76

Hypothesis 7: Tau Propagation via Muscarinic Receptor-Mediated Transcytosis Across the Blood-Brain Barrier

Mechanism: Pathological tau spreads from the periphery into the CNS and between brain regions via extracellular vesicles and a proposed "transcytosis pathway" involving M1/M3 muscarinic acetylcholine receptors. Inhibition of this pathway could block tau propagation, particularly in early disease stages before extensive neuronal loss.

Target Gene/Protein/Pathway: M1/M3 muscarinic receptors (CHRM1, CHRM3); LRP1 for tau uptake; BBB transcytosis machinery

Supporting Evidence:

• Tau is detected in peripheral blood of AD patients and correlates with CNS pathology (PMID: 26159303)
• LRP1 mediates neuronal uptake and trans-synaptic spread of tau (PMID: 29847938)
• Muscarinic receptors regulate tau secretion in cell models (PMID: 32084339)
• Anti-tau antibodies show limited efficacy in advanced disease (PMID: 34188024)

Confidence: 0.64 Cross-Cutting Themes: These hypotheses converge on protein homeostasis, glial-immune interactions, and metabolic resilience as therapeutic targets. Combination approaches (e.g., TREM2 agonism + amyloid-targeting in AD) may show synergistic efficacy.

I treat “confidence” here as confidence in the therapeutic causal hypothesis, not merely confidence that the pathway is biologically involved.

1. TDP-43 LLPS Dysregulation in ALS/FTD

Weak links: TDP-43 condensates are real, but LLPS may be an upstream stress response or assay artifact rather than the central toxic transition in vivo. Patient inclusions are end-stage, heterogeneous, and often reflect loss of nuclear TDP-43 function, RNA splicing defects, impaired nucleocytoplasmic transport, and stress-granule biology all at once. Chaperone overexpression could reduce visible aggregates while worsening proteostasis or masking toxicity.

Counter-evidence / confounds: TDP-43 pathology is common across ALS/FTD, but familial TARDBP mutations are a minority of ALS. Stress granule colocalization does not prove stress granules seed disease. FRAP after arsenite stress may model acute cellular injury more than chronic motor neuron degeneration. Recent reviews still frame LLPS-to-aggregation as plausible but not therapeutically settled. Source: [Carey & Guo 2022](https://pmc.ncbi.nlm.nih.gov/articles/PMC8847598/), [Song 2024](https://pubmed.ncbi.nlm.nih.gov/38029395/).

Falsifying experiments:
Show that restoring droplet dynamics normalizes TDP-43 nuclear function, cryptic exon suppression, axonal transport, and motor neuron survival without merely dissolving inclusions. In vivo, conditional neuron-specific modulation of TDP-43 condensate-driving domains should alter disease progression in a direction predicted by LLPS dynamics.

Revised confidence: 0.58.

2. GBA1 Loss-of-Function Exacerbates Alpha-Synuclein via Lysosomal Impairment

Weak links: This is genetically strong, but the therapeutic version is less certain. GBA1 mutations increase PD risk, yet penetrance is incomplete, and substrate accumulation, ER stress, lipid remodeling, inflammation, and alpha-synuclein handling may all contribute. TFEB activation is broad and may rescue lysosomal stress without proving the GBA1-alpha-syn feed-forward loop is primary.

Counter-evidence / confounds: Reduced GCase activity in sporadic PD substantia nigra supports convergence, but it could be secondary to neurodegeneration or alpha-syn burden rather than causal. Dat-Cre Gba1 knockout may create developmental or cell-type-specific artifacts and may not recapitulate heterozygous human GBA1-PD. Source: [Gegg et al.](https://pubmed.ncbi.nlm.nih.gov/23034917/).

Falsifying experiments:
Rescue experiments should compare GCase enzymatic restoration, substrate lowering, TFEB activation, and alpha-syn lowering side by side. If substrate normalization fails to reduce alpha-syn seeding, lysosomal impairment, or dopaminergic vulnerability in aged humanized alpha-syn models, the central feed-forward claim weakens.

Revised confidence: 0.74.

3. TREM2-SYK Axis in Microglial State Transitions

Weak links: The biology is strong; the therapeutic claim is now less secure. TREM2 can promote plaque-associated microglial responses, but “more TREM2 signaling” is not necessarily better. Microglial activation can compact plaques while increasing inflammatory injury, synapse loss, or tau-related neurodegeneration. Effects are likely disease-stage and pathology-context dependent.

Counter-evidence / confounds: TREM2 deficiency changes plaque-associated microglia and neuritic injury in mouse models, but amyloid mouse models poorly predict human cognitive benefit. Most importantly, the 2026 phase 2 AL002 trial in early AD showed target engagement and pharmacodynamic response but did not meet the primary clinical endpoint. Source: [TREM2 review](https://pmc.ncbi.nlm.nih.gov/articles/PMC9783481/), [AL002 phase 2 Nature Medicine 2026](https://www.nature.com/articles/s41591-026-04273-1).

Falsifying experiments:
A good falsifier is a longitudinal amyloid-plus-tau model where TREM2 agonism improves plaque metrics but worsens tau spread, synapse density, neurofilament, or cognition. In humans, failure to improve downstream neurodegeneration biomarkers despite robust microglial target engagement would argue the axis is not sufficient therapeutically.

Revised confidence: 0.60.

4. C9orf72 DPR Toxic Gain-of-Function

Weak links: C9orf72 disease is not only DPR toxicity. Repeat RNA foci, antisense transcripts, C9orf72 haploinsufficiency, TDP-43 pathology, immune dysfunction, nucleocytoplasmic transport, and cell-type vulnerability all compete as drivers. Poly-GR/PR toxicity in models can be strong but sometimes depends on supraphysiologic expression.

Counter-evidence / confounds: The ASO story is a major caution. BIIB078 reduced DPR biomarkers but did not show clinical benefit and was discontinued; later analyses reported persistent key CNS pathologies despite target engagement. That does not refute DPR toxicity, but it weakens “DPR reduction alone is sufficient.” Sources: [C9orf72 discovery](https://pubmed.ncbi.nlm.nih.gov/21944778/), [BIIB078 trial](https://www.sciencedirect.com/science/article/pii/S1474442224002163), [2025 molecular impact](https://pubmed.ncbi.nlm.nih.gov/40865525/).

Falsifying experiments:
If selective DPR suppression in human motor neuron-organoid or animal models leaves TDP-43 mislocalization, neurodegeneration, and functional decline unchanged despite large DPR reductions, DPRs are probably not the dominant therapeutic node. Experiments must separately suppress sense RNA, antisense RNA, and restore C9orf72 protein.

Revised confidence: 0.72.

5. Astrocyte Senescence Drives Neuroinflammation in ALS

Weak links: Senescence markers are not the same as irreversible senescence. Reactive astrocytes, DNA damage responses, aging, inflammation, oxidative stress, and terminal disease state can all induce p16/p21/SASP-like signatures. Senolytics have broad off-target effects and may remove cells that are compensatory or trophic.

Counter-evidence / confounds: SOD1G93A is an aggressive familial ALS model and may overstate astrocyte-autonomous mechanisms. ALS CSF cytokines are not cell-source-specific. Dasatinib/quercetin and navitoclax have major immune, vascular, platelet, and CNS penetration concerns. Young astrocyte rescue supports astrocyte dysfunction, not necessarily senescence as the driver.

Falsifying experiments:
Use astrocyte-specific genetic ablation of p16-high or p21-high cells, not systemic senolytics alone. If selective removal of senescent astrocytes reduces SASP but does not preserve motor neurons, NMJ integrity, or survival, the hypothesis fails therapeutically. Single-cell spatial profiling must show senescent astrocytes precede motor neuron loss.

Revised confidence: 0.45.

6. Mitophagy Induction in Sporadic Parkinson’s Disease

Weak links: Mitochondrial dysfunction is central in PD, but “induce mitophagy” is too broad. Excess mitophagy can deplete mitochondria, impair energetics, or stress already vulnerable dopaminergic neurons. PINK1/Parkin familial PD does not necessarily generalize to sporadic PD, where alpha-synuclein, lysosomes, inflammation, pesticides, aging, and vascular factors interact.

Counter-evidence / confounds: PINK1 knockout mice often have weak or late dopaminergic neurodegeneration, so rescue in that model may not translate. Nicotinamide riboside changes NAD metabolism broadly, not specifically mitophagy. Human phase I data show target engagement and short-term signals, not disease modification. Source: [NADPARK phase I](https://www.sciencedirect.com/science/article/pii/S1550413122000456).

Falsifying experiments:
In aged sporadic PD-relevant models with alpha-syn pathology, demonstrate that mitophagy flux increases in substantia nigra dopaminergic neurons and that blocking mitophagy genetically abolishes benefit. If NR/urolithin benefits persist when PINK1/Parkin-dependent mitophagy is blocked, the proposed mechanism is wrong.

Revised confidence: 0.57.

7. Tau Propagation via Muscarinic Receptor-Mediated BBB Transcytosis

Weak links: This is the least secure. LRP1-mediated tau uptake has credible evidence, but the jump to M1/M3 muscarinic receptor-mediated BBB transcytosis and peripheral-to-CNS tau propagation is speculative. Tau spread in AD is more parsimoniously explained by neuronal connectivity, local release/uptake, glial handling, LRP1/heparan sulfate pathways, and regional vulnerability.

Counter-evidence / confounds: The proposed experiment injects tau seeds directly into brain, so it does not test BBB transcytosis from periphery into CNS. CHRM1/CHRM3 double knockout has major cholinergic, vascular, cognitive, and BBB physiology confounds. Tau PET in mice is technically difficult and tracer specificity can be problematic. Source: [LRP1 tau uptake/spread](https://pubmed.ncbi.nlm.nih.gov/32296178/).

Falsifying experiments:
Use peripheral tau seed administration with endothelial-specific CHRM1/CHRM3 deletion and direct measurement of labeled tau crossing the BBB. If endothelial muscarinic loss does not reduce CNS tau entry while LRP1 or heparan sulfate perturbation does, the muscarinic transcytosis claim should be downgraded sharply.

Revised confidence: 0.32.

Overall ranking after critique: strongest mechanistic targets are GBA1/GCase and C9orf72 DPR/RNA biology, but both need better translation strategy. TREM2 remains biologically important but clinically weakened by AL002. Tau-muscarinic BBB transcytosis and astrocyte senescence as primary ALS driver need the most basic causal validation.

As of April 29, 2026, I would triage these as 2 near-term translational programs, 3 conditional/retool programs, and 2 not-ready programs.

| Rank | Hypothesis | Feasibility | Development Call |
|---|---:|---|
| 1 | GBA1/GCase in PD | High | Best drug-development case |
| 2 | C9orf72 RNA/DPR biology | Moderate-high biology, hard clinic | Worth pursuing, but not DPR-only |
| 3 | Mitophagy/NAD in PD | Moderate | Trialable, but mechanism must be sharpened |
| 4 | TDP-43 LLPS/proteostasis | Moderate-low | Strong biology, poor druggability today |
| 5 | TREM2-SYK agonism | Biologically strong, clinically weakened | Reposition or combine, not standalone AD bet |
| 6 | Astrocyte senescence in ALS | Low-moderate | Needs causal validation before clinical spend |
| 7 | Tau muscarinic BBB transcytosis | Low | Not development-ready |

1. GBA1/GCase: most feasible

Druggability is strong: enzyme enhancement, pharmacological chaperones, substrate reduction, gene therapy, and lysosomal-biogenesis strategies are all plausible. The best path is GCase restoration or substrate correction, not broad TFEB activation first. Ambroxol has human CNS exposure and target-engagement data, and a 2025 PDD trial found it safe/tolerable but not efficacy-positive on primary outcomes. A completed GBA-PD trial is listed without posted results yet, so this remains clinically alive but not proven.

Biomarkers are unusually good for neurodegeneration: genotype enrichment, CSF/plasma GCase activity, glucosylceramide/glucosylsphingosine, alpha-syn seed amplification assays, NfL, DAT-SPECT, MDS-UPDRS, and cognitive endpoints for GBA-PD/PDD. Models should include human GBA1 iPSC dopaminergic neurons, lysosomal flux assays, alpha-syn seeding models, and aged knock-in or human alpha-syn models rather than DAT-Cre knockout alone.

Clinical constraints: recruit GBA1-PD, stratify severe versus mild variants, and design for 12-24 months. Cost: roughly $10-25M for IND-enabling/phase 1 if repurposed small molecule, $40-120M through a serious phase 2, more for gene therapy. Timeline: 2-4 years to a credible phase 2 readout if using an existing CNS-penetrant agent; 6-9 years for novel biologic/gene therapy.

2. C9orf72 RNA/DPR: strong target, sobering clinical history

Druggability is real because ASOs can hit CNS RNA, and the tofersen precedent in SOD1-ALS supports genetically targeted ALS development using biomarkers such as NfL. But C9orf72 is not a simple “lower DPRs and win” program. BIIB078 and WVE-004 both showed the key warning: target/DPR biomarker movement did not translate into clinical benefit, and WVE-004 was terminated despite robust poly(GP) reduction.

The next viable approach should separate sense RNA, antisense RNA, DPRs, and C9orf72 haploinsufficiency. A pure RAN-translation small molecule like furamidine is too nonspecific and likely unattractive unless it shows clean CNS exposure and safety. Better druggability: allele-selective ASOs, RNA binders, repeat-targeted degradation, or combined RNA-lowering plus functional rescue.

Biomarkers: CSF poly(GP), NfL, repeat RNA foci in patient-derived systems, DPR immunoassays, TDP-43 mislocalization markers, neurophysiology, ALSFRS-R slope, and FTD cognitive measures. Timeline/cost: $30-80M to optimized candidate plus early ALS/FTD biomarker trial; $150-300M+ to pivotal development. Main risk is not delivery; it is whether the target node is sufficient.

3. Mitophagy/NAD in sporadic PD: trialable but mechanistically diffuse

Nicotinamide riboside and related NAD approaches are druggable, orally deliverable, and comparatively safe. NADPARK showed oral NR increased brain NAD in PD over 30 days and produced metabolic signals, but that is not disease modification. Urolithin A is similarly feasible but not yet a PD disease-modifying therapy.

The key development issue is mechanism. “Mitophagy induction” can become a catch-all. A credible program needs proof of mitophagy flux in human dopaminergic neurons, not just blood NAD metabolites. Biomarkers should include 31P-MRS or PET metabolic measures, mitochondrial DNA damage/copy number, complex I-linked signatures, alpha-syn SAA, NfL, inflammatory markers, and wearable motor measures.

Best trial: early untreated PD, 12-18 months, biomarker-enriched for mitochondrial impairment. Cost: $15-40M for a phase 2 repurposed-nutraceutical-style program; $80-150M for a regulated novel mitophagy drug. Safety is manageable for NR but less clear for potent mitophagy activators because excessive mitochondrial clearance could harm high-energy neurons.

4. TDP-43 LLPS/proteostasis: compelling but not yet product-shaped

TDP-43 is central to ALS/FTD, but LLPS restoration is hard to drug cleanly. Directly modulating condensate behavior risks broad effects on RNA granules, splicing, stress responses, and nuclear TDP-43 function. Chaperone overexpression is scientifically useful but therapeutically awkward: gene therapy dose, cell-type targeting, proteostasis overload, and chronic safety are major barriers.

Better tractable approaches: restore nuclear TDP-43 function, suppress cryptic exon pathology, improve autophagy/proteostasis selectively, or target downstream RNA-splicing consequences. Biomarkers are improving: CSF NfL, TDP-43 seeding/fragment assays if validated, cryptic exon signatures such as STMN2, neurofilament, EMG, respiratory decline, and iPSC motor-neuron survival.

Timeline: 3-5 years for a credible preclinical package; 7-10+ years to meaningful clinical proof unless a highly specific RNA or biomarker-linked approach emerges. Cost: $25-75M to candidate/IND; $150M+ to phase 2 ALS efficacy attempt.

5. TREM2-SYK: biology survives, standalone AD therapy weakened

TREM2 is druggable with antibodies and perhaps small molecules, and biomarkers are strong: CSF soluble TREM2, osteopontin, microglial PET where available, amyloid/tau PET, plasma p-tau217, NfL/GFAP, and synaptic markers. But the 2026 AL002 phase 2 trial in early AD showed CNS target engagement and pharmacodynamic response yet missed the CDR-SB primary endpoint; MRI changes resembling ARIA were common. That sharply lowers standalone feasibility.

The most realistic future is stage-specific or combination therapy, possibly with amyloid lowering, tau-directed therapy, or genotype/pathology enrichment. Safety requires careful ARIA-like MRI monitoring and inflammation/synapse-loss surveillance. Cost is high: antibody AD trials are typically $100-300M through phase 2b and $500M+ for phase 3. Timeline: 5-8 years for a redesigned clinical signal. I would not fund another broad early-AD TREM2 agonist monotherapy without a sharper responder hypothesis.

6. Astrocyte senescence in ALS: preclinical only for now

Druggability exists in the abstract: senolytics, SASP suppression, p16/p21 pathway modulation, BCL-2 family targeting, IL-6/STAT3 inhibition. But systemic senolytics are blunt, and navitoclax-class toxicity, platelet effects, immune effects, and CNS penetration are serious concerns. Removing astrocytes in ALS could also remove compensatory trophic support.

This needs causal validation before clinical development: spatial single-cell evidence that senescent astrocytes precede motor-neuron loss, astrocyte-specific genetic clearance, and rescue of NMJ integrity/survival in multiple ALS models beyond SOD1G93A. Biomarkers would include CSF SASP proteins, GFAP, NfL, inflammatory panels, PET glial imaging, and cell-source-resolved markers if available.

Timeline/cost: $5-15M for rigorous validation; $30-80M to an early CNS senolytic program. I would not start patient dosing based on current rationale.

7. Tau muscarinic BBB transcytosis: not ready

This is the weakest. LRP1/heparan sulfate tau uptake and neuronal-network propagation are plausible; muscarinic M1/M3-mediated BBB transcytosis is not yet a solid therapeutic axis. CHRM1/CHRM3 antagonism or deletion has unacceptable CNS, autonomic, vascular, and cognitive confounds for AD.

Before drug development, the required experiment is endothelial-specific perturbation with peripheral labeled tau and direct CNS-entry quantification. The proposed intracerebral seed experiment does not test the BBB claim. Biomarkers and models are also problematic: tau PET in mice is difficult, and peripheral tau does not automatically imply pathogenic CNS import.

This is a research hypothesis, not a translational program. Cost to validate: $2-8M. Timeline: 2-3 years to decide whether to kill or reframe. Clinical development should not begin.

Bottom Line

Fund first: GBA1/GCase and a redesigned C9orf72 RNA program.
Fund selectively: mitophagy/NAD in PD, but only with neuron-relevant flux biomarkers.
Keep in discovery: TDP-43 LLPS and astrocyte senescence.
Reposition: TREM2, because AL002 weakens the monotherapy case.
Do not develop yet: tau-muscarinic BBB transcytosis.

Sources checked: AL002 phase 2 Nature Medicine 2026, BIIB078 Lancet Neurology 2024, WVE-004 ClinicalTrials.gov/AFTD reports, NADPARK Cell Metabolism 2022, ambroxol PDD JAMA Neurology 2025, tofersen FDA/NINDS summaries.