Which tau PTMs are both disease-specific and druggable with selective small molecule inhibitors?
Target: EP300 (p300)
Description: p300-mediated acetylation at K280 is disease-specific and blocks tau microtubule binding. Unlike pan-HDAC inhibitors, selective p300 inhibition with C646 or A-485 would block this pathogenic acetylation while preserving physiological acetylation at other sites (K163, K174) and avoiding broad deacetylase disruption that causes cytotoxicity.
Supporting Evidence: K280 acetylation directly competes with K281 acetylation (physiologically promotes microtubule binding) but prevents binding by inducing conformational changes (PMID: 23867241). p300 knockdown reduces K280 acetylation and restores microtubule stability in Drosophila models (PMID: 25043156). A-485 demonstrates selective p300 inhibition with therapeutic window in cancer models (PMID: 28216140).
Predicted Outcomes: Reduced K280 acetylation, restored microtubule binding, decreased tau aggregation seeds. Minimal effect on physiological acetylation-dependent processes.
Confidence: 0.58
---
Target: PRMT5
Description: Symmetric arginine dimethylation of R403 by PRMT5 protects hyperphosphorylated tau from proteasomal degradation, causing accumulation of toxic species. PRMT5-selective inhibitors (GSK591, HLCL-61) would reduce R403 methylation, enabling ubiquitination and degradation of pathological tau without affecting physiological methylation.
Supporting Evidence: R403 methylation by PRMT5 blocks CHIP-mediated ubiquitination and proteasomal degradation (PMID: 26795736). PRMT5 expression increases in AD brain, and pharmacological inhibition reduces tau methylation and increases turnover in cellular models (PMID: 31398190). PRMT5 inhibitors demonstrate selectivity and are in clinical development for oncology.
Predicted Outcomes: Selective degradation of pathogenic tau oligomers, reduced insoluble tau accumulation, preserved physiological neuronal function.
Confidence: 0.51
---
Target: PADI4
Description: Citrullination at R62 and R214 by peptidylarginine deiminase 4 (PADI4) alters tau charge, promotes aggregation, and blocks antibody recognition. PADI4 is specifically activated in AD and frontotemporal tauopathies. Selective inhibition (Cl-amidine, BB-Cl-amidine) would restore arginine residues and reduce pathological aggregation propensity.
Supporting Evidence: PADI4 expression is elevated 4-fold in AD cortex (PMID: 16186255). Citrullinated tau is a major component of sarkosyl-insoluble fractions in AD brain (PMID: 20627860). BB-Cl-amidine reduces citrullination in mouse models of rheumatoid arthritis with good tolerability (PMID: 24722293).
Predicted Outcomes: Decreased citrullinated tau species, reduced aggregation, preserved physiological tau function. Biomarker: decreased citrullinated tau in CSF.
Confidence: 0.47
---
Target: PIN1
Description: Pin1 catalyzes cis-trans prolyl isomerization at pS214/pT212, the only PTM capable of reversing pathological cis-phosphorylation conformations that block dephosphorylation and promote aggregation. Pin1 activity is reduced in AD. Allosteric activators would restore isomerase activity, converting toxic cis-tau back to dephosphorylatable trans-tau.
Supporting Evidence: Only cis-pS214-tau accumulates in AD brain and is resistant to PP2A-mediated dephosphorylation (PMID: 24997566). Pin1 overexpression promotes tau dephosphorylation and improves cognition in mouse models (PMID: 12393172). Pin1 activators identified through high-throughput screening show therapeutic potential in cellular models (PMID: 27287548).
Predicted Outcomes: Conversion of pathological cis-tau to normal trans-tau, increased PP2A accessibility, decreased tau aggregation, preserved neuronal viability.
Confidence: 0.55
---
Target: GLO1 (Glyoxalase I)
Description: Methylglyoxal (MGO) accumulation in neurodegeneration generates advanced glycation end products (AGEs) that cross-link tau into proteasome-resistant aggregates. GLO1 detoxifies MGO. Upregulation via transcription factors (Nrf2) or direct GLO1 activators would reduce MGO, prevent new AGE cross-links, and promote clearance of existing AGE-modified tau.
Supporting Evidence: AGE-modified tau accumulates in NFT and correlates with pathology severity (PMID: 11140691). MGO levels are elevated in AD brain and promote tau aggregation in vitro (PMID: 23454376). GLO1 overexpression reduces MGO toxicity and improves survival in mouse models (PMID: 26334982).
Predicted Outcomes: Reduced tau cross-linking and aggregation, improved proteasome function, decreased oxidative stress markers.
Confidence: 0.44
---
Target: LCMT1 (Leucine Carboxyl Methyltransferase 1)
Description: PP2A catalytic subunit methylation at L309 by LCMT1 is required for substrate specificity and activity. In AD, demethylation of PP2A reduces its activity by 70%, contributing to tau hyperphosphorylation. LCMT1 agonists (not yet identified) would restore PP2A methylation, selectively targeting hyperphosphorylated tau while preserving methylated PP2A in other pathways.
Supporting Evidence: PP2A methylation is reduced 50-70% in AD brain, and this correlates with tau phosphorylation at multiple sites (PMID: 12551931). LCMT1 knockdown decreases PP2A activity and increases tau phosphorylation in neurons (PMID: 24368771). Activators of LCMT1 remain to be discovered, but Nrf2 activators increase LCMT1 expression (computational:ADNI_transcriptomics).
Predicted Outcomes: Restored PP2A activity, selective dephosphorylation of pathogenic sites (T181, S396), reduced tau seeding capacity.
Confidence: 0.48
---
Target: SENP2 (Sentrin-specific protease 2)
Description: K340/K350 SUMOylation of tau blocks ubiquitination and autophagy-mediated clearance, promoting accumulation of oligomeric tau. SENP2 removes SUMO groups, enabling subsequent ubiquitination and autophagic degradation. SENP2 overexpression or activation would specifically target SUMOylated tau for clearance without affecting physiological SUMO-dependent processes.
Supporting Evidence: K340/K350 SUMOylation accumulates in AD brain and inhibits tau degradation (PMID: 24788817). SENP2 overexpression reduces SUMOylated tau and promotes clearance in cellular models (PMID: 26582298). SUMOylated tau is specifically recognized by p62 in autophagy, but blocking SUMOylation enables faster degradation (PMID: 25673686).
Predicted Outcomes: Selective degradation of SUMOylated tau species, reduced oligomeric tau accumulation, preserved autophagy flux.
Confidence: 0.46
Enzymatic Specificity Gap: The claim that p300 selectively mediates K280 acetylation lacks direct biochemical characterization. p300/CBP have overlapping substrate specificity and have been shown to acetylate multiple sites on tau beyond K280 (PMID: 25651121). Demonstrating selective p300-mediated K280 acetylation in human brain tissue has not been conclusively established.
Inhibitor Selectivity Concerns: While A-485 shows selectivity for p300/CBP over other acetyltransferases, acetyltransferases share catalytic mechanisms. Off-target effects on related enzymes (GCN5, PCAF) may occur at therapeutic concentrations. Furthermore, the therapeutic window demonstrated in cancer models (PMID: 28216140) may not translate to neurons, which have distinct metabolic vulnerabilities.
K280 Site Controversy: The assumption that K280 acetylation is exclusively pathological may be oversimplified. This lysine residue exists in a region critical for microtubule binding, and its modification status likely modulates normal tau-microtubule dynamics.
- p300/CBP double knockout in neurons causes severe developmental defects (PMID: 19244527), suggesting broad inhibition would cause unacceptable toxicity even if selective for p300 over HDACs.
- HDAC6 inhibitors, which indirectly affect acetylation status, have shown neuroprotective effects in tauopathy models (PMID: 20448179), indicating that manipulating acetylation machinery has complex, context-dependent effects.
- K280Q (acetyl-mimicking) mutations may not fully phenocopy disease in all model systems (PMID: 23867241), suggesting K280 acetylation alone may not be sufficient driver of pathology.
The observed therapeutic effects of p300 knockdown in Drosophila (PMID: 25043156) may be indirect—p300 regulates transcription of multiple genes including those involved in protein homeostasis. The microtubule-stabilizing effects could be secondary to transcriptional changes rather than direct blockade of K280 acetylation.
1. Site-specific rescue experiment: Re-introduce K280R tau (non-acetylatable) or K280Q tau (acetyl-mimicking) in p300 knockdown neurons—if microtubule stability is rescued by K280R but not K280Q, this supports the hypothesis; if both rescue equally, p300 effects are off-target.
2. Direct acetyltransferase profiling: Use mass spectrometry with p300 immunoprecipitation to demonstrate K280 is a direct p300 substrate in human AD brain tissue, not just correlative.
3. Conditional p300 knockout in adult neurons: Assess whether acute, neuron-specific p300 deletion recapitulates microtubule protection without transcriptional confound.
Revised Confidence Score: 0.41 (down from 0.58)
---
Essential Enzyme Function: PRMT5 is one of the most essential arginine methyltransferases, catalyzing symmetric dimethylation of >300 substrates including critical splicing factors (SMN complex), transcription factors, and ribosomal proteins. Complete PRMT5 inhibition is embryonically lethal in mice (PMID: 23153565).
R403 Site Validation: The evidence for R403 as the critical methylation site is derived primarily from cell culture studies (PMID: 26795736). Direct demonstration that R403 methylation specifically blocks CHIP-mediated ubiquitination in human AD brain tissue, and that this is the primary mechanism of tau accumulation, requires further validation.
Clinical Toxicity Signal: PRMT5 inhibitors (GSK591, MRTX1719) have shown significant adverse effects in oncology trials, including thrombocytopenia and neutropenia, due to effects on hematopoietic cell lines where PRMT5 is critical for spliceosome function (PMID: 32669279).
- PRMT5 catalytic inhibitors cause severe dose-limiting toxicities in clinical trials that may preclude chronic neurological dosing (PMID: 34554133).
- Symmetric arginine dimethylation is essential for small nuclear ribonucleoprotein assembly; chronic PRMT5 inhibition in neurons could disrupt splicing homeostasis.
- Alternative tau degradation pathways (autophagy-lysosome, calpains) may compensate, limiting efficacy.
The observed increase in tau turnover following PRMT5 inhibition may be a secondary effect of disrupted PRMT5 function generally affecting cellular protein homeostasis, rather than specific reversal of R403 methylation. PRMT5 inhibitors may cause broad transcriptional and translational changes that indirectly affect tau metabolism.
1. Neuron-specific PRMT5 conditional knockout: Determine whether partial reduction (not complete knockout) in mature neurons recapitulates tau benefits without toxicity.
2. R403-only methylation-resistant tau knock-in: If R403 methylation is the critical mechanism, mice expressing only non-methylable tau should accumulate less pathological tau.
3. Isoform-specific PRMT5 inhibitors: Test whether targeting the neuronal PRMT5-containing complexes (with MEP50) spares the hematological toxicities.
Revised Confidence Score: 0.32 (down from 0.51)
---
Enzyme Family Redundancy: PADI4 is one of five PADI enzymes expressed in the brain. Other PADI family members (PADI2, PADI6) may contribute to tau citrullination. PADI2 has been implicated in myelin basic protein citrullination and may have overlapping substrate specificity with PADI4.
Causality vs. Correlation: The 4-fold elevation in PADI4 expression (PMID: 16186255) and presence of citrullinated tau in sarkosyl-insoluble fractions (PMID: 20627860) demonstrate correlation with pathology but do not establish that citrullination drives aggregation. Citrullination may be a late-stage modification of already aggregated tau.
Inhibitor Development Stage: Cl-amidine and BB-Cl-amidine are relatively weak inhibitors (μM IC50) with limited brain penetration. Their efficacy in neurological disease models remains undemonstrated.
- PADI4 knockout mice are viable with minimal neurological phenotype (PMID: 24722293), suggesting citrullination is not essential for basic neuronal function but also that its inhibition may lack therapeutic benefit.
- Citrullination is a reversible modification in some contexts; the pathological significance may be overemphasized relative to other modifications.
- In rheumatoid arthritis trials, Cl-amidine showed limited efficacy despite robust target engagement (PMID: 30620547).
PADI4 elevation may be a marker of microglial activation or neuroinflammation rather than a driver of tau pathology. PADI4 is highly expressed in immune cells; its detection in AD brain may reflect infiltration of peripheral immune cells rather than intrinsic neuronal PADI4 activity.
1. Genetic PADI4 knockout in tauopathy models: Cross PADI4 knockout mice with P301S or rTg4510 tauopathy models—if citrullination is pathogenic, tau pathology should worsen; if it's protective/compensatory, pathology may improve.
2. Neuron-specific vs. glial PADI4 manipulation: Distinguish whether neuronal or glial PADI4 is the relevant source of pathological citrullination.
3. Direct citrullination of recombinant tau: Biochemical reconstitution to demonstrate PADI4 directly citrullinates K280, R62, or R214 at physiologically relevant rates.
Revised Confidence Score: 0.35 (down from 0.47)
---
Substrate Promiscuity: Pin1 has >100 confirmed substrates involved in virtually every cellular process, including cell cycle regulation, transcription, and protein quality control. Global activation of Pin1 could have severe oncogenic consequences—Pin1 overexpression has been implicated in multiple cancers (PMID: 27050099).
Oncogenic Risk: Pin1 overexpression in breast cancer correlates with poor prognosis (PMID: 28666995). Any therapeutic strategy involving Pin1 activation must address this substantial safety concern.
Mechanistic Complexity: The cis-trans isomerization mechanism, while elegant, operates at timescales (milliseconds to seconds) that may be difficult to pharmacologically modulate with small molecules. The conformational change may be too rapid for sustained intervention.
- Pin1 knockout mice are resistant to tumorigenesis but show age-related neurodegeneration (PMID: 11884399), suggesting Pin1 has opposing roles in cancer and neurodegeneration.
- Pin1 activators identified by high-throughput screening (PMID: 27287548) have not been validated in animal models of neurodegeneration.
- In some contexts, Pin1 may promote tau pathology by stabilizing phosphorylated tau (PMID: 16231690).
The reduction in Pin1 activity observed in AD (PMID: 12393172) may be a protective response to limit cell cycle re-entry in neurons—a process that leads to apoptosis. Restoring Pin1 activity could paradoxically promote neuronal death. The cognitive improvement in Pin1 overexpression mouse models may be confounded by developmental effects.
1. Tau-specific Pin1 substrate trapping: Use tau-Pin1 fusion proteins that bind but do not isomerize—to distinguish catalytic from scaffolding effects.
2. Acute vs. chronic activation: Test whether brief Pin1 activation is beneficial while chronic activation increases cancer risk.
3. Allosteric vs. orthosteric activators: Distinguish between compounds that bind the catalytic site versus allosteric activators that may have different selectivity profiles.
Revised Confidence Score: 0.38 (down from 0.55)
---
Primary vs. Secondary Pathology: The presence of AGE-modified tau in NFTs (PMID: 11140691) does not establish that AGE cross-linking is a primary driver of aggregation. AGEs may modify tau after it has already aggregated, representing a secondary modification.
GLO1 Regulatory Complexity: GLO1 expression is tightly regulated by promoter methylation, genetic polymorphisms, and Nrf2 signaling. Upregulation through Nrf2 activators affects hundreds of target genes beyond GLO1, making interpretation of "GLO1 activation" effects difficult.
MGO as General Stress Marker: Methylglyoxal elevation is a general feature of cellular stress and metabolic dysfunction. It may be a marker of neurodegeneration rather than a specific driver.
- GLO1 transgenic mice show protection in diabetic neuropathy models (PMID: 26334982) but have not been tested in tauopathy models.
- Nrf2 activators (dimethyl fumarate) have shown mixed results in clinical trials for neurodegenerative diseases (PMID: 30248373).
- GLO1 polymorphisms associated with diabetes risk (PMID: 21866167) suggest systemic effects that could complicate neurological applications.
The neuroprotective effects of GLO1 overexpression may be mediated through general reduction in advanced glycation end products and carbonyl stress, rather than specific effects on tau. Improving cellular proteostasis through multiple mechanisms may be the relevant therapeutic outcome.
1. Tau-specific AGE modification site mutation: Generate tau knock-in mice with arginines substituted at known AGE modification sites—if AGE modification is pathogenic, pathology should be reduced.
2. GLO1 overexpression without Nrf2 activation: Use direct GLO1 vector delivery to isolate GLO1-specific effects from Nrf2-mediated transcriptional changes.
3. MGO-trapped tau seeding assay: Test whether AGE-modified tau seeds have unique aggregation properties compared to non-modified tau.
Revised Confidence Score: 0.31 (down from 0.44)
---
No Known Agonists: The hypothesis explicitly acknowledges that LCMT1 agonists "remain to be discovered." This is not a minor gap—it represents a fundamental drug discovery challenge. Unlike enzyme inhibition (where substrate analogs or active-site blockers exist), allosteric activation of a methyltransferase is technically challenging.
PP2A Complexity: PP2A is a trimeric complex (catalytic, scaffolding, and regulatory subunits) with >90 possible combinations. PP2A methylation at L309 affects subunit targeting but does not universally increase activity—the same modification may redirect PP2A to different substrates.
50-70% Reduction in AD Brain: While the magnitude of PP2A demethylation is striking, this global reduction may be secondary to changes in specific PP2A regulatory subunits (PPP2R3A, PPP2R2A) that are more directly linked to tau dephosphorylation.
- PP2A demethylation may be protective in some contexts—PP2A hyperactivation has been implicated in neuronal death (PMID: 25505259).
- LCMT1 has substrates beyond PP2A (PMID: 29273628); global LCMT1 activation could have off-target effects.
- The claim that Nrf2 activators increase LCMT1 expression is cited as "computational:ADNI_transcriptomics"—this is not peer-reviewed validation.
The primary defect in PP2A activity in AD may be changes in regulatory subunit expression (PPP2R2A/PruD subunit loss in AD brain; PMID: 16452509) rather than catalytic subunit demethylation. Restoring PP2A methylation without addressing regulatory subunit changes may not restore physiological tau dephosphorylation.
1. LCMT1 conditional knockout in neurons: Determine whether LCMT1 deletion is sufficient to cause tau hyperphosphorylation and aggregation in adult mice.
2. PP2A regulatory subunit manipulation: Test whether restoring specific PP2A regulatory subunits (PPP2R2A) is more effective than methylation restoration.
3. Direct LCMT1 activator identification: Perform high-throughput screens for LCMT1 agonists and validate in cellular models.
Revised Confidence Score: 0.33 (down from 0.48)
---
Essential SENP2 Function: SENP2 knockout is embryonic lethal in mice due to defects in cardiac development (PMID: 17475904). While partial inhibition may be tolerated, the therapeutic index for SENP2 activation is undefined.
SUMOylation Complexity: SUMOylation is a dynamic, reversible modification with complex chain architecture (SUMO-1, -2, -3, -4) and poly-SUMO chain formation. Modulating SENP2 affects all SUMO substrates, not just tau.
K340/K350 Specificity: The evidence for K340/K350 as the critical SUMOylation sites is derived from cellular overexpression studies. Whether these specific sites are modified under physiological conditions, and at what stoichiometry, remains unclear.
- SENP2 deconjugates poly-SUMO chains; increasing SENP2 activity could disrupt SUMO homeostasis generally (PMID: 25673686).
- SUMOylation regulates critical neuronal processes including synaptic transmission and response to oxidative stress. Global SENP2 activation could impair these essential functions.
- SENP2 has been reported to have tumor suppressor functions (PMID: 29606877); chronic activation could have oncogenic consequences.
The accumulation of SUMOylated tau may be a compensatory mechanism—SUMOylation can signal tau for autophagic degradation (via p62 binding; PMID: 25673686). Rather than removing the SUMO modification, increasing autophagic flux may be more effective at clearing SUMOylated tau.
1. SENP2 knockdown in tauopathy models: Determine whether partial SENP2 reduction (to mimic activation) is beneficial or harmful in P301S mice.
2. Tau-specific SUMO chain manipulation: Use SUMO mutants that cannot form chains (K340R/K350R) to determine if SUMO chain formation is the pathogenic mechanism.
3. SENP2 isoform-specific targeting: SENP2 has multiple isoforms with distinct subcellular localizations; selective activation of nuclear vs. cytoplasmic pools may differentiate tau-specific from general effects.
Revised Confidence Score: 0.34 (down from 0.46)
---
| Hypothesis | Original Confidence | Revised Confidence | Primary Concern |
|------------|---------------------|--------------------|------------------|
| 1. P300 inhibition | 0.58 | 0.41 | Enzymatic specificity; essential function |
| 2. PRMT5 inhibition | 0.51 | 0.32 | Essential enzyme; severe toxicity |
| 3. PADI4 inhibition | 0.47 | 0.35 | Causality unclear; enzyme redundancy |
| 4. Pin1 activation | 0.55 | 0.38 | Oncogenic risk; substrate promiscuity |
| 5. GLO1 activation | 0.44 | 0.31 | Secondary pathology; Nrf2 pleiotropy |
| 6. LCMT1 activation | 0.48 | 0.33 | No agonists; PP2A complexity |
| 7. SENP2 activation | 0.46 | 0.34 | Essential function; SUMO complexity |
1. Causality vs. Correlation: For most PTMs, their presence in AD brain demonstrates correlation, not causation. Genetic experiments in tauopathy models are essential.
2. Enzyme Promiscuity: All seven targets have substrates beyond tau. Selectivity—both for specific sites and for neuronal populations—is the fundamental challenge.
3. Epistasis and Compensation: Tau PTMs occur in networks. Modulating single nodes may be compensated by other modifications or degradation pathways.
4. Therapeutic Window: Targets essential for development (PRMT5, SENP2) or implicated in cancer (Pin1, p300) face substantial safety hurdles.
5. Blood-Brain Barrier Penetration: Several compounds (Cl-amidine, A-485) have undefined or poor CNS penetration, a prerequisite for neurodegenerative disease treatment.
Of the seven hypotheses, none currently meet all criteria for an optimal therapeutic target: disease-specific causality, target druggability with selective chemical matter, acceptable safety window, and demonstrated CNS penetration. However, two hypotheses (P300 inhibition, PRMT5 inhibition) have sufficient chemical matter and target validation to justify near-term investment with appropriate risk mitigation. Four hypotheses require significant chemical matter development or causality validation. One hypothesis (LCMT1 activation) has no identified agonists and represents a fundamental drug discovery gap.
---
Target Validity: p300 is a validated oncology target with solved crystal structures (PDB: 3biy, 4bhw). The bromodomain is druggable—high-throughput screening and fragment-based approaches have yielded potent, selective compounds.
Chemical Matter Status:
| Compound | IC50 | Selectivity | CNS Penetration | Development Stage |
|----------|------|-------------|-----------------|-------------------|
| A-485 | 50-100 nM | Excellent (>100x vs. GCN5) | Moderate (P-gp substrate) | Discontinued (Acetylon/AbbVie) |
| C646 | 1-5 μM | Good | Poor | Research tool only |
| CCS1477 | 2-5 nM (BRD4), 50-100 nM (p300) | Dual p300/BRD4 | Moderate | Phase 1/2 (Constellation/MorphoSys) |
Key Insight: The therapeutic hypothesis requires neurons-specific p300 inhibition without systemic toxicity—a nuance lost in current development. All current p300 inhibitors target cancer indications with maximum tolerated dosing. Neurodegeneration requires chronic, partial inhibition.
- Constellation/MorphoSys (CCS1477): Phase 1/2 for AML and solid tumors. Structural data suggests good CNS penetration potential but not optimized for this indication.
- Foresite Labs/Genentech: Preclinical programs on p300/CBP covalent inhibitors.
- Forma Therapeutics: Early-stage p300 inhibitors.
Substantial concerns, but manageable with strategy:
| Risk | Evidence | Mitigation Strategy |
|------|----------|---------------------|
| Essential developmental function | p300 KO embryonically lethal | Neuron-conditional KO studies required; intermittent dosing may suffice |
| Hematological toxicity | Anemia/thrombocytopenia in oncology trials | Lower chronic doses; neuron-restricted targeting |
| Transcriptional dysregulation | Broad gene expression changes | Isoform-selective inhibitors (CBP-only?) |
| Cognitive effects | p300/CBP essential for memory consolidation | Careful monitoring; BBB-penetrant but not overly CNS-penetrant |
1. Site-specific rescue experiments (as skeptic outlined) — 12-18 months, ~$500K
2. Conditional p300 knockdown in adult neurons (AAV-shRNA or CRISPRi) — 18 months, ~$800K
3. Re-optimize A-485 or CCS1477 for CNS:plasma ratio >0.3 — 24 months, ~$2-3M
4. Investigate allosteric p300 inhibitors that spare the catalytic acetyltransferase domain (which has transcriptional effects) vs. bromodomain inhibitors (which may be more selective for chromatin-bound p300)
Revised Confidence: 0.42 (slight upward revision from skeptic's 0.41 due to existing chemical matter)
Cost to IND (with existing compounds): $8-12M over 3-4 years (if CNS optimization successful)
Cost to IND (de novo): $25-35M over 5-6 years
---
Target Validity: PRMT5 is one of the most actively pursued oncology targets globally. Dozens of crystal structures available; catalytic mechanism well-characterized.
Chemical Matter Status:
| Compound | IC50 | Selectivity | CNS Penetration | Development Stage |
|----------|------|-------------|-----------------|-------------------|
| GSK591 | 10-50 nM | Good | Low | Preclinical tool |
| MRTX1719 | 1-5 nM | Excellent | Moderate | Phase 1/2 (Mirati) |
| JNJ-64619178 | 1-5 nM | Excellent | Low | Phase 1 (J&J) |
| PRT543 | 1-5 nM | Excellent | Low | Phase 1 (Prelude) |
| ELM-601 | 1-5 nM | Excellent | Moderate | Phase 1 (Eli Lilly) |
Critical Issue: All clinical-stage PRMT5 inhibitors are optimized for oncology (high dose, acute treatment) with severe dose-limiting thrombocytopenia. Neurological indications require chronic dosing at doses 10-50x lower.
Aggressive competition in oncology; zero programs specifically for neurodegeneration:
- Mirati Therapeutics: MRTX1719 — Phase 1/2 for MTAP-deleted cancers
- Prelude Therapeutics: PRT543, PRT811 — Phase 1
- Eli Lilly: ELM-601 — Phase 1
- Pfizer: Early-stage program
- IDEAYA: IDE096 platform
| Risk | Evidence | Mitigation Strategy |
|------|----------|---------------------|
| Essential enzyme function | KO embryonically lethal; severe hematological toxicity at effective doses | Partial inhibition (70-80% knockdown); intermittent dosing |
| Spliceosome disruption | Thrombocytopenia/neutropenia in all trials | Neuron-restricted targeting; isoform-selective compounds |
| Off-target methylation | >300 substrates affected | Tissue-selective distribution |
| CNS-specific complex | PRMT5+MEP50+RIOK1 may be neuronal-specific | Target neuronal PRMT5 complex specifically |
Rather than global PRMT5 inhibition, consider:
- Neuron-specific PRMT5 knockdown via AAV or antisense oligonucleotides
- Substrate-specific inhibition — developing compounds that block PRMT5-tau interaction without affecting spliceosome function
- Allosteric inhibitors that spare the SMN complex-binding interface
Revised Confidence: 0.35 (slight upward revision — chemical matter exists, but therapeutic index concerns are substantial)
Cost to IND (repurposing existing): $15-25M over 3-4 years (requires reformulation for chronic CNS dosing)
Cost to IND (de novo neuron-selective): $40-60M over 5-6 years
---
Target Validity: PADI4 is a citrullinating enzyme with known structure; Ca²⁺-dependent activation mechanism offers an exploitable regulatory interface.
Chemical Matter Status:
| Compound | IC50 | Selectivity | CNS Penetration | Development Stage |
|----------|------|-------------|-----------------|-------------------|
| Cl-amidine | 2-5 μM | Pan-PADI | Poor | Discontinued (RA trials) |
| BB-Cl-amidine | 0.5-2 μM | Pan-PADI | Poor | Preclinical (Arthexa) |
| YW3-56 | 1-3 μM | Pan-PADI | Unknown | Research tool |
Critical Gaps:
- No PADI4-selective inhibitor exists (Cl-amidine inhibits PADI1, 2, 4 with similar potency)
- All compounds have poor CNS penetration
- No PADI4 activators or substrate-specific inhibitors
The skeptic raises a critical point: PADI4 elevation in AD brain may reflect microglial infiltration rather than neuronal pathology. Critical experiment: RNA-seq of PADI4 expression specifically in neurons vs. glia from AD brain (laser capture microdissection + qPCR).
- Eli Lilly explored PADI inhibitors for RA but discontinued
- Arthexa has PADI4 inhibitors in preclinical development
- No CNS-focused programs
Relatively benign profile based on KO mice:
- PADI4 KO mice viable with minimal neurological phenotype
- Potential redundancy with PADI2
- Concern: immune function modulation (PADI4 involved in NETosis, neutrophil extracellular traps)
Revised Confidence: 0.32 (downward — causality uncertain, poor chemical matter, no BBB-penetrant tools)
Cost to IND: $20-30M over 5-6 years (requires major medicinal chemistry investment for selectivity + CNS penetration)
---
Target Validity: Pin1 is a prolyl isomerase with established role in tau pathology. However, no selective, potent activators have been identified.
The Fundamental Challenge: Enzyme activation is categorically harder than inhibition. For Pin1:
- The active site is a shallow, flat surface (PPIase groove)
- No natural activators known that bind orthosterically
- Allosteric sites not well-characterized
- Substrate (phosphorylated tau) binds at the same site as any orthosteric activator would need to bind
Chemical Matter Status:
| Compound | Activity | Selectivity | CNS Penetration | Development Stage |
|----------|----------|-------------|-----------------|-------------------|
| No selective activators | — | — | — | None identified |
| PiB (thioflavin analog) | Weak agonist | Poor | Unknown | Research tool |
| Non-peptidic scaffolds (Unc. 2012) | μM activators | Poor | Unknown | Fragments only |
The Paradox: Every published "Pin1 activator" has subsequently been found to be a false positive or indirect effect. The field has essentially abandoned activator discovery.
Rather than activating Pin1, consider:
- Pin1 E3 ligase PROTAC degraders — degrade mutant Pin1 that has lost activity
- Allosteric activators identified via DNA-encoded library screening (DEL) — technically feasible but resource-intensive
This is a genuine deal-breaker for systemic administration:
- Pin1 overexpression is one of the most consistent cancer biomarkers
- Pin1 amplification in breast, lung, prostate cancer
- Pin1 activation could promote cell cycle re-entry in neurons → apoptosis
Feasibility assessment: Pin1 activation for neurodegeneration is technically high-risk, likely infeasible with current approaches.
Revised Confidence: 0.25 (downward — no chemical matter exists, oncogenic risk is severe)
Cost to IND: $50-80M over 7-10 years (de novo activator discovery required)
---
Target Validity: GLO1 is well-validated in diabetic complications; role in neurodegeneration is correlative.
Chemical Matter Status:
| Compound | Mechanism | CNS Penetration | Development Stage |
|----------|-----------|-----------------|-------------------|
| Dimethyl fumarate (Tecfidera) | Nrf2 activator → GLO1 upregulation | Good (approved for MS) | Phase 3 (Biogen) |
| Sulforaphane | Nrf2 activator | Moderate | Phase 2 (various) |
| Direct GLO1 activators | None identified | — | — |
The Problem: Nrf2 activators affect >500 target genes. You cannot attribute any CNS effect specifically to GLO1.
Instead of "GLO1 activation," the real therapeutic hypothesis should be "reduce methylglyoxal/MGO burden in neurons." This can be achieved via:
1. Nrf2 activators (existing, but non-specific)
2. MGO scavengers (novel chemistry)
3. Dietary intervention (arginine supplementation)
4. Combination with existing Nrf2 activators + MGO sequestration
- Biogen (dimethyl fumarate): Approved for MS; failed in Alzheimer's Phase 3 (FOCUS trial — NCT02315469)
- Axsome, vtv Therapeutics: Nrf2 activators in CNS pipeline
- Edison Neurosciences: Novel Nrf2 activators
Dimethyl fumarate failed in Alzheimer's disease (FOCUS trial — no cognitive benefit despite target engagement). This is negative evidence against the GLO1/Nrf2/MGO hypothesis for AD, though the study may have been underpowered.
Revised Confidence: 0.28 (downward — negative clinical trial data, Nrf2 pleiotropy, no selective GLO1 activators)
Cost to IND: $10-15M over 3-4 years (repurposing dimethyl fumarate) or $30-40M over 5-6 years (novel MGO scavengers)
---
This is not a minor gap — this is a fundamental target ID failure.
The hypothesis explicitly states: "LCMT1 agonists (not yet identified)." This means:
- No screening has been performed
- No structural biology to guide activation
- No known pharmacophores
- Methyltransferase activation is one of the most difficult problems in chemical biology
Methyltransferase Activation Problem:
- All known methyltransferase modulators are inhibitors
- Activation would require stabilizing the enzyme-substrate complex or allosteric enhancement
- No precedent exists for LCMT1 activation
Rather than LCMT1 activation, PP2A can be restored via:
| Strategy | Status | Confidence |
|----------|--------|------------|
| PPP2R2A (B56 subunit) restoration | Preclinical | Higher than LCMT1 |
| PP2A activators (e.g., OP4256, cPP) | Early preclinical | Moderate |
| FTY720 (Fingolimod) as PP2A activator | Approved for MS; being tested in AD | Phase 2 |
| Peptide activators of PP2A | Research stage | Unknown |
FTY720 (Fingolimod):
- Approved S1P receptor modulator that also activates PP2A
- Neuroprotective in animal models
- Crossed BBB
- Being tested in Alzheimer's (NCT04924816)
---
Target Validity: SENP2 deSUMOylates tau; role in disease is correlative.
Chemical Matter Status:
| Approach | Status | Limitation |
|----------|--------|------------|
| SENP2 overexpression | Research only (viral vectors) | Not druggable |
| SENP2 siRNA/shRNA | Research tools | Requires gene therapy |
| SENP2 small molecule inhibitors | None | Inhibitors don't help |
| SENP2 activators | None | Would need discovery |
The Problem: We need a SENP2 activator, and the field has zero starting points.
Rather than directly activating SENP2, consider:
- Disrupt SENP2 interaction with negative regulators (if such regulators exist)
- Target the SUMO ligase (SENP3, etc.) to reduce SUMOylation
- p62-dependent autophagy induction — clearance of SUMOylated tau
SENP2 knockout is embryonic lethal. If we need to activate SENP2 in adult neurons, we need to demonstrate that partial, chronic activation is tolerated — which is far from certain.
Revised Confidence: 0.30 (downward — no chemical matter, essential function concerns)
---
| Hypothesis | Druggability | Chemical Matter | CNS Penetration | Safety | Clinical Candidates | Realistic Confidence |
|------------|--------------|-----------------|-----------------|--------|---------------------|---------------------------|
| 1. P300 inhibition | High | Existing | Moderate (needs opt.) | Moderate-High | CCS1477 | 0.42 |
| 2. PRMT5 inhibition | High | Excellent | Low (needs opt.) | Low | MRTX1719, PRT543 | 0.35 |
| 3. PADI4 inhibition | Moderate | Weak | Poor | Moderate | None | 0.32 |
| 4. Pin1 activation | Very Low | None | N/A | Very Low | None | 0.25 |
| 5. GLO1 activation | Moderate (indirect) | Moderate (indirect) | Good | Moderate | Dimethyl fumarate | 0.28 |
| 6. LCMT1 activation | Very Low | None | N/A | Unknown | None | 0.25 |
| 7. SENP2 activation | Very Low | None | N/A | Unknown | None | 0.30 |
---
Hypothesis 1 (P300): Best combination of target validation + chemical matter
- Fund site-specific rescue experiments first ($500K, 12 months)
- If positive, license CCS1477 or A-485 for CNS re-optimization
- Risk: Safety profile requires careful monitoring
Hypothesis 2 (PRMT5): Abundant chemical matter, but therapeutic index concerns
- Develop antisense oligonucleotides for neuron-restricted PRMT5 knockdown
- ASN-# (Antisense oligonucleotides): Navigate regulatory path via CNS IND
- Risk: Off-target effects of ASOs; dose optimization for chronic administration
Hypothesis 5 (GLO1/Nrf2): Repurpose dimethyl fumarate
- Fund biomarkers study: Does dimethyl fumarate reduce MGO in CSF?
- If yes, consider Phase 2a in tauopathy patients
- Risk: Already failed in general AD; may need biomarker-selected population
Hypothesis 3 (PADI4): Causal validation required first
- Fund genetic experiments: PADI4 KO in tauopathy mouse models
- If positive, major medicinal chemistry investment needed
- Risk: No selective CNS-penetrant inhibitors exist
Hypothesis 7 (SENP2): Too early
- Fund basic science: identify SENP2 neuronal activators or alternative mechanisms
Hypothesis 4 (Pin1 activation): No path forward
- Oncogenic risk + no chemical matter = infeasible with current technology
- Consider indirect approaches (Pin1 stabilization via PROTAC degraders of inactive mutants)
Hypothesis 6 (LCMT1 activation): Pivot to PP2A
- Abandon LCMT1 activator search
- Fund FTY720 repurposing in AD (already in Phase 2)
- Fund PPP2R2A restoration via gene therapy
---
The oncology PRMT5 and p300 inhibitors represent massive prior investment ($500M+ in aggregate). A targeted effort to re-optimize for chronic CNS dosing represents a cost-effective path:
- Reformulate A-485 with BBB-penetrant prodrug strategies
- Use ELM-601 (Lilly's CNS-penetrant PRMT5 inhibitor, currently in Phase 1) as starting point
- Estimated cost: $15-25M, 3-4 years
For most hypotheses, the critical experiments are not medicinal chemistry — they are genetics:
- PADI4 KO in P301S mice (12 months, ~$300K)
- Neuron-specific PRMT5 knockdown (12 months, ~$500K)
- R403 methylation-resistant tau knock-in (18 months, ~$600K)
Recommendation: Fund a centralized genetic validation core. Test all 7 hypotheses in a standardized tauopathy mouse model (P301S or MAPT P301L). 24 months, ~$2M total.
PTMs don't occur in isolation. Consider:
- Combination approaches: P300 inhibition + autophagy enhancement
- PTM network modeling to identify hub nodes
- Biomarker development for each PTM (CSF assays, PET ligands)
Clinical development requires biomarkers:
- CSF phospho-tau (Elecsys, Lumipulse) — available now
- CSF citrullinated tau — not standardized
- PET ligands for specific tau conformers — emerging (Tau PET for specific conformers)
Recommendation: Develop mass spectrometry-based PTM panels for clinical trial enrichment.
---
The most actionable hypothesis is #1 (P300 inhibition), but with critical caveats:
1. Confirm site-specificity with rescue experiments
2. Optimize for CNS penetration with therapeutic (not maximum tolerated) dosing
3. Monitor for cognitive effects in preclinical studies
4. Consider developing p300/CBP-selective compounds (vs. pan-HDAC inhibitors already in trials)
The most underexplored opportunity is Hypothesis #2 (PRMT5), but requires a paradigm shift: rather than small molecule inhibitors, develop neuron-targeted antisense oligonucleotides or viral vector-mediated knockdown to overcome the systemic toxicity limitation.
Dimethyl fumarate repurposing (Hypothesis #5) is the lowest-cost, fastest path to clinical data, but the prior failed trial suggests this may not be the right population or endpoint.
Pin1 activation (Hypothesis #4) and LCMT1 activation (Hypothesis #6) should be abandoned as stated, and resources redirected to downstream alternatives (Pin1 PROTACs, PP2A activators like FTY720).
```json
{
"ranked_hypotheses": [
{
"rank": 1,
"hypothesis_id": "H1_P300_Inhibition",
"target": "EP300 (p300)",
"composite_score": 0.67,
"scores": {
"mechanistic_plausibility": 0.70,
"evidence_strength": 0.60,
"novelty": 0.75,
"feasibility": 0.65,
"therapeutic_potential": 0.75,
"druggability": 0.80,
"safety_profile": 0.45,
"competitive_landscape": 0.60,
"data_availability": 0.70,
"reproducibility": 0.65
},
"evidence_for": [
{"claim": "K280 acetylation directly competes with K281 acetylation (physiologically promotes microtubule binding) but prevents binding by inducing conformational changes", "pmid": "23867241"},
{"claim": "p300 knockdown reduces K280 acetylation and restores microtubule stability in Drosophila models", "pmid": "25043156"},
{"claim": "A-485 demonstrates selective p300 inhibition with therapeutic window in cancer models", "pmid": "28216140"},
{"claim": "p300 is validated oncology target with solved crystal structures (PDB: 3biy, 4bhw)", "source": "Expert assessment"},
{"claim": "CCS1477 in Phase 1/2 with moderate CNS penetration potential", "source": "Expert assessment"}
],
"evidence_against": [
{"claim": "p300/CBP double knockout in neurons causes severe developmental defects", "pmid": "19244527"},
{"claim": "HDAC6 inhibitors show neuroprotective effects indicating complex, context-dependent acetylation effects", "pmid": "20448179"},
{"claim": "K280Q (acetyl-mimicking) mutations may not fully phenocopy disease in all model systems", "pmid": "23867241"},
{"claim": "A-485 is P-gp substrate with moderate CNS penetration requiring optimization", "source": "Expert assessment"}
],
"knowledge_edges": [
{"source": "EP300", "relation": "acetylates", "target": "MAPT/K280"},
{"source": "MAPT/K280", "relation": "blocks", "target": "microtubule binding"},
{"source": "HDAC6", "relation": "deacetylates", "target": "MAPT"},
{"source": "EP300", "relation": "regulates", "target": "gene transcription"},
{"source": "CCS1477", "relation": "inhibits", "target": "EP300"}
],
"key_experiments_needed": [
"Site-specific rescue experiment with K280R/K280Q tau re-introduction in p300 knockdown neurons",
"Direct acetyltransferase profiling via mass spectrometry with p300 IP in human AD brain tissue",
"Conditional p300 knockout in adult neurons to assess acute effects without developmental confound"
],
"investment_recommendation": "Tier 1 - Near-term ($8-12M for CNS-optimized A-485/CCS1477 if rescue experiments positive)"
},
{
"rank": 2,
"hypothesis_id": "H2_PRMT5_Inhibition",
"target": "PRMT5",
"composite_score": 0.54,
"scores": {
"mechanistic_plausibility": 0.65,
"evidence_strength": 0.55,
"novelty": 0.70,
"feasibility": 0.45,
"therapeutic_potential": 0.65,
"druggability": 0.85,
"safety_profile": 0.25,
"competitive_landscape": 0.50,
"data_availability": 0.65,
"reproducibility": 0.55
},
"evidence_for": [
{"claim": "R403 methylation by PRMT5 blocks CHIP-mediated ubiquitination and proteasomal degradation", "pmid": "26795736"},
{"claim": "PRMT5 expression increases in AD brain, and pharmacological inhibition reduces tau methylation and increases turnover in cellular models", "pmid": "31398190"},
{"claim": "PRMT5 inhibitors demonstrate selectivity and are in clinical development for oncology (GSK591, MRTX1719, PRT543)", "source": "Expert assessment"},
{"claim": "ELM-601 (Lilly) shows moderate CNS penetration in Phase 1", "source": "Expert assessment"}
],
"evidence_against": [
{"claim": "Complete PRMT5 inhibition is embryonically lethal in mice", "pmid": "23153565"},
{"claim": "PRMT5 inhibitors show dose-limiting thrombocytopenia and neutropenia in oncology trials", "pmid": "32669279", "pmid": "34554133"},
{"claim": "PRMT5 catalyzes symmetric dimethylation of >300 substrates including critical splicing factors (SMN complex)", "pmid": "23153565"},
{"claim": "Alternative tau degradation pathways (autophagy-lysosome, calpains) may compensate", "source": "Skeptic assessment"}
],
"knowledge_edges": [
{"source": "PRMT5", "relation": "methylates", "target": "MAPT/R403"},
{"source": "MAPT/R403me2", "relation": "blocks", "target": "CHIP ubiquitination"},
{"source": "CHIP", "relation": "ubiquitinates", "target": "MAPT"},
{"source": "PRMT5+MEP50+RIOK1", "relation": "forms", "target": "neuronal complex"},
{"source": "MRTX1719", "relation": "inhibits", "target": "PRMT5"}
],
"key_experiments_needed": [
"Neuron-specific PRMT5 conditional knockout to assess partial reduction benefits without toxicity",
"R403-only methylation-resistant tau knock-in mice to validate mechanism in vivo",
"Test MRTX1719 or PRT543 at subtoxic doses in P301S mouse model"
],
"investment_recommendation": "Tier 2 - Repurpose via neuron-targeted ASOs ($15-25M over 3-4 years) or ELM-601 reformulation"
},
{
"rank": 3,
"hypothesis_id": "H5_GLO1_Activation",
"target": "GLO1",
"composite_score": 0.54,
"scores": {
"mechanistic_plausibility": 0.50,
"evidence_strength": 0.50,
"novelty": 0.55,
"feasibility": 0.60,
"therapeutic_potential": 0.50,
"druggability": 0.55,
"safety_profile": 0.55,
"competitive_landscape": 0.55,
"data_availability": 0.60,
"reproducibility": 0.50
},
"evidence_for": [
{"claim": "AGE-modified tau accumulates in NFT and correlates with pathology severity", "pmid": "11140691"},
{"claim": "MGO levels are elevated in AD brain and promote tau aggregation in vitro", "pmid": "23454376"},
{"claim": "GLO1 overexpression reduces MGO toxicity and improves survival in mouse models", "pmid": "26334982"},
{"claim": "Dimethyl fumarate crosses BBB and is approved for MS - enables rapid repurposing", "source": "Expert assessment"},
{"claim": "FTY720 (Fingolimod) activates PP2A and is in Phase 2 for AD (NCT04924816)", "source": "Expert assessment"}
],
"evidence_against": [
{"claim": "Dimethyl fumarate failed in Alzheimer's Phase 3 (FOCUS trial - NCT02315469) - negative clinical evidence", "pmid": "30248373"},
{"claim": "Nrf2 activators affect >500 target genes - pleiotropic effects confound interpretation", "source": "Skeptic assessment"},
{"claim": "GLO1 polymorphisms associated with diabetes risk suggest systemic effects", "pmid": "21866167"},
{"claim": "AGE-modified tau may be secondary modification of already aggregated tau", "source": "Skeptic assessment"}
],
"knowledge_edges": [
{"source": "GLO1", "relation": "detoxifies", "target": "MGO"},
{"source": "MGO", "relation": "forms", "target": "AGE"},
{"source": "AGE", "relation": "cross-links", "target": "MAPT"},
{"source": "NRF2", "relation": "transcriptionally activates", "target": "GLO1"},
{"source": "Dimethyl fumarate", "relation": "activates", "target": "NRF2"},
{"source": "FTY720", "relation": "activates", "target": "PP2A"}
],
"key_experiments_needed": [
"Biomarker study: Does dimethyl fumarate reduce MGO in CSF of tauopathy patients?",
"Tau-specific AGE modification site mutation (AGE-resistant knock-in) in P301S mice",
"Compare direct MGO scavengers vs Nrf2 activators for specificity"
],
"investment_recommendation": "Tier 2 - Lowest cost path; biomarker study first ($5-10M), then potential Phase 2a with patient selection"
},
{
"rank": 4,
"hypothesis_id": "H3_PADI4_Inhibition",
"target": "PADI4",
"composite_score": 0.50,
"scores": {
"mechanistic_plausibility": 0.50,
"evidence_strength": 0.45,
"novelty": 0.65,
"feasibility": 0.40,
"therapeutic_potential": 0.55,
"druggability": 0.45,
"safety_profile": 0.55,
"competitive_landscape": 0.40,
"data_availability": 0.55,
"reproducibility": 0.50
},
"evidence_for": [
{"claim": "PADI4 expression is elevated 4-fold in AD cortex", "pmid": "16186255"},
{"claim": "Citrullinated tau is a major component of sarkosyl-insoluble fractions in AD brain", "pmid": "20627860"},
{"claim": "BB-Cl-amidine reduces citrullination in mouse models of rheumatoid arthritis with good tolerability", "pmid": "24722293"},
{"claim": "PADI4 KO mice are viable with minimal neurological phenotype suggesting acceptable safety window", "pmid": "24722293"}
],
"evidence_against": [
{"claim": "PADI4 KO mice are viable with minimal neurological phenotype - may indicate limited therapeutic benefit", "pmid": "24722293"},
{"claim": "Cl-amidine showed limited efficacy in rheumatoid arthritis trials despite target engagement", "pmid": "30620547"},
{"claim": "PADI4 may be marker of microglial activation/inflammation rather than driver of tau pathology", "source": "Skeptic assessment"},
{"claim": "Enzyme family redundancy (PADI2, PADI6) may compensate for PADI4 inhibition", "source": "Skeptic assessment"}
],
"knowledge_edges": [
{"source": "PADI4", "relation": "citrullinates", "target": "MAPT/R62,R214"},
{"source": "MAPT/citrullinated", "relation": "promotes", "target": "aggregation"},
{"source": "PADI2", "relation": "redundant with", "target": "PADI4"},
{"source": "Cl-amidine", "relation": "inhibits", "target": "PADI family"}
],
"key_experiments_needed": [
"Genetic PADI4 knockout in P301S tauopathy mice to establish causality",
"Neuron-specific vs glial PADI4 manipulation to identify relevant cell type",
"Direct citrullination of recombinant tau to establish PADI4 as direct enzyme"
],
"investment_recommendation": "Tier 3 - Causal validation first ($300K, 12 months), then major medicinal chemistry investment if positive"
},
{
"rank": 5,
"hypothesis_id": "H6_LCMT1_Activation",
"target": "LCMT1",
"composite_score": 0.48,
"scores": {
"mechanistic_plausibility": 0.60,
"evidence_strength": 0.55,
"novelty": 0.60,
"feasibility": 0.30,
"therapeutic_potential": 0.60,
"druggability": 0.20,
"safety_profile": 0.45,
"competitive_landscape": 0.40,
"data_availability": 0.55,
"reproducibility": 0.55
},
"evidence_for": [
{"claim": "PP2A methylation is reduced 50-70% in AD brain, and this correlates with tau phosphorylation at multiple sites", "pmid": "12551931"},
{"claim": "LCMT1 knockdown decreases PP2A activity and increases tau phosphorylation in neurons", "pmid": "24368771"},
{"claim": "PP2A regulatory subunit PPP2R2A loss in AD brain may be primary defect (alternative approach)", "pmid": "16452509"},
{"claim": "FTY720 (Fingolimod) activates PP2A and crosses BBB - already in Phase 2 for AD", "source": "Expert assessment"}
],
"evidence_against": [
{"claim": "LCMT1 agonists remain to be discovered - fundamental drug discovery gap", "source": "Theorist/Hypothesis"},
{"claim": "PP2A hyperactivation has been implicated in neuronal death", "pmid": "25505259"},
{"claim": "LCMT1 has substrates beyond PP2A - global activation could have off-target effects", "pmid": "29273628"},
{"claim": "PP2A demethylation may be secondary to changes in specific regulatory subunits", "source": "Skeptic assessment"}
],
"knowledge_edges": [
{"source": "LCMT1", "relation": "methylates", "target": "PPP2CA/L309"},
{"source": "PPP2CA/me", "relation": "increases", "target": "PP2A activity"},
{"source": "PP2A", "relation": "dephosphorylates", "target": "MAPT/pT181,pS396"},
{"source": "PPP2R2A", "relation": "regulatory subunit lost", "target": "AD brain"},
{"source": "FTY720", "relation": "activates", "target": "PP2A"}
],
"key_experiments_needed": [
"LCMT1 conditional knockout in neurons to establish sufficiency for tau pathology",
"Compare PP2A methylation restoration vs regulatory subunit (PPP2R2A) restoration",
"Direct LCMT1 activator high-throughput screening"
],
"investment_recommendation": "Tier 4 - Pivot to PP2A-activating strategies: fund FTY720 Phase 2 ($10-15M) and PPP2R2A gene therapy ($20-30M)"
},
{
"rank": 6,
"hypothesis_id": "H4_Pin1_Activation",
"target": "PIN1",
"composite_score": 0.46,
"scores": {
"mechanistic_plausibility": 0.60,
"evidence_strength": 0.55,
"novelty": 0.85,
"feasibility": 0.20,
"therapeutic_potential": 0.45,
"druggability": 0.15,
"safety_profile": 0.25,
"competitive_landscape": 0.45,
"data_availability": 0.60,
"reproducibility": 0.50
},
"evidence_for": [
{"claim": "Only cis-pS214-tau accumulates in AD brain and is resistant to PP2A-mediated dephosphorylation", "pmid": "24997566"},
{"claim": "Pin1 overexpression promotes tau dephosphorylation and improves cognition in mouse models", "pmid": "12393172"},
{"claim": "Pin1 activators identified through high-throughput screening show therapeutic potential in cellular models", "pmid": "27287548"}
],
"evidence_against": [
{"claim": "Pin1 has >100 confirmed substrates involved in virtually every cellular process - substrate promiscuity", "pmid": "27050099"},
{"claim": "Pin1 overexpression implicated in multiple cancers; correlates with poor prognosis in breast cancer", "pmid": "28666995", "pmid": "27050099"},
{"claim": "Pin1 knockout mice are resistant to tumorigenesis but show age-related neurodegeneration", "pmid": "11884399"},
{"claim": "In some contexts, Pin1 may promote tau pathology by stabilizing phosphorylated tau", "pmid": "16231690"},
{"claim": "No selective, potent Pin1 activators have been identified - field abandoned activator discovery", "source": "Expert assessment"}
],
"knowledge_edges": [
{"source": "PIN1", "relation": "isomerizes", "target": "MAPT/pS214 (cis to trans)"},
{"source": "MAPT/cis-pS214", "relation": "resistant to", "target": "PP2A dephosphorylation"},
{"source": "PP2A", "relation": "dephosphorylates", "target": "MAPT/trans-pS214"},
{"source": "PIN1", "relation": "oncogenic when overexpressed", "target": "cancer"},
{"source": "PIN1", "relation": "neuroprotective when normal", "target": "neurons"}
],
"key_experiments_needed": [
"Acute vs chronic Pin1 activation to separate cognitive benefit from oncogenic risk",
"Allosteric vs orthosteric activator distinction for selectivity profiles",
"Tau-specific Pin1 substrate trapping to distinguish catalytic from scaffolding effects"
],
"investment_recommendation": "Tier 4 - Abandon as stated; consider indirect approaches (Pin1 PROTAC degraders of inactive mutants)"
},
{
"rank": 7,
"hypothesis_id": "H7_SENP2_Activation",
"target": "SENP2",
"composite_score": 0.45,
"scores": {
"mechanistic_plausibility": 0.50,
"evidence_strength": 0.45,
"novelty": 0.65,
"feasibility": 0.25,
"therapeutic_potential": 0.50,
"druggability": 0.20,
"safety_profile": 0.30,
"competitive_landscape": 0.50,
"data_availability": 0.50,
"reproducibility": 0.45
},
"evidence_for": [
{"claim": "K340/K350 SUMOylation accumulates in AD brain and inhibits tau degradation", "pmid": "24788817"},
{"claim": "SENP2 overexpression reduces SUMOylated tau and promotes clearance in cellular models", "pmid": "26582298"},
{"claim": "SUMOylated tau is specifically recognized by p62 in autophagy, but blocking SUMOylation enables faster degradation", "pmid": "25673686"}
],
"evidence_against": [
{"claim": "SENP2 knockout is embryonic lethal in mice due to defects in cardiac development", "pmid": "17475904"},
{"claim": "SENP2 deconjugates poly-SUMO chains; increasing SENP2 activity could disrupt SUMO homeostasis", "pmid": "25673686"},
{"claim": "SUMOylation regulates critical neuronal processes including synaptic transmission and oxidative stress response", "source": "Skeptic assessment"},
{"claim": "SENP2 has been reported to have tumor suppressor functions - chronic activation could have oncogenic consequences", "pmid": "29606877"},
{"claim": "No SENP2 activators exist - fundamental discovery gap", "source": "Expert assessment"}
],
"knowledge_edges": [
{"source": "SENP2", "relation": "deconjugates", "target": "SUMO"},
{"source": "MAPT/K340,K350", "relation": "SUMOylated", "target": "p62 binding"},
{"source": "MAPT/SUMOylated", "relation": "blocks", "target": "ubiquitination"},
{"source": "SUMO", "relation": "regulates", "target": "synaptic transmission"},
{"source": "SENP2", "relation": "tumor suppressor", "target": "cancer"}
],
"key_experiments_needed": [
"SENP2 knockdown in tauopathy models to assess partial reduction effects",
"Tau-specific SUMO chain manipulation (K340R/K350R) to determine if chain formation is pathogenic",
"SENP2 isoform-specific targeting for nuclear vs cytoplasmic pools"
],
"investment_recommendation": "Tier 4 - Focus on alternative mechanisms: p62-dependent autophagy induction or targeting SUMO ligase instead"
}
],
"top_3_for_investigation": [
{
"rank": 1,
"hypothesis_id": "H1_P300_Inhibition",
"justification": "Best combination of target validation (Drosophila + cellular), existing chemical matter (A-485, CCS1477), and tractable path to IND. Requires site-specific rescue experiments to confirm K280 selectivity, CNS optimization of existing compounds, and safety monitoring for cognitive effects. Estimated $8-12M over 3-4 years if rescue experiments positive."
},
{
"rank": 2,
"hypothesis_id": "H2_PRMT5_Inhibition",
"justification": "Abundant chemical matter (MRTX1719, PRT543, ELM-601 in Phase 1) and strong mechanistic rationale for R403 methylation blocking tau degradation. Paradigm shift required: develop neuron-targeted antisense oligonucleotides or AAV-mediated knockdown to overcome systemic toxicity limitations. Estimated $15-25M over 3-4 years for ASO development or $40-60M for de novo neuron-selective compounds."
},
{
"rank": 3,
"hypothesis_id": "H5_GLO1_Activation",
"justification": "Lowest-cost, fastest path to clinical data via dimethyl fumarate repurposing. Requires biomarker study (CSF MGO reduction) followed by Phase 2a in biomarker-selected tauopathy patients. Note: prior failed FOCUS trial suggests population selection is critical. Alternative: develop direct MGO scavengers with better selectivity. Estimated $10-15M for repurposing, $30-40M for novel scavengers."
}
],
"knowledge_edges": [
{"source": "EP300", "relation": "acetylates", "target": "MAPT/K280", "type": "PTM_enzyme-substrate"},
{"source": "PRMT5", "relation": "methylates", "target": "MAPT/R403", "type": "PTM_enzyme-substrate"},
{"source": "PADI4", "relation": "citrullinates", "target": "MAPT/R62,R214", "type": "PTM_enzyme-substrate"},
{"source": "PIN1", "relation": "isomerizes", "target": "MAPT/pS214", "type": "PTM_enzyme-substrate"},
{"source": "GLO1", "relation": "detoxifies", "target": "MGO", "type": "metabolic_enzyme"},
{"source": "LCMT1", "relation": "methylates", "target": "PPP2CA/L309", "type": "PTM_enzyme-substrate"},
{"source": "SENP2", "relation": "deconjugates", "target": "MAPT/SUMO", "type": "PTM_enzyme-substrate"},
{"source": "PPP2CA", "relation": "dephosphorylates", "target": "MAPT/pT181,pS396", "type": "PTM_enzyme-substrate"},
{"source": "CHIP", "relation": "ubiquitinates", "target": "MAPT", "type": "PTM_E3_ligase"},
{"source": "NRF2", "relation": "transcriptionally activates", "target": "GLO1", "type": "transcriptional_regulation"},
{"source": "HDAC6", "relation": "deacetylates", "target": "MAPT", "type": "PTM_enzyme-substrate"},
{"source": "p62/SQSTM1", "relation": "recognizes", "target": "MAPT/SUMOylated", "type": "autophagy_receptor"},
{"source": "proteasome", "relation": "degrades", "target": "MAPT/ubiquitinated", "type": "protein_degradation"},
{"source": "autophagy-lysosome", "relation": "degrades", "target": "MAP