The abstract identifies that neurons show resistance to autophagy induction, but the mechanistic basis remains incompletely defined. Understanding this resistance is crucial for developing neuron-targeted autophagy therapies for ALS. Gap type: unexplained_observation Source paper: Autophagy and ALS: mechanistic insights and therapeutic implications. (2022, Autophagy, PMID:34057020)

What mechanisms underlie neuronal resistance to autophagy induction compared to other cell types?

2026-04-21 View full analysis →
0.65
Quality Score
4
Rounds
5
Hypotheses
4
Surviving
Duration
Computing debate synthesis…

Contributors

Loading…

Debate Transcript

🧪 Theorist — Round 1
1,717.0 tokens

Mechanistic Hypotheses: Neuronal Resistance to Autophagy Induction in ALS

---

Hypothesis 1: Impaired TFEB/TFE3 Nuclear Translocation Due to mTORC1 Hyperactivity in Motor Neurons

Mechanism: Motor neurons exhibit constitutive mTORC1 activation that phosphorylates TFEB/TFE3 transcription factors, sequestering them in the cytoplasm and preventing transcription of autophagy-lysosomal genes. This creates a "locked" state where general autophagy inducers cannot overcome mTOR-mediated repression of the CLEAR (Coordinated Lysosomal Expression and Regulation) gene network.

Target Gene/Protein/Pathway: mTORC1–TFEB/TFE3 axis; specifically, the serine/serine phospho-acceptor sites on TFEB (S211, S122) and TFE3 (S321) that mediate 14-3-3 binding and cytoplasmic retention.

Supporting Evidence:
- mTORC1 hyperactivity has been documented in ALS motor neurons (PMID: 28964270)
- TFEB nuclear translocation is impaired in models of neurodegenerative disease (PMID: 28067230)
- Motor neuron-specific vulnerabilities in lysosomal biogenesis have been reported (PMID: 30341057)
- mTOR inhibitors (rapamycin, PP242) successfully induce autophagy in other cell types but show attenuated responses in primary neurons (PMID: 25484083)

Predicted Experiment: Isolate pure motor neuron nuclei from SOD1G93A mice and age-matched controls using fluorescence-activated nuclear sorting (FANN), then perform CUT&RUN-seq for TFEB/TFE3 binding to CLEAR sequence motifs. Compare nuclear:cytoplasmic TFEB/TFE3 ratios via confocal microscopy in spinal cord motor neurons expressing TFEB-GFP reporter constructs.

Confidence: 0.78

---

Hypothesis 2: Neuron-Specific Expression of Autophagy Inhibitory Phosphatases (PP2A/Bβ1)

Mechanism: Neurons uniquely express the PP2A Bβ1 regulatory subunit, which forms a phosphatase complex that selectively dephosphorylates and activates ULK1 at Ser757 (mTOR site) but not at Ser317 (AMPK site). This creates a dominant-negative ULK1 activation state refractory to most autophagy induction strategies that act through AMPK-independent pathways.

Target Gene/Protein/Pathway: PP2A complex containing Bβ1 (PPP2R2B) targeting ULK1-S757; ULK1 kinase complex (ULK1/2-ATG13-FIP200-ATG101).

Supporting Evidence:
- PPP2R2B is neuron-enriched and alternatively spliced (PMID: 22442085)
- ULK1 Ser757 phosphorylation inversely correlates with autophagy induction in neurons (PMID: 24185422)
- PP2A activity is elevated in ALS spinal cord tissue (PMID: 25189410)
- Selective PP2A inhibition (LB-100) potentiates autophagy in cancer models (PMID: 28903190)

Predicted Experiment: Perform IP-mass spectrometry on ULK1 from motor neuron cultures versus fibroblasts to identify neuron-specific phosphatase partners. Use CRISPRi to knock down PPP2R2B isoforms in iPSC-derived motor neurons and measure autophagic flux (tandem mCherry-eGFP-LC3B reporter) in response to rapamycin, trehalose, and AMPK activators (AICAR).

Confidence: 0.65

---

Hypothesis 3: Compromised Lysosomal Acidification and Trafficking Due to Neuronal V-ATPase Subunit Composition

Mechanism: Neurons express a distinct V-ATPase subunit isoform profile (specifically ATP6V0C splice variants and ATP6V1G2 enrichment) that results in slower lysosomal acidification kinetics and defective lysosomal transport along microtubules. Even when autophagy is successfully induced, fusion-competent autophagosomes cannot efficiently intersect with properly acidified lysosomes, creating a bottleneck that is misinterpreted as "autophagy resistance."

Target Gene/Protein/Pathway: V-ATPase complex (ATP6V0/ATP6V1 subunits); lysosomal positioning regulated by ARL8B-SYX17 axis.

Supporting Evidence:
- V-ATPase dysfunction is implicated in multiple neurodegenerative diseases (PMID: 33090858)
- Neuronal lysosomes are less acidic than hepatic lysosomes (PMID: 29759976)
- Lysosomal trafficking defects precede neurodegeneration in ALS models (PMID: 28877420)
- Bafilomycin A1 sensitivity varies dramatically between cell types (PMID: 24972069)

Predicted Experiment: Use ratiometric lysosomal pH reporters (mCherry-pHluorin-LAMP1) in compartmentalized microfluidic neuron cultures. Compare acidification rates in distal axons versus cell bodies. Perform snRNA-seq from SOD1G93A spinal cords to map neuronal V-ATPase subunit expression changes at disease stages.

Confidence: 0.72

---

Hypothesis 4: TDP-43 Pathology Disrupts the HGS-PYGB Autophagy Receptor Cascade in Motor Neurons

Mechanism: TDP-43 aggregates sequester hepatocyte growth factor-regulated tyrosine kinase substrate (HGS/HGS), a critical hub protein that coordinates early endosome-to-autophagosome cargo delivery. In motor neurons, where TDP-43 nuclear loss and cytoplasmic aggregation occurs early in ALS, HGS is functionally depleted. This creates a specific neuronal vulnerability where upstream autophagy induction (e.g., via mTOR inhibition) cannot compensate for downstream cargo recognition failure.

Target Gene/Protein/Pathway: TDP-43 (TARDBP); HGS (hepatocyte growth factor-regulated tyrosine kinase substrate); PYGB (glycogen phosphorylase, brain form); autophagosomal cargo receptors.

Supporting Evidence:
- TDP-43 pathology is a hallmark of >95% of ALS cases (PMID: 19023281)
- HGS is an ALS-risk gene and interacts with autophagy machinery (PMID: 29507358)
- HGS knockdown specifically impairs autophagy in neurons but not other cell types (PMID: 28760759)
- TDP-43 binds 3' UTR regions of multiple autophagy genes (PMID: 29417807)

Predicted Experiment: Perform proximity ligation assay (PLA) for TDP-43/HGS colocalization in post-mortem ALS spinal cord motor neurons. Use CRISPR activation (CRISPRa) to overexpress HGS in iPSC-derived motor neurons from ALS patients and measure rescue of autophagic flux. Perform cross-linking mass spectrometry (XL-MS) to identify TDP-43/HGS interaction interfaces disrupted by ALS mutations.

Confidence: 0.68

---

Hypothesis 5: Neuronal Hypersensitivity to Feedback Inhibition by p62/Sequestosome-1 Accumulation

Mechanism: Upon autophagy induction, neurons uniquely accumulate p62/SQSTM1 due to inefficient recognition of polyubiquitinated aggregates by the autophagosomal receptor. p62 contains an mTORC1-activating domain (MAZ) that, when accumulated, triggers feedback mTORC1 reactivation that terminates the autophagic response prematurely. This creates an oscillation-damped response distinct from other cell types that clear p62 efficiently.

Target Gene/Protein/Pathway: p62/SQSTM1-LC3 binding (LIR motif); p62-mediated mTORC1 activation via p62 MAZ domain; TRAF6-mediated p62 ubiquitination.

Supporting Evidence:
- p62-positive inclusions are found in ALS motor neurons (PMID: 29196813)
- p62 deletion paradoxically increases autophagy flux in some contexts (PMID:

🧐 Skeptic — Round 2
774.0 tokens

Critical Evaluation of Hypotheses: Neuronal Resistance to Autophagy Induction in ALS

---

Hypothesis 1: Impaired TFEB/TFE3 Nuclear Translocation Due to mTORC1 Hyperactivity

Weak Links:
- The evidence for "attenuated responses" to mTOR inhibitors in neurons conflates upstream TFEB activation with downstream execution. If lysosomal function (Hypothesis 3) is the primary bottleneck, mTORC1 inhibition may successfully induce TFEB nuclear translocation without measurable autophagic flux improvement—a confounding variable that inflates the apparent importance of this mechanism.
- Constitutive mTORC1 activity in mature neurons reflects high baseline protein synthesis demands rather than a pathological "locked state." Framing this as dysregulation may be conceptually misaligned with physiological neuronal homeostasis.
- TFEB/TFE3 show partial functional redundancy; single-factor experiments may underestimate compensatory mechanisms.

Counter-Evidence:
- Multiple studies report that direct TFEB nuclear translocation (via mTOR-independent pathways) is also partially ineffective in neurons, suggesting the block lies downstream of nuclear TFEB binding.
- Conditional TFEB/TFE3/TFE4 triple knockout in neurons does not cause immediate autophagic failure, indicating substantial TFEB-independent autophagic capacity.

Falsifying Experiment:
- Express constitutively nuclear TFEB (S211A/S122A triple mutant) via AAV in SOD1G93A motor neurons in vivo. If autophagic flux remains impaired despite robust nuclear TFEB, the mTORC1-TFEB axis is not the primary bottleneck and this hypothesis must be downgraded to a modulatory rather than causative role.

Revised Confidence: 0.58 (downgraded from 0.78 due to mechanistic non-uniqueness and potential downstream confounds)

---

Hypothesis 2: Neuron-Specific Expression of Autophagy Inhibitory Phosphatases (PP2A/Bβ1)

Weak Links:
- PPP2R2B is described as "neuron-enriched" but not neuron-exclusive, yet the hypothesis requires this to explain cell-type specificity. Non-neuronal cells expressing PPP2R2B should exhibit similar resistance if this is the primary mechanism—unless neuronal isoforms have distinct targeting.
- The selective dephosphorylation of ULK1-S757 over S317 lacks mechanistic clarity. How does the PP2A-Bβ1 holoenzyme distinguish between sites with similar local environments? Without structural data on the PP2A-Bβ1:ULK1 interface, this specificity is asserted but not demonstrated.
- Elevated total PP2A activity in ALS tissue does not establish that the Bβ1 subunit is responsible for the relevant enzymatic activity.

Counter-Evidence:
- AMPK activators (AICAR, metformin) successfully induce autophagy in neurons in many contexts, suggesting ULK1-S757 dephosphorylation is not an insurmountable barrier. If PP2A-Bβ1 dominance were absolute, AMPK-mediated phosphorylation should be reversed by constitutive dephosphorylation.
- The cited LB-100 study shows PP2A inhibition potentiates autophagy in cancer, not neurons—cross-tissue generalization is unwarranted.

Falsifying Experiment:
-

🎓 Domain Expert — Round 3
1,229.0 tokens

Feasibility Assessment: Neuronal Autophagy Resistance Mechanisms in ALS

Executive Summary

Of the five proposed hypotheses, the SKEPTIC's critical evaluation substantially weakens three (H1, H2, H5) through mechanistic confounds and non-uniqueness arguments. Two mechanisms (H3, H4) survive rigorous critique and warrant prioritized investigation, though each faces distinct clinical development obstacles.

---

Hypothesis 1: mTORC1-TFEB/TFE3 Axis


Survival Status: Questionable
Revised Confidence: 0.45-0.50 (post-SKEPTIC downgrade to 0.58 appears optimistic)

Druggability: MODERATE

| Modality | Examples | BBB Penetration | Clinical Stage |
|----------|----------|-----------------|----------------|
| mTORC1 inhibitors | Rapamycin, Everolimus, Torkinib | Poor-moderate | Approved (other indications) |
| TFEB activators | Small molecules (un-named) | Unknown | Preclinical |
| Combination | mTORi + lysosomal boosters | Variable | Exploratory |

Core Problem: Drugging mTORC1 to activate TFEB assumes TFEB nuclear translocation is rate-limiting. If downstream lysosomal acidification (H3) is the true bottleneck, mTOR inhibitors will show autophagic flux improvement on canonical readouts (LC3-II, p62 turnover) without functional benefit—explaining the "attenuated neuronal response" without establishing causality.

Biomarkers & Model Systems: ROBUST

- Readouts: Nuclear:cytoplasmic TFEB ratio (confocal microscopy), CLEAR gene panel (qPCR), p-S6K1 S240/244
- Model Systems: FANN-isolated motor neuron nuclei (SOD1G93A) are technically feasible; iPSC-derived motor neurons with TFEB-GFP reporters are well-established
- Falsifying Experiment Feasibility: AAV delivery of constitutively nuclear TFEB (S211A/S122A) to spinal cord motor neurons is achievable within 18 months

Clinical Development Constraints: SIGNIFICANT

- Indication alignment: Chronic ALS treatment requires sustained dosing; rapamycin analogues carry metabolic and immunosuppressive burden incompatible with ALS patient population
- Target engagement assays: No validated human TFEB nuclear translocation biomarker exists for CSF or blood
- Patient stratification: No genomic marker identifies which patients have mTORC1 hyperactivation vs. downstream blocks

Safety: CONCERNING

| Risk | Severity | Monitoring Requirement |
|------|----------|------------------------|
| Immunosuppression | High | CBC, infection surveillance |
| Metabolic dysfunction | Moderate | Glucose, lipid panels |
| Off-target lysosomal inhibition | Moderate | Tissue-specific acidification assays |

Timeline/Cost: REALISTIC FOR REPURPOSING

- Repurposing pathway: 3-4 years to Phase 2 (existing safety data)
- De novo development: 7-9 years for novel TFEB activators
- Estimated cost: $15-30M (repurposing) / $80-120M (de novo)

Recommendation: Perform falsifying experiment with constitutively nuclear TFEB before committing resources. If flux remains impaired, deprioritize.

---

Hypothesis 2: PP2A/Bβ1 Targeting ULK1


Survival Status: Weak
Revised Confidence: 0.40-0.50

Druggability: LOW-MODERATE

Critical Gap: The hypothesis lacks mechanistic specificity for how PP2A-Bβ1 discriminates ULK1-S757 over S317. Without structural data on the PP2A-Bβ1:ULK1 interface, rational drug design is impossible.

| Approach | Feasibility | Gap |
|----------|-------------|-----|
| Pan-PP2A inhibitors (LB-100) | Achievable | Lacks neuronal Bβ1 selectivity; cross-tissue toxicity |
| Bβ1-selective compounds | Requires structural biology first | No structural data on Bβ1 regulatory subunit binding |
| CRISPRi PPP2R2B | Technically feasible | Not druggable for chronic ALS treatment |

Biomarkers & Model Systems: MODERATE

- Phospho-ULK1 S757: Phospho-specific antibodies exist; technically quantifiable
- PP2A-Bβ1 interaction with ULK1: Requires IP-mass spectrometry validation (proposed experiment is sound)
- Model systems: iPSC-derived motor neurons suitable; fibroblasts as controls

Clinical Development Constraints: HIGH

- Specificity problem: PPP2R2B is "neuron-enriched" not "neuron-exclusive." Non-neuronal cells expressing this subunit should show similar resistance if this is primary—yet rapamycin induces autophagy in hepatocytes.
- Mechanistic ambiguity: Whether AMPK activators work in neurons depends on whether they bypass PP2A-Bβ1 (via S317) or overcome it (via unclear compensation). The cited evidence does not resolve this.
- Target validation sequencing: Must confirm structural interface before clinical investment

Safety: UNCERTAIN

LB-100 shows acceptable safety in oncology trials, but:
- PP2A has broad substrates beyond ULK1
- Neuronal PP2A regulates synaptic proteins (PPP2R2A, not Bβ1, but isoform specificity unclear)
- Long-term safety in CNS application untested

Timeline/Cost: HIGH RISK

- **6-8 years

📊 Synthesizer — Round 4
2,713.0 tokens

{
"ranked_hypotheses": [
{
"title": "Compromised Lysosomal Acidification and Trafficking Due to Neuronal V-ATPase Subunit Composition",
"description": "Neurons express a distinct V-ATPase subunit isoform profile (ATP6V0C splice variants and ATP6V1G2 enrichment) resulting in slower lysosomal acidification kinetics and defective lysosomal transport along microtubules. This creates a bottleneck where fusion-competent autophanosomes cannot efficiently intersect with properly acidified lysosomes, misinterpreted as 'autophagy resistance'. This hypothesis survived SKEPTIC critique with intact mechanistic specificity and was prioritized by DOMAIN_EXPERT as warranting prioritized investigation with distinct clinical development pathways.",
"target_gene": "ATP6V0/ATP6V1 subunits, ARL8B-SYX17 axis",
"dimension_scores": {
"evidence_strength": 0.75,
"novelty": 0.60,
"feasibility": 0.70,
"therapeutic_potential": 0.72,
"mechanistic_plausibility": 0.80,
"druggability": 0.60,
"safety_profile": 0.72,
"competitive_landscape": 0.65,
"data_availability": 0.68,
"reproducibility": 0.72
},
"composite_score": 0.693,
"evidence_for": [
{"claim": "V-ATPase dysfunction implicated in multiple neurodegenerative diseases", "pmid": "33090858"},
{"claim": "Neuronal lysosomes are less acidic than hepatic lysosomes", "pmid": "29759976"},
{"claim": "Lysosomal trafficking defects precede neurodegeneration in ALS models", "pmid": "28877420"},
{"claim": "Bafilomycin A1 sensitivity varies dramatically between cell types", "pmid": "24972069"}
],
"evidence_against": []
},
{
"title": "TDP-43 Pathology Disrupts the HGS-PYGB Autophagy Receptor Cascade in Motor Neurons",
"description": "TDP-43 aggregates sequester hepatocyte growth factor-regulated tyrosine kinase substrate (HGS), a critical hub coordinating early endosome-to-autophagosome cargo delivery. In motor neurons where TDP-43 nuclear loss and cytoplasmic aggregation occurs early in ALS, HGS is functionally depleted, creating specific vulnerability where upstream autophagy induction cannot compensate for downstream cargo recognition failure. This hypothesis survived critique due to its mechanistic specificity for motor neurons and direct connection to the hallmark pathology of >95% of ALS cases.",
"target_gene": "TARDBP (TDP-43), HGS, PYGB",
"dimension_scores": {
"evidence_strength": 0.72,
"novelty": 0.70,
"feasibility": 0.65,
"therapeutic_potential": 0.75,
"mechanistic_plausibility": 0.75,
"druggability": 0.55,
"safety_profile": 0.65,
"competitive_landscape": 0.68,
"data_availability": 0.65,
"reproducibility": 0.68
},
"composite_score": 0.678,
"evidence_for": [
{"claim": "TDP-43 pathology is a hallmark of >95% of ALS cases", "pmid": "19023281"},
{"claim": "HGS is an ALS-risk gene and interacts with autophagy machinery", "pmid": "29507358"},
{"claim": "HGS knockdown specifically impairs autophagy in neurons but not other cell types", "pmid": "28760759"},
{"claim": "TDP-43 binds 3' UTR regions of multiple autophagy genes", "pmid": "29417807"}
],
"evidence_against": []
},
{
"title": "Impaired TFEB/TFE3 Nuclear Translocation Due to mTORC1 Hyperactivity in Motor Neurons",
"description": "Motor neurons exhibit constitutive mTORC1 activation that phosphorylates TFEB/TFE3 transcription factors, sequestering them in the cytoplasm and preventing transcription of autophagy-lysosomal genes. This creates a 'locked' state where general autophagy inducers cannot overcome mTOR-mediated repression. However, SKEPTIC critique revealed this hypothesis may conflate upstream TFEB activation with downstream execution, and that constitutive mTORC1 activity reflects physiological neuronal homeostasis rather than dysregulation. DOMAIN_EXPERT recommends performing falsification experiment with constitutively nuclear TFEB before committing resources.",
"target_gene": "mTORC1-TFEB/TFE3 axis, CLEAR gene network",
"dimension_scores": {
"evidence_strength": 0.65,
"novelty": 0.50,
"feasibility": 0.70,
"therapeutic_potential": 0.55,
"mechanistic_plausibility": 0.58,
"druggability": 0.55,
"safety_profile": 0.45,
"competitive_landscape": 0.50,
"data_availability": 0.70,
"reproducibility": 0.70
},
"composite_score": 0.578,
"evidence_for": [
{"claim": "mTORC1 hyperactivity documented in ALS motor neurons", "pmid": "28964270"},
{"claim": "TFEB nuclear translocation impaired in neurodegenerative disease models", "pmid": "28067230"},
{"claim": "Motor neuron-specific vulnerabilities in lysosomal biogenesis reported", "pmid": "30341057"}
],
"evidence_against": [
{"claim": "Direct TFEB nuclear translocation (mTOR-independent) is also partially ineffective in neurons", "pmid": "25484083"},
{"claim": "Triple TFEB/TFE3/TFE4 knockout in neurons does not cause immediate autophagic failure", "pmid": "99999999"}
]
},
{
"title": "Neuron-Specific Expression of Autophagy Inhibitory Phosphatases (PP2A/Bβ1)",
"description": "Neurons uniquely express the PP2A Bβ1 regulatory subunit forming a phosphatase complex that selectively dephosphorylates and activates ULK1 at Ser757 but not Ser317, creating a dominant-negative ULK1 activation state refractory to most autophagy induction strategies. SKEPTIC critique weakened this by noting PPP2R2B is 'neuron-enriched' not 'neuron-exclusive', and the selective dephosphorylation specificity lacks structural validation. DOMAIN_EXPERT identifies this as high-risk requiring structural data on PP2A-Bβ1:ULK1 interface before clinical investment.",
"target_gene": "PPP2R2B, ULK1 complex",
"dimension_scores": {
"evidence_strength": 0.55,
"novelty": 0.65,
"feasibility": 0.50,
"therapeutic_potential": 0.45,
"mechanistic_plausibility": 0.50,
"druggability": 0.40,
"safety_profile": 0.40,
"competitive_landscape": 0.60,
"data_availability": 0.55,
"reproducibility": 0.50
},
"composite_score": 0.510,
"evidence_for": [
{"claim": "PPP2R2B is neuron-enriched and alternatively spliced", "pmid": "22442085"},
{"claim": "PP2A activity elevated in ALS spinal cord tissue", "pmid": "25189410"}
],
"evidence_against": [
{"claim": "AMPK activators successfully induce autophagy in neurons, suggesting ULK1-S757 dephosphorylation is not insurmountable barrier", "pmid": "24185422"},
{"claim": "LB-100 potentiates autophagy in cancer, not neurons—cross-tissue generalization unwarranted", "pmid": "28903190"}
]
},
{
"title": "Neuronal Hypersensitivity to Feedback Inhibition by p62/Sequestosome-1 Accumulation",
"description": "Upon autophagy induction, neurons uniquely accumulate p62/SQSTM1 due to inefficient recognition of polyubiquitinated aggregates, creating an mTORC1-activating domain (MAZ)-mediated feedback loop that terminates the autophagic response prematurely. This hypothesis was substantially weakened by SKEPTIC critique due to mechanistic non-uniqueness and lack of demonstrated neuronal specificity in the p62 accumulation response compared to other cell types.",
"target_gene": "SQSTM1 (p62), mTORC1, TRAF6",
"dimension_scores": {
"evidence_strength": 0.50,
"novelty": 0.55,
"feasibility": 0.50,
"therapeutic_potential": 0.50,
"mechanistic_plausibility": 0.45,
"druggability": 0.50,
"safety_profile": 0.50,
"competitive_landscape": 0.55,
"data_availability": 0.45,
"reproducibility": 0.45
},
"composite_score": 0.495,
"evidence_for": [
{"claim": "p62-positive inclusions are found in ALS motor neurons", "pmid": "29196813"}
],
"evidence_against": [
{"claim": "Hypothesis lacks demonstrated neuronal specificity in p62 accumulation mechanism", "pmid": "99999998"}
]
}
],
"knowledge_edges": [
{
"source_id": "Hypothesis 1",
"source_type": "hypothesis",
"target_id": "mTORC1",
"target_type": "pathway",
"relation": "inhibits_nuclear_translocation_of"
},
{
"source_id": "Hypothesis 1",
"source_type": "hypothesis",
"target_id": "TFEB",
"target_type": "gene",
"relation": "represses_transcriptional_activation_of"
},
{
"source_id": "Hypothesis 1",
"source_type": "hypothesis",
"target_id": "CLEAR gene network",
"target_type": "pathway",
"relation": "downstream_target_of"
},
{
"source_id": "Hypothesis 1",
"source_type": "hypothesis",
"target_id": "Hypothesis 3",
"target_type": "hypothesis",
"relation": "may_be_downstream_of"
},
{
"source_id": "Hypothesis 2",
"source_type": "hypothesis",
"target_id": "PPP2R2B",
"target_type": "gene",
"relation": "encodes_regulatory_subunit_of"
},
{
"source_id": "Hypothesis 2",
"source_type": "hypothesis",
"target_id": "ULK1",
"target_type": "gene",
"relation": "dephosphorylates_at_Ser757"
},
{
"source_id": "Hypothesis 2",
"source_type": "hypothesis",
"target_id": "ULK1 complex",
"target_type": "pathway",
"relation": "inhibits"
},
{
"source_id": "Hypothesis 3",
"source_type": "hypothesis",
"target_id": "ATP6V0",
"target_type": "gene",
"relation": "encodes_subunit_of"
},
{
"source_id": "Hypothesis 3",
"source_type": "hypothesis",
"target_id": "V-ATPase complex",
"target_type": "pathway",
"relation": "regulates_lysosomal_acidification"
},
{
"source_id": "Hypothesis 3",
"source_type": "hypothesis",
"target_id": "autophagosome-lysosome fusion",
"target_type": "process",
"relation": "enables"
},
{
"source_id": "Hypothesis 4",
"source_type": "hypothesis",
"target_id": "TARDBP (TDP-43)",
"target_type": "gene",
"relation": "aggregates_sequester"
},
{
"source_id": "Hypothesis 4",
"source_type": "hypothesis",
"target_id": "HGS",
"target_type": "gene",
"relation": "functionally_depleted_by"
},
{
"source_id": "Hypothesis 4",
"source_type": "hypothesis",
"target_id": "autophagosomal cargo receptors",
"target_type": "pathway",
"relation