# Compound Dive: Approved and Late-Stage Clinical Agents Targeting TDP-43 Proteostasis, Autophagy, and Stress Granule Dynamics in ALS

**Date:** May 4, 2026  
**Focus:** Compounds with an established clinical/regulatory history or late-stage ALS trial data that modulate TDP-43 proteostasis, macroautophagy/autophagy-lysosome pathways, or stress granule (SG) condensation dynamics.

---

## 1. Executive Summary

Cytoplasmic mislocalization and aggregation of TDP-43 is the defining neuropathology in >95% of amyotrophic lateral sclerosis (ALS) cases. TDP-43 dysfunction converges on three druggable axes: **(i)** proteostasis (protein folding, chaperoning, ER stress, and clearance), **(ii)** autophagy (macroautophagy and selective autophagy), and **(iii)** stress granule dynamics (liquid-liquid phase separation and persistent condensate formation). Several approved or late-stage clinical compounds have been positioned against these axes. The late-stage ALS clinical landscape has been sobering: large Phase 2/3 trials of trehalose, TUDCA/AMX0035, colchicine, and arimoclomol have all returned negative primary outcomes in broad ALS populations. However, three agents retain the strongest translational rationale for repurposing or precision-development: **rapamycin**, because it has ALS Phase 2 safety data and recent, robust human preclinical evidence of TDP-43 rescue; **edaravone**, because it is already FDA-approved for ALS and newly linked to TDP-43 proteostasis via the SIRT1–XBP1 axis; and **α-lipoic acid / lipoamide**, because of a striking 2025 preclinical package showing direct dissolution of stress granules and rescue of TDP-43 and FUS mutant phenotypes across species, combined with a long-standing supplement/drug safety record.

---

## 2. Background: Why These Pathways Matter

### TDP-43 Proteostasis
TDP-43 mislocalization from the nucleus to the cytoplasm leads to both **loss of nuclear function** (cryptic splicing, RNA-processing defects) and **gain of toxic function** (cytosolic aggregation). Protein quality control pathways—the ubiquitin-proteasome system, cytosolic chaperones (e.g., HSP70, HSPB8, BAG3), and ER stress responses—are critical for preventing the accumulation of misfolded TDP-43.

### Autophagy
Macroautophagy is the principal route for clearing large TDP-43 aggregates, oligomers, and damaged organelles. The autophagy-lysosome pathway (ALP) is often impaired in ALS motor neurons. mTOR inhibition (e.g., by rapamycin), chemical chaperones (e.g., TUDCA), and disaccharides (e.g., trehalose) have all been shown to upregulate autophagic flux in preclinical models.

### Stress Granule Dynamics
Under stress, cytoplasmic RNA-protein condensates called stress granules form transiently. ALS-linked mutations in TDP-43 and FUS promote aberrant, persistent SGs that mature into solid aggregates. Compounds that modulate SG dynamics—particularly by targeting redox-sensitive intrinsically disordered regions (IDRs) in SG proteins such as SFPQ—can prevent downstream TDP-43 aggregation.

---

## 3. Compound Profiles

### 3.1 Rapamycin (Sirolimus)
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Immunosuppression (organ transplant), lymphangioleiomyomatosis |
| **Regulatory status** | Approved (FDA, EMA) |
| **ALS clinical stage** | Phase 2 completed (RAP-ALS; NCT03359538) |
| **Target pathway** | Autophagy (mTOR inhibition) |
| **Mechanism** | Rapamycin inhibits the mechanistic target of rapamycin complex 1 (mTORC1), relieving the brake on macroautophagy and promoting clearance of protein aggregates and damaged organelles. In ALS models, mTOR inhibition also dampens neuroinflammation and expands regulatory T-cell populations. |
| **Published ALS evidence** | Mandrioli et al. (2023, *Nature Communications*) reported a multicenter, randomized, double-blind Phase 2 trial (n=63) in ALS. The primary endpoint (≥30% increase in Tregs from baseline) was not met. Rapamycin was **safe and well tolerated**. Secondary analyses showed decreased IL-18 mRNA and protein, increased classical monocytes, and modulation of S6-ribosomal protein phosphorylation. Separately, Casiraghi et al. (2025, *Experimental Neurology*) demonstrated that rapamycin **prevented TDP-43 loss of splicing activity and cytoplasmic aggregation** in chronic oxidative stress models using human neuroblastoma cells, ALS patient fibroblasts, iPSC-derived motor neurons, and 3D brain organoids. |
| **Assessment** | Strong mechanistic and safety profile; lacks a definitive clinical efficacy signal in a broad ALS population. The preclinical TDP-43 rescue data are compelling and argue for a biomarker-enriched, precision trial rather than a broad-population study. |

### 3.2 Edaravone (Radicava / Radicava ORS)
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | ALS (FDA approved 2017) |
| **Regulatory status** | Approved for ALS |
| **ALS clinical stage** | Approved; post-hoc analyses and new preclinical mechanistic studies ongoing |
| **Target pathway** | TDP-43 proteostasis / ER stress |
| **Mechanism** | Originally developed as a free-radical scavenger. Recent transcriptomic and mechanistic studies reveal that edaravone also modulates protein quality control pathways. It activates the SIRT1–XBP1 axis, a key regulator of the ER-stress response, and thereby mitigates TDP-43 mislocalization in motor neurons. |
| **Published ALS evidence** | A 2025 preclinical study (published in *Neuropharmacology* / *FASEB Journal*; PMID 40010009) using iPSC-derived motor neurons showed that edaravone treatment **concentration-dependently rescued TDP-43 mislocalization induced by oxidative stress** and normalized the proteostasis network, protecting against neurotoxicity. RNA-seq identified over 1,000 differentially expressed genes and implicated XBP1 as a critical mediator. |
| **Assessment** | Already approved for ALS, so safety and regulatory pathways are solved. The translational opportunity lies in **precision deployment**: using emerging TDP-43 biomarkers (e.g., CSF pTDP-43, UNC13A cryptic exons) to test whether edaravone preferentially benefits TDP-43 proteinopathy patients. |

### 3.3 Trehalose
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Food-grade stabilizer / disaccharide; orphan drug applications in neurodegeneration |
| **Regulatory status** | Generally Recognized As Safe (GRAS) |
| **ALS clinical stage** | Phase 2/3 completed (HEALEY ALS Platform Trial; NCT04297683, Regimen E) |
| **Target pathway** | Autophagy (mTOR-independent) |
| **Mechanism** | Trehalose is an autophagy activator that promotes clearance of misfolded and aggregated proteins via an mTOR-independent route. In animal models of ALS, it has prolonged survival and attenuated disease signs. |
| **Published ALS evidence** | The HEALEY ALS Platform Trial regimen for trehalose (dose 0.75 g/kg IV weekly) enrolled 161 participants. The published results (2025, *The Lancet Neurology*, PMID 40409314) showed a disease-rate ratio of **0.87 (95% credible interval 0.665–1.102)** for the composite of ALSFRS-R decline and survival over 24 weeks. There was **no statistically significant benefit** on primary or secondary endpoints, including biomarker measures. Serious adverse events were higher in the trehalose arm (16% vs. 7%). |
| **Assessment** | A definitive late-stage negative result. The dose, route (IV), and broad population design may have contributed to failure, but the trial provides little justification for ALS repurposing as currently formulated. |

### 3.4 Tauroursodeoxycholic Acid (TUDCA)
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Cholestatic liver diseases (UDCA is FDA-approved; TUDCA is a conjugate available as a supplement and investigational drug) |
| **Regulatory status** | Component of AMX0035 (Relyvrio), which was withdrawn from the market |
| **ALS clinical stage** | Phase 3 completed (TUDCA-ALS; NCT03800524) |
| **Target pathway** | Proteostasis / ER stress / autophagy (chemical chaperone) |
| **Mechanism** | TUDCA is a hydrophilic bile acid that acts as a chemical chaperone, reducing ER stress, stabilizing mitochondria, and modulating apoptosis. Combined with sodium phenylbutyrate (AMX0035), it was hypothesized to improve proteostasis and reduce neuronal death. |
| **Published ALS evidence** | A Phase 2b pilot study suggested TUDCA + riluzole slowed ALSFRS-R decline compared with riluzole alone (PMID 25664595). AMX0035 received accelerated FDA approval in September 2022 based on the CENTAUR Phase 2 trial. However, the confirmatory **PHOENIX Phase 3 trial failed**, and Amylyx voluntarily withdrew AMX0035 in **April 2024**. The standalone European Phase 3 TUDCA-ALS trial (n=337) also **failed to meet its primary endpoint** (PMID 38053196; results released ~March 2024). |
| **Assessment** | Late-stage clinical evidence is negative in both combination and monotherapy settings. Without compelling new mechanistic insights or biomarker-defined subgroups, repurposing trials are not currently justified. |

### 3.5 Colchicine
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Gout, familial Mediterranean fever, pericarditis |
| **Regulatory status** | Approved |
| **ALS clinical stage** | Phase 2 completed (Co-ALS; NCT03693781) |
| **Target pathway** | Proteostasis / autophagy (HSPB8–BAG3–HSP70 axis) |
| **Mechanism** | Colchicine upregulates the small heat-shock protein HSPB8 and its co-chaperone BAG3, driving selective autophagy of misfolded proteins. It attenuates stress granule-mediated TDP-43 aggregate formation and promotes cytosolic clearance of TDP-43. |
| **Published ALS evidence** | The Co-ALS Phase 2 trial was predicated on strong preclinical data in TDP-43 models. Results reported in 2023/2024 showed that **colchicine did not slow disease progression** in ALS. |
| **Assessment** | A clean Phase 2 failure. Despite elegant cell-biology rationale, efficacy was not observed in a broad ALS population. Repurposing is not supported unless a specific biomarker-defined responder subset can be retrospectively identified. |

### 3.6 Arimoclomol
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Niemann-Pick disease Type C (approved in EU) |
| **Regulatory status** | Approved (EU) |
| **ALS clinical stage** | Phase 3 completed (ORARIALS-01) |
| **Target pathway** | Proteostasis (heat shock response amplification) |
| **Mechanism** | Arimoclomol is a co-inducer of heat shock factor 1 (HSF1), amplifying the expression of chaperones (e.g., HSP70, HSP90). It stabilizes lysosomal membranes and facilitates autophagic degradation of aggregated proteins. It also regulates SQSTM1/p62 phosphorylation and Atg gene expression. |
| **Published ALS evidence** | Arimoclomol showed efficacy in SOD1(G93A) and TDP-43 mouse models. The ORARIALS-01 Phase 3 trial in early sporadic ALS, however, **failed to meet its primary endpoint** (reported ~May 2024). |
| **Assessment** | A plausible proteostasis mechanism was not translated into clinical benefit in a broad ALS cohort. Given the definitive Phase 3 result, standalone ALS repurposing is unjustified. |

### 3.7 α-Lipoic Acid / Lipoamide
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Diabetic neuropathy (Rx formulations); dietary supplement; antioxidant |
| **Regulatory status** | Approved for diabetic neuropathy (in some jurisdictions); GRAS as supplement |
| **ALS clinical stage** | No ALS-specific clinical trials published to date |
| **Target pathway** | Stress granule dynamics (redox modulation of SG proteins) |
| **Mechanism** | The dithiolane ring of lipoamide is redox-active and modulates the condensation of intrinsically disordered stress granule proteins, particularly SFPQ. By stabilizing the reduced/redox state of SFPQ, lipoamide **prevents aberrant cytoplasmic condensation** of SGs. This intercepts the pathological cascade where persistent SGs nucleate TDP-43 aggregation. |
| **Published ALS evidence** | Uechi et al. (2025, *Nature Chemical Biology*, PMID 40369342) identified lipoamide in a screen of 1,600 compounds and demonstrated that it: (i) specifically dissolves cytoplasmic SG condensates; (ii) ameliorates aging-associated SG protein aggregation in *C. elegans*; (iii) improves neuronal morphology; and (iv) **recovers motor defects in Drosophila and human iPSC motor-neuron models of TDP-43 and FUS mutants**. Thermal proteome profiling confirmed stabilization of IDR-containing proteins including SRSF1 and SFPQ. |
| **Assessment** | The most compelling **new mechanism** in this survey. It is not late-stage for ALS, but the compound class (α-lipoic acid is the reduced, bioavailable precursor) has extensive human safety data. The translational leap required is smaller than for a de novo chemical entity, but an ALS-specific proof-of-concept trial is essential. |

### 3.8 Enoxacin
| Attribute | Detail |
|-----------|--------|
| **Primary indication(s)** | Fluoroquinolone antibacterial |
| **Regulatory status** | Approved (antibiotic) |
| **ALS clinical stage** | Preclinical |
| **Target pathway** | DNA damage response / TDP-43 proteostasis (DICER/DROSHA axis) |
| **Mechanism** | In TDP-43 and mutant FUS models, cytoplasmic inclusions co-localizing with stress granules impair the DNA damage response (DDR). Enoxacin stimulates DICER enzymatic activity, which restores DDR signaling at DNA double-strand breaks and reduces DNA damage accumulation. |
| **Published ALS evidence** | Modafferi et al. (2025, *Cell Death and Differentiation*, PMID 40437235) showed that enoxacin **restores a proficient DDR and reduces DNA damage** in cell cultures with TDP-43/FUS inclusions and in a murine ALS model. Dicer-2 overexpression in *Drosophila* rescued TDP-43-mediated retinal degeneration. |
| **Assessment** | Fascinating mechanistic angle, but the ALS evidence is restricted to preclinical models. Repurposing an antibiotic for chronic ALS therapy carries significant safety and tolerability concerns (tendon, CNS side effects). Not prioritized over rapamycin or lipoic acid/lipoamide. |

---

## 4. Top Candidates: Most Promising Translational Profiles

The following two (with a third emerging candidate) are judged to have the best balance of mechanistic relevance to TDP-43 pathways, human safety/clinical data, and actionable next steps for repurposing.

### 4.1 Rapamycin (Sirolimus) — *Best balance of clinical maturity + autophagy/TDP-43 mechanism*
**Why it stands out:**  
Rapamycin is the only compound in this survey with **both** a completed ALS Phase 2 safety/tolerability dataset and strong, recent **human cellular** evidence of TDP-43 rescue. Its mechanism—mTORC1 inhibition driving macroautophagy—is among the best-validated routes for clearance of TDP-43 aggregates. Unlike trehalose or TUDCA, rapamycin has a well-established pharmacology, including CNS penetration and biomarkers of target engagement (S6 phosphorylation, autophagy markers).

**What would need to be true for a repurposing trial to be justified:**
1. **Biomarker enrichment:** A trial must stratify or enrich for patients with confirmed TDP-43 proteinopathy (e.g., using CSF pTDP-43, UNC13A cryptic exon detection, or validated TDP-43 PET/blood assays). Broad, all-comers ALS trials have repeatedly failed for pathway-targeted agents.
2. **Optimized dosing:** The RAP-ALS trial used a standard immunosuppressive regimen. An intermittent or pulsed low-dose schedule may be required to maximize autophagic flux while minimizing immunosuppression and metabolic side effects. Pharmacodynamic confirmation in CSF (e.g., decreased p-mTOR/p-S6, increased LC3-II or p62 dynamics) is essential.
3. **TDP-43 target engagement in humans:** Biomarker studies must demonstrate that rapamycin at the chosen dose reduces TDP-43 mislocalization or aggregation markers in patient biofluids or iPSC-derived neurons.
4. **Regulatory pathway clarity:** Because rapamycin is approved for other indications but showed no efficacy signal in the prior ALS Phase 2, a new trial would require a compelling biomarker-linked hypothesis and likely a Phase 1b/2a target-engagement study design rather than a traditional efficacy trial.

### 4.2 Edaravone (Radicava) — *Approved for ALS with a newly validated TDP-43 proteostasis mechanism*
**Why it stands out:**  
Edaravone is already FDA-approved for ALS and has a favorable intravenous and oral formulation profile. The conventional view is that it is a nonspecific antioxidant. The 2025 preclinical discovery that it rescues TDP-43 mislocalization via the SIRT1–XBP1 pathway reframes it as a **proteostasis-modifying therapy**. If this mechanism is operative in humans, it could explain variability in edaravone response and suggests that TDP-43 proteinopathy patients—rather than all ALS patients—may benefit most.

**What would need to be true for a repurposing/precision trial to be justified:**
1. **Human target confirmation:** Data must show that edaravone treatment in ALS patients (or patient-derived cells at clinical concentrations) modulates XBP1 splicing, SIRT1 activity, and/or nuclear TDP-43 retention.
2. **Biomarker stratification:** A prospective or retrospective analysis must link TDP-43 pathology status (e.g., CSF TDP-43 levels, cryptic exon burden) to clinical response. If edaravone only benefits the TDP-43-positive subgroup, a precision-enrichment trial would be the logical next step.
3. **Exposure adequacy:** The standard edaravone dosing regimen must achieve CNS levels sufficient to engage the proteostasis pathway, as suggested by iPSC-MN data.
4. **Outcome validation:** A registrational trial would need to show a clinically meaningful separation in a TDP-43–enriched cohort on a validated endpoint (e.g., ALSFRS-R slope, survival, or neurofilament light chain dynamics).

### 4.3 α-Lipoic Acid (precursor to lipoamide) — *Most innovative stress-granule mechanism, but preclinical-stage for ALS*
**Why it is notable:**  
The 2025 *Nature Chemical Biology* report on lipoamide represents a conceptual advance: direct **dissolution of persistent stress granules** via redox modulation of SFPQ. Because persistent SGs are an upstream nucleation site for TDP-43 aggregation, an SG-targeting agent is mechanistically distinct from autophagy inducers or antioxidants. α-Lipoic acid is the clinically available, well-tolerated precursor metabolite. It has Phase 2 data in multiple sclerosis and geographic atrophy (safety), though no published ALS trial.

**What would need to be true for a repurposing trial to be justified:**
1. **ALS-specific Phase 1b/2 trial:** A dedicated clinical study in ALS is non-negotiable. The existing preclinical data are strong but entirely preclinical.
2. **Pharmacokinetic bridging:** Human dosing must achieve CNS levels comparable to the concentrations used in the iPSC motor neuron and fly rescue experiments. Brain/plasma partitioning data in humans are needed.
3. **Target engagement biomarker:** A measurable biomarker of SG dissolution (e.g., CSF or blood levels of SG-resident proteins, or imaging of condensate dynamics in accessible cells) must be validated and shown to respond to treatment.
4. **Safety at neuroprotective doses:** α-Lipoic acid is generally safe, but the doses required for redox modulation of SG proteins may exceed typical supplement ranges. A formal dose-escalation study in ALS patients must confirm tolerability.

---

## 5. What Would Need to Be True for Repurposing Trials in This Space

Across all candidates, several cross-cutting prerequisites have emerged from the recent string of late-stage ALS failures:

| Prerequisite | Rationale |
|--------------|-----------|
| **TDP-43 biomarker enrichment** | Broad ALS populations are pathologically heterogeneous (TDP-43, SOD1, FUS, C9orf72, etc.). Agents targeting TDP-43 proteostasis or stress granule dynamics must be tested in patients with confirmed TDP-43 pathology to avoid signal dilution. |
| **Pharmacodynamic proof of mechanism** | Trials must include target-engagement biomarkers (autophagy markers, XBP1 splicing, SG dissolution, S6 phosphorylation) in patient CSF or accessible cells to prove that the drug is hitting the intended pathway at the tested dose. |
| **CNS exposure validation** | Many agents with strong in vitro data fail because human CNS penetration is insufficient. PET tracers, CSF drug levels, or imaging surrogates are needed. |
| **Adaptive / biomarker-driven trial designs** | Given the rapid progression of ALS and heterogeneity of disease, platform trials (e.g., HEALEY) or adaptive basket trials with integrated biomarker readouts are preferable to fixed traditional Phase 3 designs. |
| **Differentiation from prior failures** | If a mechanism (e.g., generic autophagy activation) has already failed in a large trial (trehalose, TUDCA), a new trial must justify why the agent, dose, population, or endpoint strategy is meaningfully different. |

---

## 6. Conclusion

The field has generated an impressive preclinical toolkit of compounds that modulate TDP-43 proteostasis, autophagy, and stress granule dynamics. However, translation to broad ALS populations has been poor: **trehalose, TUDCA/AMX0035, colchicine, arimoclomol, and BIIB105** have all returned negative results in Phase 2 or Phase 3 trials. The most viable near-term repurposing opportunities are **rapamycin** (strong autophagy/TDP-43 data, ALS Phase 2 safety) and **edaravone** (already approved for ALS, with a newly validated TDP-43 mechanism that supports precision enrichment). **α-Lipoic acid/lipoamide** offers a novel stress-granule mechanism but requires an ALS-specific proof-of-concept trial before it can be considered a viable repurposing candidate. The central lesson across all agents is that **clinical trial design matters as much as target biology**: without TDP-43 biomarker enrichment and pharmacodynamic confirmation in humans, even mechanistically elegant therapies are unlikely to succeed in ALS.

---

## 7. Sources and Key References

1. Mandrioli et al. (2023). *Randomized, double-blind, placebo-controlled trial of rapamycin in amyotrophic lateral sclerosis.* **Nature Communications**. 
2. Casiraghi et al. (2025). *Modeling of TDP-43 proteinopathy by chronic oxidative stress identifies rapamycin as beneficial in ALS patient-derived 2D and 3D iPSC models.* **Experimental Neurology**. 
3. HEALEY ALS Platform Trial Study Group (2025). *Safety and efficacy of trehalose in amyotrophic lateral sclerosis (HEALEY ALS Platform Trial).* **The Lancet Neurology**. 
4. TUDCA-ALS Study Group / European Horizon 2020 trial (NCT03800524); results reported ~March 2024.
5. Amylyx Pharmaceuticals / AMX0035 withdrawal (April 2024); PHOENIX Phase 3 failure.
6. Colchicine Co-ALS trial results (NCT03693781); Mandrioli et al., reported 2023/2024.
7. Arimoclomol ORARIALS-01 Phase 3 results (reported ~May 2024).
8. Uechi et al. (2025). *Small-molecule dissolution of stress granules by redox modulation benefits ALS models.* **Nature Chemical Biology**. 
9. Modafferi et al. (2025). *DNA damage response defects induced by the formation of TDP-43 and mutant FUS cytoplasmic inclusions and their pharmacological rescue.* **Cell Death and Differentiation**. 
10. Preclinical edaravone/TDP-43 mechanism (2025, **Neuropharmacology** / **FASEB Journal**; PMID 40010009, 39887552).
11. Biogen/Ionis BIIB105 termination announcement (May 2024).
12. EKZ-438 / EKZ-102 HDAC6 inhibitor preclinical data (2025); Eikonizo/Otsuka collaboration.

---

*Disclaimer: This document is for scientific research and strategic planning purposes only. It does not constitute clinical, regulatory, or investment advice. Off-label use of approved drugs should only be considered within the context of controlled clinical trials.*