Vodobatinib in dementia with Lewy bodies: a c-Abl program stopped by the readout next door

What was tried

Georgetown University ran a randomized, double-blind, placebo-controlled Phase 2 study (NCT03996460) of vodobatinib (K0706) in dementia with Lewy bodies. Vodobatinib is an orally available c-Abl tyrosine kinase inhibitor (ChEMBL CHEMBL4130229, ABL tyrosine-protein kinase inhibitor and Bcr/Abl inhibitor by ChEMBL mechanism annotation). The design compared 192 mg and 384 mg daily powder formulations against matching placebo over 12 weeks, followed by a 4-week washout, with quadruple masking across participants, care providers, investigators, and outcomes assessors. The protocol set treatment-emergent adverse events as the primary outcome, framing the study as a safety, tolerability, and pharmacodynamic readout with exploratory clinical measures. Actual enrollment reached 29 participants aged 25 to 90 before the study was terminated, with a primary completion date of 22 April 2024 (ClinicalTrials.gov, NCT03996460).

The biological hypothesis

The rationale rested on c-Abl as a stress-activated kinase that drives alpha-synuclein pathology. Karuppagounder and colleagues showed that activation of c-Abl contributes to alpha-synuclein-induced neurodegeneration, linking the kinase to the same protein that aggregates in both Parkinson disease and dementia with Lewy bodies (Journal of Clinical Investigation, 2016). Inhibiting c-Abl was proposed to restore autophagic clearance of alpha-synuclein and reduce dopaminergic neuron loss. Because dementia with Lewy bodies and Parkinson disease share alpha-synuclein aggregation as a core lesion, a c-Abl inhibitor that worked in one synucleinopathy was expected to read across to the other. Open Targets places the ABL1 to Lewy body dementia association at 0.0988, with contributions from literature (0.074) and clinical evidence (0.159) and no human genetic association datatype score. The kinase is highly druggable, with ABL1 carrying small-molecule tractability buckets up to Approved Drug, but druggability is not disease validation.

What actually happened

The dementia with Lewy bodies study was not stopped on its own data. It was halted because the larger sibling program failed. The posted reason states that an interim analysis of 442 patients who completed Part I of the PROSEEK study (CLR_18_06) in Parkinson disease did not show evidence of treatment benefit in patients receiving vodobatinib (ClinicalTrials.gov, NCT03996460). With the lead synucleinopathy indication showing no signal at interim, the smaller dementia with Lewy bodies study lost its justification and was terminated at 29 participants. This pattern mirrors the earlier nilotinib experience. The NILO-PD randomized Phase 2 study of nilotinib, another Abl inhibitor, did not support advancement in moderately advanced Parkinson disease (JAMA Neurology, 2021), following earlier safety and biomarker work (JAMA Neurology, 2020) and open-label observations in Parkinson disease and dementia with Lewy bodies (Journal of Parkinson's Disease, 2016). Vodobatinib carries no reports in OpenFDA FAERS, consistent with an investigational compound that never reached market.

Failure mechanism, best guess

This was a translational mismatch rather than a safety or operational failure. The c-Abl neuroprotection hypothesis was built on rodent models where kinase inhibition reduced alpha-synuclein burden and rescued neurons. The clinical record across two distinct c-Abl inhibitors now points the other way. Three gaps are plausible. First, central nervous system target engagement at tolerated doses may have been insufficient, a recurring problem for kinase inhibitors that must cross the blood-brain barrier and act in neurons rather than in dividing leukemic cells. Second, c-Abl activation may be a downstream marker of neurodegeneration rather than an upstream driver in established human disease, so late intervention cannot reverse an accumulated lesion. Third, the ABL1 to disease link lacks human genetic anchoring; the Open Targets score of 0.0988 reflects literature and clinical inference, not genetics. A review of the field framed targeting alpha-synuclein and c-Abl as biologically reasonable but clinically unproven (Trends in Molecular Medicine, 2023), and a parallel preclinical c-Abl inhibitor program reported neuroprotection in mouse models that has yet to translate (Science Translational Medicine, 2023).

How to prevent this next time

Two quantitative tools would have changed the decision calculus. First, a Bayesian go/no-go framework anchored on the PROSEEK interim would have let the dementia with Lewy bodies team treat the Parkinson disease readout as a prior rather than as a surprise. The posterior probability of success for a downstream synucleinopathy study is:

PP(success)=∫θ​P(final success∣θ,interim data)⋅π(θ∣interim data)dθ

With a futile interim in the lead indication, the posterior mass for the related indication collapses, which is exactly what the sponsor acted on. Pre-registering that dependency would have avoided enrolling 29 participants into a study whose fate was tied to external data. Second, a target-engagement gate with explicit power would have helped. A small cerebrospinal fluid pharmacodynamic study powered at 80 percent to detect a prespecified reduction in a c-Abl activity marker, run before any efficacy cohort, would have established whether tolerated doses inhibit the kinase in human brain at all. Without that gate, a null clinical result cannot distinguish a wrong hypothesis from inadequate exposure. The single highest leverage change would have been a blood-brain-barrier target-engagement biomarker gate that had to be cleared before any synucleinopathy efficacy cohort, in any indication, was opened.

What this means for similar programs

The c-Abl axis has now absorbed repeated clinical disappointment across nilotinib and vodobatinib without a positive efficacy readout in a synucleinopathy. ChEMBL annotates 18 distinct molecules with a mechanism of action on ABL1, so the target is heavily worked, mostly in oncology where the biology and the dosing logic differ from neurodegeneration. The Claidex Mechanism Risk Score for ABL1 is 35 of 100 (yellow), driven by a weak genetic deficit component (13.52 of 15) alongside the documented Phase 2 burden (10.74 of 40). Neurodegeneration programs that repurpose oncology kinase inhibitors should assume the burden of proof has shifted, and lead with human target engagement rather than another efficacy bet.

Open questions

Did vodobatinib achieve measurable c-Abl inhibition in the human central nervous system at the doses tested, and was pharmacodynamic data from the 29 enrolled participants analyzed before termination? Would an earlier-stage synucleinopathy population behave differently from a late clinical cohort? Is c-Abl activation a cause or a consequence of human Lewy pathology, a question that current clinical data cannot resolve?

Sources

1. • ClinicalTrials.gov, NCT03996460, vodobatinib (K0706) in dementia with Lewy bodies, study record and termination reason. https://clinicaltrials.gov/study/NCT03996460 - Karuppagounder SS et al. Activation of tyrosine kinase c-Abl contributes to alpha-synuclein-induced neurodegeneration. Journal of Clinical Investigation, 2016. https://- Pagan FL et al. Nilotinib effects on safety, tolerability, and potential biomarkers in Parkinson disease. JAMA Neurology, 2020. https://- Simuni T et al. Efficacy of nilotinib in patients with moderately advanced Parkinson disease (NILO-PD). JAMA Neurology, 2021. https://- Pagan F et al. Nilotinib effects in Parkinson's disease and dementia with Lewy bodies. Journal of Parkinson's Disease, 2016. https://- Karim MR et al. Targeting alpha-synuclein and c-Abl in Parkinson's disease. Trends in Molecular Medicine, 2023. https://- Hebron M et al. The c-Abl inhibitor IkT-148009 suppresses neurodegeneration in mouse models. Science Translational Medicine, 2023. https://- Open Targets Platform, ABL1 (ENSG00000097007) to Lewy body dementia (EFO_0006792) association, score 0.0988. https://platform.opentargets.org - ChEMBL, vodobatinib (CHEMBL4130229), maximum phase 2; ABL1 target (CHEMBL1862), 18 molecules with annotated mechanism of action. https://www.ebi.ac.uk/chembl - OpenFDA FAERS, drug event query for vodobatinib, no reports returned. https://api.fda.gov/drug/event.json.