A systemic myosin inhibitor met a focal disease, and the index was set before dosing

What was tried

Motric Bio ran NCT06830642, a Phase 2a, randomized, quadruple masked, placebo controlled study of oral MTR-601 in cervical dystonia. Enrollment was 38 across 17 locations, dosed 80 mg daily for weeks 1 and 2 then 160 mg for weeks 3 and 4. The study ran from 28 February 2025 to 18 December 2025, and the record moved to TERMINATED on 17 September 2026 with the reason "Pending further safety evaluation".

Entry required a confirmed diagnosis, prior botulinum toxin exposure not discontinued for lack of efficacy, and a Toronto Western Spasmodic Torticollis Rating Scale total of at least 20, with botulinum toxin withheld for the duration. Co-primary endpoints were treatment emergent adverse events through week 6 and change in TWSTRS overall score to week 4. Results have been posted. MTR-601 is also listed as MPH-220, originated by Motor Pharma. No ChEMBL record resolves to either identifier, and openFDA FAERS returned none.

The biological hypothesis

Cervical dystonia is driven by involuntary, sustained contraction of neck muscles under aberrant central motor commands. Botulinum toxin A, the reference treatment, acts at the neuromuscular junction of muscles a clinician selects, so its effect is focal.

MTR-601 proposed the opposite geometry. Gyimesi and colleagues reported in 2020 that skeletal and cardiac myosin-2 differ at a key residue in the communication center between their functional regions, and used that difference to design MPH-220, a direct myosin inhibitor selective for skeletal over cardiac myosin, with relaxation reported without cardiovascular side effects and the bound structure deposited as PDB 6YSY. The company describes MTR-601 as a selective inhibitor of myosin 2 in fast twitch skeletal muscle.

The fast twitch myosin heavy chains in adult human limb and neck muscle are MYH1 and MYH2. Published characterization of MPH-220 resolves selectivity against cardiac myosin rather than between fast skeletal isoforms, so MYH2 is the representative gene here, not a claim of isoform-specific binding. That gene has a documented human loss of function phenotype: Tajsharghi and colleagues described in 2010 the first patients lacking fast type 2A fibres entirely, from truncating MYH2 mutations, with early onset generalized weakness and extraocular muscle involvement.

In Open Targets, queried on 18 September 2026, MYH2 has no recorded association with cervical dystonia of any evidence type, so the score is 0.0, and it carries no drug or clinical candidate records. The rationale was pharmacological and symptomatic, not genetic.

What actually happened

Thirteen of 17 MTR-601 participants and 14 of 21 placebo participants completed. Treatment emergent adverse events were reported in 14 of 17 and 14 of 21. One serious adverse event occurred, hypertransaminasaemia, in 1 of 17 on MTR-601 against 0 of 21, and no deaths.

The non-serious profile separated on nervous system terms. Dizziness was reported in 7 of 17 on MTR-601 against 2 of 21 on placebo, somnolence in 5 of 17 against 1 of 21, and fatigue in 3 of 17 against 1 of 21. Blood creatine phosphokinase rose in 2 of 17 against 0 of 21. Single events on MTR-601 with none on placebo included balance disorder, diplopia, a fall and aspartate aminotransferase increase.

The posted least squares mean change in TWSTRS overall score to week 4 was -7.77 (standard error 2.978) on MTR-601 and -5.55 (2.899) on placebo, on 17 and 21 participants. Taking the arms as independent under a normal approximation, the difference is -2.22 with a standard error of 4.16 and an approximate 95 percent interval of -10.37 to 5.93. On the week 4 subscales, severity favored MTR-601 by 1.43 points and disability by 1.60, while pain favored placebo by 0.47.

Published minimal clinically important change thresholds for TWSTRS total are -11.9 (95 percent CI -13.9 to -10.0) for an average trial population and at least 8 points in the CD PROBE registry analysis. The observed 2.22 point difference falls short of both.

Failure mechanism, best guess

The posted reason is safety, and the safety data support that reading. The deeper problem is that the therapeutic index of a systemic myosin inhibitor in a focal disease is set by muscles the drug was never meant to reach. Botulinum toxin works because a clinician chooses which muscles to weaken. An oral fast skeletal myosin inhibitor weakens every fast twitch muscle at once, so the dose that relaxes the dystonic sternocleidomastoid also acts on postural, ocular and respiratory musculature. The selectivity engineered into MPH-220 is skeletal against cardiac, which protects the heart and does nothing for the index within skeletal muscle.

The human MYH2 null phenotype predicts what full target engagement looks like: generalized weakness with extraocular involvement. The active arm reported balance disorder, a fall, diplopia and raised creatine phosphokinase. Individually these are single events in 17 participants, inseparable from chance. As a pattern against a known loss of function phenotype, it is what the mechanism predicts.

How to prevent this next time

Endpoint level data are posted here, so a quantitative check is available.

Illustrative: scaling the posted standard errors by the square root of arm size gives implied standard deviations of 12.28 and 13.28, pooling to 12.85. At that variance, 80 percent power at two sided alpha 0.05 needs about 41 participants per arm to detect an 8 point difference and about 19 per arm to detect 11.9 points. The trial as run had roughly 82 percent power against 11.9 points and roughly 49 percent against 8. Assumptions: independent arms, normal approximation, no covariate adjustment. These are bounds, not the sponsor's numbers. The trial was sized against the optimistic end of the published threshold.

Measure the therapeutic index directly. A mechanism whose on-target toxicity is weakness should carry a pre-specified quantitative weakness endpoint, myometry in an untargeted muscle group, read against the TWSTRS change in the same participants. Without it, a safety stop is a judgment call rather than a measurement. Extraocular involvement in MYH2 null patients points directly at diplopia as a sentinel event.

The single highest leverage change would have been to pre-specify a paired weakness and efficacy readout, so the study measured the ratio between neck relaxation and systemic weakness instead of testing whether either one existed.

What this means for similar programs

MYH2 now carries a Mechanistic Risk Score of 38, band yellow. The genetic deficit term sits at its maximum of 15 of 15, because Open Targets holds no MYH2 to cervical dystonia association. Phase burden is 8.05 of 30, archetype severity 8.52 of 25, recency 3.89 and saturation 2.72.

The transferable point is about effector targeting in focal disease. Inhibiting the final common effector of a symptom is mechanistically clean and pharmacologically blunt. When the disease is focal and the effector is systemic, the achievable index is bounded before the first dose, and selectivity against a different tissue does not change that bound. Spasticity, where MTR-601 is also in development, is a generalized target and needs different arithmetic.

Open questions

Which finding triggered "Pending further safety evaluation", given that only one serious adverse event is posted?

Did the hypertransaminasaemia case resolve, and is a hepatic signal separable from the muscle enzyme rise in 2 of 17?

Is MTR-601 isoform-resolved between MYH1 and MYH2?

Sources

1. • ClinicalTrials.gov, NCT06830642, record and posted results as of 17 September 2026. https://clinicaltrials.gov/study/NCT06830642 - Motric Bio, MTR-601 program page, accessed 18 September 2026. https://www.motricbio.com/mtr-601 - AdisInsight, MPH 220 drug record, accessed 18 September 2026. https://adisinsight.springer.com/drugs/800060863 - Gyimesi M, et al. Single Residue Variation in Skeletal Muscle Myosin Enables Direct and Selective Drug Targeting for Spasticity and Muscle Stiffness. Cell, 2020. https://- RCSB PDB 6YSY, skeletal myosin bound to MPH-220, MgADP-VO4. https://- Tajsharghi H, et al. Human disease caused by loss of fast IIa myosin heavy chain due to recessive MYH2 mutations. Brain, 2010. https://- Martinsson T, et al. Myosin heavy chain IIa gene mutation E706K is pathogenic and its expression increases with age. Neurology, 2002. https://- Tajsharghi H, et al. Mutations and sequence variation in the human myosin heavy chain IIa gene (MYH2). European Journal of Human Genetics, 2005. https://- Dominantly inherited myosin IIa myopathy caused by aberrant splicing of MYH2. BMC Neurology, 2022. https://- Marques RH, et al. Botulinum toxin type A therapy for cervical dystonia. Cochrane Database of Systematic Reviews, 2020. https://- Minimal clinically important change in the Toronto Western Spasmodic Torticollis Rating Scale. Parkinsonism and Related Disorders, 2018. https://- Minimal clinically important change in patients with cervical dystonia: Results from the CD PROBE study. Journal of the Neurological Sciences, 2019. https://- Open Targets Platform, MYH2 (ENSG00000125414) against cervical dystonia (MONDO_0000481), queried 18 September 2026. https://platform.opentargets.org/target/ENSG00000125414 - openFDA drug adverse event API, queried 18 September 2026. https://api.fda.gov/drug/event.json.