SMA Research Platform

Dual-Target Molecules

EXPLORATORY

8 candidates

Phase 6.1 (exploratory hypothesis) — Compounds predicted to both modify SMN2 splicing AND modulate ion channels. This page tests an internal hypothesis: that SMN restoration alone may not fully recover motor-neuron function, and that co-modulating motor-neuron excitability could help. This is a computational hypothesis only — none of the combinations below is wet-lab validated, and the bioelectric premise itself is unproven.

Why dual-target matters for SMASMA motor neurons face two issues. (1) Low SMN protein from mis-spliced SMN2 — well established, and the basis of risdiplam/nusinersen/onasemnogene. (2) A hypothesis we are testing: that surviving motor neurons may also have altered excitability, so even SMN-restored neurons might fire sub-optimally. The literature does show SMN-restoring therapies leave residual motor deficit, but the cause is multifactorial and is NOT established to be electrical dormancy — that is our working hypothesis, not a clinical finding. A dual-target compound would address both axes at once. Treat every entry below as a computational prediction or rational-design concept, not a therapeutic recommendation.

Score interpretation: SMN2 Score — predicted effect on SMN2 exon 7 inclusion (0–1, higher = more inclusion). Channel Score — predicted modulation of the target ion channel (0–1). Composite — weighted combination prioritizing compounds that score well on both axes simultaneously, with BBB permeability as a requirement. All scores are model predictions, not measured potencies.

Compound	SMN2 Score	Ion Channel	Channel Score	BBB	Composite	Status
Risdiplam + 4-AP combination concept	0.95	KCNA2	0.80	90%	0.87	Neither approved for SMA combination, but both individually available. Clinical trial of combination is immediately feasible.
Valproic acid (VPA)	0.45	SCN1A	0.50	90%	0.60	Approved for epilepsy. Clinical trials in SMA showed modest benefit.
Riluzole	0.35	SCN1A	0.70	85%	0.58	Approved for ALS. Off-label SMA use explored.
Lamotrigine	0.25	HCN1	0.65	90%	0.53	Approved for epilepsy/bipolar. Good safety profile.
Retigabine (ezogabine)	0.20	KCNQ2	0.75	80%	0.50	Was FDA-approved for epilepsy (withdrawn — blue skin discoloration). Low-dose SMA application could avoid this.
4-Aminopyridine (Dalfampridine)	0.15	KCNA2	0.80	95%	0.50	FDA-approved for MS (walking improvement). Excellent CNS penetration.
Roscovitine (Seliciclib)	0.30	CACNA1A	0.60	70%	0.49	Phase 2 for cancer. Repurposing for NMJ disorders explored.
GV-58	0.10	CACNA1A	0.85	60%	0.42	Preclinical. Shown to improve NMJ transmission in SMA mice.

Note: All candidate combinations above are computational predictions or rational-design concepts — none has been tested in a wet lab. The bioelectric hypothesis (electrical dormancy in SMA motor neurons) is unproven.

Multi-Target Docking Hits

DIFFDOCK

5 drugs

Drugs with at least one positive DiffDock pose (best_confidence > 0) against two or more SMA-relevant targets. A shared predicted pose can indicate polypharmacology — or an off-target liability — and says nothing about whether the compound is active in cells. Several of these targets sit on the ROCK2-LIMK-CFL axis, which our own analysis down-weights (the muscle-restricted ROCK pathway showed no motor/innervation benefit in published mouse work; Bowerman 2010/2012), so a positive pose against ROCK2/LIMK2 should not be read as on-mechanism. Every row here is an in-silico docking prediction requiring orthogonal validation.

Drug	Targets Hit	# Targets	Max Confidence	Avg Confidence
4-Aminopyridine	LIMK1 · MAPK14 · MDM2 · ROCK2 · SMN2 · UBA1	6	0.640	0.296
Riluzole	CFL2 · LIMK1 · LIMK2 · MAPK14 · ROCK2 · SMN2	6	0.443	0.215
Y-27632	LIMK1 · LIMK2 · ROCK2	3	0.465	0.278
Ripasudil	LIMK2 · ROCK2	2	0.491	0.267
Palbociclib	LIMK1 · MAPK14	2	0.246	0.174