Description

MECHANISM: SMA is caused by loss-of-function mutations in SMN1, leaving SMN2 as the sole source of survival motor neuron (SMN) protein. A synonymous C-to-T transition at position +6 of SMN2 exon 7 disrupts an exonic splicing enhancer (ESE) recognized by SRSF1 and simultaneously creates a binding site for the splicing repressor hnRNP A1/A2, resulting in ~90% skipping of exon 7 and production of a truncated, rapidly degraded SMNΔ7 protein. The remaining ~10% full-length SMN protein is insufficient to maintain snRNP biogenesis and RNA splicing fidelity in motor neurons, which are uniquely vulnerable due to their high metabolic demand and axonal length. SMN copy number is the primary disease modifier because each additional SMN2 copy contributes an additive increment of full-length SMN protein; patients with higher copy numbers present with less severe phenotypes. Nusinersen (spinraza), an antisense oligonucleotide (ASO), exploits this mechanism by binding the intronic splicing silencer N1 (ISS-N1) in SMN2 intron 7, sterically blocking hnRNP A1/A2 binding and shifting splicing toward exon 7 inclusion, thereby increasing full-length SMN2 protein production (PMIDs: 29689734, 35533607, 33501671). EVIDENCE CONVERGENCE: Multiple independent clinical studies converge on robust, sustained improvement in motor function following nusinersen-mediated SMN2 splicing correction. In SMA type 1 patients, 77% achieved a clinically meaningful ≥4-point improvement in CHOP-INTEND scores at 6 months, with a mean change of 9.0±8.0 points, and improvements in motor milestones (HINE-2) were documented in 31.1% of patients (PMID: 29689734). Longitudinal data over 3 years confirm gradual, sustained motor improvement in both type 1 and type 2 patients (PMID: 38192577). Critically, treatment timing is a dominant modifier: earlier initiation (≤3 months from symptom onset) consistently yields superior motor and feeding outcomes, consistent with the mechanistic prediction that irreversible motor neuron loss must be prevented before it occurs (PMIDs: 38192577, 41757676). Combination approaches layering nusinersen with gene replacement (onasemnogene abeparvovec) and small-molecule splicing modifiers (risdiplam) have achieved milestones such as independent walking in SMA type 0 (PMID: 41795276), further validating that incremental restoration of SMN protein above a threshold drives proportional functional rescue. CONTRADICTIONS: Despite strong evidence of benefit in type 1 and type 2 SMA, clinical responses in type 2–4 patients are heterogeneous, with some showing stabilization rather than improvement and a subset experiencing deterioration (PMID: 35533607). This suggests that a threshold of irreversible motor neuron loss exists beyond which SMN restoration is insufficient to recover function, and that non-motor pathologies (e.g., cardiac, autonomic, metabolic) may persist. Furthermore, all clinical evidence is derived from nusinersen as a proxy for SMN2 splicing correction; direct quantification of exon 7 inclusion rates and full-length SMN protein levels in human CNS tissue in treated patients remains technically limited, introducing uncertainty about the precise protein threshold required for protection. The contribution of peripheral vs. central SMN restoration to motor outcomes is also incompletely resolved. THERAPEUTIC ANGLE: The validated mechanism supports continued development of SMN2 splicing-correction modalities, with ASOs (intrathecal delivery to CNS) representing the most clinically advanced approach. The principle that SMN2 copy number and splicing efficiency set a molecular ceiling on functional SMN protein production argues for combination strategies that simultaneously address splicing (ASO or small molecule) and absolute gene dosage (AAV9-SMN1 gene replacement), particularly in severe low-copy-number patients. Pre-symptomatic or neonatal screening-driven treatment initiation is the highest-leverage intervention given the time-dependency of motor neuron survival. Future refinements include engineering next-generation ASOs with improved CNS distribution, longer durability, and the possibility of non-invasive delivery, as well as identifying the minimum effective SMN protein threshold in human motor neurons to rationalize dosing strategies.

Key questions

• What is the minimum percentage of SMN2 exon 7 inclusion (and corresponding full-length SMN protein level) required in human spinal motor neurons to prevent axonal degeneration, and does this threshold differ between SMA types?
• Does pre-symptomatic initiation of nusinersen in SMN2 copy-number-stratified newborns identified by universal screening result in complete prevention of motor neuron loss as measured by neurofilament light chain biomarkers and neuroimaging, compared to post-symptomatic treatment?
• In SMN-depleted iPSC-derived motor neurons, does restoration of full-length SMN protein to 20%, 50%, or 100% of wild-type levels differentially rescue snRNP assembly, axonal transport fidelity, and neuromuscular junction formation in a dose-dependent manner?
• Does combinatorial blockade of both ISS-N1 and the hnRNP A1 binding site within SMN2 exon 7 via bifunctional ASOs produce supra-additive exon 7 inclusion compared to ISS-N1 targeting alone in SMA mouse models?
• In SMA type 2–4 patients who show deterioration despite nusinersen treatment, is motor decline associated with insufficient CSF drug exposure, incomplete exon 7 splicing correction, or loss of motor neurons prior to treatment, and can PET or CSF biomarker profiling distinguish these mechanisms?

Supporting evidence (269)

• 69d82e27-f57a-8b18-bf59-fff508a0a0c6
• 8d3779a2-415c-7e31-b2dd-869f67bb2933
• 6fa44d05-1998-16aa-bc2b-ffdaa64e696a
• b52998a3-b3fb-b02b-40cf-2aadd94e6740
• 2ee72abd-cb13-806a-0887-74a3bdafde50
• 0352b994-9864-e817-d888-8e4f81887c25
• b71112b6-43d9-5076-34d5-935dd5d1e0a2
• e1ab7f1f-092c-0e08-19a2-d24f9c98569a
• 31438298-7990-73b8-0f3e-d001ab94a50e
• 13d164a3-bb46-ecad-45dd-4ed0404b5866
• 682d1544-e79b-069c-2c1e-7c57c7e0507b
• 84f50e30-e373-b91f-57e4-cdde18039960
• db67fc0d-b323-2d81-bd9a-07346f3d53e3
• 393eb226-ca0f-89de-baf1-fad607d0db43
• c46249d5-40f3-efba-af10-bfe3f7e89924
• 161502a4-9837-8577-f198-fa8eba1e490b
• 579c5ce0-8e6a-04e1-31b4-b692fdb1597d
• 1928ace8-955c-4ce7-b110-53640b3c36d0
• 06f512ed-f250-4090-b869-dbc140a43fa4
• ffe116c5-7efb-4429-b21a-39145a3d734b
• … and 249 more

Related claims (20)