Serum Myostatin Tracks Disease Severity in Spinal Muscular Atrophy, Small Study Finds

Researchers in Milan have mapped how two blood-borne proteins behave across the spinal muscular atrophy (SMA) spectrum, and their findings point to myostatin as a candidate marker of how severely the disease has struck an individual patient. In a cohort of 37 clinically and genetically confirmed pediatric and adult SMA patients followed at the Fondazione IRCCS Istituto Neurologico Carlo Besta, serum levels of myostatin, a potent negative regulator of skeletal muscle mass, rose in step with better preserved motor function, while the downstream signaling molecule Smad3 showed a more complicated and less encouraging profile. The study, published in Annals of Clinical and Translational Neurology, is among the first to track both proteins longitudinally in patients already receiving approved disease-modifying therapies.

SMA is a rare genetic neuromuscular condition in which progressive muscle wasting and weakness lead to loss of movement. Its overall prevalence ranges between 1 and 2 per 100,000 people, and roughly 1 in 11,000 births is affected. About 96 percent of cases stem from mutations or homozygous deletions in the SMN1 gene on chromosome 5q13.2, which encodes most functional full-length survival motor neuron protein, a molecule critical for motor neuron viability. A nearly identical human gene, SMN2, carries a single C-to-T substitution in exon 7 that alters RNA splicing; most SMN2 transcripts therefore lack exon 7 and yield a truncated, non-functional protein, with only about 10 to 15 percent spliced to include exon 7 and produce a small amount of full-length protein, insufficient to prevent disease. Complete absence of SMN is embryonically lethal, while diminished SMN causes anterior horn cells to degenerate, producing motor neuron loss and subsequent skeletal muscle atrophy and weakness. The number of SMN2 copies, from 0 to 4 or more, acts as a disease modifier, with higher copy numbers usually associated with milder phenotypes.

The last decade has brought three approved therapies that upregulate SMN protein: risdiplam and nusinersen enhance exon 7 inclusion in mature mRNA, while onasemnogene abeparvovec delivers gene replacement via AAV9. Beyond these SMN-targeted approaches, myostatin inhibitors are under investigation as potential therapeutic agents in SMA, because myostatin normally suppresses skeletal muscle growth. Apitegromab selectively binds the latent form of myostatin, stabilizing its conformation and preventing access to proteolytic cleavage sites required for activation; a Phase 3 trial reported improved motor function in nonambulatory SMA type II and III patients with generally good tolerability. GYM-329/emugrobart similarly targets latent myostatin and blocks its conversion into the mature bioactive form, whereas taldefgrobep is directed against mature myostatin, inhibiting its interaction with activin type II receptors to promote skeletal muscle growth and function.

Understanding myostatin biology explains why the researchers looked at it in the blood. Like other members of the TGF-beta family, myostatin is secreted as an inactive precursor and becomes active after cleavage of its propeptides. Mature myostatin then binds activin receptors on myoblasts, triggering receptor dimerization, activation of ALK4 or ALK5, phosphorylation of Smad2 and Smad3 which form the Smad4 complex, nuclear translocation of Smad4, suppression of myogenesis genes such as Pax7, MyoD, MyoG and MyHC, and increased expression of atrophy-related genes including Atrogin1 and MuRF1. Earlier work had shown that serum myostatin levels are reduced in SMA patients compared with healthy subjects, reflecting the extent of motor neuron degeneration, and that ambulant SMA type III patients show higher circulating myostatin than non-walkers, though still lower than healthy individuals.

The Milan team enrolled 37 patients, 18 males and 19 females, with a median age at onset of 2 years and a median baseline age of 17 years; 19 were pediatric and 18 adults. The cohort comprised one SMA type I patient, 19 with type II and 17 with type III, classified at baseline as 6 non-sitters, 15 sitters and 16 walkers. Fifteen patients were treated with nusinersen and 15 with risdiplam as monotherapy, while 7 were enrolled in anti-myostatin pharmacological trials. Serum samples were drawn at baseline and again after 6 months in pediatric patients and 14 months in adults, the longer interval reflecting the smaller and slower effect of disease-modifying therapies in adults. Myostatin and total Smad3 were measured with colorimetric ELISA kits, and motor function was assessed with the Hammersmith Functional Motor Scale Expanded (HFMSE, maximum 66 points) and the Revised Upper Limb Module (RULM, maximum 37 points), with improvements of at least 3 points on HFMSE or 2 points on RULM considered clinically meaningful.

At baseline, myostatin levels were significantly higher in male patients (915.0 plus or minus 592.5 pg/mL) than in female patients (593.4 plus or minus 231.4 pg/mL, p = 0.03), in SMA type III (952.2 plus or minus 598.9 pg/mL) compared with type II (541.1 plus or minus 326.4 pg/mL, p = 0.04), and in walkers (1047.6 plus or minus 548.9 pg/mL) compared with non-sitters (374.6 plus or minus 188.5 pg/mL, p = 0.003). Notably, these sex differences in myostatin were not mirrored by differences in disease severity, since HFMSE and RULM scores did not differ significantly between males and females. Smad3 levels, by contrast, showed no significant differences by gender, SMA type or motor function status, and neither protein differed by SMN2 copy number, age at baseline or disease duration. Baseline creatine phosphokinase, available in 25 patients, correlated significantly with myostatin (r = 0.44, p = 0.03), a relationship that held only in the adult subgroup (r = 0.59, p = 0.04). Body weight, BMI and creatinine showed no significant correlations with myostatin, and follow-up changes in body weight were likewise unrelated to myostatin changes.

The correlation analyses anchored the study’s central claim. Across the entire cohort, serum myostatin correlated positively with HFMSE scores both at baseline (r = 0.56, p = 0.0004) and after treatment (r = 0.60, p = 0.0006). In the adult subgroup, this positive correlation persisted at baseline (r = 0.58, p < 0.011) and at 14 months (r = 0.67, p < 0.003). The pediatric picture was different and more provocative: myostatin fold change, the ratio of post-treatment to baseline values, correlated negatively with HFMSE (r = -0.75, p < 0.008), meaning that children whose myostatin fell during therapy tended to show motor improvement, independent of absolute baseline values. Smad3 told a complementary story in children, with significant negative correlations at baseline with HFMSE (r = -0.50, p < 0.03) and at 6 months with RULM (r = -0.65, p < 0.03), suggesting that higher Smad3 levels accompany more compromised motor function in the pediatric cohort.

During treatment itself, however, the biomarkers behaved inconsistently. Statistical analysis revealed no significant changes in serum myostatin or Smad3 at the group level in either age group. Using fold-change thresholds, myostatin decreased in 5 of 12 pediatric patients and increased in 6, while Smad3 decreased in 7 of 12; in adults, myostatin decreased in 9 of 18 and increased in 9, and Smad3 decreased in 11 of 18 and increased in 7. Strikingly, Smad3 fell in all non-sitter patients of both age groups, and myostatin rose in all adult non-sitters, a pattern the authors interpret cautiously given the small numbers. No significant differences emerged between nusinersen- and risdiplam-treated patients, no correlation was found between myostatin and Smad3 levels at any time point, and neither biomarker’s fold change associated with responder status, despite clinically meaningful motor responses in roughly 54.5 percent of pediatric and 27.8 percent of adult patients. The authors attribute the absence of treatment-related correlations to high inter-individual variability over time, noting that relevant modifications in previous studies were limited to presymptomatic patients treated in the neonatal period.

The authors are explicit that the study is exploratory and hypothesis-generating. The small, heterogeneous cohort limits statistical power, and the lack of pediatric healthy controls restricts what can be said about baseline expression levels. On Smad3, they conclude that serum measurements are biologically problematic: unlike secreted myostatin, Smad3 is primarily an intracellular signaling mediator, its circulating levels may not reflect activation status or tissue-specific activity in skeletal muscle, and it can be activated by other TGF-beta superfamily ligands involved in inflammation and fibrosis. Their data therefore do not support serum Smad3 as a clinically applicable biomarker at this stage. Myostatin, in contrast, emerges as a marker of disease severity and functional status in adults and of disease progression dynamics in children, regardless of ongoing treatment, though the pediatric-adult differences may reflect disease stage, subtype distribution and treatment timing rather than age-dependent biology. Validation in larger, independent, ideally multicenter longitudinal cohorts, the team writes, will be essential before these observations change clinical practice.

Subject of Research: Serum myostatin and Smad3 as biomarkers of disease severity and treatment response in spinal muscular atrophy

Article Title: Myostatin/Smad3 in Spinal Muscular Atrophy: Biomarkers of Disease Severity and Treatment Response

Article References: Riolo, G., Dosi, C., Parravicini, S., Zanin, R., Cheli, M., Grigoli, E. G., Canciello, A., Andreetta, F., Masson, R., Bonanno, S., Marcuzzo, S., & Maggi, L. (2026). Myostatin/Smad3 in Spinal Muscular Atrophy: Biomarkers of Disease Severity and Treatment Response. Annals of Clinical and Translational Neurology, Article acn3.70564. https://doi.org/10.1002/acn3.70564

Image Credits: AI Generated

DOI: 10.1002/acn3.70564

Keywords: spinal muscular atrophy, myostatin, Smad3, biomarkers, nusinersen, risdiplam, HFMSE, RULM, motor function, disease-modifying therapy, neuromuscular disease, ELISA