🧬 Aging Wiki

aav-follistatin

Properties1
aliasesAAV-FS344, AAV-follistatin, follistatin gene therapy, AAV1-FS344

12 min read

AAV-Follistatin gene therapy

Intramuscular delivery of an adeno-associated virus (AAV) encoding follistatin (FST), a naturally secreted glycoprotein that binds and neutralises myostatin (GDF-8) and related activins in the extracellular space. Myostatin is the dominant negative regulator of skeletal muscle mass in mammals; its blockade drives muscle fibre hypertrophy and satellite cell activation 1. AAV-mediated follistatin overexpression has completed a Phase 1 trial in inclusion body myositis (IBM) and Becker muscular dystrophy (NCT01519349), with modest functional improvements and no serious adverse events 2. As of 2026-05-06, no completed Phase 2 trial exists for any indication; a Phase 1/2a sarcopenia-specific trial (NCT07443826) is actively recruiting. Antibody-based myostatin antagonists (bimagrumab, trevogrumab) are the competing clinical modality with a simpler regulatory path, but the gene-therapy approach offers the potential for durable single-administration effect.

Follistatin biology

Follistatin (FST) is a monomeric glycoprotein encoded by the FST gene (chromosome 5q11.2 in humans) that was originally characterised as an activin-binding protein in the pituitary. Its broader role as a pan-TGF-β ligand trap became clear when Lee & McPherron demonstrated that FST overexpression in transgenic mice produced muscle mass increases that met or exceeded those seen in myostatin knockout mice — implying that follistatin sequesters additional growth-inhibiting ligands beyond myostatin alone 1.

Key binding targets of follistatin relevant to muscle:

LigandFamilyEffect on muscle when follistatin blocks it
Myostatin (GDF-8)TGF-βRemoves primary negative regulator of myofibre hypertrophy
Activin ATGF-βRemoves catabolic signal; activin A rises with age and disease
Activin BTGF-βSecondary catabolic signal; less muscle-specific than activin A
BMP-11 (GDF-11)TGF-βControversial; may oppose muscle growth in some contexts

Three isoforms exist (FST-288, FST-303, FST-315), with FST-315 being the predominant circulating form. The FS344 vector (AAV1 serotype) encodes a secreted follistatin isoform designed for local paracrine action after intramuscular injection, with limited systemic distribution intended to confine hypertrophic effects to injected muscles.

The mechanism is upstream of SMAD2/3 signalling: follistatin sequesters myostatin before receptor engagement, preventing ALK4/ALK5 receptor activation and the SMAD2/3 transcriptional complex that represses muscle anabolic genes. See tgf-beta for pathway-level SMAD signalling detail.

Human genetic proof-of-concept — Schuelke 2004

The strongest human genetics evidence that myostatin loss-of-function drives muscle hypertrophy is a NEJM case report of a boy born with an inactivating myostatin mutation who displayed extraordinary muscle development with no adverse health effects at age 4.5 years 3. This established that complete myostatin blockade is safe in humans at paediatric ages and produces the expected hypertrophy phenotype. This is a single case; long-term safety of lifelong myostatin deficiency in adulthood is unknown needs-replication long-term-unknown.

Mendell 2017 Phase 1 trial (IBM + Becker MD) — NCT01519349

The foundational human gene therapy trial. Note on timing: the trial completed accrual circa 2014–2015; a conference abstract appeared in 2016 4; the full peer-reviewed publication appeared in 2017 2. The user brief refers to this as “Mendell 2015”; the citable primary source is the 2017 Molecular Therapy paper.

Design: Open-label, single-group Phase 1 (NCT01519349; n=15 total: 9 sIBM + 6 BMD; 3 sIBM received unilateral dosing only and are not analysed in this paper). rAAV1.CMV.huFS344, 6×10¹¹ vg/kg per leg, delivered by bilateral IM injection to the quadriceps in 12 sites per leg under EMG/MRI guidance. Registered primary endpoint: safety (Grade III or higher treatment-related toxicities at 2 years); the 6MWT comparison vs an external clinic cohort is a post-hoc efficacy analysis on a non-pre-specified endpoint.

IBM cohort results (from Mendell 2017):

MeasureOutcome
Serious adverse eventsNone reported
6MWT4 of 6 participants improved (absolute change +58 m to +153 m); 2 minimally improved (+5 m and +23 m); annualised median change +56.0 m/yr vs −25.8 m/yr in matched untreated sIBM controls (n=8), p=0.01
Muscle biopsyDecreased fibrosis (reduced TGF-β, Col1A, fibronectin expression); increased muscle fibre number and normalised fibre size distribution; mTORC1 activation (increased phospho-S6P and phospho-4E-BP1); no reduced myostatin protein expression was reported
ImmunogenicityNo consistent pattern of T-cell immunity to AAV1 capsid; serum anti-follistatin antibody levels remained below 1:50 titers throughout; no adverse immunogenicity-related loss of expression documented

Limitations of the trial:

  • n=6 in the sIBM cohort (the full trial NCT01519349 enrolled 15 subjects: 9 sIBM + 6 Becker MD; 3 sIBM received unilateral dosing only and are not analysed in this paper); no placebo arm; functional effect sizes are uninterpretable without a concurrent control arm
  • Registered primary endpoint was safety (Grade III or higher treatment-related toxicities at 2 years); the 6MWT comparison vs clinic controls is a post-hoc-defined efficacy analysis on a non-pre-specified endpoint — a point specifically criticised by Greenberg 2017 5
  • Subjects received ≥4 co-interventions (AAV-FS344 gene therapy, 60 days high-dose prednisone, structured exercise programme, open-label placebo effects); efficacy cannot be attributed to gene therapy alone
  • IM delivery to quadriceps only; no systemic distribution data; cannot address sarcopenia as a whole-body condition
  • Only 2 of 6 subjects have 6MWT data at the registered 2-year endpoint in this publication
DimensionStatusNotes
Pathway conserved in humans?yesMyostatin–FST axis is well-characterised in humans; human genetics supports it (Schuelke 2004)
Phenotype conserved in humans?partialIBM is an inflammatory myopathy, not sarcopenia; results may not extrapolate to age-related muscle loss
Replicated in humans?in-progressNCT07443826 is a Phase 1/2a trial in age-related sarcopenia; results pending

needs-replication (no controlled trial in sarcopenia population)

Aging-relevance rationale

Myostatin is a member of the TGF-β superfamily that functions as a circulating negative regulator of muscle mass throughout adult life. Its relevance to aging-related muscle decline is supported by several lines of evidence:

  1. Rising myostatin with age. Circulating and intramuscular myostatin expression increases with advancing age in humans 6, coinciding with progressive sarcopenic muscle loss. Whether this rise is causal or adaptive is not definitively resolved. no-mechanism

  2. Satellite cell regulation. Myostatin suppresses satellite cell activation and self-renewal via SMAD2/3, contributing to reduced regenerative capacity of aged muscle. This maps to the stem-cell-exhaustion hallmark (satellite cells are the resident muscle stem cells — see satellite-cells).

  3. Paracrine and endocrine signalling. Myostatin is secreted by muscle and acts both locally and systemically; its antagonism by follistatin is thus relevant to altered-intercellular-communication (myostatin as an aging-amplifying secreted signal). Aged muscle has an altered secretome relative to young muscle that includes elevated myostatin and activin A 6.

  4. Functional consequence. Sarcopenia (defined as loss of muscle mass + function) affects ~10–15% of adults over 65 and predicts fall risk, hospitalisation, and mortality. See sarcopenia for the full epidemiology and mechanistic breakdown.

The therapeutic rationale is therefore: AAV-mediated follistatin overexpression could provide durable local suppression of the myostatin/activin brake, promoting satellite cell activation and fibre hypertrophy in targeted muscle groups — potentially reversing or retarding the sarcopenic trajectory with a single administration.

Competing modalities: antibody-based myostatin antagonists

Several protein-based myostatin and activin antagonists have advanced further in clinical development than AAV-follistatin for sarcopenia:

AgentMechanismStatus (2026)Developer
BimagrumabAnti-ActRIIA/IIB antibody (blocks activin + myostatin receptor)Phase 2 completed (sarcopenia, obesity); Phase 3 plannedNovartis / Scholar Rock
TrevogrumabAnti-myostatin monoclonal antibodyPhase 2/3 for FSHD; sarcopenia programme activeRegeneron
ApitegromabAnti-pro-myostatin/latent-myostatin antibodyPhase 2/3 SMAScholar Rock

Relative to AAV-follistatin, antibody-based antagonists have:

  • Reversible pharmacology (safer for unknown long-term effects)
  • Systemic distribution (addresses whole-body sarcopenia, not one muscle group)
  • A more established regulatory precedent
  • Larger clinical databases

The gene therapy advantage is the potential for permanent or very long-lived effect from a single injection, which could be relevant if long-term antibody dosing is cost- or compliance-limited. This advantage has not yet been demonstrated to translate to meaningful superiority in any head-to-head comparison. needs-replication

Translation barriers

  1. IM-only delivery. The AAV1.FS344 construct is delivered intramuscularly to targeted muscle groups, not systemically. This is appropriate for focal dystrophies (Becker MD, IBM) but limits utility for sarcopenia, which requires whole-body muscle mass restoration. Systemic IV delivery of higher-titre AAV9 or AAV-rh74 encoding follistatin might achieve broader distribution, but safety data for systemic follistatin overexpression (liver, ovary, bone) are limited and ovarian follicle development would be a concern (follistatin is a key regulator of reproductive biology via activin A). long-term-unknown

  2. Permanent overexpression safety. Unlike an antibody with a 2–4-week half-life, AAV-driven gene expression is intended to be long-lived (years in post-mitotic cells). Chronic supraphysiological follistatin may dysregulate activin-mediated bone remodelling, cardiac function, and reproductive axis in ways not captured in short-term trials. long-term-unknown

  3. AAV immunogenicity. Pre-existing neutralising antibodies against AAV1 in a subset of the population will reduce transduction efficiency or require pre-screening. Re-dosing is impractical with the same serotype.

  4. No active sarcopenia IND (as of 2026-05-06). The NCT07443826 trial is early-stage; no Phase 2 data exist for aging-related sarcopenia as a primary indication. The regulatory path requires validated sarcopenia endpoints (lean mass + functional composite), which are still being standardised.

  5. Competing antibody modality. If bimagrumab or trevogrumab receive regulatory approval for sarcopenia — a feasible near-term scenario — the commercial rationale for a more complex gene therapy approach narrows substantially.

Active clinical trials (as of 2026-05-06)

NCTTitlePhaseStatusConditionsSponsor
NCT01519349Follistatin Gene Transfer to Patients With Becker Muscular Dystrophy and Sporadic Inclusion Body Myositis1CompletedBecker MD, IBMNationwide Children’s Hospital
NCT07443826CALM-AF-AI: Counteracting Age-related Loss of Muscle With AAV-Follistatin (±VEGF plasmid)1/2aRecruitingSarcopenia, age-related muscle decline—

clinical-trials-active: 1 (NCT07443826 recruiting as of 2026-05-06; confirmed via ClinicalTrials.gov v2 API).

No trials specifically in aging-healthy-adult or longevity endpoints exist as of this date.

Gaps and limitations

  • No Phase 2 controlled data. The entire efficacy evidence base in humans is an open-label n=6 IBM cohort with no control arm, subject to significant natural-history confounds. needs-replication

  • IBM ≠ sarcopenia. IBM is an immune-mediated inflammatory myopathy. Generalising the IBM trial result to age-related sarcopenia requires at minimum a dedicated sarcopenia cohort — now initiating in NCT07443826. needs-human-replication

  • Mendell 2017 disputed. A formal published letter (Greenberg 2017) challenged the Mendell 2017 paper on multiple grounds: the reported primary outcome (6MWT) was not the registered primary outcome (safety); the “annualised 6MWT” is a post-hoc-defined metric; co-administration of prednisone + structured exercise + open-label placebo effects makes AAV attribution impossible; and the AE reporting was misleading (fall rate 33%, not the 6% stated in the paper) 5. No published rebuttal from the Mendell group has appeared as of wiki creation 2026-05-06. This is a live scientific dispute. contradictory-evidence

  • Long-term safety of persistent FST overexpression unknown. Follistatin regulates bone metabolism (activin/BMP axis), ovarian folliculogenesis, and cardiac hypertrophic signalling. Lifelong overexpression in muscle (with some paracrine/endocrine spillover) has not been characterised in any long-term animal study. long-term-unknown

  • Dose-response undefined. A single dose level per cohort was tested in NCT01519349; optimal dose for muscle hypertrophy vs. acceptable systemic exposure has not been established. dose-response-unclear

  • myostatin protein page is a stub. This page links heavily to myostatin (the protein), which does not yet have a dedicated page. Given that four separate claims on this page depend on myostatin biology details, promoting myostatin to R24b is recommended (see summary).

Cross-references

  • myostatin (implicit stub — no page; R24b promotion candidate; this page depends on it)
  • tgf-beta (R20 verified-partial) — pathway; SMAD2/3 signalling downstream of myostatin receptor ALK4/ALK5
  • sarcopenia (existing page) — target phenotype for aging application
  • satellite-cells (verified-partial) — resident muscle stem cells; myostatin suppresses their activation
  • stem-cell-exhaustion (hallmark) — satellite cell depletion is a key sarcopenia driver
  • altered-intercellular-communication (hallmark) — myostatin as aging-amplifying secreted paracrine/endocrine signal
  • aav-tert (verified R23b sibling) — sibling gene therapy page; compare translation-gap strategies
  • aav-klotho (R23b sibling) — another systemic aging gene therapy
  • aav-osk (R23b sibling) — partial-reprogramming approach; different mechanism
  • crispr-base-editing-pcsk9 (R23b sibling) — cardiovascular gene therapy; regulatory precedent
  • hallmarks-of-aging — stem-cell-exhaustion and altered-intercellular-communication
  • sens-damage-categories — applicable to intracellular junk / extracellular junk (indirect)

Footnotes

Footnotes

  1. doi:10.1073/pnas.151270098 · Lee SJ, McPherron AC · in-vivo · Proc Natl Acad Sci USA 2001 · 98(16):9306–9311 · model: mouse (transgenic FST overexpression; hybrid SJL/C57BL/6 founders × wild-type C57BL/6 F1 offspring) · follistatin overexpression (human short-form isoform driven by myosin light chain promoter) produced muscle mass increases of 194–327% in best founder (F3), exceeding those seen in myostatin null mice in comparable backgrounds; authors attribute the excess effect to blockade of additional follistatin-sensitive ligands beyond myostatin (e.g., GDF-11, activins); defines follistatin as an in vivo muscle-growth promoter and functional myostatin antagonist · archive: local PDF available ↩ ↩2

  2. doi:10.1016/j.ymthe.2017.02.015 · Mendell JR et al. (Sahenk Z, Al-Zaidy S, Rodino-Klapac LR, Lowes LP, Alfano LN, Berry K, Miller N, Yalvac M, Dvorchik I, Moore-Clingenpeel M, Flanigan KM, Church K, Shontz K, Curry C, Lewis S, McColly M, Hogan MJ, Kaspar BK) · phase-1 · Mol Ther 2017 · 25(4):870–879 · model: human sIBM (n=6 analysed; full trial NCT01519349 n=15 including Becker MD); rAAV1.CMV.huFS344, 6×10¹¹ vg/kg per leg, bilateral IM injection to quadriceps · registered primary endpoint: safety (Grade III+ toxicities at 2 years); 6MWT efficacy analysis is post-hoc vs n=8 matched untreated clinic controls; annualised median 6MWT +56.0 m/yr treated vs −25.8 m/yr untreated (p=0.01); absolute changes ranged from +5 m to +153 m across 6 subjects; biopsy: decreased fibrosis, mTORC1 activation; no consistent T-cell immunity; anti-follistatin antibodies remained <1:50; no treatment-related AEs or SAEs · archive: local PDF available ↩ ↩2

  3. doi:10.1056/NEJMoa040933 · Schuelke M, Wagner KR, Stolz LE, Hübner C, Riebel T, Kömen W, Braun T, Tobin JF, Lee SJ · case-report · N Engl J Med 2004 · 350(26):2682–2688 · model: human (single paediatric case); homozygous loss-of-function myostatin mutation · gross skeletal muscle hypertrophy from birth; increased strength; no adverse effects at 4.5 years; proof-of-concept that myostatin loss of function is safe in humans · archive: local PDF available ↩

  4. doi:10.1016/s1525-0016(16)33306-8 · Mendell JR et al. · conference abstract · Mol Ther 2016 · 24:S197–S198 · preliminary 6MWT improvement data from NCT01519349; superseded by full paper 2 for quantitative claims ↩

  5. doi:10.1016/j.ymthe.2017.09.002 · Greenberg SA · letter/commentary · Mol Ther 2017 · 25(10):2235–2237 · critical analysis of Mendell 2017 2; specific criticisms: (1) the published primary outcome (6MWT) contradicts the ClinicalTrials.gov-registered primary outcome (safety at 2 years); (2) the “annualised 6MWT” is a post-hoc-defined imaginary metric; (3) the treated group received 4 co-interventions (AAV-FS344, prednisone, exercise, placebo effect), making AAV attribution impossible; (4) AE frequency is misrepresented (fall rate was 33%, not 6%); (5) the conclusion that this is “the first trial to show clear evidence of treatment benefit in sIBM” neglects prior randomised trials; does not present new primary data; no published rebuttal from the Mendell group as of wiki creation 2026-05-06 · archive: local PDF available ↩ ↩2

  6. doi:10.1159/000356740 · White TA, LeBrasseur NK · review · Gerontology 2014 · 60(2):89–97 · model: review of human + rodent data · summarises evidence for myostatin rise with aging; links myostatin to sarcopenic phenotype; not OA · archive: not_oa (no local PDF) — use with caution; review-level evidence only ↩ ↩2

Graph View

Backlinks