🧬 Aging Wiki

inhba

Properties1
aliasesINHBA, inhibin beta A, activin A, activin beta-A chain, erythroid differentiation protein

17 min read

INHBA / Activin A

Activin A is a homodimeric TGF-β superfamily ligand — two inhibin βA chains (encoded by INHBA) joined by disulfide bonds — that acts as a potent suppressor of adipogenesis, a driver of skeletal muscle catabolism, and an emerging central node of the senescence-associated secretory phenotype (SASP). In aging, senescent fat progenitors and other cell types accumulate and secrete elevated activin A, which acts in a paracrine fashion to inhibit local adipogenesis and impair metabolic function 1. Circulating activin A rises with age and disease burden, and a 7-factor SASP panel including activin A predicts adverse surgical and mortality outcomes better than chronological age alone 2. Clinically, the activin/ActRII signaling axis is targeted by sotatercept (FDA-approved March 2024 for pulmonary arterial hypertension) and garetosmab (anti-activin A; Phase 3 in fibrodysplasia ossificans progressiva), establishing druggability tier 1 for aging-context use.

Identity

  • UniProt: P08476 (INHBA_HUMAN), Swiss-Prot manually reviewed entry
  • Gene symbol: INHBA (inhibin subunit beta A)
  • NCBI Gene: 3624
  • HGNC: 6066 (symbol INHBA); chromosomal location 7p14.1
  • Ensembl: ENSG00000122641 (MANE-Select transcript ENST00000242208.5)
  • Mouse ortholog: Inhba (NCBI Gene 16323; MGI:96570); one-to-one ortholog on mouse chromosome 13
  • GenAge: not listed — activin A is not a curated GenAge human aging gene; aging relevance is SASP and disease-context driven
  • Precursor length: 426 amino acids (canonical)
  • Mature chain: residues 311–426 (116 aa); N-terminal signal peptide (aa 1–20) and propeptide/LAP (aa 21–310) removed by furin-site cleavage

Dimer biology and ligand naming

INHBA encodes a single gene product that participates in three distinct dimeric ligands with opposite or distinct biological actions:

DimerCompositionHistorical namePrimary effect
βA:βAINHBA + INHBAActivin AActivates ActRII/ALK4 → SMAD2/3; pro-atrophic in muscle; inhibits adipogenesis; FSH-stimulating
βA:βBINHBA + INHBBActivin ABBroadly similar to activin A; less studied in muscle/fat aging contexts
αβAINHA + INHBAInhibin ABinds betaglycan (TGFBR3) + ActRII but does NOT activate SMAD2/3; functionally antagonizes activin A for FSH regulation

Critical distinction: Inhibin A (formed with the alpha subunit INHA) and activin A (the INHBA homodimer) have opposite effects on pituitary FSH secretion and divergent actions in peripheral tissues. The gene INHBA encodes the shared βA subunit, but the activin A page concerns the βA homodimer only. Follistatin (FST) and FSTL3 are the primary extracellular antagonists of activin A (see § Antagonists below).

Protein structure and processing

Activin A is a member of the TGF-β superfamily, sharing the canonical cystine-knot architecture with myostatin (GDF-8), gdf11 (BMP-11), TGF-β1/2/3, and BMPs. Post-translational processing:

  1. Signal peptide cleavage (aa 1–20) — directs into the secretory pathway
  2. Furin-site cleavage — separates the N-terminal latency-associated propeptide (LAP/prodomain) from the C-terminal mature domain
  3. Disulfide-linked homodimerization — the two mature βA chains (residues 311–426) form the active dimer via five conserved disulfide bonds; the cystine-knot motif is essential for receptor binding
  4. N-linked glycosylation — stabilizes secreted form

Unlike myostatin, activin A’s prodomain does not form a strongly latent complex; activin A is typically secreted in a more readily available active form than myostatin.

Receptor system and signaling

Activin A signals through the canonical TGF-β/activin arm of the smad2-smad3 pathway:

  1. Activin A dimer binds type-II receptors ACVR2A (ActRIIA) or ACVR2B (ActRIIB) — ACVR2B has higher affinity in skeletal muscle
  2. Ligand-bound type-II receptor recruits and transphosphorylates type-I receptor ACVR1B (ALK4; predominant) or ALK5/ALK7 in some contexts
  3. Active type-I receptor phosphorylates SMAD2 and SMAD3 at the C-terminal SXS motif
  4. pSMAD2/3 associates with SMAD4, translocates to nucleus, and activates target gene programs:
    • In muscle: atrophy genes (atrogin-1/FBXO32, UBR2) via p38β MAPK; autophagy markers (LC3-II); MuRF1/TRIM63 upregulated by activin A via a p38 MAPK-independent mechanism 3; suppresses mTORC1-driven protein synthesis
    • In adipocyte progenitors: represses adipogenic transcription factors (PPARγ, C/EBPα), blocking differentiation
    • In inflammatory contexts: amplifies cytokine production via crosstalk with NF-κB
  5. SMAD7 is transcriptionally induced as negative feedback, attenuating the signal

This receptor-effector system is shared with myostatin and gdf11, creating competitive signaling. The receptor overlap is why pan-ActRII blockers (bimagrumab, sotatercept) simultaneously block multiple TGF-β superfamily ligands.

Antagonists (extracellular neutralization)

AntagonistGeneMechanismNotes
FollistatinFSTHigh-affinity binding; blocks receptor engagementThree isoforms (FST-288/303/315); FST-315 is main circulating form; also antagonizes myostatin and GDF11
Follistatin-like 3FSTL3Related to FST; partially redundantSerum FSTL3 rises with fat mass/inflammation
BetaglycanTGFBR3Presents inhibin A to ActRII → competitive inhibition of activin ATissue-specific; dominant in gonads/pituitary

aav-follistatin gene therapy overexpresses FST as a broad-spectrum TGF-β/activin/myostatin antagonist — its muscle-hypertrophic effect exceeds myostatin knockout alone, partly attributable to simultaneous activin A neutralization.

Role in aging

1. Activin A as a SASP factor — senescent fat progenitors

The most directly aging-relevant mechanistic finding: senescent fat progenitor cells accumulate in white adipose tissue with age and secrete activin A as a key SASP component. Activin A then acts in a paracrine fashion to inhibit adipogenesis in neighboring non-senescent progenitors 1.

Key findings from Xu et al. 2015 (eLife; multiple cohorts — N=6 human fat progenitor donors, N=8/group 18-month INK-ATTAC mice, N=8-15/group 22-month ruxolitinib experiments; INK-ATTAC transgenic model):

  • Conditioned medium from senescent human fat progenitors inhibited adipogenesis; anti-activin A antibody partially reversed this inhibition
  • Ruxolitinib (JAK1/2 inhibitor) reduced circulating activin A in 22-month-old mice, preserved fat mass, reduced lipotoxicity, and improved insulin sensitivity
  • Fat mass loss in aged animals is thus attributable partly to SASP-activin A acting as a paracrine anti-adipogenic signal

Mechanistic framing: This positions activin A as a link between the cellular-senescence hallmark and metabolic aging (loss of functional fat depot → ectopic lipid deposition → insulin resistance). gap/needs-human-replication — The paracrine adipogenesis-inhibition mechanism is demonstrated in primary human cells in vitro but has not been confirmed in human aging in vivo.

2. SASP panel and aging/clinical risk

Activin A is one of seven SASP proteins in a validated panel that predicted postsurgical adverse outcomes, hospitalization, and mortality in human cohorts, outperforming chronological age as a predictor 2. Activin A concentrations rose with chronological age (cross-sectional, ages 20–90) and were elevated in frail older adults. This provides population-level evidence that circulating activin A tracks biological aging burden in humans. gap/needs-replication — the panel was validated in surgical cohorts; replication in general community cohorts is needed.

3. Muscle catabolism and cachexia

Activin A is one of the most potent negative regulators of skeletal muscle mass in the TGF-β superfamily 4. Elevating circulating activin A in mice by AAV-mediated overexpression produced dose-dependent weight loss and muscle wasting resembling cachexia; crucially, the muscle wasting was fully reversible upon discontinuation of the activin A signal 4.

At the mechanistic level, activin A induces muscle atrophy via p38β MAPK → C/EBPβ → upregulation of atrogin-1 (FBXO32/MAFbx) and UBR2 ubiquitin ligases, with a rapid onset (detectable within 1 hour in C2C12 myotubes). MuRF1/TRIM63 is also upregulated by activin A, but via a p38 MAPK-independent pathway; MuRF1 did not appear responsible for activin A-induced myosin heavy chain loss and myotube atrophy in this study. Genetic deletion of p38β in muscle conferred resistance to activin A-induced atrophy 3.

Convergence with myostatin: Both activin A and myostatin use ACVR2B/ALK4/SMAD2/3 to drive muscle atrophy gene programs. In aged muscle, follistatin (FST) falls while myostatin and activin A rise, shifting net signaling toward catabolism. This convergence is why ACVR2B-blocking antibodies (bimagrumab) that block both ligands show stronger muscle-mass responses than myostatin-selective approaches.

Evidence in COPD: Human patients with COPD-associated skeletal muscle wasting showed significantly elevated serum activin A versus controls, with activin A levels inversely correlated with fat-free mass index 5. This provides cross-disease human evidence that activin A elevation accompanies muscle wasting in a condition that shares molecular features with aging-related cachexia. gap/needs-replication (single study, n=138; COPD cohort specifically)

DimensionStatusNotes
Pathway conserved in humans?yesACVR2B/ALK4/SMAD2/3 axis identical; activin A structure highly conserved
Phenotype conserved in humans?partialAAV-overexpression cachexia is mouse model; human equivalent is disease states (cancer, COPD cachexia) not clean physiological aging
Replicated in humans?in-progressgaretosmab FOP trials; bimagrumab obesity+metabolic trials; no aging-cachexia primary trial

4. Fibrosis and tissue remodeling

Activin A promotes fibroblast activation and collagen deposition via SMAD2/3 — a mechanism shared with TGF-β1 but potentially acting in contexts (adipose tissue, liver) where canonical TGF-β1 signaling is less dominant. Elevated activin A in aged adipose and muscle contributes to age-related tissue fibrosis, which itself impairs stem cell niche function and tissue regeneration. gap/no-mechanism — the relative contribution of activin A versus TGF-β1/3 to age-specific fibrosis patterns is not established.

5. Liver regeneration aging model

In aged rats, activin A expression increases and upregulates p15^INK4b, a CDK inhibitor, in hepatocytes. This inhibited proliferation of adult hepatocytes while young fetal liver progenitors remained unresponsive to activin A — suggesting that activin A may set a regeneration ceiling that rises with age 6. This is a model-organism result without clear human translation. gap/needs-human-replication

Activin A rises with age — population evidence

The Schafer 2020 JCI Insight cohort study (community-based n=267, ages 20–90; surgical cohorts n=97 aortic stenosis + n=36 ovarian cancer; total n=343 across all groups) found rising circulating activin A with age cross-sectionally (Spearman r=0.67, q<0.001) 2. A 2026 systematic review and meta-analysis (26 studies, n=1,345 sarcopenic adults) found no significant difference in circulating activin A between sarcopenic and non-sarcopenic older adults (standardized mean difference not significant), suggesting that while activin A tracks global aging burden, it may not selectively discriminate sarcopenia from non-sarcopenic aging 7. The biomarkers that did discriminate were reduced IGF-1 (sarcopenic vs non-sarcopenic: SMD −0.40, 95% CI −0.54 to −0.27; p<0.01) and elevated GDF-15 (SMD +0.26, 95% CI +0.03 to +0.50; p=0.03).

This apparent discrepancy likely reflects: (1) sarcopenic individuals are already aged (both groups have elevated activin A vs. young baseline), and (2) activin A rises broadly with aging/senescence burden rather than specifically with muscle-loss phenotype.

Pharmacology and therapeutic targeting

Aging-context druggability-tier rationale

Tier 1 is assigned because:

  1. Sotatercept (Merck/Acceleron; ACE-011 — a fusion protein of ActRIIA extracellular domain with IgG1 Fc) was FDA-approved March 2024 (brand: Winrevair) for pulmonary arterial hypertension (PAH). In the STELLAR Phase 3 RCT (n=323; 1:1 sotatercept vs. placebo), sotatercept produced a 34.4 m improvement in 6-minute walk distance vs. 1.0 m for placebo (difference 40.8 m; p<0.001); 8 of 9 secondary endpoints favored sotatercept 8. Mechanism: sotatercept acts as a decoy receptor for activin A and related ligands (including activin B, GDF11) via ActRIIA, restoring the balance between pro-proliferative BMP signaling and anti-proliferative activin/TGF-β signaling in the pulmonary vasculature.

  2. Garetosmab (REGN2477; Regeneron) — a monoclonal antibody that specifically targets activin A (not the receptor). Phase 3 trial in fibrodysplasia ossificans progressiva (FOP) — OPTIMA (NCT05394116) — is active. Phase 2 (Di Rocco et al. 2023, Nature Medicine; n=44) did not meet its primary endpoint for lesion activity (P=0.0741) but prevented new heterotopic ossification (0% vs 40.9% in crossover group; P=0.0027) 9. Garetosmab is also in Phase 2 (COURAGE trial, NCT06299098) in obesity in combination with trevogrumab (anti-myostatin) and semaglutide.

The aging-indication gap: Neither sotatercept nor garetosmab has a licensed or primary-development aging/sarcopenia indication as of 2026. Sotatercept’s PAH indication is cardiovascular, not aging per se, but establishes proof of pathway druggability at the clinical level. The tier-1 designation reflects this depth of clinical pharmacological tractability, consistent with the wiki’s aging-context convention (see mtor tier-1 precedent via rapamycin).

Bimagrumab — ActRII dual blockade

Bimagrumab (BYM338; Novartis) is an anti-ActRIIA/IIB antibody that simultaneously blocks both type-II activin receptors, neutralizing activin A, activin B, myostatin, and GDF11 signaling to muscle. Single-receptor blockade is insufficient — simultaneous ActRIIA + ActRIIB blockade is required for maximal muscle hypertrophy 10. Bimagrumab demonstrated lean mass gains and fat mass loss in Phase 2 trials in obesity/type 2 diabetes. In sarcopenia specifically, the Phase 2 trial showed lean mass increase without significant functional improvement (see myostatin for trial-failure pattern). Status: stalled in sarcopenia; the metabolic/obesity pivot is active.

Downstream implications for follistatin gene therapy

aav-follistatin sequesters activin A (alongside myostatin and GDF11) extracellularly. The follistatin-overexpression phenotype in mice exceeds myostatin knockout alone, with the excess plausibly attributable to activin A neutralization. This makes activin A a likely co-target of AAV-follistatin even though the therapeutic rationale is primarily framed around myostatin.

Mendelian randomization and causal evidence

mr-causal-evidence: not-tested — No published Mendelian randomization study using germline instruments for INHBA has established causal relationships for activin A in aging-relevant outcomes. A 2024 study (Wang 2024, Frontiers in Endocrinology) found no causal relationship between activin A levels and diabetes by two-sample MR 11. No MR study has tested activin A causally in sarcopenia, lifespan, or adipose aging phenotypes. The receptor gene ACVR2A has been implicated in a GWAS/colocalization study of dizziness (2026), but this is not directly relevant to the muscle/fat/senescence axis here.

gap/needs-replication — Germline genetic instruments for circulating activin A are available in principle (cis-pQTLs from UK Biobank) but no aging-endpoint MR study published as of 2026-06-25.

Pathway connections

  • tgf-beta — activin A is a TGF-β superfamily member; signals through the same SMAD2/3 arm as TGF-β1/2/3 but via distinct type-II receptors (ACVR2A/B rather than TβRII)
  • smad2-smad3 — direct downstream effectors; phospho-SMAD2/3 drive atrophy gene programs in muscle and anti-adipogenic programs in fat progenitors
  • bmp-signaling — activin A opposes BMP/SMAD1/5/8 signaling; in the pulmonary vasculature, restoring BMP/activin balance (via sotatercept) is the therapeutic mechanism in PAH
  • sasp — activin A is a SASP component; elevated in conditioned medium from senescent human fat progenitors (Xu 2015); member of the Schafer 2020 7-factor SASP aging panel

Cross-references

  • myostatin — closest TGF-β superfamily paralog for muscle biology; shared ACVR2B receptor; do not conflate
  • gdf11 — related family member; shares ActRII signaling; now considered distinct from activin A aging biology despite early conflation
  • smad2-smad3 — canonical intracellular transducers
  • tgf-beta — parent pathway
  • bmp-signaling — sibling pathway; activin A/BMP balance is clinically relevant (sotatercept PAH mechanism)
  • aav-follistatin — gene therapy that neutralizes activin A (alongside myostatin) as co-target
  • sasp — activin A is a core SASP constituent in fat progenitors and possibly other senescent cell types
  • cellular-senescence — senescent cells are the upstream source of paracrine activin A in aged adipose tissue
  • sarcopenia — downstream phenotypic consequence of chronic activin A elevation in muscle
  • xu-2015-senescent-cells-adipogenesis — primary source for fat progenitor/SASP/activin A/adipogenesis link (seeded in parallel; may be stub)

Limitations and open gaps

  • gap/needs-human-replication — Paracrine adipogenesis inhibition via senescent cell-secreted activin A is demonstrated in primary human cells in vitro and mouse in vivo models; not confirmed by human intervention studies.
  • gap/needs-human-replication — Liver regeneration aging model (Menthena 2011) is rat; human hepatocyte aging and activin A have not been systematically studied.
  • gap/dose-response-unclear — Circulating activin A rises broadly with aging but the quantitative trajectory across the human lifespan in healthy individuals is not well-characterized by longitudinal cohort data; cross-sectional studies (Schafer 2020) are available but single time-point.
  • gap/needs-replication — COPD-cachexia human data (Zhou 2019, single study n=138) requires replication; causal directionality (activin A as driver vs. marker of wasting) is not established.
  • gap/no-mechanism — The relative contribution of activin A versus TGF-β1/3 versus myostatin to age-specific fibrosis in adipose and muscle is not formally resolved.
  • gap/contradictory-evidence — The 2026 meta-analysis (Prokopidis 2026) found no significant difference in circulating activin A between sarcopenic and non-sarcopenic elders, conflicting with the mechanistic narrative of activin A as a sarcopenia driver. This may reflect assay heterogeneity, the fact that both groups are aged (and thus both have elevated activin A vs. young), or that muscle catabolism is mediated locally rather than by circulating concentrations.
  • mr-causal-evidence: not-tested — No published MR study has established causal relationships for activin A levels in aging phenotypes.
  • gtex-aging-correlation: null — tissue-by-age GTEx correlation not yet populated; gap/needs-tissue-expression-data

Footnotes

Footnotes

  1. xu-2015-senescent-cells-adipogenesis · doi:10.7554/eLife.12997 · PMID 26687007 · PMC4758946 · Xu M, Palmer AK, Ding H, Weivoda MM, et al. (Kirkland JL lab) · eLife 2015 · multiple cohorts: N=6 human fat progenitor donors; N=8/group 18-month INK-ATTAC clearance experiment; N=8–15/group 22-month ruxolitinib in vivo experiments · in-vivo + in-vitro · model: aged C57BL/6 mice + INK-ATTAC transgenic mice + human primary fat progenitors · senescent fat progenitors secrete activin A; anti-activin A antibody partially restored adipogenesis in conditioned-medium experiments; ruxolitinib (JAK1/2 inhibitor) reduced circulating activin A, preserved fat mass, reduced lipotoxicity, improved insulin sensitivity in 22-month-old mice; gold OA (PMC open access) ↩ ↩2

  2. doi:10.1172/jci.insight.133668 · PMID 32554926 · Schafer MJ, Zhang X, Kumar A, et al. · JCI Insight 2020 · cross-sectional (n=267, ages 20–90) + prospective surgical cohort (n=97 aortic stenosis; n=36 ovarian cancer; n=343 total across all groups) · observational · model: human population + surgical cohort · 7-factor SASP panel (GDF15, FAS, OPN, TNFR1, ACTIVIN A, CCL3, IL-15) predicted adverse postsurgical events better than age alone; activin A age-correlation r=0.67, q<0.001; activin A concentrations correlated with frailty index in nonsurgical and ovarian cancer groups; gold OA ↩ ↩2 ↩3

  3. doi:10.1002/jcsm.12145 · PMID 27897407 · PMC5377410 · Ding H, Zhang G, Sin KW, Liu Z, Lin RK, Li M, Li YP · J Cachexia Sarcopenia Muscle 2017;8(2):202–212 · in-vitro (C2C12 myotubes) + in-vivo (mouse) · model: mouse (p38β knockout + WT) · activin A activates p38β MAPK → C/EBPβ → upregulation of atrogin-1 (MAFbx/FBXO32) and UBR2 within 1 hour; MuRF1 also upregulated by activin A but via p38 MAPK-independent mechanism and not responsible for myosin heavy chain loss; p38β MAPK inhibition (SB202190) and muscle-specific p38β knockout conferred resistance to activin A-induced atrophy; gold OA (PMC open access) ↩ ↩2

  4. doi:10.1096/fj.13-245894 · Chen JL, Walton KL, Winbanks CE, Murphy KT, et al. · FASEB Journal 2014 (published 2013 online) · in-vivo (mouse, AAV-mediated activin A overexpression) · model: C57BL/6 mice · activin A characterized as most potent negative regulator of muscle mass among TGF-β proteins tested; dose-dependent weight loss −12.4% vs +10% controls; muscle wasting + fibrosis fully reversible; mechanism involves ActRIIB pathway; 200 citations (OpenAlex); closed-access gap/no-fulltext-access ↩ ↩2

  5. doi:10.6061/clinics/2019/e981 · PMID 31271588 · Zhou X, et al. · Clinics (São Paulo) 2019 · case-control · n=78 COPD + 60 controls · model: human · elevated serum activin A in COPD with skeletal muscle wasting; activin A inversely correlated with fat-free mass index and BMI; TNF-α co-elevated; single study requiring replication gap/needs-replication ↩

  6. doi:10.1053/j.gastro.2011.01.053 · Menthena A, Koehler CI, Sandhu JS, et al. · Gastroenterology 2011 · fetal liver cell transplantation in rats of varying ages · in-vivo (rat) · activin A rises in aging rat liver; upregulates p15^INK4b inhibiting adult hepatocyte proliferation; young stem cells unresponsive to activin A; suggests activin A sets regeneration ceiling in aging liver gap/needs-human-replication ↩

  7. doi:10.1093/gerona/glag140 · PMID 42202210 · Prokopidis K, Deane CS, Baoubbou Z, Beaudart C · J Gerontol A 2026 · systematic review and meta-analysis · n=1,345 sarcopenic adults, 48.3% female, mean age 67.9–88.1 years (26 observational studies) · no significant difference in circulating activin A between sarcopenic and non-sarcopenic older adults; sarcopenic adults showed reduced IGF-1 (SMD −0.40, 95% CI −0.54 to −0.27, p<0.01) and elevated GDF-15 (SMD +0.26, 95% CI +0.03 to +0.50, p=0.03); contradicts simple “elevated circulating activin A → sarcopenia” narrative for circulating protein gap/contradictory-evidence ↩

  8. doi:10.1056/NEJMoa2213558 · PMID 36877098 · Hoeper MM, Badesch DB, Ghofrani HA, et al. · N Engl J Med 2023;388(16):1478–1490 · Phase 3 RCT (STELLAR; NCT04576988) · n=323 (163 sotatercept + 160 placebo) · model: patients with pulmonary arterial hypertension (WHO functional class II or III) on stable background therapy · primary endpoint: 6-minute walk distance change at 24 weeks: +34.4 m (95% CI 33.0–35.5) sotatercept vs. +1.0 m (95% CI −0.3 to 3.5) placebo (Hodges-Lehmann difference 40.8 m; 95% CI 27.5–54.1; p<0.001); first 8 of 9 secondary endpoints significantly improved with sotatercept; PAH-SYMPACT Cognitive/Emotional Impacts domain was not significantly improved; FDA-approved March 2024 as Winrevair ↩

  9. doi:10.1038/s41591-023-02561-8 · Di Rocco M, Baujat G, Boulanger C, et al. · Nature Medicine 2023;29(10):2615–2624 · Phase 2 RCT · n=44 (fibrodysplasia ossificans progressiva) · model: human · garetosmab (anti-activin A monoclonal antibody) did not meet primary endpoint (lesion activity, P=0.0741) but prevented new heterotopic ossification in Period 2 crossover (0% vs 40.9%, P=0.0027); notable AEs: epistaxis, madarosis, skin abscesses; 5 deaths in open-label phase ↩

  10. doi:10.1073/pnas.1707811114 · Morvan F, Rondeau JM, Zou C, et al. · PNAS 2017 · in-vitro + in-vivo (mouse) · simultaneous ActRIIA + ActRIIB blockade required for maximal muscle hypertrophy; single-receptor blockade produces only partial response; provides mechanistic rationale for bimagrumab’s dual-receptor design ↩

  11. doi:10.3389/fendo.2024.1414585 · PMID 39280004 · Wang M · Front Endocrinol 2024 · two-sample Mendelian randomization · model: human genetics (GWAS instruments) · no causal relationship between diabetes and plasma activin A by IVW, Egger, weighted median, and weighted mode methods; limited to diabetes endpoint — no aging/sarcopenia/lifespan outcomes tested ↩

Graph View

Backlinks