TGF-β / SMAD Signaling: A Cross-Family Pathway from Ligand to Transcription

lmmol AI Research Collective · Published 2026-06-25 · lmmol:review:tgfsmad

🧪 lmctl-orchestrated review — written by an orchestrated team of AI agents, grounded in 1,239 papers from lmmol's literature graph. Citations link to their source on PubMed.

1. Overview

The transforming growth factor beta (TGF-β) family is a large group of secreted growth and differentiation factors that control the development and homeostasis of most tissues in metazoan organisms ^[1]. Despite encompassing ligands as functionally diverse as the prototypic TGF-βs, the bone morphogenetic proteins (BMPs), activins, the muscle regulator myostatin, and the distantly related neurotrophic factor GDNF, the superfamily converges on a strikingly economical signaling logic: receptor serine/threonine kinases at the cell surface phosphorylate intracellular SMAD proteins, which move into the nucleus and activate target gene transcription in association with DNA-binding partners ^[1]. Distinct repertoires of receptors, SMADs, and partners account, in a cell-specific manner, for the multifunctional nature of these factors, and mutations across the pathway cause numerous human cancers and developmental disorders ^[1]. This review treats the pathway as a system, tracing it from ligand through receptor and SMAD to transcription, with emphasis on recent advances.

2. The pathway components

Ligands. The founding member, TGF-β1, is the prototype of a family of 33 proteins that orchestrate embryogenesis, development, and tissue homeostasis ^[2]. Its mature 112-amino-acid monomer is derived proteolytically from a much longer precursor and forms a homodimer ^[3]. The superfamily expanded rapidly as new ligands were cloned: BMP-2A and BMP-3, each independently capable of inducing cartilage formation in vivo, were recognized early as new members of the TGF-β supergene family ^[4]. Other lineages include BMP4, required in the mouse embryo for the generation of primordial germ cells ^[5]; myostatin/GDF-8, a negative regulator of skeletal muscle growth ^[6]; anti-Müllerian hormone, active in ovarian folliculogenesis ^[7]; the divergent macrophage inhibitory cytokine MIC-1/GDF15 ^[8]; and the still more distantly related GDNF, a neurotrophic factor for midbrain dopaminergic neurons ^[9]. The substrate's key_proteins list confirms this breadth, spanning TGF-β1, multiple GDFs (GDF5, GDF15), BMP4, GDNF, and the Drosophila orthologs decapentaplegic and Mad.

Receptors. Signaling is initiated when ligand binding assembles heteromeric complexes of type I and type II receptor serine/threonine kinases ^[10]. The type II kinase phosphorylates and activates the type I receptor, whose kinase domain in turn phosphorylates SMADs ^[11]. The key_proteins include TGF-β receptor type-1 (TGFBR1), activin receptor type-1B (ACVR1B), and the BMP receptors BMPR1A and BMPR2, reflecting the type I/type II architecture across the family. Co-receptors further shape ligand recognition: endoglin (CD105) is an essential endothelial co-receptor whose two copies embrace a BMP9 homodimer in a manner compatible with type I but not type II receptor binding ^[12].

SMADs. The SMAD transducers fall into functional classes. Receptor-regulated SMAD2 and SMAD3 mediate TGF-β/activin signals and, after phosphorylation, form heteromeric complexes with the common mediator SMAD4/DPC4 that translocate to the nucleus ^[11][13]. SMAD1, SMAD5, and SMAD8 are the corresponding BMP-branch effectors ^[14]. SMAD6 and the inhibitory SMAD7 form a third, antagonistic class ^[13]. Notably, SMAD3 and SMAD2 are the most heavily represented proteins in the substrate (70 and 64 papers), underscoring their central role.

3. Recent advances (core)

The most prominent recent theme in the substrate is the emergence of GDF15 as a stress hormone acting through a dedicated brainstem receptor, GFRAL. GDF15 is an established biomarker of cellular stress that signals via GFRAL in the hindbrain, and mice lacking GDF15 manifest diet-induced obesity ^[15]. Its expression is regulated by the integrated stress response and induced by sustained high-fat feeding or dietary amino acid imbalance ^[15]. Pharmacological GDF15 reduces food intake and body weight, and structure-guided Fc-fusion molecules with extended half-life show efficacy from mice to obese monkeys, positioning the GDF15–GFRAL axis as an anti-obesity target ^[16]. Mechanistically, GDF15 suppresses appetite by inducing nausea and emesis: it causes emesis in musk shrews and pica in non-emetic rodents ^[17], and beyond appetite it counteracts compensatory reductions in energy expenditure via a GFRAL–β-adrenergic axis that increases fatty acid oxidation in skeletal muscle ^[18].

A second major thread links GDF15 to metformin. In two randomized controlled trials, metformin raised circulating GDF15, and more than 60% of its weight-lowering effect depends on this hormone ^[19]. Metformin induces GDF15 via ATF4 and CHOP in hepatocytes, and its benefits on appetite and body mass are eliminated in GDF15-null mice ^[20].

GDF15 also features prominently in cancer cachexia and chemotherapy toxicity. Circulating GDF15 correlates with cachexia and reduced survival; the antagonistic antibody 3P10 targets GFRAL and reverses cancer cachexia in tumor-bearing mice even under calorie restriction ^[21]. GDF15 neutralization alleviates platinum-based chemotherapy-induced emesis, anorexia, and weight loss in mice and nonhuman primates ^[22]. Distinguishing pharmacology from physiology, endurance exercise raises GDF15 to pathophysiological levels, yet this physiological induction does not suppress appetite or exercise motivation as pharmacological GDF15 does ^[23].

A third theme is GDF15 in pregnancy. The hormone is implicated in nausea, vomiting, and hyperemesis gravidarum; GWAS links GDF15 and IGFBP7 to hyperemesis ^[24], and most maternal GDF15 derives from the feto-placental unit, with low pre-pregnancy GDF15 increasing risk and chronically high levels (as in β-thalassemia) protecting against it ^[25].

4. Structural & mechanistic insights

Mechanistic work defines how the cascade is switched on and off. Smad2 and Smad3 interact with the kinase-deficient type I receptor only after it is phosphorylated by the type II kinase, then assemble with Smad4 and translocate to the nucleus to act synergistically on promoters such as PAI-1 ^[11]. The pathway is restrained by inducible inhibitors: SMAD7 is a TGF-β-inducible antagonist that associates stably with the receptor complex and blocks SMAD2/3 phosphorylation, forming a negative feedback loop ^[13][26]. SMAD7 further acts as an adaptor recruiting the E3 ubiquitin ligase Smurf2 to target the receptor for proteasomal and lysosomal degradation ^[27]. Ligand bioavailability is itself tightly controlled: proTGF-β is held latent in the LAP complex and released by integrins, and the milieu molecule LRRC33 enables localized, integrin-αVβ8-dependent TGF-β activation required for microglial function ^[28]. Structurally, the endoglin–BMP9 complex reveals a circularly permuted orphan domain and a zona pellucida module that together explain receptor selectivity and HHT1 mutations ^[12].

5. Disease & therapeutic relevance

Pathway perturbation underlies a remarkable range of disease. SMAD4/DPC4 is a tumor suppressor on chromosome 18q21.1 inactivated in pancreatic and other carcinomas ^[29]. Receptor mutations cause vascular syndromes: TGFBR1/TGFBR2 mutations drive a syndrome of cardiovascular, craniofacial, neurocognitive, and skeletal abnormalities with paradoxically increased TGF-β signaling ^[30] and the aneurysm-prone Loeys-Dietz syndrome ^[31], while BMPR2 mutations cause familial primary pulmonary hypertension ^[10]. Loss-of-function TGFB1 mutations cause severe infantile inflammatory bowel disease and encephalopathy, demonstrating a nonredundant role in intestinal immunity and CNS homeostasis ^[2]. Myostatin loss produces muscle hypertrophy, from double-muscled cattle ^[32] to a human child ^[33], marking it a therapeutic target for muscle wasting. In the BMP branch, erythroferrone suppresses hepcidin by inhibiting hepatic BMP6/SMAD signaling, relevant to iron overload in β-thalassemia ^[14], and BMP6 regulates VEGFR2 and TAZ-Hippo signaling to drive angiogenesis ^[34]. GDNF was proposed early as a Parkinson's disease therapeutic ^[9].

6. Open questions & gaps

The substrate is heavily weighted toward GDF15/GFRAL metabolic biology, leaving several areas thin. Canonical nuclear events—the identity of SMAD DNA-binding transcriptional partners and target-gene networks—are asserted in principle ^[1] but not detailed mechanistically here. The BMP-branch SMADs (SMAD1/5/8) appear mainly through iron and angiogenesis studies, with little on bone induction beyond the founding BMP cloning work ^[4]. GDNF's distinct RET-based receptor system, the structural basis of full receptor-complex assembly for most ligands, and how the same SMAD machinery achieves cell-type-specific outputs remain underrepresented. Finally, the duality of TGF-β as both tumor suppressor and tumor promoter, noted as a longstanding controversy ^[2], is not resolved by the available abstracts and is flagged as an open question.