Relaxin

A two-chain peptide hormone of the insulin superfamily that signals through the RXFP1 receptor to produce vasodilation and antifibrotic effects, and whose recombinant form, serelaxin, failed the RELAX-AHF-2 phase 3 trial in acute heart failure.

Also referenced as: Relaxin-2, Relaxin H2, H2 relaxin, Human relaxin-2, Recombinant human relaxin-2, Serelaxin, RLX030, Reasanz, RLN2

Also appears in: Hormone · Tissue repair

COAPublic product evidenceSearch the public certificate ledger for this compoundNo exact compound records are currently indexed under this profile name.→

---

What is relaxin?

Relaxin is a peptide hormone of the insulin superfamily. The form that carries essentially all of the pharmacology is relaxin-2 (also written H2 relaxin, gene RLN2): a roughly 6 kDa two-chain molecule built from a 29-residue B chain and a 24-residue A chain, joined by two interchain disulfide bonds with a third disulfide inside the A chain — the same architecture as insulin, not the single linear chain of a typical research peptide. It is secreted mainly by the corpus luteum, and additionally by placenta and decidua in pregnancy. It was identified in 1926 as an ovarian factor that relaxed the pubic ligament of the guinea pig, which is where the name comes from, and the reproductive biology is now the least interesting part of its pharmacology.

It reaches the research-compound market on the strength of two literatures: a large preclinical antifibrotic record across heart, kidney, lung, and liver, and a 2013 acute heart failure trial that reported a 37% reduction in 180-day mortality. That mortality figure is still widely quoted. The confirmatory trial that was built to test it failed.

Because relaxin-2 is a recombinant two-chain protein rather than a synthesized single-chain peptide, most catalog material for it is life-science reagent stock rather than research-peptide stock, and the two are packaged and characterized very differently.

How it works

Relaxin-2 acts on RXFP1 (relaxin family peptide receptor 1, formerly LGR7), a leucine-rich-repeat G-protein-coupled receptor. Serelaxin is recombinant relaxin-2 and shares this mechanism exactly; the newer clinical agents are engineered RXFP1 agonists rather than the hormone itself.

• RXFP1 is the receptor. The long-elusive relaxin receptor was identified as the orphan GPCRs LGR7 and LGR8, signaling through a cAMP-dependent pathway distinct from the insulin/IGF receptors despite the ligand’s insulin-like structure (Hsu et al., Science, 2002;295(5555):671–674). RXFP1 activates a broad spectrum of downstream pathways including cAMP and nitric oxide, while the related receptors RXFP2, RXFP3, and RXFP4 serve different ligands and different biology (Bathgate et al., Physiological Reviews, 2013;93(1):405–480).
• Vasodilation runs through nitric oxide, in two phases. A rapid phase couples Gα(i/o) to PI3-kinase/Akt-dependent phosphorylation and activation of endothelial nitric oxide synthase; a sustained phase works through vascular endothelial and placental growth factors and increased arterial gelatinase activity, which cleaves big endothelin to ET(1–32) and activates the endothelial ET-B receptor/nitric oxide pathway (Conrad, American Journal of Physiology-Regulatory, Integrative and Comparative Physiology, 2011;301(2):R267–R275).
• The renal effect is measurable in humans. Eleven healthy volunteers given a 5-hour intravenous infusion of recombinant human relaxin at concentrations comparable to early pregnancy showed a 47% increase in renal plasma flow regardless of sex (P < 0.0001), with no significant change in glomerular filtration rate (Smith et al., Journal of the American Society of Nephrology, 2006;17(11):3192–3197). This is the cleanest human pharmacodynamic result in the whole program.
• Antifibrotic action is TGF-β1/Smad2 interference, not general growth suppression. Relaxin requires RXFP1 and an nNOS–NO–cGMP pathway to inhibit Smad2 phosphorylation, myofibroblast differentiation, and collagen production in injured-kidney fibroblast cultures (Mookerjee et al., FASEB Journal, 2009;23(4):1219–1229). Across organs it selectively opposes pro-fibrotic cytokine-driven fibroblast activity rather than unstimulated fibroblast activity, and also augments matrix degradation by upregulating matrix metalloproteinases and downregulating their tissue inhibitors (Samuel et al., British Journal of Pharmacology, 2017;174(10):962–976).

Research status

Relaxin has one of the most complete — and most instructive — clinical records of any compound on the research market: a promising phase 2, a phase 3 that split its primary endpoint, a widely publicized mortality signal, rejection by regulators on both sides of the Atlantic, and a large confirmatory trial that flatly failed.

Phase 2b dose-finding. Pre-RELAX-AHF randomized 234 acute heart failure patients with systolic blood pressure above 125 mm Hg to 48-hour intravenous infusions of placebo or relaxin at 10, 30, 100, or 250 µg/kg/day. The 30 µg/kg/day dose improved dyspnea by Likert scale (40% versus 23% moderately or markedly improved; p = 0.044) and was carried forward (Teerlink et al., Lancet, 2009;373(9673):1429–1439).

Phase 3, split result. RELAX-AHF (NCT00520806) randomized 1,161 patients to a 48-hour infusion of serelaxin at 30 µg/kg/day or placebo. It had two co-primary dyspnea endpoints and met only one: the visual analogue scale area under the curve improved (448 mm × h, 95% CI 120–775; p = 0.007), while the Likert co-primary did not (27% versus 26%; p = 0.70). There was no effect on cardiovascular death or rehospitalization for heart or renal failure (HR 1.02, 95% CI 0.74–1.41; p = 0.89). A prespecified additional endpoint produced the number everyone remembers: 42 deaths at day 180 versus 65 on placebo (HR 0.63, 95% CI 0.42–0.93; p = 0.019) (Teerlink et al., Lancet, 2013;381(9860):29–39).

Both regulators said no. The FDA’s Cardiovascular and Renal Drugs Advisory Committee met on 27 March 2014 to review serelaxin injection — filed by Novartis as a biologics license application, BLA 125468, not as a small-molecule drug application — and voted against approval; the agency never licensed it. In Europe, the CHMP adopted a negative opinion on Reasanz (serelaxin) on 23 January 2014, confirmed the refusal on re-examination on 22 May 2014, and the European Commission made the refusal final on 5 August 2014, on the grounds that a single main study could not establish effectiveness and that differences in background treatment between the study groups clouded the analysis.

The confirmatory phase 3 failed. RELAX-AHF-2 (NCT01870778) repeated the identical regimen — a 48-hour infusion of 30 µg/kg/day started within 16 hours of presentation — in 6,545 patients analyzed by intention to treat, with two primary endpoints. Cardiovascular death at 180 days occurred in 8.7% on serelaxin versus 8.9% on placebo (HR 0.98, 95% CI 0.83–1.15; P = 0.77). Worsening heart failure at day 5 occurred in 6.9% versus 7.7% (HR 0.89, 95% CI 0.75–1.07; P = 0.19). There were no significant differences in all-cause death at 180 days, in the death-or-rehospitalization composite, or in length of stay, and adverse event rates were similar (Metra et al., New England Journal of Medicine, 2019;381(8):716–726). This is the decisive result for serelaxin in acute heart failure.

The meta-analysis is weaker evidence than the failure it followed. Pooling six randomized trials — 6,105 patients on serelaxin against 5,254 controls — found a reduction in 5-day worsening heart failure (6.0% versus 8.1%; HR 0.77, 95% CI 0.67–0.89; P = 0.0002) and in all-cause mortality through last follow-up (HR 0.87, 95% CI 0.77–0.98; P = 0.0261), but again no effect on cardiovascular death or rehospitalization (Teerlink et al., European Journal of Heart Failure, 2020;22(2):315–329). It includes the hypothesis-generating trials alongside the confirmatory one, so it does not overturn RELAX-AHF-2.

Scleroderma failed earlier, with a safety signal. A randomized, double-blind, placebo-controlled trial gave recombinant human relaxin at 10 or 25 µg/kg/day by continuous subcutaneous infusion for 24 weeks in diffuse cutaneous systemic sclerosis. The primary endpoint, modified Rodnan skin thickness score, was no different from placebo at any timepoint, functional disability was unchanged, and forced vital capacity decreased significantly in the relaxin groups. Stopping the infusion at week 24 was followed by statistically significant declines in creatinine clearance and serious renal adverse events in seven relaxin-treated patients and none on placebo; the authors wrote that any other therapeutic use of relaxin requires close monitoring of blood pressure and renal function (Khanna et al., Arthritis & Rheumatism, 2009;60(4):1102–1111).

Cirrhosis: mechanism confirmed, outcomes untested. A randomized, open-label, parallel-group phase 2 study assigned 40 patients with alcohol-related cirrhosis and portal hypertension 1:1 to serelaxin or to terlipressin as an active comparator (NCT01640964). Across a 120-minute infusion — 60 minutes at 80 µg/kg/day, then 60 minutes at 30 µg/kg/day — serelaxin increased total renal artery blood flow by 65% from baseline (95% CI 40–95%; p < 0.001) without harming systemic blood pressure or hepatic perfusion. It was a hemodynamic study in a small, stable, well-compensated population, not an efficacy trial (Snowdon et al., PLoS Medicine, 2017;14(2):e1002248).

Newer RXFP1 agonists are not doing better. Volenrelaxin (LY3540378), a long-acting form of human relaxin dosed 25, 50, or 100 mg subcutaneously once weekly, was tested in worsening chronic HFpEF (NCT05592275). The sponsor stopped the trial early after 332 participants because of evidence for worsening congestion. Only the 25 mg dose improved left atrial reservoir strain (+3.9%, 95% CI 1.1–6.6, P = 0.006), with no effect at 50 or 100 mg; pooled across doses NT-proBNP rose 24.5% (95% CI 2.0–51.8), eGFR was unchanged, heart failure hospitalization was numerically higher (HR 2.64, 95% CI 0.93–7.56, P = 0.070), and cardiovascular and renal serious adverse events trended up (OR 2.52, 95% CI 0.95–6.68, P = 0.056) (Borlaug et al., Nature Medicine, 2025;31(11):3853–3861). Separately, AZD3427 — a fusion of a single relaxin-2 with the Fc fragment of human IgG1, with a 13- to 14-day terminal half-life in heart failure patients (Connolly et al., Journal of the American Heart Association, 2024;13(15):e034067) and non-human primate work showing improved ejection fraction, cardiac output, and stroke volume over 21 weeks of dosing (Papworth et al., Cardiovascular Research, 2025;121(6):871–881) — completed a 260-patient phase 2 trial in heart failure with WHO group 2 pulmonary hypertension in August 2025 (NCT05737940), with no results posted at the time of writing. Its phase 1b renal-perfusion study (NCT06611423) was terminated after 10 participants, but the sponsor recorded the reason as recruitment difficulty and business considerations explicitly not affecting the benefit–risk balance — not a safety finding. A separate relaxin agonist, R2R01, is in phase 2 for acute kidney injury indications, including hepatorenal syndrome and cardiac-surgery-associated acute kidney injury.

No regulatory authority has approved relaxin, serelaxin, or any relaxin receptor agonist for any indication.

Common dosage forms

• Life-science reagent vials. Recombinant human relaxin-2 is most commonly sold as a lyophilized protein in microgram quantities — 5 µg, 20 µg, and similar — labeled research use only, supplied with reconstitution buffer and cold-chain storage instructions rather than clinical directions, and intended for cell culture and receptor assays.
• Milligram-scale vials. Some research-peptide sellers list relaxin-2 in milligram quantities for reconstitution, a presentation that follows research-peptide packaging convention rather than anything in the clinical record.
• Fragment products that are not the hormone. Catalogs also list “relaxin-2 C-peptide” sequences such as residues 56–129 and 132–157. These fall inside the connecting peptide — residues 56–157 of the 185-residue preprohormone — which is excised and discarded when the prohormone is processed. They are not the active two-chain molecule and do not bind RXFP1.
• The clinical formats have never been consumer formats. Every serelaxin efficacy trial used a continuous 48-hour intravenous infusion at 30 µg/kg/day in a monitored hospital setting. The scleroderma program used a continuous subcutaneous infusion pump for 24 weeks. The current once-weekly subcutaneous agents are engineered Fc-fusion analogs, not relaxin-2.

This section describes formats only and is not dosing guidance.

Key considerations

• The phase 3 failed cleanly, at scale, and on both endpoints. RELAX-AHF-2 ran the same drug, dose, route, duration, and patient population as the trial that generated the excitement, in more than five times as many patients, and found nothing on either primary endpoint. Any material citing a “37% mortality reduction” is quoting a prespecified additional endpoint from the 2013 trial, which the confirmatory trial did not reproduce. The compound is not an open question at the stage it was tested; it is a negative result.
• Neither major regulator approved it. The FDA’s advisory committee voted against serelaxin in March 2014 and the agency never licensed it; the EMA’s CHMP refused marketing authorisation for Reasanz, confirmed that refusal on re-examination, and the European Commission made it final in August 2014 — the European decision citing reliance on a single main study and uncertainty about benefit. Nothing approved anywhere contains relaxin.
• The chronic-dosing safety signals point the wrong way. The two programs that dosed relaxin or a relaxin agonist for months on end both produced adverse findings: significant falls in creatinine clearance and serious renal adverse events after stopping subcutaneous infusion in scleroderma, along with a significant decline in forced vital capacity on drug; and worsening congestion, a ~24.5% rise in NT-proBNP, and numerically more heart failure hospitalizations and cardiovascular/renal serious adverse events with weekly volenrelaxin, which was stopped early for that reason. Repeated dosing is the regimen least supported by the favorable data and most associated with harm signals.
• Route, duration, and molecule all mismatch the research market. The human efficacy data come from short intravenous infusions; serelaxin’s terminal half-life is roughly 6.5–8.8 hours in renal-impairment pharmacokinetics and 7–8 hours in hepatic-impairment pharmacokinetics (Dahlke et al., Journal of Clinical Pharmacology, 2016;56(4):474–483; Kobalava et al., British Journal of Clinical Pharmacology, 2015;79(6):937–945). The weekly subcutaneous data come from Fc-fusion analogs engineered specifically because native relaxin does not last. A subcutaneous injection of native relaxin-2 has no counterpart anywhere in the clinical record.
• Identity verification is harder here than for ordinary peptides. Relaxin-2 is a recombinant, two-chain, three-disulfide protein. A total-mass or HPLC purity figure from a routine peptide panel does not establish correct A-chain/B-chain pairing or disulfide topology — a misfolded or mispaired preparation can match on mass and fail on function — and E. coli-expressed material raises endotoxin and host-cell-protein questions that single-chain synthetic peptide certificates are not designed to answer.
• The naming is a minefield. Relaxin-2 (RLN2, H2 relaxin) is the circulating hormone and the subject of every trial above. Relaxin-1 (RLN1, H1) is a separate gene product. Relaxin-3 (INSL7) is primarily a brain neuropeptide acting at RXFP3 with stress and metabolic biology, not the vascular hormone. INSL3, sometimes called relaxin-like factor, is the RXFP2 ligand. Serelaxin, RLX030, and Reasanz are all the same recombinant relaxin-2. Volenrelaxin (LY3540378), AZD3427, and R2R01 are engineered RXFP1 agonists whose results — good or bad — are results for those molecules, not for relaxin-2.