What's Next After GLP-1: 6 Peptide-Receptor Systems Driving the Next Decade of Obesity Drugs

GLP-1 receptor agonists work. Wegovy delivers 15-17% mean weight loss. Tirzepatide delivers 17-22%. Retatrutide's Phase 3 TRIUMPH-4 readout hit 28.7%. The numbers are real, sustained over 68-72 weeks of treatment, and accompanied by genuine cardiovascular benefit in the SELECT cohort.

So why is the field talking openly about what comes after?

Two essays published this spring frame the answer. Bloomgarden's Journal of Diabetes commentary (February 10, 2026) emphasized two limits: 15-30% of patients can't tolerate the GI side effects at therapeutic doses depending on agent and titration, and 10-20% are non-responders who fail to reach the expected weight-loss range. With ~12% of US adults estimated to have used a GLP-1 at some point, those failure modes affect a lot of people in absolute terms. The Frontiers in Drug Discovery review (April 10, 2026) cataloged the candidate peptide-receptor systems beyond GLP-1, arguing the underexplored peptides 'failed' previously not for biological reasons but for technical and conceptual barriers solvable with modern peptide engineering.

The practical reading of these essays: GLP-1 has been the breakthrough that opens the field. The science of peptide therapeutics is now mature enough — stabilization chemistry, fusion proteins, lipidation, oral delivery platforms — that several other endogenous peptide systems are clinically tractable for the first time. The next 5-10 years will see at least some of them reach human trials. Most won't make it to approval; one or two might.

Apelin is the endogenous ligand for the APJ receptor, a G-protein-coupled receptor with widespread expression in cardiac, vascular, adipose, and hypothalamic tissues. The peptide has been studied since 1998, mostly in cardiovascular disease — apelin signaling produces positive inotropy without arrhythmia, improves endothelial function, and protects against ischemia-reperfusion injury.

The metabolic-disease story is more recent and less developed. Apelin signaling improves insulin sensitivity, increases glucose uptake into skeletal muscle, and modulates lipid metabolism in ways that resemble the desirable end of the GLP-1 effect — but through a pathway independent of incretin signaling. In animal models, apelin or apelin analogs produce modest body-weight reduction (much smaller than GLP-1s) with concurrent improvement in cardiometabolic markers.

The clinical translation has been slow because endogenous apelin has a half-life measured in minutes — the C-terminal sequence is rapidly degraded by angiotensin-converting enzyme 2 (ACE2). Stabilized apelin analogs are the technical fix, and at least three companies have been working on them. None has reached late-stage trials yet. The Frontiers review's argument is that the engineering is ready and the indication portfolio (heart failure with preserved ejection fraction, metabolic syndrome, possibly obesity as an adjuvant) is large enough to justify continued development. The path is plausible; the clinical timeline is 5+ years.

Spexin is a 14-amino-acid peptide identified bioinformatically in 2007 and characterized over the following decade. It's expressed in hypothalamus, stomach, adipose tissue, and other metabolic-relevant sites. Both galanin receptors (GALR2 and GALR3) appear to be ligands, with spexin acting as an inhibitor of food intake through galanin-receptor-mediated signaling.

The interesting feature of spexin is the direction of effect. Plasma spexin levels are inversely correlated with body weight in human cross-sectional studies — obese individuals have lower spexin levels, and bariatric surgery raises them. Whether that's cause or correlation isn't fully resolved. The clinical case for spexin as a therapeutic depends on what happens when you give exogenous spexin to a patient with low endogenous levels: animal models support the idea that supplementation suppresses food intake and improves glucose tolerance, but human trials have been small and inconclusive.

Where this could go: a stabilized spexin analog or a small-molecule galanin-receptor agonist that mimics spexin's effects in the metabolic axis without the broader CNS effects of full galanin activation. The Frontiers review treats spexin as a high-value pre-clinical target with a plausible mechanism story; the operational reality is that no major pharma company has it in active clinical development as of mid-2026. Academic groups in Korea, Poland, and Canada are the primary venues for current spexin work.

Phoenixin is a 14-residue peptide identified in 2013, with two cleavage products (PNX-14 and PNX-20) showing different biological profiles. Its receptor was contested for years; current consensus identifies GPR173 as the primary endogenous receptor. Tissues of expression include hypothalamus, adipose, pancreas, and gonads.

The metabolic-disease relevance has emerged gradually. Phoenixin signaling appears to modulate insulin secretion (positively), food intake (the data is mixed, possibly biphasic), and reproductive function. The cross-talk between metabolism and reproductive function — between energy availability and fertility — is the kind of physiology that has historically been difficult to drug because it sits across multiple organ systems and the desired clinical effect depends on which axis the drug is being aimed at.

For phoenixin specifically, the most plausible near-term therapeutic case is in polycystic ovary syndrome (PCOS), where the combined metabolic and reproductive dysfunction is the central pathology. A drug that targets one axis without disrupting the other has theoretical appeal. The realistic timeline for a clinical-stage phoenixin analog is mid-to-late 2020s; the work is still mostly in academic preclinical labs. Worth tracking because the PCOS indication is large and underserved by current pharmacology.

Relaxin-3 is a 24-amino-acid neuropeptide that's a member of the broader relaxin/insulin superfamily. It's produced almost exclusively in the nucleus incertus, a small brainstem nucleus, and signals through RXFP3 (a G-protein-coupled receptor) across multiple brain regions including hypothalamus, septum, and amygdala.

Relaxin-3 sits at an interesting intersection of behaviors: feeding, arousal, stress response, and reward. RXFP3 agonism in animal models increases food intake; antagonism reduces it. The molecule's behavioral profile maps somewhat onto orexin/hypocretin (the neuropeptide system that drives wakefulness and influences appetite), but the receptor pharmacology is distinct.

Clinical relevance for metabolic disease: a small-molecule or peptide RXFP3 antagonist could suppress appetite through a mechanism that's complementary to GLP-1 signaling, and might be effective in patients who don't respond to incretins. The mechanistic differentiation from GLP-1 (which works through gut-brain signaling) versus relaxin-3 (which works through nucleus-incertus-to-hypothalamic projections) is real and meaningful.

Where it stands: stronger preclinical evidence than spexin or phoenixin, but no clinical candidates yet. A handful of academic groups (Tregear and colleagues in Australia, the Vagena lab in the US) have been working on this for a decade. The Frontiers review treats relaxin-3 as one of the more biologically credible candidates in the 'beyond GLP-1' portfolio.

Osteocalcin is well-established in the bone-metabolism literature as a marker of bone turnover, but the discovery that the protein has endocrine functions independent of bone — secreted by osteoblasts, acting on pancreas, brain, testes, and adipose — only came together over the last 15 years. Undercarboxylated osteocalcin appears to be the active form for the endocrine effects.

The metabolic-disease case rests on osteocalcin's effects on pancreatic beta-cell function and insulin sensitivity. Animal models with osteocalcin deficiency show insulin resistance and glucose intolerance; supplementing osteocalcin improves metabolic phenotype in mice with diet-induced obesity. The receptor identified to date is GPRC6A, a G-protein-coupled receptor with broad tissue expression.

Clinical translation is at an early stage. Human cross-sectional data shows inverse correlation between osteocalcin levels and metabolic syndrome features, but causality is hard to establish. A few academic groups have been working on osteocalcin or GPRC6A agonists for type 2 diabetes; none has reached clinical trials. The interesting feature of osteocalcin as a therapeutic target is the bone-metabolic crosstalk — a drug that improves metabolic disease while also improving bone density would be valuable, especially in the perimenopausal and postmenopausal cohorts where both indications converge.

The path forward is unclear. Osteocalcin's pharmacology is complex enough that engineering a clean therapeutic agonist is harder than for some of the other peptides in this list. The Frontiers review acknowledges this; it treats osteocalcin as a promising target with significant medicinal-chemistry barriers.

Irisin is the peptide that gets the most popular-press coverage of anything in this list, and the one with the most contested clinical evidence. Discovered by Bruce Spiegelman's lab at Dana-Farber in 2012, irisin is a fragment cleaved from FNDC5 (fibronectin type III domain-containing protein 5) under conditions including exercise. The initial Spiegelman lab paper showed that irisin signaling drove white adipose to a brown-fat-like phenotype with increased thermogenesis — the 'exercise mimetic' framing.

The last 12 years have produced a complicated picture. Some labs reproduce the original findings; others find that human irisin levels are difficult to measure reliably, that the FNDC5 gene has a non-canonical start codon that affects translation, and that the cleavage and circulation kinetics of irisin in humans are messier than the original framing suggested. The biology is probably real; the clinical translation has been slow.

Where it stands in 2026: no clinical-stage irisin analog. Academic and biotech groups continue to work on FNDC5-pathway compounds, with mechanistic targets including the integrin αV/β5 receptor (a candidate irisin receptor identified in the last few years). The Frontiers review's framing — that the underlying biology is sound but the technical translation has been hard — is fair. Whether irisin or an irisin-mimetic becomes a real drug is genuinely uncertain. Bet on it cautiously; the popular-press 'exercise in a pill' narrative is well ahead of the clinical pipeline.

Of all the 'beyond GLP-1' candidates, the one with the most concrete near-term clinical path is Stanford's BRP (BRINP2-related peptide), a 12-amino-acid peptide identified through AI-screening of the human genome in 2024-2025. Published in Nature, BRP reduces food intake by up to 50% in mice within an hour and produces fat-selective weight loss without nausea or muscle loss — the side-effect profile that has dogged the GLP-1 class.

The mechanism is distinct from GLP-1. BRP appears to act directly on hypothalamic appetite-control circuits, with effects independent of leptin, GLP-1 receptor, and melanocortin-4 receptor. Animal data has been replicated in minipigs, which is the standard pre-human translational step.

First-in-human Phase 1 trials are planned. The Stanford team has been working with a biotech partner on stabilized analogs and formulation for human dosing. Timeline-realistic for a Phase 1 readout: late 2026 or 2027. If the human data confirms the animal findings, BRP is the most plausible 'beyond GLP-1' candidate in the field — better tolerability than the GLP-1 class with potentially similar efficacy, and a mechanism that's complementary rather than redundant.

The caveats are real. AI-screened peptides have had their share of false starts when human pharmacology differs from rodent pharmacology. The BRP receptor isn't yet definitively identified (a fact that has held up other peptide programs in the past). And the discovery-to-clinic timeline for novel peptide drugs is typically 8-12 years; BRP being on a faster path doesn't mean it gets to approval on a faster path.

What to watch: the first-in-human Phase 1 results. If they replicate the animal weight-loss and tolerability profile, BRP becomes the most-watched non-incretin candidate in the field. If they don't, the broader 'beyond GLP-1' research program continues but with one of its most promising candidates removed.

Taking all of this together, the realistic shape of the next-generation obesity-pharmacology pipeline:

In the near term (next 24 months): incretins continue to dominate. Retatrutide (Phase 3 TRIUMPH program with 7 readouts in 2026), CagriSema (NDA filed, FDA review ongoing), AstraZeneca's ASCEND (AZD9550 + AZD6234), survodutide, and Pfizer's MET-097i all approach approval or further commercial expansion. The next-generation incretins will define what's clinically available for most patients through ~2028.

In the medium term (3-5 years): combination therapy with non-incretin mechanisms enters the clinic. Lilly's quintuple agonist (GLP-1/GIP/glucagon/amylin/calcitonin in a single peptide) presented at ADA 2026 in late May points toward combinations; Roche/Zealand's petrelintide + enicepatide is the most mature combination program. Apelin or relaxin-3 analogs could plausibly reach Phase 2 in this window if companies commit serious development resources.

In the longer term (5-10 years): if BRP's human data holds, a fundamentally different mechanism class becomes available. If apelin, spexin, or osteocalcin analogs reach later-stage trials, they fill specific niches (cardiometabolic, PCOS, perimenopause-bone-metabolism) where incretins don't dominate.

The single biggest unknown: whether the 'beyond GLP-1' pipeline produces drugs with materially better tolerability than the incretin class. The 8.1% three-year persistence number for current GLP-1s is a tolerability and access failure, not an efficacy failure. A peptide drug with 12-15% mean weight loss and 50% three-year persistence would be commercially more important than another 25% weight-loss compound with 8% persistence. The field is starting to recognize this, but the trial endpoints haven't caught up. Designs that prioritize tolerability and long-term adherence — rather than maximum 72-week weight loss — would change which compounds actually make it through development.