Beyond Semaglutide: The Next Wave of GLP-1 Combinations and Genetic Response Predictors

Clinicians who prescribe GLP-1s have known for years that response is uneven. Trial averages — the 15% or 20% weight-loss figures that make magazine covers — hide a wide distribution. Some patients sit at the top of the curve. Others lose almost nothing despite dutiful adherence, and end up wondering whether the failure is theirs.

A 2025 analysis published in Diabetes, Obesity & Metabolism offers one of the first credible biological explanations. Researchers mining the NIH All of Us cohort of more than 6,500 GLP-1 users built a genetic score around neurobeachin (NBEA), a gene encoding a protein kinase A anchor protein expressed heavily in the brain. People whose score crossed the "responsive" threshold were 82% more likely to be among the top weight-loss responders on liraglutide, a finding that held up when tested again in the UK Biobank. The same score also flagged a non-response signal: liraglutide users on the other end of the distribution were about 50% more likely to lose no weight at all.

The semaglutide signal was directionally similar but weaker, with odds ratios of roughly 1.6 in discovery and 2.2 in validation. That nuance matters. It suggests the genetic effect is real but drug-specific, and that NBEA alone will not become a clean yes-or-no test. What it does provide is a foothold — a first plausible mechanism for why the same dose lands so differently in different bodies.

A second 2025 study, this one a randomized trial of 100 adults with obesity, asks a quieter but equally important question: when GLP-1s improve metabolic markers, is the drug doing the work, or is the weight loss?

Participants were randomized to liraglutide plus metformin or metformin alone for 24 weeks, with standardized diet and exercise. Both groups improved; the GLP-1 group improved more. But when researchers looked specifically at proinsulin processing — a marker of how cleanly the pancreas converts its precursor hormone into mature insulin — the change tracked body-weight and visceral-fat reduction rather than GLP-1 exposure itself, even after adjusting for age, sex, baseline BMI and treatment arm.

The implication is subtle but useful. GLP-1s may be best understood, at least for some endpoints, as efficient delivery vehicles for weight loss rather than as drugs with broad standalone metabolic magic. That framing aligns with what many endocrinologists already suspected, and it sharpens expectations: a patient who loses little weight on a GLP-1 is unlikely to capture the full downstream metabolic benefit, regardless of how the drug is theoretically supposed to work.

If single-agent GLP-1s have a ceiling — set partly by genetics, partly by tolerability — the obvious next move is to stack mechanisms. Tirzepatide, which combines GLP-1 and GIP receptor activity in one molecule, already hints at the upside. The early pipeline is going further.

One example: VB-87531, a small-molecule inhibitor of MOGAT2, an enzyme involved in resynthesizing dietary triglycerides for transport to the liver. In diet-induced obese mice, VB-87531 reduced weight and food intake on its own. More interesting, when paired with either semaglutide or tirzepatide, the combination produced greater weight loss and appetite suppression than the MOGAT2 inhibitor alone, alongside lower insulin and leptin and higher FGF21 and PYY — a hormonal profile that, on paper, looks favorable.

The caveat is large and worth stating plainly: this is preclinical work in mice. The history of obesity pharmacology is littered with compounds that dazzled in rodents and disappointed in humans. VB-87531 is not a treatment; it is a hypothesis. But the underlying logic — combine a centrally acting appetite drug with a peripherally acting fat-handling drug — is exactly the direction the field is moving, and several similar pairings are in earlier human exploration.

None of this changes a prescription today. There is no NBEA test you can order, no MOGAT2 combination on a pharmacy shelf, and no guidance that the proinsulin finding should alter how a clinician chooses between liraglutide, semaglutide or tirzepatide. What is changing is the framing.

For readers already on a GLP-1, the most practical takeaway is that uneven response is biological, not a personal failing — and that the metabolic gains tend to follow the weight loss, which is why sticking with lifestyle scaffolding still matters even on a powerful drug. For readers considering one, it is worth asking your clinician how response will be monitored, what a reasonable trial period looks like, and what the plan is if you fall on the lower end of the curve. Those are the questions the next few years of research are designed to answer.

The GLP-1 story started as a blockbuster. It is quietly turning into something more interesting: a precision-medicine story, in which the right drug, the right combination and the right expectations may eventually be matched to the right patient.