Skip to content
glp1.howSign in

Using GLP-1s

A new pancreas study on GLP-1 drugs: what it actually shows

A March 2026 Salk Institute study found a molecular switch, phosphorylation of a protein called Med14, that helps pancreatic beta cells carry out a broad gene response to prolonged GLP-1 exposure. It's a peer-reviewed finding from rat cells and mouse islets, with no human confirmation yet. What it found, what it means, and what it doesn't tell us.

By the glp1.how editorial teamReviewed 3 min read

Evidence-backed

Basic-science papers sometimes make it into GLP-1 news under headlines like "GLP-1 flips a molecular switch in your pancreas." In March 2026, a team at the Salk Institute (Van de Velde et al., published in Proceedings of the National Academy of Sciences) reported a finding along those lines. The paper is more limited than the coverage suggested.

What the study found

GLP-1 medications work partly through the beta cells in your pancreas, the cells that make insulin. Researchers already knew that when beta cells are exposed to a GLP-1 receptor agonist for a long time, beyond a single dose, they release extra insulin in the moment and also switch on a broader, longer-lasting gene program that seems to help them stay healthy and resist stress.

What wasn't known was how that switch gets flipped. The Salk team pointed to a protein called Med14, part of a larger machine inside the cell (the Mediator complex) that helps control which genes get turned on. Sustained exposure to a GLP-1 receptor agonist causes a single spot on Med14, a serine called "Ser983," to get phosphorylated, or chemically tagged. In the lab cells, prolonged exposure changed the activity of about 1,000 genes (1,065 in the study's cell line), including genes that support insulin secretion, stress resistance and cell health. When the researchers blocked the tag, much of that response was blunted.

In a mouse model with that one spot on Med14 mutated, beta cells lost much of the gene response and the ratio of alpha cells to beta cells shifted unfavorably. Block the switch, and the pancreas doesn't get the same reprogramming from GLP-1 exposure.

What it was tested in

The findings come from rat insulinoma cells (INS-1, a lab cell line) and mouse islet cells, not from people. The researchers say human confirmation is still ahead. In Salk's news release, co-author Reuben Shaw lists validating the findings in human tissues, and seeing whether Med14 has a similar role in other cells and organs, among the questions still to answer, and the team names fat as one tissue it's curious about. First author Sam Van de Velde has also mentioned the part of the brain that controls appetite (Times of San Diego, March 2026). None of that was tested in this paper.

What this does and doesn't mean for you

What it plausibly supports is a biological explanation for why beta cells seem to tolerate, and may even benefit from, sustained GLP-1 exposure without wearing out. That fits with separate human evidence on pancreatic cancer. A 2026 meta-analysis of 48 randomized placebo-controlled trials found GLP-1 drugs probably have little or no effect on pancreatic cancer risk (Ko et al., Annals of Internal Medicine, 2026, DOI (external link)). Another 2026 meta-analysis of 60 studies found no increase in risk, with a possible inverse association seen mainly in observational cohorts (Lauri et al., Pancreatology, 2026, DOI (external link)). In one high-risk group, people with chronic pancreatitis, GLP-1 use has been linked to lower rates (Ailawadi et al., Cancers, 2026, DOI (external link)). Those studies answer a different question from the Med14 paper, using human population data, but the two lines of evidence point the same reassuring way.

What it doesn't support is any claim that this mechanism has been shown in your pancreas, or that it changes anything about dosing, side effects, or how long you should stay on a GLP-1. It's a mechanism paper in mouse and rat cells. It helps researchers understand why these drugs behave as they do at the cellular level, and it says nothing yet about patient outcomes.

The bigger picture

If you've wondered whether years on a GLP-1 might be hard on your pancreas: the human safety data so far haven't shown that pattern, and this study offers a plausible biological reason beta cells might hold up well under sustained exposure. That's encouraging, and the picture is still being filled in. For the basics of how these medications work day to day, see how GLP-1 medications work, in plain language.

Study findings above are attributed to their authors and were located via PubMed. This is general education, not medical advice. Talk to your prescriber about any concerns specific to your health history.

pancreasbeta-cellsmechanismresearchlong-term-safetyhow-it-works

Common questions

Questions people often ask about this topic.

  • Was this study done in humans?

    No. It used a rat pancreatic cell line (INS-1) and mouse islet cells, including a mouse model with the key protein mutated. The Salk team lists validating the findings in human tissues as one of the questions still to answer; this paper didn't do that.

  • What is Med14 and what does it do?

    Med14 is a scaffolding subunit of the Mediator complex, a group of proteins that helps control which genes get switched on inside a cell. When beta cells are exposed to a GLP-1 receptor agonist for a sustained period, a specific spot on Med14 (Ser983) gets phosphorylated. Blocking that tag blunted much of the broad gene response that prolonged GLP-1 exposure normally produces in beta cells.

  • Does this mean GLP-1 medications protect my pancreas?

    It's a plausible biological explanation for why beta cells might hold up well under long-term GLP-1 exposure without wearing out, and it lines up with separate human data: 2026 meta-analyses of randomized trials and observational studies found no increased pancreatic cancer risk with these drugs. But the mechanism itself hasn't been confirmed in humans, so it isn't proof that your pancreas is being protected. It's one piece of a picture still being filled in.

  • Does this study change anything about how I should take my GLP-1?

    No. It's a cell-and-mouse mechanism study, not a clinical trial, and it doesn't address dosing, side effects, or how long to stay on treatment. Nothing here is a reason to change your regimen. Talk to your prescriber about any pancreas-related concerns specific to you.

Evidence

The sources behind this article, labeled by type and strength.

Supporting6

Mixed findings1

Revision history

Updated 6 times since Sep 3, 2026 · show changes
  1. v7

    Evidence sources 5 → 7

  2. v6

    Common questions revised

  3. v5

    Body revised

  4. v4

    Evidence sources 3 → 5

  5. v3

    Common questions revised

  6. v2

    Summary rewritten; Body revised

  7. v1

    Change tracking began

Related guides

  • Can you boost GLP-1 naturally? How food and lifestyle compare to the medications · You can raise your own GLP-1 with protein, fermentable fiber, and exercise, and that part is real. What doesn't hold up is the implied equivalence with the medications. This guide covers what nudges your natural GLP-1, why your body clears it in minutes while the drugs hold it for a week, and why "natural boosting" works as a foundation for treatment and can't replace it.
  • How GLP-1 medications work, in plain language · Semaglutide and tirzepatide copy hormones your gut already makes after a meal. This guide explains what those hormones do, why the medications quiet hunger and 'food noise,' what the second hormone in tirzepatide adds, and how the same signal that cuts appetite can also cause nausea.
  • Every key GLP-1 clinical trial: what it tested, who paid for it, and what it found · STEP 1, SURMOUNT-4, SELECT, LEADER: GLP-1 trial names get quoted everywhere, usually without context. This reference page lists 205 named, completed GLP-1 trials in one place, grouped by drug and by question, with what each one tested, who sponsored it, its headline result, and a link to the published paper. It also lists the trials that finished but have not reported, and shows who paid for the research.
  • Do GLP-1 receptors "burn out"? What long-term use actually does · A common worry is that GLP-1 receptors "burn out" with long-term use: that the drug stops working or permanently damages your natural system. Where the idea comes from, what two-year trial data show, which effect really does fade (the slowing of stomach emptying), why a plateau is a new equilibrium, and what happens to your own appetite when you stop.
  • Comparing the GLP-1 pills: Ozempic tablets (Rybelsus), the Wegovy pill, and Foundayo · There are now three GLP-1 pills in the US, and they aren't interchangeable: oral semaglutide for diabetes (Rybelsus, now sold as Ozempic tablets), oral semaglutide for weight (the Wegovy pill), and orforglipron (Foundayo). Here's how they compare on approval, how you take them, what the trials found, side effects, cost, and what you give up versus a weekly injection.
  • Obesity phenotypes: why "hungry gut" vs "hungry brain" may change your GLP-1 results · Researchers now describe four "phenotypes" of obesity (hungry brain, hungry gut, emotional hunger, and slow burn), and early evidence suggests they help explain why the same GLP-1 produces very different results in different people. One newly described "hungry gut" subtype lost nearly twice as much weight on tirzepatide. What the phenotypes are, what the studies show, and what to know before paying for a phenotype test.

Related terms

  • GLP-1 · A gut hormone (glucagon-like peptide-1) that stimulates insulin release, slows digestion, and reduces appetite; the target of drugs like semaglutide and liraglutide.
  • Receptor agonist · A drug that binds to and activates a specific receptor, mimicking the effect of the body's natural hormone.
  • Agonist · A general pharmacology term for any substance that activates a receptor to produce a biological response.
  • Pancreatitis · Inflammation of the pancreas — a rare but serious potential side effect of GLP-1 medications requiring immediate medical attention.

Was this helpful?

Improve this page