Amylin may shape anxiety and sociability, differently by sex

Amylin has spent most of its public life in the shadow of insulin.

It is a peptide hormone made by pancreatic beta cells and released alongside insulin, and it has a clean, practical reputation: it helps shape appetite and post-meal physiology. But as obesity pharmacology has shifted from “weight loss is hard” to “weight loss is scalable,” the questions around what else these peptides do have gotten sharper.

A new paper in Translational Psychiatry puts that issue in plain terms. In rats, systemic amylin didn’t just touch feeding related biology. It shifted anxiety-like behavior, depression-like behavior, aggression, and sociability, and it did so in a way that depended strongly on sex (PubMed).

Why this is suddenly a practical question

Amylin is not an exotic newcomer. The field has known about it for decades, and an amylin analog (pramlintide) has been used clinically as an adjunct to insulin for type 1 and type 2 diabetes.

What feels new is the direction of travel.

Amylin based drugs are now part of the “next wave” conversation in obesity, including combinations that aim to move beyond the current plateau of glucagon-like peptide 1 (GLP-1) receptor agonists. A 2026 Nature Metabolism review captures that arc: from discovery (1986), to pramlintide approval, to a renewed push for longer acting amylin therapies and combinations (PubMed).

When exposure moves from a relatively narrow clinical niche into broad population use, side effects that once felt academic start to matter. And obesity drug history has made the industry unusually sensitive to signals involving mood and mental health. The withdrawal of the cannabinoid 1 (CB1) receptor antagonist rimonabant in 2008 is a reminder that “effective weight loss” can still be a commercial dead end if psychiatric effects are not manageable. One post-authorization monitoring study in England followed 10,011 rimonabant users and underscored how psychiatric signals can become the deciding variable for an otherwise plausible mechanism (PubMed).

What the new study actually tested

Byun and colleagues asked a pretty direct question: if you give amylin, do you change emotionality and social behavior, and does sex matter?

They dosed male and female rats with amylin in two ways:

Systemically, to approximate a whole-body exposure.
Directly into the central amygdala, a brain region heavily involved in threat processing, valence, and stress responsive behaviors.

Then they ran multiple behavioral assays.

The details are worth reading in the paper, but the headline is easy to summarize: amylin produced consistent, sex divergent effects on anxiety-like behavior across two different behavioral readouts, and the central amygdala was sufficient to drive that pattern.

A striking split: anxiolytic in males, anxiogenic in females

In both the elevated plus maze and the acoustic startle response, the direction of change differed by sex.

In males, systemic amylin and central amygdala amylin both looked anxiolytic.
In females, those same manipulations looked anxiogenic.

If you read the paper with an “obesity pipeline” lens, this split is the point. It is less about whether any single rodent task maps cleanly to a human diagnosis, and more about the implication that a therapy class could have behavioral liabilities that are not symmetric across sexes, even when the intervention is pharmacologically similar.

Depression-like behavior and the central amygdala

The forced swim test is often treated as a blunt instrument, but it is still widely used as a behavioral probe.

Here, intra-amygdala amylin increased depression-like behavior in females only. Systemic administration did not produce the same clear signal in that specific assay, which is interesting on its own: it suggests the circuit level manipulation may reveal effects that get diluted, compensated, or redistributed when the entire body is exposed.

That does not automatically mean “the amygdala is the whole story.” But it does strengthen the argument that amylin is not just a peripheral satiety peptide. It can act on central nodes that matter for affect.

Aggression went down in both sexes, but sociability was more complicated

The cleanest bidirectional finding in the paper may actually be aggression.

In a resident intruder test, aggressive behavior decreased in both males and females, whether amylin was given systemically or into the central amygdala.

Sociability did not follow the same pattern. Systemic amylin increased social interaction, but intra-amygdala amylin did not. That divergence is a quiet but important clue: if the effect is real, it may rely on distributed circuits, peripheral signals, or a different brain region than the central amygdala.

What this does and does not mean for amylin based obesity drugs

This is not a human safety readout, and it should not be oversold.

But it is a useful “watch this” paper for two reasons.

First, it frames a risk that is easy to ignore when a drug class is talked about primarily in metabolic terms: peptide hormones that alter appetite often have brain effects, because appetite is a brain mediated behavior.

Second, the study is explicit about sex as a variable that changes the direction of the behavioral phenotype. In other words, it is not just “maybe there is an effect.” It is “the effect might look reassuring in one group and concerning in another.”

That is the kind of signal developers want to catch early, because it changes how you design trials, how you monitor post-marketing data, and how you talk about benefits and risks.

The skeptical read: translation is the bottleneck

If you want to be maximally cautious, there are several reasons to resist strong conclusions.

Rodent behavioral paradigms have known limitations, and acute dosing does not tell you what happens with chronic exposure, where tolerance, learning, and hormonal adaptation can dominate. The paper also uses amylin itself rather than a specific long-acting clinical candidate, which complicates “drug class” inference.

The most productive next steps would be:

replication with chronic dosing,
comparison across amylin analogs under development,
and mechanistic work that explains why sex changes the sign of the effect.

If the field eventually sees a consistent story across models, then the human question becomes more concrete: are there measurable differences in anxiety, sleep, irritability, or social functioning signals in real-world amylin exposure? That is an empirical question, and the datasets will only get better as these agents move into larger populations.