Tirzepatide defined as an 'imbalanced and biased' dual agonist

A team from Eli Lilly, Duke University and the University of Copenhagen published a laboratory study in JCI Insight describing tirzepatide — then known by the code name LY3298176 — as an "imbalanced and biased" dual GIP and GLP-1 receptor agonist ^[1][2]. The paper was received May 26, 2020, accepted July 22, 2020, and carries a September 3, 2020 collection date ^[1].

The study's main contribution was a method. The authors wrote that they "establish methodology for calculating occupancy of each receptor for clinically efficacious doses of the drug" ^[1]. Applying that method, they reported "a greater degree of engagement of tirzepatide for the GIP receptor than the GLP-1 receptor, corroborating an imbalanced mechanism of action" ^[1]. In plain terms: at doses being tested in people, tirzepatide appears to occupy more GIP receptors than GLP-1 receptors — it is not a 50-50 drug.

What "biased" means here

Receptors do not have a single on-switch. After a drug binds, a cell can respond through several internal pathways. The researchers found that tirzepatide "mimics the actions of native GIP at the GIP receptor" but at the GLP-1 receptor shows bias "to favor cAMP generation over β-arrestin recruitment" ^[1]. They also reported that tirzepatide has a weaker ability to drive GLP-1 receptor internalization — the process by which a receptor is pulled inside the cell after activation — compared with GLP-1 itself ^[1].

Why might that matter biologically? In experiments on primary islets, the tissue that contains insulin-producing beta cells, the team found that β-arrestin1 limits the insulin response to GLP-1, but not to GIP or tirzepatide ^[1]. The authors suggested "that the biased agonism of tirzepatide enhances insulin secretion" and concluded that imbalance toward the GIP receptor, combined with distinct signaling at the GLP-1 receptor, "together may account for the promising efficacy of this investigational agent" ^[1]. That is a hypothesis the authors put forward, not a proven cause-and-effect link in patients.

The paper frames the question using clinical data available at the time. In a 26-week phase 2b trial in type 2 diabetes, about 30% of patients receiving the 15 mg dose reached normoglycemia, defined as an HbA1c below 5.7% by the American Diabetes Association standard, and 1 in 4 participants lost at least 15% of their body weight ^[1]. Tirzepatide also strongly lowered fasting triglycerides, and HOMA2-IR analysis pointed to improved insulin sensitivity that was only partly explained by weight loss ^[1]. At the time of publication, tirzepatide was described as under development for type 2 diabetes, obesity and nonalcoholic steatohepatitis ^[1].

Why it matters for patients

This is a mechanism paper, not a clinical trial, and it does not change how any medicine is used. Its practical value is vocabulary. The terms "imbalanced" (more GIP receptor occupancy than GLP-1 receptor occupancy) and "biased" (favoring cAMP over β-arrestin at the GLP-1 receptor) come from this work and are now the common shorthand in discussions of how tirzepatide differs from single-target GLP-1 drugs such as semaglutide ^[1].

It also explains why researchers do not treat "GLP-1 medicines" as interchangeable. Two drugs can hit overlapping targets and still produce different cellular responses, because what happens after binding differs ^[1]. That is one reason head-to-head clinical trials, rather than reasoning from mechanism, remain the way differences in results get settled.

Several things are not answered by this study. It does not measure side effects, nausea, durability of weight loss, or long-term outcomes, and it does not show that the signaling bias is what produces the clinical results — the authors say only that it "may account for" the efficacy seen ^[1]. Readers should also note the study's origin: most senior authors are employees of Eli Lilly Research Laboratories, the company developing tirzepatide, working with academic collaborators at Duke and Copenhagen ^[1][2]. The paper is open access under a Creative Commons license ^[1].