Biased Agonism: Same Receptor, Different Effects
Peptide Receptor Signaling
15 Class B GPCRs
Fifteen Class B G protein-coupled receptors respond to peptide hormones, and biased agonism at these receptors is reshaping drug development for diabetes, obesity, and cardiovascular disease.
Tasma et al., Pharmacological Reviews, 2025
Tasma et al., Pharmacological Reviews, 2025
View as imageWhen a peptide binds a receptor, it does not simply switch the receptor "on." It stabilizes one of several possible receptor conformations, each of which activates different intracellular signaling cascades. This is biased agonism: the ability of different ligands to selectively activate specific pathways through the same receptor. The concept has fundamentally changed how researchers think about G protein-coupled receptor signaling, moving from a simple on/off model to one where the quality of the signal matters as much as its strength.
A 2025 review in Pharmacological Reviews established the current state of the field: all 15 Class B GPCRs that respond to peptide hormones display some degree of biased signaling, and this property is now being deliberately engineered into next-generation therapeutics for diabetes, obesity, migraine, and cardiovascular disease.[1]
Key Takeaways
- All 15 Class B GPCRs that respond to peptide hormones display biased signaling between G protein and beta-arrestin pathways (Tasma et al., 2025)
- CT-388, a signal-biased dual GLP-1/GIP agonist, produced superior weight loss compared to unbiased agonism in preclinical models (Chakravarthy et al., 2026)
- SAL0112, a G protein-biased GLP-1R agonist, achieved potent receptor activation without the desensitization seen with balanced agonists (Sun et al., 2024)
- Tirzepatide shows biased agonism at the GIP receptor, preferentially activating the cAMP pathway while weakly recruiting beta-arrestin (Rees et al., 2024)
- A tunable biased AT1R agonist enhanced cardiac contractility with limited blood pressure increase in preclinical heart failure models (Hadjadj et al., 2026)
- Endogenous opioid peptides show variant-specific pharmacological profiles across mu opioid receptor splice variants, demonstrating natural biased signaling (Abrimian et al., 2021)
Two Pathways, One Receptor
GPCRs, the largest family of drug targets in medicine, signal primarily through two mechanisms. When a ligand binds, the receptor can couple to G proteins (Gs, Gi, Gq, and others) that activate second messenger cascades like cAMP production or calcium release. Alternatively, the receptor can recruit beta-arrestin proteins, which trigger a separate set of downstream effects including receptor internalization, desensitization, and their own signaling cascades through MAP kinases and other pathways.[10]
A "balanced" agonist activates both pathways proportionally. A "biased" agonist selectively favors one pathway over the other. G protein-biased agonists preferentially activate G protein signaling while minimizing beta-arrestin recruitment. Beta-arrestin-biased agonists do the opposite.
The therapeutic significance is that different pathways often mediate different physiological outcomes. At many receptors, the desired therapeutic effect flows through one pathway while side effects flow through the other. If you can design a ligand that activates only the beneficial pathway, you get the therapeutic effect without the adverse events. This is a cleaner approach than developing a balanced agonist and then managing side effects with additional drugs.
The GLP-1 Receptor: A Model System for Biased Agonism
The glucagon-like peptide-1 receptor (GLP-1R) has become the primary model for studying biased agonism in peptide therapeutics, driven by the enormous commercial success of GLP-1 agonists for diabetes and obesity.
Douros et al. (2024) published a detailed analysis of the GLP-1R as a template for understanding and exploiting biased agonism. Different GLP-1R agonists already on the market show varying degrees of bias between cAMP production (via Gs protein) and beta-arrestin recruitment, and these differences correlate with clinical profiles including efficacy, side effects, and dosing requirements.[2]
A 2026 review by Kowalska et al. mapped the complete GLP-1 signaling architecture: receptor binding triggers G protein/beta-arrestin dual signaling, but the relative activation of each pathway determines whether downstream effects favor insulin secretion, appetite suppression, gastric slowing, or nausea.[10] Nausea, the most common side effect of GLP-1 agonists, appears to correlate with beta-arrestin-mediated receptor internalization and desensitization. A G protein-biased agonist could theoretically preserve the metabolic benefits while reducing gastrointestinal side effects.
Engineering Bias into GLP-1 Agonists
Several groups are testing this hypothesis. Douros et al. (2025) designed a GLP-1 analogue optimized for cAMP-biased signaling (favoring the Gs pathway) and demonstrated improved weight loss in obese mouse models compared to balanced agonists.[4]
Sun et al. (2024) reported on SAL0112, a G protein-biased GLP-1R agonist that achieved potent Gs pathway activation without the receptor desensitization that limits balanced agonists. SAL0112 also showed superior metabolic stability compared to danuglipron (a small-molecule GLP-1R agonist) in human and rat liver microsomes.[6]
Tran et al. (2026) took a structural approach, using C-terminal mutations of the GLP-1 receptor to model G protein-biased agonism and identify which receptor regions control the balance between G protein and beta-arrestin coupling.[12]
Modder et al. (2025) tested the other direction: what happens when you use oppositely biased GLP-1R agonists? They found that a G protein-biased agonist (acyl-ExF1), given peripherally for six weeks, prevented body weight gain and reduced plasma glucose. A beta-arrestin-biased agonist did not produce the same metabolic benefits.[11] This provides functional evidence that the metabolic benefits of GLP-1R activation flow primarily through G protein signaling.
Biased Dual and Triple Agonists
The next generation of metabolic drugs combines activity at multiple receptors, and biased agonism is being built into these multi-receptor molecules.
CT-388 is a once-weekly signal-biased dual GLP-1/GIP receptor agonist. Chakravarthy et al. (2026) reported that its signal-biased design produced superior weight loss compared to unbiased dual agonism in preclinical models, with early clinical data supporting the approach.[3]
Tirzepatide, which is already FDA-approved for diabetes and obesity, shows biased agonism at the GIP receptor. Rees et al. (2024) demonstrated that tirzepatide preferentially activates the Gs/cAMP pathway at the GIP receptor while having weaker effects on IP1 accumulation and beta-arrestin recruitment.[7] This GIP receptor bias may partly explain tirzepatide's superior efficacy compared to pure GLP-1 agonists, though the precise contribution of biased versus balanced GIP signaling to clinical outcomes is not fully established.
The differences between short-acting and long-acting GLP-1 agonists may partly reflect different signaling bias profiles at the GLP-1 receptor, a hypothesis under active investigation.
Beyond Metabolic Disease: Cardiovascular Applications
Biased agonism is not limited to metabolic receptors. Hadjadj et al. (2026) designed tunable biased ligands for the angiotensin II type 1 receptor (AT1R) and demonstrated a compelling therapeutic split. Their lead compound (Compound 12) showed low Gq activity but potent beta-arrestin activity at AT1R. In a heart failure model, this translated to enhanced left ventricular ejection fraction (improved cardiac pumping) with a limited pressor response (minimal blood pressure increase).[5]
This is the biased agonism promise in action. At AT1R, the Gq pathway mediates vasoconstriction (raising blood pressure), while beta-arrestin signaling mediates cardioprotective effects. A beta-arrestin-biased AT1R agonist gives you the cardiac benefit without the hypertension, a therapeutic profile impossible to achieve with a balanced agonist.
The study used the term "tunable" because the researchers could dial the ratio of Gq to beta-arrestin signaling by modifying specific residues in their peptide ligand, demonstrating that bias is not a binary property but a continuous spectrum that can be engineered.
Natural Biased Signaling: The Body Already Does This
Biased agonism is not purely a pharmaceutical invention. The body uses it naturally.
Abrimian et al. (2021) demonstrated that endogenous opioid peptides (enkephalins, endorphins, dynorphins) show variant-specific pharmacological profiles across different splice variants of the mu opioid receptor (OPRM1). Different natural peptides acting at the same receptor class produce distinct patterns of receptor activation, G protein coupling, and beta-arrestin recruitment depending on which receptor variant they bind.[8]
This means the body already uses biased signaling to produce different outcomes from the same peptide-receptor family. The multiple splice variants of OPRM1 effectively create a menu of signaling options, and different endogenous peptides select different items from that menu. Understanding this natural bias is relevant to how endogenous opioid peptides like casomorphins interact with opioid receptors.
Karaki et al. (2021) identified another example: a putative beta-arrestin-biased superagonist of the growth hormone secretagogue receptor (GHSR, the ghrelin receptor). Discovered through dual screening (calcium antagonism combined with ERK1/2 phosphorylation), this compound preferentially activated beta-arrestin pathways over G protein signaling at GHSR.[9] Since GHSR mediates both appetite stimulation and growth hormone release through partially separable pathways, a biased agonist could potentially stimulate growth hormone without increasing appetite.
Why Receptor Desensitization Matters for Bias
Biased agonism connects directly to receptor desensitization. Beta-arrestin recruitment to GPCRs serves a dual purpose: it initiates its own signaling cascades and it also triggers receptor internalization and desensitization. A G protein-biased agonist that minimizes beta-arrestin recruitment should, in theory, produce less receptor desensitization, maintaining efficacy over time without dose escalation.
SAL0112 demonstrated this principle at the GLP-1R: its G protein bias corresponded to reduced receptor desensitization compared to balanced agonists.[6] If this translates clinically, biased agonists could offer more sustained therapeutic effects with less tachyphylaxis (loss of response over time).
The allosteric modulation approach offers a complementary strategy: rather than designing biased orthosteric ligands, allosteric modulators can shift the receptor's conformational landscape to favor one signaling pathway when the natural ligand binds.
Limitations and Open Questions
The field faces several unresolved challenges.
Measurement bias. Quantifying signaling bias requires comparing pathway activation in carefully controlled assay systems. Different cell types, receptor expression levels, and assay conditions can change the apparent bias of a ligand. Results from HEK293T cells (the standard lab system) may not reflect signaling in native tissues where receptor expression, G protein availability, and scaffolding proteins differ.
In vivo complexity. Most bias measurements come from cell-based assays. Whether a ligand that appears G protein-biased in a dish maintains that bias in a living organism, where receptor populations are heterogeneous and pharmacokinetics influence receptor occupancy over time, is often unclear. The Modder et al. (2025) study showing differential metabolic effects of oppositely biased GLP-1R agonists in vivo is encouraging but used mouse models.[11]
Clinical translation. No drug has been approved specifically because of its biased agonism profile, though tirzepatide's GIP receptor bias may contribute to its clinical superiority. The gap between demonstrating bias in preclinical systems and proving that bias produces better clinical outcomes in humans remains large. CT-388's clinical development will be an important test case.
Pathway oversimplification. The G protein versus beta-arrestin framework, while useful, oversimplifies reality. GPCRs couple to multiple G protein subtypes (Gs, Gi, Gq, G12/13), and each produces different effects. Beta-arrestin itself activates multiple downstream pathways. A truly rational biased agonist might need to modulate three or four pathways simultaneously with different degrees of activation, a design challenge that exceeds current capabilities.
The Bottom Line
Biased agonism describes how different ligands can selectively activate specific signaling pathways through the same receptor. The GLP-1 receptor has emerged as the primary model system, with G protein-biased agonists showing improved metabolic effects and reduced desensitization compared to balanced agonists in preclinical studies. Biased dual agonists like CT-388 and the inherent bias of tirzepatide at the GIP receptor suggest this approach has clinical relevance. The principle extends beyond metabolic disease to cardiovascular applications and occurs naturally with endogenous opioid peptides. Significant gaps remain between in vitro bias measurements and clinical outcomes.