Tirzepatide: Biased Dual Agonism and the SURPASS/SURMOUNT Record

Tirzepatide is a synthetic peptide of 39 amino acids with the molecular formula C225H348N48O68 and a molecular weight of roughly 4,814 g/mol. It resolves in PubChem under CID 166567236 and carries CAS 2023788-19-2, which places it among the minority of compounds in this library with a fully public structural record.

Its backbone is derived from native glucose-dependent insulinotropic polypeptide (GIP) rather than from GLP-1. That choice is the single most consequential fact about the molecule, because everything asymmetric about its receptor behaviour follows from it. Two α-aminoisobutyric acid (Aib) substitutions confer resistance to dipeptidyl peptidase-4, the enzyme that clears native incretins within minutes.

A C20 fatty diacid is conjugated to a lysine residue via a linker. This moiety does not change what the peptide does at a receptor; it changes how long the peptide remains available to do it, by binding reversibly to circulating albumin and creating a slowly released depot. The same strategy - different chain length - appears in semaglutide. Extended half-life in this class is a chemistry decision, not a receptor decision.

The shorthand description of tirzepatide as 'GIP plus GLP-1 in one molecule' is convenient and slightly wrong. Willard and colleagues examined its behaviour at both receptors and found two departures from that picture, either of which would be notable on its own.

The first is imbalance. Relative to each native hormone, tirzepatide's potency is skewed toward the GIP receptor; it does not sit at the midpoint between the two activities. The second is bias. At the GLP-1 receptor, the compound preferentially drives cAMP accumulation over β-arrestin recruitment, and correspondingly produces less receptor internalisation than a balanced agonist would. Because β-arrestin recruitment is the principal route to receptor desensitisation and downregulation, a ligand that under-recruits β-arrestin may sustain signalling through a receptor population that remains at the cell surface.

It is worth stating what this does and does not establish. Biased agonism at a cloned receptor in a cell line is a measured property of the ligand; that it explains the clinical separation between tirzepatide and pure GLP-1 receptor agonists is a hypothesis the in vitro data support but do not prove. The clinical difference is documented in SURPASS-2. The causal link between that difference and the signalling bias is inference.

Tirzepatide's human evidence base is unusually strong for a peptide discussed in research-material contexts, because it was developed through a conventional registrational programme. Two families of trials carry most of the weight: SURPASS in type 2 diabetes and SURMOUNT in obesity.

SURPASS-2, reported by Frías and colleagues, is the most informative single trial in the programme for comparative purposes. It set tirzepatide directly against once-weekly semaglutide over 40 weeks in adults with type 2 diabetes and reported greater reductions in both glycated haemoglobin and body weight. Head-to-head trials between active agents are uncommon in this field; most comparisons rest on placing separate placebo-controlled trials side by side, which conflates differences in population, duration and escalation schedule with differences in drug. SURPASS-2 does not have that problem.

SURMOUNT-1, reported by Jastreboff and colleagues, addressed obesity as an indication in its own right. Over 72 weeks in adults with obesity or overweight with a weight-related complication, the highest-dose group reached a mean body-weight reduction approaching 21%. The 72-week duration is a meaningful feature of the design rather than a detail - many obesity trials end while the weight curve is still falling, which makes their endpoints difficult to interpret as steady states.

Across both programmes, gastrointestinal adverse events were the most frequently reported findings: nausea, diarrhoea and vomiting, generally mild to moderate, concentrated during dose escalation and dose-related. This is a class pattern shared with GLP-1 receptor agonists rather than something distinctive to dual agonism.

Pharmacokinetics receive far less attention than efficacy in the popular discussion of this class, but they explain several practical properties. Tirzepatide's albumin binding slows distribution and clearance enough to support weekly rather than daily administration intervals, and the peptide is cleared by proteolytic catabolism rather than by hepatic cytochrome P450 metabolism, which limits the classical drug-interaction pathways.

The review by Min and colleagues sets out the interaction profile that does matter, and it is mostly indirect. Incretin receptor agonists slow gastric emptying, and slowed gastric emptying changes the absorption of co-administered oral drugs - an effect that is most pronounced during dose escalation, when the delay is greatest. This is a pharmacodynamic interaction rather than a metabolic one, and it applies across the class.

The role of the GIP receptor arm is still genuinely contested. GIP receptor agonism and GIP receptor antagonism have both been reported to produce weight loss in preclinical models, which is an uncomfortable finding for any simple account of why tirzepatide works. The literature has not resolved this, and anyone claiming a settled mechanism for the GIP contribution is ahead of the evidence.

Durability and discontinuation are the second open area. Trials of this class consistently show weight regain after withdrawal, and the practical question - whether the effect requires indefinite continuation - is answered more by ongoing observation than by the registrational trials, which were not designed to address it.

Body composition is a third. Weight loss of the magnitude reported in SURMOUNT-1 necessarily includes lean tissue, and the proportion of lean to fat mass lost is reported less thoroughly than total weight across the programme. Finally, tirzepatide's cardiovascular outcome evidence is younger and thinner than semaglutide's, so comparisons on that endpoint rest on a weaker footing than comparisons on glycaemia or weight.