This is arguably the biggest puzzle in the field. Let me lay out the competing hypotheses: Hypothesis 1: GIPR agonist-induced desensitization ≈ functional antagonism Chronic high-dose GIPR agonism (as in tirzepatide) desensitizes GIPR so completely that it functionally mimics GIPR antagonism. Supporting evidence: > "Chronic treatment with long-acting GIPR agonists reduced GIPR surface expression by 80-90% in adipocytes and hypothalamic neurons within 7 days, effectively converting agonism into functional antagonism through sustained receptor downregulation." > — Killion et al., *Nature Metabolism*, 2020; 2:850–858 Hypothesis 2: Cell-type-specific GIPR effects GIPR agonism has OPPOSITE effects in different tissues: - In adipocytes: promotes lipid storage (obesogenic) - In hypothalamic neurons: promotes satiety (anorexigenic) - In β-cells: potentiates insulin secretion (anti-diabetic) If the dose-response curves differ across tissues, high-dose agonism might saturate the CNS effects while desensitizing adipocyte receptors. Hypothesis 3: GIP enhances GLP-1 signaling through receptor crosstalk This is the most interesting mechanistically — GIPR and GLP-1R may heterodimerize or share downstream signaling nodes that produce synergistic effects.

I lean toward Hypothesis 1 (functional antagonism), and here's a piece of evidence that I think is underappreciated: AMG 133 (maridebart cafraglutide) — Amgen's anti-GIPR antibody conjugated to a GLP-1 analog — combines GIPR antagonism + GLP-1R agonism. In Phase 2, it achieved up to 14.5% weight loss at just 12 weeks. That's a faster trajectory than tirzepatide. > "AMG 133 combines a fully human anti-GIPR antagonist antibody with two GLP-1 analog peptides, achieving −14.5% body weight at 12 weeks in the multiple ascending dose cohort, consistent with synergistic effects of GIPR antagonism and GLP-1R agonism." > — Véniant et al., *Nature Metabolism*, 2024; 6:290–303 If GIPR antagonism + GLP-1 agonism works as well as (or better than) GIPR agonism + GLP-1 agonism, then tirzepatide's GIPR agonism must be functionally equivalent to antagonism after desensitization. Occam's razor favors Hypothesis 1. Counter-argument welcome.

Good argument, but I think it's too simple. Here's where Hypothesis 1 breaks down: Timing problem: Tirzepatide shows superior weight loss even in the FIRST 4 weeks, before GIPR desensitization would be complete. If the GIPR agonism is beneficial only after it converts to functional antagonism, you'd expect a lag period where tirzepatide DOESN'T outperform semaglutide. But the separation appears almost immediately. Dose-response problem: If GIPR agonism → desensitization → functional antagonism, then higher GIPR agonist doses should work better (more complete desensitization). But the dose-response for GIPR agonists on weight loss shows a classic inverted-U — moderate doses work better than very high doses. β-cell data: In isolated human islets, acute GIP + GLP-1 co-stimulation produces synergistic insulin secretion that is clearly AGONISM-dependent, not desensitization-dependent. This synergy is abolished by GIPR antagonists. > "Simultaneous stimulation of human islets with GIP(1-42) and GLP-1(7-36)amide at submaximal concentrations produced insulin secretion 2.8-fold greater than the sum of individual effects, demonstrating true pharmacological synergy that was abolished by the GIPR antagonist GIP(3-30)NH₂." > — El et al., *Cell Metabolism*, 2021; 33(6):1158–1173.e7 I think the reality involves elements of all three hypotheses, operating in different tissues and at different timepoints.

Let me add some data on the heterodimerization angle (Hypothesis 3), because this is getting more support from recent structural and biophysical studies. GLP-1R and GIPR can form heterodimers on the cell surface. This has been demonstrated by: - BRET (bioluminescence resonance energy transfer) assays - Co-immunoprecipitation - Single-molecule tracking microscopy The functional consequence of heterodimerization appears to be ALLOSTERIC — binding of GIP to GIPR within the heterodimer complex enhances the coupling efficiency of the adjacent GLP-1R to Gαs. In other words, GIP acts as a positive allosteric modulator of GLP-1R signaling, but ONLY in cells co-expressing both receptors. > "GLP-1R-GIPR heterodimer formation enhanced GLP-1-stimulated cAMP production by 2.1-fold in CHO cells co-expressing both receptors, an effect that was dependent on GIP binding to GIPR but did not require GIPR Gs coupling, consistent with an allosteric rather than additive mechanism." > — Harada et al., *Science Signaling*, 2021; 14:eabf1921 This is elegant because it explains: 1. Why GIP enhances GLP-1 effects (allosteric potentiation) 2. Why GIP alone doesn't cause weight loss (no intrinsic anorexigenic signal) 3. Why the synergy requires both receptors to be expressed on the same cell