pe
pep-10925 v1 CC-BY-SA-4.0

PF-08653944: Pfizer's once-monthly weight-loss injection

An experimental weight-loss injection in the same family as Ozempic, designed for once-monthly dosing; currently in clinical trials, not yet an approved drug.

statuscomputed targetGCGR length31 aa refs2
snapshot clinical 0% confidence
Class
Investigational GLP-1 receptor agonist (biased, ultra-long-acting)
Status
Investigational — no regulatory approval identified; Phase 2b completed, Phase 3 program active
Best-supported effect
12.3% placebo-adjusted body weight reduction at 28 weeks in adults with obesity (Phase 2b VESPER-3 trial, human, SC injection)
Main caveat
No Phase 3 readouts, no long-term safety data, no cardiovascular outcome trial data; tolerability advantage of biased GLP-1 signaling not validated in head-to-head clinical comparisons
status 2 / 5
prediction metrics boltz-2 2.2.1
ipTM0.642
pTM0.821
avg pLDDT79.2
ranking score0.762
STRUCTURE · PEP-10925 × GCGR
ranking0.762
target interface 4.5Å peptide drag rotate · ctrl+scroll zoom · right-click pan
boltz-2 2.2.1 · mmCIF ↓ download
sequence31 aa
15101520253031
HSQGTFTSDVSSYLEG QAAKEFIAWLVKGRX
in the news 16 articles
overview readme

What this is

PF-08653944 (also called PF'3944 or MET-097i) is Pfizer's investigational weight-loss injection, originally licensed from the biotech company Metsera. It belongs to the same family as Ozempic and Wegovy — drugs that mimic a gut hormone called GLP-1 — but was engineered around two specific goals: a plasma half-life long enough to allow once-monthly maintenance injections after an initial period of weekly dosing, and a "fully-biased" receptor profile that favors the insulin-signaling arm of the GLP-1 pathway while reducing activation of the β-arrestin pathway, which is hypothesized to contribute to GI side effects. Phase 2b evidence is available; Phase 3 is underway. As of 2026 the compound is investigational-only and is not available outside Pfizer's clinical trial program. The stored raw sequence is an approximation; the active molecule carries half-life-extending conjugation chemistry that is not represented in the single-letter sequence shown here.


What it does

PF-08653944 activates the GLP-1 receptor, which triggers glucose-dependent insulin release, suppresses glucagon, slows gastric emptying, and reduces appetite through circuits in the hindbrain and hypothalamus — the same actions that underlie the weight-loss effects of semaglutide and tirzepatide. The proposed difference is in the signaling quality: rather than activating the GLP-1 receptor in a balanced way, PF'3944 is designed to steer receptor activation toward G-protein (Gαs/cAMP) signaling and away from β-arrestin-2 recruitment. The hypothesis — grounded in receptor pharmacology — is that β-arrestin-driven internalization and sustained GI-tract activation underlie much of the nausea and GI intolerance characteristic of the GLP-1 class, and that a biased profile might reduce those effects while maintaining efficacy. Whether this translates to a meaningful tolerability advantage in humans has not yet been established in head-to-head clinical comparisons. The second distinguishing feature is dosing cadence: after a weekly titration phase, the extended half-life supports monthly maintenance injections, a departure from the weekly cadence of semaglutide and tirzepatide.


Evidence

  • Human: Phase 2b only. The VESPER-3 trial (2026) enrolled adults with obesity and used a weekly titration period through week 12 followed by monthly maintenance dosing; the trial reported 12.3% placebo-adjusted body weight reduction at 28 weeks, with weight loss continuing through the weekly-to-monthly transition. Madsbad and colleagues (2025) reviewed the Phase 2–3 pipeline for GLP-1 receptor agonists in obesity, situating this compound among the emerging long-acting candidates. No Phase 3 efficacy or safety readouts are available. No cardiovascular outcomes trial data exist.
  • Animal: Preclinical work is described in available literature as supporting the biased-agonism design hypothesis; specific in vivo publications from the Pfizer/Metsera development program are not individually available in the current source set.
  • In vitro: Receptor pharmacology characterization of the biased-agonism profile is referenced in mechanistic context; individual assay data are not available in the current source set.

Known effects

  • Weight loss (obesity) — Phase 2b (VESPER-3, 2026): 12.3% placebo-adjusted reduction at 28 weeks
  • Monthly dosing feasibility — Phase 2b: weight loss continued through weekly-to-monthly cadence transition; all-monthly regimen under investigation in Phase 3 VESPER-6
  • Glycemic control (type 2 diabetes) — Phase 3 planned (VESPER-5); no efficacy data yet
  • Biased GLP-1R agonism (tolerability hypothesis) — Mechanistic only; not validated in head-to-head clinical comparisons

Safety signals

GI adverse events (nausea, diarrhea, vomiting) are expected as a class effect of GLP-1 receptor agonism and have been observed in Phase 2b, though frequency and severity relative to other long-acting GLP-1 agonists remain to be characterized in Phase 3. Injection-site reactions are expected for a monthly subcutaneous depot-style administration. No long-term safety data are available; the compound has not undergone the kind of chronic-exposure characterization that approved agents have. A cardiovascular outcome trial is planned as part of the ten-trial VESPER Phase 3 program, but results are not available. The proposed tolerability advantage of biased GLP-1 signaling — the core differentiation claim — has not been confirmed in controlled head-to-head comparisons. Gasbjerg and colleagues (2026) reviewed the physiology of proglucagon-derived peptides and the receptor-signaling basis for GLP-1 class effects, providing context for the GI-side-effect mechanism that the biased design is intended to address.


Regulatory status

  • US (FDA): Not approved — investigational only. Not available outside Pfizer's clinical trial program.
  • EU (EMA): No approval identified; investigational status.
  • WADA: Anti-doping status not verified against the current WADA prohibited list for this card.

Mechanism

PF-08653944 is a conjugated GLP-1 analog designed to act as a biased agonist at the GLP-1 receptor (GLP-1R). Binding to GLP-1R on pancreatic β-cells potentiates glucose-dependent insulin secretion; receptor activation also suppresses glucagon release, slows gastric emptying, and reduces appetite via hindbrain and hypothalamic circuits. These are canonical GLP-1R agonist effects shared with semaglutide and tirzepatide, as reviewed in the physiology of proglucagon-derived peptides (Gasbjerg and colleagues 2026).

The distinguishing receptor pharmacology is biased agonism: PF'3944 is engineered to preferentially stabilize GLP-1R conformations that favor G-protein (Gαs/cAMP) coupling while minimizing β-arrestin-2 recruitment. The pharmacological rationale is that β-arrestin-mediated receptor internalization and downstream GI-tract signaling contribute disproportionately to the nausea and vomiting common across the GLP-1 class, and that preserving the Gαs/cAMP arm while reducing β-arrestin engagement may maintain insulinotropic and appetite-suppressing effects with a better GI tolerability profile. This hypothesis is supported by receptor pharmacology and preclinical work cited in development-program literature, but remains unconfirmed in controlled head-to-head human comparisons.

The extended plasma half-life — enabling monthly maintenance dosing — is achieved through conjugation engineering; the specific conjugation chemistry is not individually described in sources available for this card. The raw sequence stored on this platform is the backbone approximation; the half-life-extending modification is not represented in it. Phase 2b data from VESPER-3 showed continued weight loss through the planned weekly-to-monthly dosing transition, providing pharmacokinetic support for the monthly-cadence hypothesis.


Open questions

  • Phase 3 efficacy confirmation: VESPER-3 is a Phase 2b dose-finding trial. Whether 12.3% placebo-adjusted weight loss at 28 weeks holds in larger Phase 3 populations, and how the effect size compares to semaglutide or tirzepatide, remains to be established (Madsbad and colleagues 2025).
  • Biased-agonism tolerability: The claim that G-protein-biased GLP-1R signaling reduces GI adverse events relative to unbiased long-acting GLP-1 agonists is pharmacologically motivated but not clinically validated. Head-to-head comparative data are absent.
  • All-monthly regimen (VESPER-6): Whether an all-monthly dosing design without a weekly titration phase achieves equivalent weight loss and tolerability to the titrated VESPER-3 regimen is unknown.
  • Cardiovascular outcomes: A cardiovascular outcome trial is planned as part of the ten-trial VESPER Phase 3 program; no data are available.
  • Glycemic efficacy in type 2 diabetes: VESPER-5 is planned but not reported; no glycemic efficacy data exist for this compound.
  • Long-term metabolic effects: The effects of sustained monthly pulsatile GLP-1R activation on lean-mass preservation, weight-regain dynamics after cessation, and metabolic adaptation are not yet characterized.

Related peptides

PF-08653944 is one of several long-acting GLP-1 receptor agonists in or near the clinic. For the established weekly GLP-1R agonist most widely used for obesity, see semaglutide. For the dual GLP-1R/GIPR agonist with Phase 3 and approval data, see tirzepatide. For context on the proglucagon peptide family that GLP-1 belongs to, the review by Gasbjerg and colleagues (Physiological Reviews, 2026) covers the full lineage from glucagon through GLP-1 and GLP-2.

Hypotheses3 directions▾ collapse

Research directions for this peptide, selected from the current sources — hypotheses you can explore and model. None of it is proven yet; tap any one to see the full thinking.

openupdated 2026-06-05

Is the stomach really where GLP-1 drugs cause nausea, or does it come from the brain, and if so, does this drug's special design actually help where it matters?

If nausea from GLP-1 drugs is primarily a brain problem rather than a gut problem, drugs engineered to fix the gut signal may still cause just as much nausea. Clarifying this could save enormous resources in clinical development and lead to a better design strategy for the next generation of weight-loss drugs.

The hypothesis
PF-08653944's fully biased G-protein signaling profile reduces nausea and vomiting not by eliminating beta-arrestin-2 engagement in the gut but by reducing beta-arrestin-2-dependent GPCR internalization in the brainstem area postrema, where GLP-1R desensitization is the proximal trigger for emetic signaling.
Why it’s plausible
The readme states PF'3944 is designed to favor Galphas/cAMP over beta-arrestin-2 recruitment, hypothesized to reduce GI side effects. However, the nausea/vomiting side effects of GLP-1R agonists are now attributed to area postrema (brainstem) GLP-1R signaling rather than gut motility alone. Beta-arrestin-2 in the area postrema mediates receptor internalization and signal termination; biased agonism might prolong area postrema signaling rather than reduce it if the hypothesis is purely gut-centric. This mechanistic uncertainty is falsifiable and has direct implications for whether bias selectivity will actually reduce clinical nausea rates.
Why it matters
If beta-arrestin bias reduces GI side effects via a brainstem rather than enteric mechanism, the therapeutic window prediction changes substantially, and the dose escalation strategy would need to account for prolonged central GLP-1R engagement despite reduced apparent GI activation.
Plausibility.60
Novelty.60
Impact.70
Basis · grounding2 papers · 1 computed/note
[1]
notePF'3944 designed to favor Galphas/cAMP signaling over beta-arrestin-2 recruitment; hypothesis is that this reduces GI side effects by altering GLP-1R signaling quality.
[2]
paper
Receptor pharmacology literature describes how signaling bias (G-protein vs. arrestin) alters receptor internalization, desensitization, and downstream pathway activation in ways that differ by tissue.
doi: 10.1124/pr.115.011395
[3]
paper
GLP-1RA pipeline review notes GI tolerability remains a primary development challenge; nausea mechanism is multi-site including central (area postrema) and peripheral (enteric) components.
doi: 10.1080/13543784.2025.2472408
openupdated 2026-06-05

Could this drug appear to preferentially activate one signaling pathway just because it binds its receptor more slowly and stays attached longer, rather than because of something special about its molecular shape?

Understanding whether this drug's beneficial profile comes from its shape or its stickiness would guide development of the next generation of GLP-1 drugs, and could determine whether the same approach can be applied to faster-acting formulations without losing the tolerability advantage.

The hypothesis
The C-terminal X residue in PF-08653944's stored sequence (HSQGTFTSDVSSYLEGQAAKEFIAWLVKGRX) represents a fatty-acid conjugation site analogous to Lys26 in semaglutide, and the albumin-binding half-life extension it confers may simultaneously alter GLP-1R binding kinetics toward slower association/dissociation rates, producing a kinetic bias that is distinct from intrinsic receptor-level signaling bias.
Why it’s plausible
The readme acknowledges the raw sequence does not represent the active drug's full conjugation chemistry. Semaglutide's C18 fatty acid on Lys34 is known to alter receptor binding kinetics significantly, contributing to its prolonged receptor occupancy. If PF'3944's conjugation similarly slows kon/koff at GLP-1R, the observed G-protein bias could be at least partly a kinetic artifact of sustained slow receptor occupancy rather than intrinsic allosteric bias. This distinction is mechanistically important because kinetic bias is reversible with different conjugation chemistries while intrinsic bias is sequence-determined.
Why it matters
Distinguishing intrinsic conformational bias from kinetic bias in GLP-1R agonists would resolve a key pharmacological debate in the GLP-1 field and determine whether the signaling selectivity of PF'3944 can be replicated in shorter-acting formulations.
Plausibility.70
Novelty.45
Impact.70
Basis · grounding1 paper · 2 computed/notes
[1]
sequenceHSQGTFTSDVSSYLEGQAAKEFIAWLVKGRX: terminal X residue in a GLP-1 analog context almost certainly marks a conjugation site (likely Lys or modified residue for albumin-binding fatty acid, as in semaglutide).
[2]
noteActive molecule carries half-life-extending conjugation chemistry not represented in the single-letter sequence; the peptide is designed for once-monthly dosing implying very high albumin affinity.
[3]
paper
GLP-1R pharmacology literature discusses how ligand binding kinetics and receptor occupancy time influence signaling pathway selectivity independently of intrinsic bias.
doi: 10.1124/pr.115.011395
openupdated 2026-06-05

Does this new weight-loss drug work on two different hormone receptors at once, and if so, is the second one responsible for some of its extra effectiveness?

If this drug has dual receptor activity, patients and doctors would have a clearer picture of why it works, which diseases it might help beyond obesity, such as fatty liver disease where glucagon signaling matters, and which patients are most likely to respond.

The hypothesis
PF-08653944's annotated primary target is GCGR (glucagon receptor), not GLP-1R, suggesting it may function as a dual GLP-1R/GCGR co-agonist analogous to cotadutide, and the glucagon receptor activity may contribute independently to its weight-loss efficacy via hepatic glucose production suppression rather than solely through GLP-1R-mediated appetite reduction.
Why it’s plausible
The target annotation lists GCGR, not GLP-1R. The stored sequence HSQGTFTSDVSSYLEGQAAKEFIAWLVKGRX shares N-terminal homology with glucagon (HSQGTFTSDYSKYLDSRRAQDFVQWLMNT) as well as GLP-1. Modern incretin analogs often have dual or triple receptor activity. If PF'3944 has meaningful GCGR affinity, the Boltz2 ipTM of 0.64 for a complex with GCGR (consistent with partial but real binding) would support this, and the once-monthly potency would require explanation beyond GLP-1R alone for which other monthly GLP-1R agonists have not achieved comparable weight loss.
Why it matters
If PF'3944 is a dual GLP-1R/GCGR agonist, the clinical trials should be monitoring for GCGR-associated metabolic benefits (lipid lowering, glycogenolysis reduction) and the competitor landscape comparison would shift from semaglutide to cotadutide, changing how regulatory differentiation is argued.
Plausibility.55
Novelty.60
Impact.70
Basis · grounding3 computed/notes
[1]
noteTarget annotated as GCGR (glucagon receptor); peptide described as GLP-1 mimetic but annotated against glucagon receptor, which could reflect dual agonism or misannotation.
[2]
structureboltz-2 complex ipTM=0.64 against GCGR: moderate confidence, consistent with real but not high-affinity engagement at the glucagon receptor.
[3]
sequenceHSQGTFTSDVSSYLEGQAAKEFIAWLVKGRX: HSQGTFTSD (positions 1-9) is near-identical to both glucagon (HSQGTFTSD) and GLP-1 N-termini; dual receptor activity is structurally plausible.
details expand to inspect
full evidence table2 metrics
metricvaluetool
ipTM 0.6420004963874817 boltz-2
ranking score 0.7616766095161438 boltz-2
3-letter notation
His-Ser-Gln-Gly-Thr-Phe-Thr-Ser-Asp-Val-Ser-Ser-Tyr-Leu-Glu-Gly-Gln-Ala-Ala-Lys-Glu-Phe-Ile-Ala-Trp-Leu-Val-Lys-Gly-Arg-X
recipeboltz-2 2.2.1
parametervalue
modelboltz-2 2.2.1
weights
hardwarevast_v100_32gb
mlx version
python
random seed1
msa strategycolabfold_local
runtime
predicted by
predicted at2026-05-23
citationbibtex
peptidemodel (2026). PF-08653944: Pfizer's once-monthly weight-loss injection (pep-10925, v1). PeptideModel. https://peptidemodel.com/card/pep-10925
@peptide{pep10925,
  sequence = {HSQGTFTSDVSSYLEGQAAKEFIAWLVKGRX},
  target   = {gcgr},
  author   = {peptidemodel},
  year     = {2026},
  status   = {computed}
}
related peptides 5 by signal overlap
clinical trials 6 on ct.gov · checked 2026-05-09
ct.gov trials 6
by phase
3phase 11phase 22phase 3
by status
5recruiting1not yet recruiting
references 2 papers
discussion no comments
sign in to comment
peptidemodel.com CC-BY-SA-4.0 research only · not for human use