CT-388

Roche's investigational once-weekly subcutaneous dual GLP-1/GIP receptor agonist peptide (INN enicepatide), engineered for cAMP-biased signaling with minimal beta-arrestin recruitment, now in phase 3 obesity trials after 48-week phase 2 results of up to 22.7% weight loss.

Also referenced as: Enicepatide, CT388, RO7795068, RG6640

Also appears in: Metabolic

COA Public product evidence No certificate records are bound to a tracked CT-388 product page A record counts here only when a vendor publishes a certificate against one of its own tracked CT-388 product pages. Other certificates naming this compound may still exist — search the wider certificate database. →

---

What is CT-388?

CT-388 is a synthetic peptide that activates both the glucagon-like peptide-1 (GLP-1) receptor and the glucose-dependent insulinotropic polypeptide (GIP) receptor, the same receptor pair targeted by tirzepatide, but it was deliberately engineered so that both receptors signal through cyclic AMP with almost no beta-arrestin recruitment or receptor internalization (Chakravarthy et al., Molecular Metabolism, 2026;103:102291). It has the molecular formula C226H345FN48O67 and a molecular weight of 4,825 g/mol. The registered substance record (FDA UNII MTP9QR7GYF) describes a 38-residue amidated peptide with an exendin-like C-terminal tail, an aminoisobutyric acid substitution at position 12, a non-natural N-terminal cap built from a 3-cyano-5-fluorophenyl group and beta-alanine, and a C20 fatty diacid attached to the lysine at position 19 through a gamma-glutamate and two short polyethylene-glycol spacers, the half-life extension chemistry also used in semaglutide and tirzepatide. The compound was discovered by Carmot Therapeutics of Berkeley, California, using its Chemotype Evolution screening platform, and passed to Roche, which acquired Carmot in January 2024. Its international nonproprietary name, enicepatide (WHO proposed INN list 133), is used on the phase 3 registry records posted in January 2026, and Roche’s internal codes are RO7795068 and RG6640.

CT-388 draws attention because its 24-week phase 1b and 48-week phase 2 results in 2024 and 2026 were among the largest reported in phase 2 for an incretin agonist, and because, unlike tirzepatide, it was designed from the start for signaling bias at both receptors. It has begun to appear in research-reagent listings under its code and INN, though far less widely than retatrutide or tirzepatide.

How it works

• GLP-1 receptor agonism. CT-388 is a potent full agonist for cAMP production at the human GLP-1 receptor (EC50 0.087 nM, 99% of the maximal GLP-1 response), four-fold less potent than native GLP-1. GLP-1 receptor activation slows gastric emptying, reduces appetite through hypothalamic and hindbrain circuits and increases insulin secretion in a glucose-dependent manner, the actions that account for most of the weight loss seen with semaglutide and tirzepatide (Chakravarthy et al., Molecular Metabolism, 2026; Samms et al., Trends in Endocrinology and Metabolism, 2020;31(6):410–421).
• GIP receptor agonism, skewed toward GLP-1. At the human GIP receptor CT-388 produces cAMP with an EC50 of 2.47 nM (92% of maximal), 42-fold less potent than native GIP, so its activity is weighted toward the GLP-1 receptor, the opposite of tirzepatide, whose engagement is weighted toward the GIP receptor (Willard et al., JCI Insight, 2020;5(17):e140532). GIP receptor activation on its own has only modest effects on body weight in humans, but combined with GLP-1 receptor agonism it appears to add weight loss and to blunt nausea, possibly through central GIP receptors (Samms et al., Trends in Endocrinology and Metabolism, 2020).
• Biased signaling with minimal receptor internalization. In cell assays CT-388 recruited almost no beta-arrestin-2 to the GLP-1 receptor (4% of the GLP-1 response) and little to the GIP receptor (28% of the GIP response); GLP-1 receptor internalization was 6% of the native response and GIP receptor internalization was below the detection limit. Beta-arrestin recruitment normally drives receptor desensitization and internalization; GLP-1 receptor agonists with reduced recruitment produce more sustained insulin secretion and weight loss in mice than balanced agonists (Jones et al., Nature Communications, 2018;9(1):1602). Whether this bias explains the clinical results is a hypothesis, not an established fact; tirzepatide is also cAMP-biased at the GLP-1 receptor but unbiased at the GIP receptor (Willard et al., JCI Insight, 2020).
• Weekly pharmacokinetics. The terminal half-life in the phase 1 study was 123 to 139 hours after single doses and 130 to 151 hours after repeated doses (about five to six days), consistent with albumin binding by the fatty diacid side chain, and clearance and volume of distribution scaled with body weight (Chakravarthy et al., Molecular Metabolism, 2026).

Research status

Preclinical and phase 1 (NCT04838405, CT-388-101). The single peer-reviewed report (Chakravarthy et al., Molecular Metabolism, 2026;103:102291, funded by Carmot and written by Genentech employees, former Carmot employees and investigators at contracted sites) covers mouse studies and the first two parts of the phase 1 trial. In diet-induced obese mice on a steatohepatitis-inducing diet an 8-week course reduced body weight by about 20% versus 1% with vehicle and improved liver histology. The single-ascending-dose part (40 participants, Australia, 0.5 to 7.5 mg) and the multiple-ascending-dose part (24 participants, Mexico, four weekly doses titrated to 7.5 or 12 mg) enrolled adults with overweight or obesity and no diabetes. By day 29 body weight had fallen 4.7% to 8.0% across the weekly dose groups versus 0.5% with placebo, with improved fasting glucose, insulin and oral glucose tolerance. Decreased appetite occurred in every treated participant in the repeat-dose part, gastrointestinal events in 83%, and one of the two 12 mg cohorts showed a transient mean heart-rate rise of 15.8 beats per minute at day 29. The trial enrolled 129 participants in total across its parts and completed on August 2, 2024.

24-week cohort. Roche announced on May 16, 2024 that a later cohort of the same trial, treated for 24 weeks, reached a placebo-adjusted weight loss of 18.8% (p < 0.001), with every participant who had prediabetes at baseline normoglycemic by week 24. The release gave neither the dose nor the cohort size, and these data have not been published in a peer-reviewed journal.

12-week type 2 diabetes cohort. A further cohort randomized 19 adults with obesity and type 2 diabetes (HbA1c 7.0% to 10.0%, on diet and exercise alone or with metformin) 4:1 to 22 mg weekly or placebo for 12 weeks. HbA1c fell 3.0 percentage points with CT-388 versus 0.2 with placebo (placebo-adjusted 2.8, 95% CI 2.1 to 3.6) from a baseline of 8.7%, and weight fell 8.6% versus 1.2%. There were no serious adverse events or discontinuations for adverse events (Steinberg et al., Diabetes, 2025;74(Supplement 1):763-P, an American Diabetes Association abstract).

Phase 2 in obesity (NCT06525935, CT-388-103). A randomized, double-blind, placebo-controlled trial at 34 United States sites enrolled 469 adults (467 randomized) with a BMI of at least 30 kg/m², or 27 to 30 with a weight-related comorbidity, and no diabetes, from August 16, 2024 to December 8, 2025. Participants were assigned equally to placebo or to target doses of 4, 8, 12, 16 or 24 mg weekly for 48 weeks; the mean age was 49, the mean BMI 38 kg/m², 68% were women and about half had prediabetes. Genentech and Roche announced top-line results on January 26 and 27, 2026, and Lingvay and colleagues presented the results at the American Diabetes Association’s 86th Scientific Sessions in June 2026 (late-breaking abstract 2813-LB). Weight loss was dose-dependent and had not plateaued at week 48, reaching 22.7% at 24 mg, a placebo-adjusted difference of 22.5% on the efficacy estimand; on the treatment-regimen estimand, which counts participants who stopped treatment, the 24 mg difference was 18.3%. In the 24 mg arm 54% ended with a BMI below 30 versus 13% on placebo, and 73% of those with prediabetes at baseline were normoglycemic at week 48 versus 7.5% on placebo. Gastrointestinal events, mostly mild to moderate, were the most frequent adverse events, and 5.9% of all treated participants versus 1.3% on placebo discontinued treatment because of an adverse event. Arm-by-arm weight change, adverse-event rates and serious adverse events have appeared only on the congress poster; no peer-reviewed publication had appeared by November 2026.

Phase 2 in type 2 diabetes (NCT06628362, CT-388-104). A 48-week, randomized, double-blind, placebo-controlled trial enrolled 447 adults with type 2 diabetes (HbA1c 7% to 10.5%, BMI 25 kg/m² or higher, on diet, metformin or an SGLT2 inhibitor) at sites in the United States, Puerto Rico, Canada, Mexico, Argentina and New Zealand between November 21, 2024 and October 2, 2026, with co-primary endpoints of weight and HbA1c change at week 36. Genentech and Roche announced on September 21 and 22, 2026 that both endpoints were met: at 24 mg, HbA1c fell 2.65 percentage points from a baseline of 8.1% and weight fell 15.5% at week 48 without a plateau; 2.0% of treated participants and none on placebo discontinued because of adverse events. Placebo-arm changes, lower-dose results and the estimand were not disclosed, and no publication or presentation had followed by November 2026.

Phase 3. The ENITH program began in March 2026 with two 72-week, randomized, double-blind, placebo-controlled trials comparing three enicepatide dosing regimens with placebo on percent change in body weight: ENITH-1 (NCT07351045, 2,000 adults with obesity or overweight and no diabetes, started March 16, 2026, at about 210 sites in 21 countries) and ENITH-2 (NCT07351058, 1,600 adults with obesity or overweight and type 2 diabetes with HbA1c 6.5% to 10%, started March 23, 2026). Primary completion is listed for July 24 and August 7, 2028. A 52-week phase 3 trial in 300 Chinese adults with or without type 2 diabetes (NCT07670416) started on July 3, 2026 alongside a 36-participant pharmacokinetic study in healthy Chinese adults (NCT07626515), and ENI-SPIRE-2 (NCT07889648), a 72-week open-label phase 3 trial comparing enicepatide with tirzepatide on HbA1c in 948 adults with type 2 diabetes inadequately controlled on oral agents, was first posted on October 22, 2026 with an estimated start of December 1, 2026. ZYNERGY (NCT07589686), a 40-week phase 2 trial of enicepatide co-administered with the amylin analog petrelintide in 486 adults, is expected to start on November 30, 2026.

Regulatory status. No marketing application for CT-388 has been filed anywhere, and the compound has no approval, no FDA label and no entry in Drugs@FDA. Genentech describes it as a “fast-track asset” within its own portfolio, but no FDA Fast Track, Breakthrough Therapy or other expedited designation has been publicly announced. No trial of CT-388 has been terminated or reported a failed primary endpoint as of November 2026.

Common dosage forms

• Investigational subcutaneous injection, supplied to clinical-trial sites for use by trial participants. The device, whether a vial with syringe or a prefilled pen, has not been publicly described.
• No approved or compounded product exists in any country. CT-388 is not an FDA-approved active ingredient, so it cannot be legally compounded in the United States. The only non-trial presentation is lyophilized powder in research vials labeled for laboratory use.

Key considerations

• Regulatory status. The Investigational status reflects an active phase 3 program with no approval, no filed application and no announced FDA expedited designation anywhere as of November 2026. Primary completion is registered for July and August 2028.
• Evidence gaps. Peer-reviewed data cover only 4 weeks of dosing in 64 participants; the 24-week, 12-week and 48-week results exist as press releases and conference abstracts, the diabetes trial without placebo figures. There is no cardiovascular outcomes trial, no completed head-to-head comparison with tirzepatide or semaglutide, no data beyond 48 weeks and no information on weight regain after stopping.
• Safety signals. Tolerability follows the incretin class, with dose-related nausea, vomiting, diarrhea and constipation concentrated during escalation and a transient heart-rate rise during titration. Serious adverse event rates and arm-level tolerability have not been published, whether pancreatitis, gallbladder disease, lean-mass loss or anti-drug antibodies differ from approved agents is unknown, and no data in people over 75, in pregnancy or with renal or hepatic impairment have been reported.
• Naming. CT-388, CT388, enicepatide, RO7795068 and RG6640 refer to the same peptide. The code RO7790121, which sometimes circulates alongside CT-388, belongs to afimkibart, a Roche anti-TL1A antibody studied in inflammatory bowel disease, and should not be used for this compound. CT-388 is also distinct from Carmot’s other incretins, CT-868 (acmopatide, RO7795074, a daily dual agonist for type 1 diabetes) and CT-996 (an oral small-molecule GLP-1 receptor agonist), and “dual GLP-1/GIP agonist” is a class description shared with tirzepatide and VK2735, which are different peptides.
• Reference material. No analytical reference standard for CT-388 has been published, so material offered under the name outside the trials cannot be matched to the clinical compound from the literature.