AI-Designed Weight-Loss Peptide Enters Human Trials: Hanmi Pharmaceutical Targets Fat Loss Alongside Muscle Gain

Existing weight-loss drugs can effectively reduce body weight, but may also be accompanied by a loss of lean tissue. South Korea’s Hanmi Pharmaceutical is trying a different approach: using artificial intelligence to help design a long-acting protein drug intended to increase or preserve muscle while reducing fat. The candidate drug HM17321 has entered a Phase 1 clinical trial in the United States, putting the concept to its first test of safety and pharmacological effects in humans.

In July, Hanmi Pharmaceutical presented its proprietary HARP-pSAR platform at the ISMB 2026 conference in Washington. According to the company, the platform is used to predict how pharmacological activity may change after subtle alterations to a protein’s amino acid sequence, and then to screen multiple designs for more promising candidates. The tool primarily narrows the experimental search space; it does not directly predict a drug’s efficacy and safety in humans.

HM17321, also known as LA-UCN2, is a long-acting urocortin 2 analog that selectively targets corticotropin-releasing factor receptor type 2 (CRF2). Hanmi Pharmaceutical says it was optimized using AI and structural modeling technologies. Obesity studies in animals disclosed by the company showed that the candidate reduced body weight and fat mass while increasing lean tissue. However, the evidence currently comes mainly from company research and conference posters and cannot yet demonstrate that changes in body composition of the same magnitude or nature will occur in humans.

Clinical trial registry data show that this randomized, placebo-controlled Phase 1 study is expected to enroll approximately 90 participants. The first part will test single ascending doses in approximately 40 healthy adults; the second part plans to administer HM17321 or placebo once weekly for 12 consecutive weeks to approximately 50 adults with obesity but without diabetes. The study will primarily assess safety, tolerability, pharmacokinetics, and pharmacodynamics, rather than directly verify long-term weight-loss efficacy.

The trial began in November 2025, is listed as recruiting, and is estimated to be completed in March 2027. Its design can provide preliminary observations of pharmacodynamic signals such as body weight and body composition, but the sample size and follow-up period are insufficient to determine durability of efficacy or fully characterize rarer adverse reactions. The CRF2 pathway is involved in muscle, metabolic, and cardiovascular physiology, and systemic effects as the dose increases will be an important safety issue in early development.

The significance of the program lies not only in another weight-loss drug candidate, but also in testing whether AI protein design can move beyond laboratory optimization and translate into a usable drug for humans. Publicly available information on HARP-pSAR remains insufficient to assess the scale of its training data, its external-validation performance, and its actual advantages over traditional design processes. Only if future clinical data simultaneously demonstrate acceptable safety, reduced fat, and improved muscle function could they support the claim that it can “improve the quality of weight loss.”