Roche’s potential $2.3bn HM17321 deal puts muscle preservation in focus

What the source reports

Hanmi Pharmaceutical announced on 24 August 2026 that Genentech, a member of the Roche Group, had licensed HM17321, an experimental obesity and metabolic-disease treatment designed to reduce fat while preserving or improving lean mass. Genentech receives exclusive development, manufacturing and commercial rights worldwide except South Korea.

Hanmi will receive $190 million upfront. Development, regulatory and commercial milestones could bring the disclosed deal value to approximately $2.3 billion, while the agreement also provides for tiered royalties if a product eventually reaches the market.

The large headline value is therefore mostly conditional rather than cash paid today. HM17321 remains in Phase I, has no published human efficacy result and is not an approved treatment.

What the $2.3 billion figure means

Only the $190 million upfront payment is fixed at signing. The rest depends on future events such as clinical progress, regulatory decisions and commercial milestones that may take years and may never occur.

A licensing agreement shows that Genentech sees scientific and commercial potential worth funding. It does not show that regulators have accepted the proposed mechanism, that human trials have preserved muscle, or that the candidate will become a medicine.

Hanmi is due to complete the ongoing Phase I programme. Genentech is expected to take responsibility from Phase II, when larger studies could begin testing whether the body-composition effects reported in animals translate to people.

What HM17321 is designed to do

HM17321 is a long-acting analogue of urocortin-2, usually shortened to UCN2. Hanmi describes it as selective for the corticotropin-releasing factor receptor 2, or CRFR2, and as a non-incretin treatment because it does not work through GLP-1, GIP or another incretin receptor.

The development idea is to influence fat and muscle biology directly: reduce fat mass while retaining or increasing lean tissue and potentially improving muscle function. Hanmi is studying HM17321 both as a stand-alone injection and as a possible partner for incretin-based medicines.

That is a proposed profile, not an established human effect. The company’s supporting body-composition results come from animal models and scientific posters; the current human trial is primarily designed to establish safety, tolerability and how the drug behaves in the body.

The Phase I study is still an early test

ClinicalTrials.gov lists a randomized, blinded, placebo-controlled Phase I study with approximately 90 planned participants. It includes a single-dose section in about 40 healthy adults and a multiple-dose section in about 50 adults with obesity.

The obesity cohorts are expected to receive HM17321 or placebo once weekly for 12 weeks. The primary outcome is the incidence and severity of treatment-emergent adverse events; pharmacokinetic measurements are secondary, while body-weight change is listed as another exploratory outcome.

The registry estimates completion in March 2027. No results are posted. The design may provide an initial human signal, but a small dose-escalation study cannot establish durable weight loss, muscle preservation, improved strength, uncommon adverse effects or comparative benefit against an approved medicine.

Why lean mass is becoming an industry target

Substantial weight reduction usually includes losses of both fat and lean tissue. That has pushed obesity-drug research beyond the number on a scale towards body composition, strength, mobility and physical function—especially for older adults and people already vulnerable to frailty.

Lean mass is not identical to skeletal muscle. Body-composition scans count water, organs and other non-fat tissue as well as muscle, so retaining lean mass on a scan does not automatically mean greater strength, easier movement or better long-term health.

A useful obesity treatment would ultimately need to show that any body-composition difference matters to people. Trials therefore need direct functional measures and longer follow-up, not only percentages of weight lost from fat and lean compartments.

The wider muscle-preservation pipeline

HM17321 joins several different attempts to change the quality of weight loss. These include myostatin-directed treatments such as apitegromab, activin-receptor approaches such as bimagrumab, and other muscle-directed candidates being tested alongside semaglutide or tirzepatide.

A 24-week Phase II trial of apitegromab with tirzepatide reported less lean-mass loss than tirzepatide with placebo, but it did not show a clear difference in exploratory physical-function measures. HM17321 uses a different mechanism and has not produced comparable human data.

These programmes should not be treated as one class or ranked using separate early studies. Some aim mainly to retain tissue during incretin treatment; HM17321 is also being positioned as a possible stand-alone fat-loss treatment. Each claim requires its own randomized evidence.

Why Roche may want another mechanism

Roche already has incretin and amylin programmes in its obesity portfolio. Licensing a non-incretin candidate gives the company another biological route that could be studied alone or in combination rather than relying entirely on appetite-related mechanisms.

Combination potential may be commercially attractive, but adding a second medicine could also add adverse effects, monitoring, complexity and cost. Those trade-offs cannot be evaluated until appropriate human studies report results.

The transaction is best understood as a portfolio decision: Roche is paying for the right to investigate a differentiated hypothesis. It is not a clinical verdict that selective fat loss has been achieved in people.

Preclinical evidence has strict limits

Hanmi has reported reduced body weight and fat mass alongside increased lean mass in diet-induced obesity models, including studies of HM17321 alone and with other experimental or incretin treatments. Such work can support dose selection and a biological rationale for human trials.

Animal metabolism, dosing and muscle biology do not reproduce the full complexity of human obesity or long-term treatment. Positive preclinical findings frequently fail to become useful medicines because effectiveness is smaller than expected, safety limits dosing or the target does not translate.

Until controlled human results are available, descriptions such as muscle-preserving, muscle-improving or selective fat loss should be read as development aims supported by nonclinical experiments—not benefits that a patient can expect.

What remains uncertain

The immediate questions are whether repeated HM17321 dosing is acceptably safe and tolerable in people, what exposure can be achieved, and whether the Phase I study shows a credible signal in weight or body composition.

Later trials would need to determine how much weight comes from fat rather than lean tissue, whether any lean-mass effect reflects skeletal muscle, and whether strength, mobility, metabolic health or quality of life improve. Durability and results after stopping would also matter.

It is also unknown whether HM17321 works better alone or with an incretin medicine, which population might benefit, and whether any added benefit would justify treatment burden and cost. The licensing milestones do not answer those clinical questions.

Bottom line

Genentech’s $190 million upfront commitment—and a deal that could reach approximately $2.3 billion—shows how valuable muscle preservation has become as an obesity-research strategy. HM17321 is notable because it targets a non-incretin UCN2 pathway and is intended to reduce fat without sacrificing lean tissue.

The evidence is much earlier than the transaction headline suggests. Human Phase I testing is ongoing, its primary focus is safety, and the proposed lean-mass benefit has not yet been demonstrated in people. For now, this is a major investment in a hypothesis rather than proof of a new obesity treatment.