← Back to Home

Adding IL-18 “Armor” to T Cells: First Patient Dosed with CAR T Therapy for Ovarian Cancer

GB-5267 aims to enable immune cells to keep fighting within solid tumors while omitting lymphodepleting chemotherapy before infusion. With the first patient treated, the next questions are what dose is safe and whether this design can work in humans.

By SURL BioNews

For ovarian cancer that platinum-based chemotherapy can no longer adequately control, enabling immune cells to recognize cancer cells is only the beginning of the treatment challenge; they must also reach the tumor and continue working in an environment that suppresses immune responses. On October 1, 2026, Roswell Park Comprehensive Cancer Center in the United States and Generate Biomedicines announced that the first patient had been dosed in the first-in-human trial of GB-5267, a CAR T-cell therapy designed to address these barriers. The announcement has not yet provided efficacy or safety results for the patient.

GB-5267 is made from the patient’s own T cells: after collection, the cells undergo genetic modification and expansion at Roswell Park before being infused back into the body. Its chimeric antigen receptor (CAR) recognizes MUC16; according to the development team, it targets the form that remains on the tumor cell membrane. The cells also incorporate a design featuring the cytokine interleukin-18 (IL-18), creating so-called “armor” intended to help T cells resist suppression by the tumor microenvironment. This proposed mechanism remains to be validated and cannot yet be considered effective in human tumors.

The Phase 1 trial NCT07489287, led by gynecologic cancer specialist Emese Zsiros, enrolls eligible adults with platinum-resistant ovarian, peritoneal, or fallopian tube cancer. Registry information from the U.S. National Cancer Institute indicates that the study uses dose escalation to assess safety, tolerability, and the maximum tolerated dose; preliminary antitumor activity is a secondary objective. Dosing the first patient marks a step forward in clinical testing, but substantial work remains before treatment benefit can be determined.

The route of administration is also part of the study. Ovarian cancer often spreads within the peritoneal cavity, so the team hopes to explore the possibility of delivering cells closer to the lesions. The registered protocol distinguishes between two groups: one receives intravenous infusion only, while the other receives intravenous infusion followed by intraperitoneal administration. This is more specific than the announcement’s general reference to “two routes” and means that not every participant will receive intraperitoneal treatment.

Another design feature is that lymphodepleting chemotherapy is not required before infusion; this type of conditioning is typically used to reduce existing immune cells and create more favorable conditions for the infused cells. Omitting this step avoids the associated chemotherapy exposure, but does not establish that the overall treatment regimen is safer. The development team said the study will monitor cytokine release syndrome and immune effector cell-associated neurotoxicity, and that a data and safety monitoring board must review the available data before dose escalation; the registered protocol also provides for up to 15 years of follow-up.

Generate uses artificial intelligence to design and screen CAR structures, evaluating factors including cancer-cell killing ability, cell proliferation, and duration of activity, rather than estimating performance solely from antigen-binding strength. However, the announcement and accompanying interviews did not provide complete training data, the scale of candidate designs, or quantitative preclinical results sufficient for independent comparison. AI therefore serves as a research and development tool here; whether its designs confer a clinical advantage remains a question for human data to answer.

### Background Context

The results achieved by CAR T therapies in blood cancers have encouraged researchers to continue pursuing solid tumors, but barriers to entering tumor tissue and immune suppression make cell behavior beyond target recognition equally critical. GB-5267 combines IL-18 armor, the site of administration, and the omission of lymphodepleting conditioning in a single strategy. This early trial must first establish whether the cells can be reliably manufactured, safely infused, and expand and persist in patients; whether tumors shrink and how long responses last will require evidence to be built through subsequent results.

References

  1. Roswell Park Comprehensive Cancer Center
  2. Generate Biomedicines
  3. National Cancer Institute
  4. Generate Biomedicines