Mutation predicts chemotherapy resistance and better responses to immunotherapy in advanced bladder cancer

MD Anderson Research News January 09, 2026

• Researchers studied the KDM6A gene, which is altered in about 26% of advanced bladder cancers
• Tumors with KDM6A mutations did not respond well to chemotherapy but demonstrated enhanced sensitivity to anti-PD1 immune checkpoint therapy
• KDM6A mutations could serve as a biomarker to guide treatment selection

Researchers at The University of Texas MD Anderson Cancer Center have determined that mutations in the KDM6A gene – common in advanced bladder cancer – may serve as a predictive biomarker to determine which patients are less likely to benefit from standard chemotherapy but may respond better to immunotherapy.

This study, published in Nature Communications and led by Sangeeta Goswami, M.D., Ph.D., associate professor of Genitourinary Medical Oncology and assistant member of the James P. Allison Institute™, underscores the potential for integrating genomic testing into bladder cancer care treatment.

“Our goal is to move beyond one-size-fits-all treatments,” Goswami said. “KDM6A gives us a clinically actionable signal and one that may spare patients from ineffective treatment and improve outcomes.”

What is KDM6A and what does it do?
KDM6A is an epigenetic regulator, turning genes on and off to control cell development and function. In a healthy cell, it removes specific chemical marks from histones, helping keep growth-related genes in balance.

KDM6A is also one of the most mutated genes in urothelial carcinoma, and loss-of-function mutations are frequent in advanced disease. When KDM6A is inactivated, tumor cells develop widespread changes that drive uncontrolled growth, genomic instability and treatment resistance.

How can KDM6A be a significant biomarker for bladder cancer?
In this study, researchers used engineered lab models to demonstrate that the loss of KDM6A drives the formation of extrachromosomal circular DNA. This circular DNA carries genes that provide resistance to chemotherapy, helping tumors evade standard chemotherapy treatments. At the same time, KDM6A-deficient tumors are unable to repair DNA, meaning these cancer cells accumulate damage that makes them more sensitive to immune checkpoint inhibition.

KDM6A loss also rewired tumor metabolism, forcing tumors to slow down energy consumption, leading to a reduction in lactate. This metabolic shift reduces histone lactylation in key immunosuppressive genes in regulatory T cells, weakening their ability to restrain immune responses and allowing cancer-fighting immune cells to act more effectively weakening their ability to restrain immune responses and allowing cancer-fighting immune cells to act more effectively. This is consistent with earlier work from the Goswami laboratory, published in Nature Immunology, showing that histone lactylation plays an important role in T cell function.

Together, these genomic, metabolic and epigenetic changes likely help explain the improved responses to immune checkpoint inhibitors seen in patients with KDM6A-deficient tumors.

“This dual effect – resistance to chemotherapy but heightened responsiveness to immunotherapy – helps explain previously conflicting clinical outcomes and gives us a roadmap for improved precision treatment strategies,” Goswami said.

Why is this study important?
These findings establish KDM6A as a key epigenetic regulator controlling genomic stability, tumor metabolic state and immune function that drives differences in treatment response in aggressive bladder cancer. The results suggest that KDM6A status at diagnosis may help identify patients more likely to respond to certain treatments, enabling more precise therapeutic decision-making in KDM6A-deficient bladder cancers.

***

This research was supported by the James P. Allison Institute Assistant Member Fund, the MD Anderson Physician Scientist Award, and the National Institutes of Health (R37 CA279192-1510 01). For a full list of collaborating authors, disclosures and funding sources, see the full paper in Nature Communications.