PD-L1 is best known as the molecular "off switch" cancer cells use to disable the immune system at their surface. In recent years, however, it has become clear that PD-L1 can also enter the nucleus, yet what drives it there, and what it does once inside, has remained largely a mystery. Researchers from Zhejiang University and the University of Macau now provide an answer: DDB1, a protein adaptor, tags PD-L1 with K63-linked ubiquitin chains at the K185 site—a modification that does two things at once. It removes an acetylation mark that would otherwise keep PD-L1 out of the nucleus, and it enables the nuclear protein to bind DNA and switch on a program of immunosuppressive genes. In breast cancer patients resistant to immunotherapy, this nuclear form of PD-L1 was markedly elevated. Remarkably, thalidomide, an old drug already used for multiple myeloma, disrupts the DDB1–PD-L1 partnership. Combined with anti-PD-1 antibody in mouse models, the therapy shrank tumors dramatically, boosted infiltration of cancer-killing T cells, and extended survival. Published in Science Bulletin, the study uncovers a non-canonical route to immune evasion and offers a "repurposed" combination strategy poised for clinical testing.

Researchers found that four immune system genes bear signatures of thousands of years of human evolution, identifying rare variants in a particular gene that were associated with severe COVID-19 and with prior organ transplants. The findings indicated that immune system genes shaped by ancient infectious diseases could eventually point to new strategies for disease prevention and treatment.