Immune checkpoint blockade (ICB) targeting the PD-1/PD-L1 axis has transformed cancer therapy, delivering remarkable and often durable responses across a range of malignancies. Yet in unselected solid tumors such as breast cancer, its benefits remain frustratingly limited—and, more puzzling still, many patients whose tumors express high levels of PD-L1 fail to respond at all. This paradox has long troubled oncologists: if PD-L1 is the very target of therapy, why does its abundance so often fail to predict success? A new study published in Science Bulletin suggests the answer lies not on the surface of the cancer cell, but deep within its nucleus.
The researchers found that PD-L1 can slip into the nucleus and moonlight as a transcriptional regulator, switching on an entire network of immunosuppressive genes that drive resistance to anti-PD-1 therapy. Tumors from patients who did not respond to ICB harbored significantly higher levels of nuclear PD-L1 than those from responders—implicating this non-classical, gene-regulating role as a hidden engine of treatment failure.
At the heart of this process, the team identified DDB1, an adaptor subunit of the CUL4A E3 ubiquitin ligase complex, as the master switch. DDB1 catalyzes K63-linked polyubiquitination of PD-L1 at the K185 residue, and this single modification proves to be doubly decisive. First, it counteracts acetylation at K263—a mark that normally bars PD-L1 from entering the nucleus—thereby freeing PD-L1 to bind vimentin and hitch a ride into the nuclear compartment. Second, and more surprisingly, the ubiquitin chain itself is structurally indispensable for nuclear PD-L1 to recognize and bind specific DNA sequences at the promoters of immune checkpoint genes (CD276, CD273) and NF-κB pathway genes (TRAF1, BIRC3, RELB). When the researchers mutated the K185 site (K185R), PD-L1 could no longer be ubiquitinated, failed to accumulate in the nucleus, and lost the ability to launch these transcriptional programs altogether. Cut&Tag and ChIP-qPCR analyses confirmed that the K185R mutant no longer bound DNA at these key genomic loci.
Crucially, the discovery points directly to a therapeutic opportunity. Thalidomide, an immunomodulatory drug already approved for multiple myeloma, is known to act on the CRL4 E3 ubiquitin ligase complex. The team showed that thalidomide disrupts the DDB1–PD-L1 interaction, suppresses K63-linked ubiquitination of PD-L1, and curtails its nuclear localization—effectively closing the door that lets PD-L1 into the nucleus.
The consequences in vivo were striking. In a 4T1 breast cancer mouse model, combining thalidomide with an anti-PD-1 antibody caused tumors to shrink dramatically, with complete regression in some animals—far outperforming either agent alone. Flow cytometry revealed that the combination reshaped the tumor microenvironment, boosting infiltration of Granzyme B⁺ cytotoxic T cells while depleting exhausted TIM3⁺PD-1⁺ T cells. Overall survival was significantly prolonged in the combination group.
"Our findings reveal that DDB1-mediated K63-linked ubiquitination acts as a dual switch for PD-L1—controlling both its entry into the nucleus and its function once it gets there," said Professor Lin. "By repurposing thalidomide to break this axis, we can attack PD-L1 on two fronts at once: blocking its classical role at the cell membrane with anti-PD-1 antibody, while silencing its hidden nuclear transcriptional program with thalidomide."
Science Bulletin