Activation and exhaustion of CD8 T cells in patients with chronic lymphocytic leukemia treated with ibrutinib and pembrolizumab.

Immune checkpoint blockade has been shown to restore anti-tumor T-cell function and elicit durable responses in select solid and hematopoietic malignancies. However, single-agent anti-programmed death 1 (PD-1) antibodies proved less efficacious in patients with chronic lymphocytic leukemia (CLL). In patients with high-risk or relapsed/refractory CLL, we conducted a phase 2 study testing the combination of lead-in ibrutinib and up to 2 cycles of fludarabine, followed by continuous therapy with ibrutinib and 17 cycles of pembrolizumab administered every 3 weeks. A total of 15 patients were enrolled. In 10 patients evaluable for response, we observed 1 complete response and 9 partial responses. There was no discernible benefit of the combination beyond what is expected from ibrutinib monotherapy. However, 3 weeks after the first dose of pembrolizumab, we detected CD8 T-cell proliferation in a subset of patients, whom we called "immune responders." In the responders, CD27-expressing CD8 T cells were relatively increased over immune nonresponders. Paired single-cell RNA and TCR sequencing revealed clonal expansion of activated GZMK+ CD8 effector memory and terminally differentiated effector cells. After 6 months of pembrolizumab treatment, the proportion of activated and proliferating CD8 T cells returned to baseline levels. Similarly, most novel clonotypes identified after 1 cycle of pembrolizumab decreased in frequency on long-term treatment. In summary, pembrolizumab did not improve the clinical response over ibrutinib monotherapy but transiently activated distinct clonotypes of CD8 T cells in a subset of CLL patients.