In a study of patients with metastatic melanoma, researchers reported that higher numbers of cytotoxic T cells positive for both PD-1 and CTLA-4 in the tumor predicted response to anti–PD-1 therapy and improved progression-free survival.

Immune checkpoint blockers pose a vexing problem: They don't work in many eligible patients. However, according to a recent study in melanoma, patients whose tumors have an abundant population of CD8+ cytotoxic T cells expressing both PD-1 and CTLA-4 are more likely to respond to these drugs (J Clin Invest 2016;126:3447–52).

Normally CD8+ T cells express little PD-1. However, persistent antigen stimulation drives up PD-1 expression in tumor-associated T cells. Continuous signaling through PD-1 can leave T cells “exhausted” and unable to function. Anti–PD-1 therapies expose tumors to immune attack by blocking interactions between PD-1 and its ligands expressed in the tumor microenvironment.

Oncologists can gauge whether patients are good candidates for anti–PD-1 therapy by assessing their tumors' PD-1 and PD-L1 expression levels through immunohistochemistry. The trouble is that histology is sometimes subjective: A protein's expression may be judged “high” by one pathologist and “medium” by another. “It can come down to an opinion,” says study leader Michael Rosenblum, MD, PhD, of the University of California, San Francisco. Furthermore, histology analyzes only a thin section of the tumor. Flow cytometry is more objective, quantitative, and comprehensive.

Rosenblum's team performed flow cytometry on fresh metastatic melanoma samples from 20 patients, assessing the relative quantities of tumor-infiltrating effector, regulatory, and cytotoxic T cells, as well as PD-1, PD-L1, CTLA-4, and MHC class II expression in these subsets. The patients were then started on the PD-1 inhibitors pembrolizumab (Keytruda; Merck) or nivolumab (Opdivo; Bristol-Myers Squibb). Two years later, the researchers correlated the patients' tumor immune profiles with their clinical responses. In this discovery cohort, the combination of PD-1 and CTLA-4 emerged as the best predictor of response. Patients whose tumors had at least 20% of CD8+ T cells expressing these two markers had a median progression-free survival of 31.6 months, versus 9.6 months for patients with less than 20% of this subtype among their tumor CD8+ cells.

The predictive value of PD-1 and CTLA-4 held up in a separate validation cohort of 20 more patients with melanoma. If less than 20% of their tumor's CD8+ cells expressed these two markers, the patient “had no chance of responding” to monotherapy with pembrolizumab or nivolumab, says Rosenblum. Meanwhile, individuals “had about an 80% chance of responding” if at least 30% of their tumor-infiltrating CD8+ cells were positive for both markers. As such, patients whose PD-1 and CTLA-4 levels fall below 30% could be given pembrolizumab plus ipilimumab (Yervoy; Bristol-Myers Squibb) “right off the bat,” Rosenblum adds. “The risk–benefit there probably favors double therapy.”

In functional assays, this PD-1– and CTLA-4–high population behaved like “partially exhausted” T cells that can produce some cytokines but not others. The researchers verified that these T cells could not produce TNFα and IL2 and that checkpoint blockade successfully reactivated the cells.

The findings “have the potential to be very useful,” says Roy Herbst, MD, PhD, of Yale School of Medicine in New Haven, CT, who previously reported that CTLA-4 mRNA expression in tumors predicted response to PD-L1 blockade. However, he adds, it's not always possible to obtain fresh tumor tissue—a requirement for flow cytometry—and the technique demands a high level of expertise. –Esther Landhuis

For more news on cancer research, visit Cancer Discovery online at