Researchers at the University of California, Los Angeles, have shed light on mechanisms by which melanoma cells become resistant to PD-1 blockade with pembrolizumab. One mechanism involves loss-of-function mutations in JAK1 or JAK2 that abolish interferon-γ signaling; the other involves a truncating mutation in B2M that results in defective antigen presentation.

A study from researchers at the University of California, Los Angeles (UCLA), sheds light on mechanisms by which melanoma cells become resistant to PD-1 blockade (N Engl J Med 2016 July 13 [Epub ahead of print]).

“Delayed relapse, or secondary resistance, occurs in about 25% of patients who initially respond well to immune checkpoint inhibition,” says senior author Antoni Ribas, MD, PhD, director of UCLA's tumor immunology program. “Why this happens is largely unknown.”

Ribas's group studied four patients with metastatic melanoma who achieved objective responses to pembrolizumab (Keytruda; Merck), but who experienced tumor recurrence after a median time of nearly 1.8 years, despite continuous therapy. The researchers carried out whole-exome sequencing of tumor tissue obtained from these patients before treatment and after disease progression. The recurrent tumors of two patients had loss-of-function mutations in JAK1 and JAK2, respectively; a third had a truncating mutation in B2M.

“These were newly acquired, homozygous mutations,” says Jesse Zaretsky, the study's first author. “To mutate one allele, lose the other wild-type copy, and duplicate the mutated allele involved considerable work on the tumor's part, suggesting strong selective pressure after checkpoint inhibition.” None of these alterations were found in the fourth patient's recurrent tumor, he adds.

In the first two patients, melanoma cells became “selectively deaf” to antiproliferative signaling through interferon-γ, which requires intact JAK1 and JAK2, Zaretsky says. On balance, such growth-curbing effects are usually acceptable to a tumor because interferon-γ signaling also upregulates PD-L1, enabling evasion of immune scrutiny, Ribas says. However, with pembrolizumab added to the picture, “continued PD-L1 production would be futile,” he explains, “and the other associated effects of interferon-γ signaling would become a liability. We think these tumors figured out that abolishing this pathway's activity altogether, by inactivating JAK1 or JAK2, was more advantageous.”

B2M's protein product, beta-2-microglobulin, is necessary for the proper folding and transport of MHC class I molecules, which present foreign antigens to the immune system, to the cell surface. Consequently, cytotoxic T cells with mutant B2M no longer recognized and targeted the melanoma cells for destruction, Zaretsky says, even with continued PD-1 blockade exposing the tumor to immune surveillance.

Although the researchers did not find any mutations in the fourth patient that were clearly linked to secondary resistance, Ribas points out that other nongenetic immune escape mechanisms may exist. Meanwhile, a larger sample size is needed to determine the frequency of JAK1/2 and B2M mutations in pembrolizumab-resistant patients, he adds.

Thomas Gajewski, MD, PhD, an immunologist at the University of Chicago in Illinois, thinks this study raises “several interesting next-level questions,” including precisely how the genetic variants identified by Ribas's group arise, and whether secondary resistance mechanisms leading to new lesions—not just recurrence at existing metastatic sites—are similar or different.

“Oncolytic viruses could be a therapeutic option in the JAK1/2-mutant setting; they're thought to preferentially replicate in tumor cells with blunted interferon-mediated signaling,” Gajewski observes. “Overall, though, it's important to consider combination immunotherapy early in the disease course. Eliminating the majority of tumor cells quickly, before extensive genetic variants emerge, may prevent secondary resistance from ever happening.” –Alissa Poh

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