Explaining Resistance to CAR T Cells

A recent study identifies T-cell characteristics that predict whether patients with chronic lymphocytic leukemia (CLL) will respond to chimeric antigen receptor (CAR) T-cell therapy (Nat Med 2018;24:563–71). The results could help researchers refine CAR T-cell manufacturing and select patients most likely to benefit.

CAR T cells induce complete remission in more than 90% of patients with relapsed or refractory acute lymphoblastic leukemia (ALL) but only 26% of patients with relapsed or refractory CLL. Other studies have found that loss of CD19, CAR T cells' target, is a major cause of resistance in ALL. Why so many patients with CLL are treatment resistant has been unclear.

To find out, J. Joseph Melenhorst, PhD, of the University of Pennsylvania (Penn) in Philadelphia, and colleagues studied 41 patients with advanced CLL who had received CAR T-cell therapy. Their initial analysis ruled out several possible factors, including patients' age, prior therapy, and p53 status.

Next, patients were divided into four categories based on treatment results: complete remission, partial response, no response, or partial response after which the cancer evolved into aggressive B-cell lymphoma. The scientists then measured gene expression in T cells engineered to express the CAR protein and were ready to be infused into the patients.

T cells from patients who had complete remission and those whose cancer transformed into B-cell lymphoma showed increased expression of genes that favor early differentiation into memory cells. Why the response was so strong in patients whose cancers evolved isn't clear. In patients who had partial responses or did not respond, however, expression of genes involved in effector T-cell differentiation went up. Also, the activity of genes that indicate exhaustion or that promote apoptosis and aerobic glycolysis, which provides much of the energy for effector cells, also increased. These differences in gene expression were not the result of the CAR T-cell manufacturing process, the researchers determined.

The scientists found further evidence that an early memory T-cell phenotype correlates with a stronger response when they analyzed STAT3, which spurs memory cell differentiation and maintenance. STAT3 activity was higher in patients who had a complete remission or whose cancers evolved into B-cell lymphoma than in patients with partial or no responses. “We think that STAT3 signaling is really important for the cells' behavior in vivo,” says co-author Joseph Fraietta, PhD, also of Penn.

Another sign that the CAR T-cell treatment was likely to be successful was the presence of CD8+ T cells that carry CD27 but lack CD45RO, the scientists determined. These cells show characteristics of long-lived memory cells and were more than twice as abundant in patients who had complete remissions or whose disease evolved than in the other two groups.

The study is “a step forward in trying to understand what is a good CAR T product and what is a good starting [T-cell] population,” says Sattva Neelapu, MD, of The University of Texas MD Anderson Cancer Center in Houston, who wasn't connected to the research. “It provides some strategies to improve efficacy in CLL patients.”

Those strategies, Fraietta and Melenhorst note, could include testing patients' T cells before they undergo CAR T-cell treatment to select the most likely responders. In addition, modifying CAR T-cell manufacturing might increase the number of cells that promote a response. For example, it might be possible to bolster this subpopulation by modifying the mixture of cytokines used to stimulate them. –Mitch Leslie

For more news on cancer research, visit Cancer Discovery online at http://cancerdiscovery.aacrjournals.org/CDNews.