Researchers may have discovered why some patients with advanced chronic lymphocytic leukemia (CLL) don’t respond to chimeric antigen receptor (CAR) T-cell therapy.

The team found that CLL patients with elevated levels of “early memory” T cells prior to receiving CAR T-cell therapy had a partial or complete response to treatment, while patients with lower levels of these T cells did not respond.

The early memory T cells were marked by the expression of CD8 and CD27, as well as the absence of CD45RO.

The researchers validated the association between the early memory T cells and response in a small group of patients, predicting with 100% accuracy which patients would achieve a complete response.

Joseph A. Fraietta, PhD, of the University of Pennsylvania in Philadelphia, and his colleagues reported these findings in Nature Medicine. This research was supported, in part, by Novartis.

For this study, the researchers retrospectively analyzed 41 patients with advanced, heavily pretreated, high-risk CLL who received at least 1 dose of CD19-directed CAR T cells.

Consistent with the team’s previously reported findings, they were not able to identify patient or disease-specific factors that predict who responds best to the therapy.

Therefore, the researchers compared the gene expression profiles and phenotypes of T cells in patients who had a complete response, partial response, or no response to therapy.

The CAR T cells that persisted and expanded in complete responders were enriched in genes that regulate early memory and effector T cells and possess the IL-6/STAT3 signature.

Non-responders, on the other hand, expressed genes involved in late T-cell differentiation, glycolysis, exhaustion, and apoptosis. These characteristics make for a weaker set of T cells to persist, expand, and fight the CLL.

“Pre-existing T-cell qualities have previously been associated with poor clinical response to cancer therapy, as well differentiation in the T cells,” Dr Fraietta said. “What is special about what we have done here is finding that critical cell subset and signature.”

Elevated levels of the IL-6/STAT3 signaling pathway in these early T cells correlated with clinical responses to CAR T-cell therapy.

To validate these findings, the researchers screened for the early memory T cells in a group of 8 CLL patients, before and after CAR T-cell therapy. The team identified the complete responders with 100% specificity and sensitivity.

“With a very robust biomarker like this, we can take a blood sample, measure the frequency of this T-cell population, and decide with a degree of confidence whether we can apply this therapy and know the patient would have a response,” Dr Fraietta said.

“The ability to select patients most likely to respond would have tremendous clinical impact, as this therapy would be applied only to patients most likely to benefit, allowing patients unlikely to respond to pursue other options.”

These findings also suggest the possibility of improving CAR T-cell therapy by selecting for cell manufacturing the subpopulation of T cells responsible for driving responses. However, this approach would come with challenges.

“What we’ve seen in these non-responders is that the frequency of these T cells is low, so it would be very hard to infuse them as starting populations,” said study author J. Joseph Melenhorst, PhD, also of the University of Pennsylvania.

“But one way to potentially boost their efficacy is by adding checkpoint inhibitors with the therapy to block the negative regulation prior to CAR T-cell therapy, which a past, separate study has shown can help elicit responses in these patients.”

The researchers also noted that it’s unclear why some patients’ T cells are suboptimal prior to treatment. However, the team believes this could have to do with prior therapies.

Future studies with a larger group of CLL patients should be conducted to help answer these questions and validate the findings from this study, the researchers said.

Predicting response to CAR T-cell therapy in CLL