This article has been corrected to remove chemotherapy from the SKYSCRAPER-01 protocol.
NEW YORK – Researchers from the Mayo Clinic and Case Western University may have found a way to halt the falling remission rates seen with CD19-directed CAR T-cell therapy in refractory non-Hodgkin lymphoma patients.
According to a study published earlier this month in Cancer Discovery, the high rates of long-term relapse — around 60 to 70 percent — observed in real-world NHL patients on CD19-directed CAR T-cell therapies are likely due to exhaustion of the cells.
In the study, researchers led by David Wald, a professor of pathology at Case Western Reserve University, and Tae Hyun Hwang, a professor of artificial intelligence at Mayo Clinic, identified a T-cell immunoreceptor, TIGIT, as a biomarker for CAR T-cell exhaustion that if blocked could improve remission rates and even offer a way to identify which patients are likely to relapse on CAR T-cell treatment.
Hwang said the problem of diminishing remission rates with CAR T-cell therapy is particularly challenging because when patients relapse on these treatments, which typically cost over $400,000, there's nothing left to give them. "Once patients relapse with this cellular therapy, because this is the last resort, it's hard to think about how to treat them," said Hwang.
The US Food and Drug Administration granted accelerated approval to Gilead Sciences' Yescarta (axicabtagene ciloleucel) in 2017 as a treatment for patients with certain types of refractory large B-cell lymphomas, including diffuse LBCL, based on results from the ZUMA-1 Phase I/II trial. At the time of accelerated approval, 51 percent of patients had a complete response rate on Yescarta. Two-year follow-up data from ZUMA-1 pushed up the complete response rate up to 58 percent, and the objective response rate was 82 percent.
Even longer follow-up data on another CAR T-cell therapy approved for relapsed or refractory LBCL, Novartis' Kymriah (tisagenlecleucel), showed that 60 percent of patients with refractory DLBCL and follicular lymphoma had a sustained response at five years.
Wald, Hwang, and colleagues wanted to figure out why more patients weren't experiencing long-term remissions on these treatments. They decided to look for resistance mechanisms by comparing transcriptional and phenotypic profiles of CAR T cells that previously treated refractory NHL patients received within a clinical trial. They performed single-cell RNA sequencing on 24 cryopreserved samples from 13 patients taken 14 and 30 days after CAR T-cell infusion.
The analysis generated 133,000 single-cell expression profiles adding up to over 30 terabytes of data. Using computational approaches, they were able to link RNA and protein expression patterns associated with CAR T-cell treatment at the single-cell level. Hwang said that while sequencing samples from 13 patients might not seem like a "big data" approach, it definitely was. "We're not talking about just the number of patients, we're talking about the number of cells," Hwang said. "The complexity of this data is very high compared to traditional bulk RNA-seq or bulk exome sequencing."
The study revealed that upregulation of several transcription factors and two inhibitory checkpoint receptors, TIGIT and PD-1, were involved in T-cell exhaustion in patients who turned out to be poor responders compared to those who responded well. The marker that best separated responders and nonresponders was TIGIT.
T-cell exhaustion is a well-characterized process in which T cells progressively lose function in response to chronic antigen stimulation. First described in the context of viral infection, it has more recently been observed in connection with tumors. This study suggests that TIGIT and other markers of exhaustion are already present in CAR T cells 14 or 30 days after infusion and may be useful for identifying which patients will eventually relapse.
The researchers went on to investigate whether blocking TIGIT could improve the tumor-fighting function of cells using in vitro and in vivo models. An in vitro model that recapitulates the process of exhaustion in CAR T cells demonstrated that TIGIT blockade was effective. A study of NHL further showed that CAR T-cell efficacy improved with the addition of a TIGIT antibody.
The investigators are now hoping to soon launch a clinical trial to explore combining a TIGIT blocking antibody with CAR T-cell therapy to prolong the effectiveness of the treatment. A number of anti-TIGIT cancer therapies are in development, but so far have not had much success. In March, a combination of Roche's anti-TIGIT drug tiragolumab and PD-L1 inhibitor Tecentriq (atezolizumab) failed to beat Tecentriq in the Phase III SKYSCRAPER-01 trial in patients with PD-L1-high advanced non-small cell lung cancer.
The in-depth single-cell analysis "is actually bringing some good news to the field," said Hwang. "Our study can give hope that a TIGIT and CD19 CAR-T combinatorial therapy could benefit the patient." He and his team still need to discuss details of a potential clinical trial with a pharmaceutical industry partner.
Meanwhile, Mayo Clinic and Case Western researchers are also planning a larger prospective study to validate TIGIT as a marker of CAR T-cell exhaustion. The aim there is to generate single-cell sequencing data from about 100 patients from multiple institutions and gauge whether TIGIT can predict if they'll relapse to CAR T-cell therapy. If that study is successful, there's potential to develop TIGIT as a prognostic or predictive biomarker in NHL.
The recently published data also points to the broader utility of single-cell sequencing not only as a clinical research technology but also as a precise patient management tool. Single-cell sequencing is important for detecting a signal of cancer relapse in the present context, Hwang explained, because all of a patient's CAR T cells would not necessarily express high levels of TIGIT, but there may be subclonal populations of CAR T cells that do.
Ultimately, TIGIT is just one possible predictive marker among many possible markers. "There are many other interesting potential genes that are possibly associated with CAR T-cell dysfunction or exhaustion," Hwang said. "That's our ongoing work."