Reasons for CAR T-cell therapy failure and potential management strategies

Chimeric antigen receptor (CAR) T-cell therapy has been in the spotlight due to its ability to achieve lasting remissions in 40-50% of patients with large B-cell lymphoma (DLBCL), who relapse after³ two lines of conventional therapy.¹ To date, most of the clinical research has focused on patients who achieve remission after CAR T-cell therapy and not on the remaining 50-60% who relapse following treatment failure.¹ Little is known about the reasons why CAR T-cell therapy fails in patients with large B-cell lymphoma and even less about how to manage these patients.

Michael Byrne from Vanderbilt School of Medicine, Nashville, US, and colleagues published in Biology of Blood and Marrow Transplantation¹ a comprehensive review exploring the potential reasons for relapse following CAR-T and the management strategies for this population of patients with large B-cell lymphoma.

According to the authors, the main reasons for CAR T-cell failure can be classified into three groups: a) tumor intrinsic factors b) host-related factors and c) inadequacy of CAR T-cell therapy, which are described below.

Tumor intrinsic factors

Changes in epitope expression on the surface of DLBCL cells is one potential reason for CAR T failure. For CAR T cells to work they need to bind to the CD19 epitope on tumor cells. The loss of CD19 from the cell surface of tumor cells has been reported in 10-20% of patients with acute lymphoblastic leukemia (ALL) and occurs via multiple mechanisms, including the acquisition of secondary CD19 mutations or exon variants that render CD19 non-functional. Interestingly, this CD19 epitope loss was detected in 27% of patients with DLBCL that were treated with the CAR-T product axicabtagene ciloleucel (axi-cel) during the ZUMA-1 clinical trial.²

Apart from epitope loss, the microenvironment is crucial for CAR T cell proliferation and survival. There is a possibility of elevated expression of inhibitory signals around CAR T cells, which can cause suppression and render the therapy ineffective. In support of this theory, patients with high-grade lymphoma that express high levels of the ligand of the T cell surface inhibitor receptor programmed death 1 (PD-1; PD-L1) have poor clinical outcomes. Moreover, accumulating evidence indicates that solid tumors upregulate inhibitory proteins that lead to T cell suppression.

Lastly, high tumor burden has been associated with CAR T failure in patients with ALL but not in patients with chronic lymphocytic leukemia (CLL). Preliminary data from ZUMA-1 indicate that high tumor burden might be linked to inferior outcome in patients with DLBCL.

Host-related factors

Patient baseline characteristics, including number or type of previous lines of treatment, might influence patients’ capacity to respond to CAR T treatment. However, both in the ZUMA-1 and JULIET (tisagenlecleucel³) trials no statistically significant differences in response outcomes were observed between patients with different numbers of prior lines of treatment or in different molecular subgroups, respectively. Nevertheless, the sample sizes of these trials are considered small to reach an interpretable conclusion.

Currently, there are different conditioning chemotherapy protocols used for T cell depletion prior to CAR T cell infusion. There is evidence that conditioning with cyclophosphamide and fludarabine is associated with more favorable cytokine profiles and a lower risk of progression-free survival events. This indicates that indeed the type of conditioning therapy may affect CAR T response outcomes. Nevertheless, whether or not CAR T cell persistence affects remission duration is still unknown as around 75% of long-term responders experience B-cell recovery.

Another host-related factor that the authors consider in their review is CAR T cell kinetics. Evidence from ZUMA-1 indicates that high numbers of CAR T cells (higher expansion kinetics) are associated with better response outcomes. However, this was not verified by the JULIET trial, where no significant differences between CAR-T cell levels and responses were observed. Interestingly, ALL and CLL patients who respond to CAR T-cell therapy seem to have high transgene levels in the periphery and higher maximum serum transgene levels.

Inadequacy of CAR T-cell therapy

The structure of the CAR T construct along with its qualitative characteristics may also influence response outcomes. CAR T manufacturing, changes in tumor microenvironment, previous treatments or the effects of neighbouring cells can cause ‘CAR T cell exhaustion’. Data indicate that ‘exhausted’ CAR T cells are not as proliferative or potent as their ‘non-exhausted’ counterparts. This ultimately affects their efficacy and thus the patient’s response to therapy. Moreover, the translational profile of the infused CAR T cells could contribute to response outcomes. It has been reported that CAR T cells from ALL patients who are able to respond to the therapy express a higher number of immunity-related genes.

Proposed treatment strategies for patients who do not respond or relapse following CAR-T

In this review, different treatment strategies are recommended for the management of patients who do not respond to CAR T-cell therapy. The authors classify the three following patient subgroups of ‘non-responders to CAR T’ and propose management strategies accordingly: a) patients that do not respond at all to CAR T infusion (primary resistance), b) patients who respond initially but relapse within < 3 months of infusion (early relapse), and c) patients who respond initially and relapse late, 3 months after infusion (late relapse).

Primary resistance

 For these patients with progressive disease, the authors recommend further treatment with cytotoxic chemotherapy and allogeneic stem cell transplantation for those achieving remission. Supportive care can be given for those with multiple prior treatments. Moreover, enrolment in clinical trials that are designed for high-risk patients should be considered for this patient population. The best management strategy for this patient population remains unclear.

Early relapse

Patients who respond to CAR T treatment but relapse early within <3 months following CAR T infusion are advised to follow-up with standard salvage strategies like chemotherapy, radiation and/or immunomodulatory drugs (IMiDs).  Off-label administration of the anti-PD-1 antibodies pembrolizumab and nivolumab has shown promising activity in this patient subgroup when given within a month after CAR-T infusion. Interestingly, both inhibitors improved the expansion and function of the infused CAR T cells and increased immunity, leading to good clinical responses. Another study reported an overall response rate (ORR) of 27% following pembrolizumab administration in patients with CAR T failure. Clinical enrolment in an appropriate, well-designed trial should also be considered.

Late relapse

Patients who achieve a good response to CAR-T therapy but relapse ³3 months after infusion are advised to firstly undergo a biopsy to determine whether CD19 is still an available target. If CD19 is still persistent, then a repeated CAR-T infusion may lead to clinical responses, as seen in nine late-relapsing patients in ZUMA-1. Nevertheless, further clinical studies are needed to validate the safety and efficacy of a second CAR T infusion of the same or another transgene. Standard salvage regimens like radiation, IMiDs or ibrutinib treatment should also be considered depending on tumor characteristics and location. In patients with CLL, ibrutinib administration prior to CAR T infusion led to better clinical outcomes and higher CAR T expansion rates. Pre-clinical data from animal models also indicate that IMiD therapy with lenalidomide might enhance CAR T efficacy. Last but not least, accumulating preclinical and clinical evidence points to the safe use of radiation prior to CAR T infusion as an effective disease control strategy. Nevertheless, its use after CAR T failure remains uncertain.

Conclusion

To date, most of the clinical research has focused on patients responding to CAR T-cell therapy and not on the ones for which treatment fails. There is inadequate information on why patients do not respond to CAR T-cell therapy and little is known about how to best manage these patients. Future clinical trials designed for this patient population are needed in order to identify the key influencers of CAR T success and the best treatment strategies to address this unmet medical need.

For more details on the pivotal CAR-T trials for DLBCL read here.