Rest, then rejuvenate: CD19 boost to improve CAR persistence Free

In this issue of Blood, Annesley et al demonstrate durable leukemia-free remissions after SCRI-CAR19 (autologous CD19.4-1BB chimeric antigen receptor T cell) in a multicenter phase 2 trial, with particularly striking results in patients with low disease burden. The authors also show that a novel approach to boost CD19 antigen can improve chimeric antigen receptor T-cell (CAR-T) activity in those predicted to have short persistence.¹ 

Despite the practice-changing successes of autologous CD19.CAR-T for relapsed and refractory (r/r) B-cell acute lymphoblastic leukemia, the ability of these cells to eradicate leukemia and maintain remission is contingent, regardless of the construct, on persistence.² How best to enhance persistence or reinduce CAR-T expansion when endogenous costimulation fails is one unknown that continues to plague the field. This question is especially pertinent given the widespread practice of treating patients with “low antigen load,” whether due to minimal leukemia burden, minimal residual disease (MRD) negativity, and/or prior bispecific engager or antibody-induced depletion of healthy B cells.

Multiple groups are investigating methods of boosting expansion and persistence, including providing exogenous CD19 to prevent early loss of B-cell aplasia (BCA, a commonly used surrogate for CD19.CAR-T persistence). To date, strategies to enhance persistence can be grouped into intrinsic modifications to the CAR itself or extrinsic modifications to improve costimulation and antigen presentation such as (1) cytokine/chemokine expression, (2) gene editing to regulate CAR-T activation/exhaustion, (3) small molecule regulation of CAR-T differentiation, (4) CAR stimulation through the native T-cell receptor, (5) vaccination, and (6) provision of additional costimulation through alternative antigen presentation.^3-5 

In their study, the authors infused up to 6 doses of autologous artificial antigen-presenting cells (APCs) possessing a truncated CD19 (CD19t-APC) every 4 weeks to enhance antigen presentation in a subset of patients with loss of CAR-T persistence after infusion of SCRI-CAR19. Patients receiving CD19t-APCs had either low CD19 marrow antigen burden prior to start of lymphodepletion or rapid early contraction of SCRI-CAR19, both of which were associated with a higher risk of early CAR loss in phase 1 trials of this product.⁶ A third group included patients undergoing reinfusion of SCRI-CAR19 for prior CAR loss. Half of those treated with SCRI-CAR19 had low disease burden, including 26% without measurable leukemia.

Outcomes for the study were outstanding: 85% of patients with measurable disease achieved MRD negativity; central nervous system disease was cleared in 14 of 14 patients; and leukemia-free survival (LFS) in the group with low disease burden exceeded that of previous reports.⁷ Despite lower response rates in patients with high disease burden, 82% achieved MRD⁻ complete response (vs 100% MRD⁻ complete response rate in patients with low disease burden). Improved LFS and EFS in the 12 primary refractory subjects was likely related to lower disease burden and perhaps earlier treatment of this cohort compared to those with r/r disease. Thirty-eight percent of patients treated in this trial went on to consolidative allo-hematopoietic stem cell transplantation (allo-HSCT) after SCRI-CAR19, most (19 of 27 patients) due to loss of BCA. Twelve of the 18 relapses were CD19⁻ (median time to relapse 5 months), whereas median time to CD19⁺ relapse was 12 months.

Compared with the prior Pediatric and Young Adult Leukemia Adoptive Therapy (PLAT) studies using the same construct, subjects in this cohort had lower disease burden, muddling the effect of the minor manufacturing change⁸ on improved outcomes. However, better understanding of the factors that contribute to relapse after CAR-T likely influenced evidence-based allocation of patients to consolidative allo-HSCT, contributing to superior LFS overall. We also posit that the longer median persistence of CAR-T (1 year vs the 3 months in the previously reported phase 1 cohort) could be related to omission of the CAR selection step, which may have enriched for a more exhausted CAR phenotype that limited persistence.⁹ 

Although the PLAT-02 efficacy outcomes are remarkable, the novelty of this investigation lies in the PLAT-03 cohort. Introducing exogenous antigen to enhance CAR activation, expansion, and persistence has been explored clinically, but infusing engineered APCs expressing the CAR antigen as a separate product to patients meeting predefined criteria is both novel and attractive. Although not designed to compare rates of persistence between groups, the authors compared outcomes with and without T-APC boost. Manufacture and infusion of CD19t-APC were feasible and well-tolerated; however, it is important to note that even repeat CD19t-APC infusion did not significantly improve persistence for subjects with low antigen burden. Nevertheless, those who received CD19t-APCs had a <20% incidence of early CAR loss, compared with 46% in patients who did not. In addition, an ad hoc comparison of patients who met the criteria for but did not receive CD19t-APCs demonstrated a significant improvement in persistence. The authors also report a decreased probability of early CAR loss that bordered on statistical significance (P = .05) in patients with rapid early contraction who received T-APCs.

These exciting data raise additional questions about the ideal time to “recharge” CAR-T with exogenous antigen. Patients in the study who received CD19t-APCs later (“rapid early contraction” cohort) had more encouraging results than those who received them earlier (low antigen burden). Because patients were required to have ongoing BCA to be eligible for CD19t-APCs, circulating SCRI-CAR19 may have been less “responsive” to antigen boost, perhaps exhibiting an exhaustion phenotype. Did CAR-T exhibiting longer persistence possess a “more responsive” memory phenotype primed for re-expansion upon restimulation? How much B-cell antigen is enough? Is it possible that providing “too much” exogenous antigen leads to antigen-induced CAR-T exhaustion and limits persistence? The 4 weeks between CD19t-APC infusion likely provided periods of rest and rejuvenation, preventing CAR-T cells from entering a state of exhaustion. Would delivery of T-APCs be more beneficial for a CAR-T product with less antigen-negative relapse? Could CD19t-APCs obviate the need for consolidative transplant or other “maintenance” therapies in patients at high risk of relapse due to early loss of BCA? A future randomized study in which the decision to administer T-APC is made prior to CAR infusion could answer these remaining questions. In addition, correlative analyses further defining the kinetics of tAPC-induced persistence, phenotype of reinvigorated SCRI-CAR19, and impact on cytokine secretion will be vital to understanding how much, for whom, and when exogenous antigen stimulation is required.

Annesley and colleagues have presented a valuable proof-of-principal for the use of exogenous antigen as a mechanism to reinvigorate CAR-T therapy. Although the most practical application of this novel engineered approach remains to be determined, the possibilities are endless.

Conflict-of-interest disclosure: R.H.R. has received research funding from the Leukemia Lymphoma Society and the National Institutes of Health, prior honoraria from Novartis, and consulting fees from Pfizer.