Humanizing CAR T-cell therapy in B-ALL

In this issue of Blood Advances, Gauthier et al¹ report on a phase 1 clinical trial conducted at Fred Hutchinson Cancer Research Center which tested a humanized, CD19-directed chimeric antigen receptor (CAR) T-cell therapy in adults with relapsed or refractory B-cell acute lymphoblastic leukemia (B-ALL). They found that this CAR was safe but associated with limited durable remissions.

CD19-directed CAR T-cell therapy has expanded therapeutic opportunities for children and adults with relapsed/refractory (R/R) B-ALL.^2,3 Despite excellent response rates, prolonged remissions are observed in fewer than half of adults, underscoring the need for novel approaches to improve durability. One avenue involves modifications of the CD19 single chain variable fragment (scFv) binder. Brexucabtagene autoleucel and tisagenleucleucel, both approved in the United States for R/R B-ALL, use an identical high affinity murine FMC63 scFv,^2,4 whereas, obecabtagene autoleucel, a CD19 CAR T-cell therapy recently approved in the United States for R/R B-ALL in adults, uses a novel “fast-off” murine scFv that was shown to reduce T-cell exhaustion and promote CAR persistence in preclinical models.^5,6 

JCAR021 is an autologous CD19 CAR T cell with a 4-1BB costimulatory domain which is administered at a defined ratio of CD4⁺ and CD8⁺ CAR T cells.^1,7 The novelty of the construct is that rather than employing the traditional murine scFv, this CAR uses a fully humanized scFv which is generated by isolating anti-human CD19 scFv from human antibody/DNA libraries. The trial enrolled 23 adults with R/R B-ALL with a median age of 45 (range, 36-58) years. The median blast count was 7.7% at enrollment and patients had received a median of 3 prior lines of therapy including allogeneic stem cell transplant (alloHCT) (65%), blinatumomab (57%), and/or inotuzumab (52%). Patients received escalating CAR T-cell dose ranging between 7 × 10⁴ and 7 × 10⁶; median vein-to-vein time was 20 days. The median relapse free survival (RFS) and overall survival were 5.5 and 23 months, respectively. Cytokine release syndrome (CRS) and neurotoxicity occurred in 83% and 52% of patients, respectively. None of the patients developed grade ≥3 CRS, whereas 35% developed grade ≥3 neurotoxicity.

This phase 1 clinical trial builds on the growing literature of novel CAR T-cell therapies designed to improve persistence and durability of response. The investigators, and others in the field, have hypothesized that use of a humanized scFv may overcome immune rejection observed in CARs utilizing murine scFv, thus improve CAR T-cell function and persistence.^8-10 It is difficult to discern whether the investigators proved their hypothesis in this relatively small trial. CAR persistence was not clearly improved relative to this group’s prior CAR19 approaches in B-ALL,⁷ durable remissions were rare, and most relapses were CD19⁺, suggesting lack of CAR durability and ongoing efficacy. Other reports of humanized CARs in B-ALL have demonstrated varying results. In a phase I study of humanized CAR T of 52 patients with R/R B-ALL, Shi et al¹⁰ observed comparable efficacy to murine CAR T-cell constructs, with a median event free survival of 10 months with the majority of relapses expressing CD19. In contrast, Myers et al⁹ reported more encouraging results of humanized CAR T-cell therapy in a single-arm pilot study in children and young adults with R/R B-ALL. Unlike the Gauthier and Shi trials, this study included patients who were previously treated with murine-derived CD19 CAR T cells. Among patients who achieved complete response/complete response with incomplete count recovery in the retreatment cohort (n = 33; 45%), the 12 months RFS and cumulative incidence of relapse were 74% and 24%, respectively. However, 10 of these patients received repeated infusions which may have contributed to CAR T-cell persistence and the trial cohort was predominantly children and adolescents rather than adults.⁹ 

The collective data suggest that it is too early to know whether humanized CAR T cells offer benefits over murine constructs in B-ALL. To overcome relapse seen with autologous CD19 CAR T cells, other strategies have been investigated, including multiantigen targeting CAR T cells,^11,12 CD19 followed by CD22 CAR T-cell approaches,¹³ reinfusion of CD19 CAR T cells to promote durable B-cell aplasia (ClinicalTrials.gov identifier: NCT04605530), novel targets including B-cell activating factor receptor and thymic stromal lymphopoietin receptor, post–CAR T-cell stimulation with cytokines,¹⁴ “off the shelf” allogeneic CAR T cells, and donor derived allogeneic CAR T-cell constructs (ClinicalTrials.gov identifier: NCT05507827). None of these approaches have been studied in large comparative trials, and small, single center reports are inconclusive. Additional fundamental questions remain, including whether reinfusion of CAR T cells (humanized or not) provide genuine clinical benefit in B-ALL, and if so, whether functional measurements of persistence such as loss of B-cell aplasia can be used to guide reinfusions strategies prevent relapse.

This has been a truly exciting time for B-ALL with the availability of several effective agents in both the upfront and R/R setting. Yet, most adults with relapsed B-ALL will ultimately relapse again after salvage therapy, indicating a true need for durable immunotherapies in the R/R setting. CAR T cells offer a potential advantage in this respect, but refinements are required to realize this potential. Gauthier et al demonstrate that manufacturing a fully humanized scFv CAR T-cell product is feasible with a reasonable safety profile. Future studies will certainly build on these efforts and determine whether humanized CAR T cells are advantageous.

Conflict-of-interest: The authors declare no competing financial interests.