Enhancing CD19 Chimeric Antigen Receptor T Cells Through Memory-Enriched T Cells

In this issue of Clinical Cancer Research, Aldoss and colleagues demonstrate the excellent safety profile and therapeutic response of CD19 chimeric antigen receptor T (CAR-T) cells generated from naïve, stem, and central memory T cells in a phase I/II clinical trial in patients with relapsed/refractory (R/R) B-cell acute lymphoblastic leukemia (B-ALL; Fig. 1; ref. 1). These promising findings highlight the potential use of CD19 CAR-T cells generated from this T-cell population for high-risk patients that failed prior therapies. Moreover, these studies underscore the need to explore the use of CD19 CAR-T cell therapy early in the treatment course of B-ALL.

The treatment of ALL has traditionally involved three stages: induction of remission with standard chemotherapy, consolidation, and long-term maintenance. Optimization of chemotherapy regimens over the past 4–5 decades has made the treatment of pediatric B-ALL a success story (2, 3). More than 80% of children with B-ALL are now cured and 5-year survival rates have increased from 10% in the 1960s to 80%–90% at present (4–6). Unfortunately, these same treatment strategies have not recapitulated similar success for adults with B-ALL. In spite of the high response to induction chemotherapy, only 30%–40% of adults with B-ALL achieve remission with a reported 5-year survival rate of 68.6% in the United States (7–9). Furthermore, 50% of adults who achieve remission will eventually relapse (10, 11). However, advances in our understanding of disease pathogenesis, optimization of patient risk stratification, and introduction of targeted therapy and immunotherapy have improved therapeutic outcomes and overall survival (OS).

The propensity of ALL to relapse in adults after complete remission (CR) has been a major challenge. Treatment options have been historically limited to conventional chemotherapy leading to dismal response after salvage therapy. CR ranges from 35% to 41% after first salvage and decreases to 21% and 11% in the second and third/greater salvage, respectively. These poor outcomes have generated a need for novel therapies.

Fortunately, antibody-based and CAR T-cell therapies have made massive improvements in the management of B-ALL. The vast majority of mature and precursor B cells express CD19 and more than 90% express CD22. mAbs targeting these surface antigens can induce complement-dependent cytotoxicity, antibody-dependent cell-mediated cytotoxicity, or direct phagocytosis of malignant cells. Blinatumomab is a bispecific T cell–engaging antibody (BiTE) that targets CD3 T cells to CD19 B cells. This allows for cytokine release, T-cell proliferation, and generation of cytolytic T cells that induce B-cell apoptosis. Data from the phase III TOWER trial compared blinatumomab with standard-of-care chemotherapy in R/R Philadelphia-chromosome negative (Ph-negative) ALL confirmed the clinical benefit of this therapy (12). Blinatumomab resulted in higher overall response rates (ORR), OS, and remission in patients with R/R Ph-negative ALL compared with standard chemotherapy (12). Blinatumomab also conferred a survival benefit when used as a salvage therapy compared with chemotherapy. This trial was followed by the phase III ALCANTARA trial which expanded the use of this therapy to R/R Ph-positive ALL by demonstrating a response rate benefit in this ALL subtype (13)].

Inotuzumab Ozogamicin is a calicheamicin-linked antibody that targets CD22. Data from the phase III INO-VATE trial compared Inotuzumab Ozogamicin versus chemotherapy in patients with R/R Ph-negative ALL in salvage 1 or 2. CR, ORR, and PFS were significantly higher in the Inotuzumab Ozogamicin group and more patients proceeded to allogenic stem cell transplant.

Beyond these antibody and BiTE therapies, CAR T cells have further transformed the therapy of R/R B-cell ALL. The currently FDA-approved CAR T-cells used for B-cell ALL contain an extracellular domain that targets CD19 linked to a CD3 zeta intracellular domain, in addition to a costimulatory domain (CD28 or 4-1BB; Fig. 1; ref. 14).

Several clinical trials have been conducted using CAR-T cells for the treatment of ALL. These have demonstrated the robust therapeutic benefit of CAR-T cells, especially in patients that have failed multiple therapies, ultimately leading to the approval of two CAR-T cell therapies for R/R B-ALL: tisagenlecleucel (Kymriah) and brexucabtagene autoleucel (Tecartus). Current data from use of these products demonstrate durable CRs in patients with R/R disease after a single CAR-T cell infusion (14). Despite the significant CR achieved with these agents, approximately 50% of patients with R/R B-ALL eventually experience relapse (14). Moreover, CAR-T cell therapy is not without toxicities. These include two main toxicities, cytokine release syndrome and immune effector cell–associated neurotoxicity, each of which are more common in adult patients with ALL (14).

The success of tisagenlecleucel and brexucabtagene autoleucel has led to concerted efforts to improve the activity and safety of CD19 CAR-T cells. Two main strategies have been pursued: (i) changes to the antigen-presenting and costimulatory domains within CAR-T constructs as well as (ii) alterations in the T-cell populations transduced to generate CAR-T products. For this latter point, several T-cell populations have been used as a starting point for CAR transduction in clinical trials including unselected CD3⁺ T cells or more purified populations of central memory T cells, naïve T cells, and/or memory stem T cells (14).

In the current study, Aldoss and colleagues demonstrate that careful selection of the T-cell population and CD19-targeted CAR construct may play a role in improving the therapeutic efficacy and toxicity of CD19 CAR-T therapy (1). The authors conducted a phase I trial initially in two stages: first using CD19 CAR-T construct transduction of purified central memory T cells, and second using a less purified pool of naïve, stem, and central memory T cells. Importantly, the specific use of central memory-derived CAR-T cells yielded poor expansion and persistence and this portion of the trial was closed early due to lack of efficacy. In contrast, therapeutic efficacy and safety profile was observed with their use of a broader less differentiated T-cell population and 46 patients were ultimately treated on a phase II trial with the recommended phase II dose of this unique CD19 CAR-T product. These data thereby demonstrate that less differentiated T cells have an antileukemic effect similar to other populations used in CAR-T cell manufacturing in prior studies. At the same, there is theoretical potential for greater persistence of the antileukemic effect in patients compared with previously described clinical CD19 CAR-T products. Finally, the possible lower toxicity rate of the autologous CAR-T product here may imply that T-cell population selection for CAR-T cell manufacturing plays a role in CAR-T cell therapy side-effect profile.

The current study provides encouraging results for the therapeutic benefit of using naïve, stem, and central memory T cells to generate CD19 CAR-T cells in patients with high-risk R/R B-ALL including older adults. Currently, the only available CAR-T cell therapy for adults >25 years of age with R/R B-ALL is brexucabtagene autoleucel. The current study demonstrates that almost all adults with R/R B-ALL >50 achieved CR and few developed toxicity with the unique described CAR-T product. These findings could expand the arsenal of CAR-T cell therapies available to the adult B-ALL population. Furthermore, the high response observed in patients with high-risk cytogenetics and extramedullary site involvement is promising. Finally, the encouraging safety data demonstrate the potential use of CD19 CAR-T cells for salvage or even consolidation therapy in adults with B-ALL. Overall, this study argues that selection of the starting T-cell population used to generate CAR-T therapies may be a critical factor in future generation of clinical CAR T-cell therapies.

O. Abdel-Wahab reports grants from Loxo/Lilly and Nurix Therapeutics and personal fees from Pfizer Boulder, Envisagenics Inc., Harmonic Discovery Inc., and AIChemy outside the submitted work. No disclosures were reported by the other author.

O. Abdel-Wahab is supported in part by the Edward P. Evans Foundation, NIH/NCI (R01 CA251138 and R01 CA242020), NIH/NHLBI (R01 HL128239), NIH/NCI P50 CA254838-01, and the Leukemia & Lymphoma Society.