CAR-DLI for ALL, but not without lymphodepletion Free

In this issue of Blood, Roddie et al¹ present data from a phase 1 trial demonstrating that lymphodepletion (LD) enhances the efficacy of subsequent matched donor-derived anti-CD19 chimeric antigen receptor T cells (CAR-Ts) in adult patients with B-cell acute lymphoblastic leukemia (B-ALL) relapsed after allogeneic hematopoietic stem cell transplantation (alloHSCT).

Despite advances with the addition of therapeutic antibodies and tyrosine kinase inhibitors (TKI) for primary treatment as well as CAR-T therapy for relapsed disease, alloHSCT still remains a mainstay for relapsed or refractory B-ALL. And then again, despite improved post-alloHSCT approaches like TKI maintenance or preemptive donor lymphocyte infusions (DLIs), relapse after alloHSCT poses a risk for a substantial number of patients. Strategies for improving treatment in these patients are needed.² 

This need is addressed in the trial presented by Roddie et al in which 14 patients with B-ALL relapse after alloHSCT were treated with anti-CD19 CAR-Ts, termed as CAR-DLI by the authors. The study stands out for several reasons (see figure).

For commercial anti-CD19 CAR-T products, LD with fludarabine/cyclophosphamide was mandatory in the approval trials. However, in patients who underwent alloHSCT, LD may result in higher graft-versus-host disease (GVHD) risk as well as prolonged cytopenia. Previously, CAR-T therapy with prior LD for B-ALL relapse after alloHSCT had been evaluated in pediatric³ as well as adult⁴ patients yielding promising results. The study by Roddie et al is the first to directly compare the role of LD in adult patients in this situation. Half of the patients received only CAR-DLIs whereas the other half were treated with LD prior to CAR-DLIs, which is an extremely interesting aspect of this study. Application of LD was not only associated with substantially improved CAR-T engraftment and persistence. It also resulted in improved overall survival at 12 months of 83.3% for LD-treated patients compared to 28.6% for patients treated with CAR-DLIs alone. Moreover, no relevant GVHD or immune effector cell-associated neurotoxicity syndrome was observed in both groups and only 3 cases of >1° cytokine release syndrome, interestingly all in the CAR-DLI-alone group, were seen. Despite the small number of patients included, these data are sufficiently encouraging to routinely apply LD prior to matched donor-derived anti-CD19 CAR-Ts in adult patients with B-ALL relapse after alloHSCT, given the extreme difficulty in conducting randomized trials in this patient population.

For the study, anti-CD19 CAR-Ts were produced at an academic center in a point-of-care (POC) approach. T cells were apheresed not from the patients themselves—as performed for all available commercial products—but from the respective original allogeneic donor. In this sense, it rightfully resembles traditional DLIs. This is relevant because apheresis from the healthy allogeneic donor avoids risk of blast contamination as well as harvest of exhausted T cells, affected by post-SCT toxicities and relapse itself. However, reapproaching the original donor for a second apheresis poses a significant logistic effort, particularly if a nonrelated donor is concerned. Yet the data provided by Roddie et al argue that such an approach is feasible because more than half of patients in their study had an unrelated donor.

Given the low incidence of the disease and the changes in B-ALL treatment, it seems unlikely that commercial products will soon become available for such an approach. Yet, alloHSCT will be relevant for B-ALL treatment in the near future and so will be treatment of post-alloHSCT relapse. Thus, the study by Roddie et al not only underscores the feasibility of POC production of CAR-Ts but also argues for regulatory support of such academic approaches, not only in pediatric but also adult patients.⁵ 

Besides these main findings, the data are interesting in several more aspects. For instance, accompanying analyses point toward the role of interleukin-15 in mediating the LD-related better outcome, as also recently and prominently suggested by others.⁶ 

Not surprisingly, many remaining questions are encouraged to be answered. Although 14 patients received treatment, 3 others died before receiving CAR-DLI and the median interval between registration and treatment was 2 months. This points to the need for faster production of CAR-Ts. Here an optimized POC approach, such as the use of preapheresed T cells from the original donor, as done for regular DLIs by many centers, or the use of universal allogeneic CAR-Ts, may provide solutions.

The lower long-term survival in the study by Roddie et al compared to the reported data from pediatric³ and adult⁴ patients may be related to many variables including higher disease load prior to CAR-DLIs up to different vector design. All this will likely not be answered by future randomized trials but multicenter register analysis may aid future development.

The study also examined the feasibility of repeated CAR-DLIs in escalating doses. However, no response was observed in the 8 patients receiving a second CAR-DLI and only 1 patient received a third CAR-DLI. The second CAR-DLI were given at 2 to 16 months after first dose and it remains speculative whether a stricter time schedule may have produced better results. So far, the data by Roddie et al do not encourage repeated CAR-T therapy in case of incomplete response after first CAR-DLI. Rather, the observed rate of 93% measurable (or minimal) residual disease (MRD)–negative complete remission (CR) at 6 weeks after CAR-DLI followed later by a 78% relapse rate, argues for a maintenance therapy with prophylactic DLIs⁷ or TKI⁸ or parameter-driven preemptive intervention⁹ in this high-risk population. This notion is supported by the single patient in the study who received a third CAR-DLI following MRD-relapse after second CAR-DLI: after CR was achieved, a TKI maintenance was initiated and resulted in a survival of around 2.5 years.

In summary, the study comprehensively analyzed an allogeneic CAR-T therapy–based approach for B-ALL relapse after alloHSCT demonstrating the need for LD in these patients. And last but not least, 2 patients treated in the study are alive and in remission for more than 5 years after treatment. This encourages all efforts we can muster to improve the therapeutic options, thus aiming to help our patients in this situation.

Conflict-of-interest disclosure: L.P.M. reports an advisory role or expert testimony for Pfizer, Gilead, Novartis, and Amgen and honoraria/travel grants from AbbVie, Bristol Myers Squibb, and Gilead.