A CAR against CAR for unintended consequences

In this issue of Blood, Ruella and colleagues report the development of a chimeric antigen receptor (CAR) that could target anti-CD19 CAR T cells (CAR19⁺ T) used for therapy of B-cell malignancies as well as leukemic B cells inadvertently transduced with anti-CD19 CAR (CAR19⁺ B).¹ 

Tisagenlecleucel, an autologous anti-CD19 CAR T-cell therapy, is manufactured by transducing peripheral blood T cells with a CAR molecule that consists of a single-chain variable fragment (scFv) derived from the CD19-specific monoclonal antibody, FMC63, and CD137 (4-1BB) and CD3ζ signaling domains. It has been demonstrated to induce durable remissions in pediatric patients with relapsed or refractory acute lymphoblastic leukemia²  as well as adult patients with relapsed or refractory large B-cell lymphoma³  leading to its approval for these indications. Axicabtagene ciloleucel (axi-cel), an autologous anti-CD19 CAR T-cell therapy using scFv from FMC63 but CD28 and CD3ζ signaling domains, has also been shown to induce durable remissions and is approved for adult patients with relapsed or refractory large B-cell lymphoma.^4,5  Although these therapies offer hope for patients who previously had no effective therapeutic options, they may also be associated with undesirable on-target off-tumor effects. Specifically, they can induce B-cell aplasia because CD19 is also expressed on normal B cells and lead to hypogammaglobulinemia and increased risk of sinopulmonary and other infections. It is possible that the B-cell aplasia and hypogammaglobulinemia may persist lifelong in some patients because CAR T cells can survive long term as memory T cells. Although immunoglobulin replacement therapy may be used to treat hypogammaglobulinemia and minimizes risk of infections, it could add up to significant expense over the lifetime of the patients especially for children treated with anti-CD19 CAR T-cell therapy. Therefore, a strategy to eliminate the CAR T cells would be desirable once it is determined that the patient is likely cured from the malignancy.

In addition to on-target off-tumor effects, autologous CAR T-cell therapies may also be associated with unintentional transduction of tumor cells especially in patients who have high-circulating tumor cell count. In an unusual case that Ruella and colleagues reported in a prior publication, transduction of tisagenlecleucel into a leukemic B cell led to masking of CD19 on the surface of the leukemic cell due to in cis binding between the transduced anti-CD19 CAR molecule and endogenous CD19 (see figure).⁶  This prevented elimination of the transduced leukemic B cell by anti-CD19 CAR T cells and resulted in a relapse due to immune escape.

To address the above challenges, Ruella and colleagues designed a CAR (anti-CAR19) using the scFv derived from an anti-idiotype monoclonal antibody (clone 136.20.1)⁷  generated previously against the antigen-recognition domain of FMC63 used in the development of tisagenlecleucel and axi-cel. They demonstrated that T cells transduced with anti-CAR19 (anti-CAR19-CART) induced lysis of CAR19⁺ leukemic B cells. More importantly, anti-CAR19-CART showed significant disease control in immunodeficient mice engrafted with primary CAR19⁺ leukemic blasts. In contrast, CAR19⁺ T cells developed using a CAR construct similar to tisagenlecleucel failed to lyse the CAR19⁺ leukemic B cells in vitro or control the disease in vivo. The anti-CAR19-CART were also shown to efficiently and specifically kill CAR19⁺ T cells.

This anti-CAR19-CART could potentially be used for treatment of the rare patient who may develop a relapse following infusion of CAR19⁺ leukemic B cells. In addition, it may be used for treatment of another potential but not yet observed unintended consequence of CAR T-cell therapy: the risk of insertional mutagenesis and development of T-cell leukemia (see figure). Although it is also appealing to use the anti-CAR19-CART to ablate CAR19⁺ T cells and reverse B-cell aplasia to restore normal immunoglobulin levels, the practical challenge is to determine when a patient can be considered cured of the B-cell malignancy in order to apply such therapy. Moreover, the need for ablation of CAR19⁺ T cells may be lower in adults with large B-cell lymphoma both because adults have long-lived plasma cells that can continue to produce protective immunoglobulins after anti-CD19 CAR T-cell therapy and because the available evidence so far suggests that long-term functional persistence of CAR T cells may not be needed to maintain durability of the remissions.^3,5  Nevertheless, this study provides an important proof of concept that it is feasible to develop a CAR against CAR to mitigate unintended consequences. Such an approach may be even more important in malignancies where the antigens are shared between the tumor cells and their normal counterpart, and short-term persistence of CAR T cells is desirable to prevent long-term toxicity. For example, one of the challenges in developing CAR T-cell therapy for T-cell malignancies is the risk of inducing T-cell aplasia and fatal opportunistic infections. Similarly, CARs targeting acute myeloid leukemia may induce bone marrow aplasia because of shared antigens with hematopoietic progenitor cells. However, if the CAR T cells targeting these malignancies could be eliminated using an anti-CAR-CART after induction of remission, it may be more feasible to develop CAR T-cell therapy approaches against these malignancies. Although the anti-CAR19-CART described here is useful for treatment of unintended consequences for the current approved products of tisagenlecleucel and axi-cel, alternative strategies may be explored in the future, including safety switches such as inducible caspase 9 and CAR T cells that persist for a finite period as observed with CARs using CD28 vs 4-1BB costimulatory domain in certain models.^8-10