CD19 Epitope Escape after 4SCAR19 T Cell Therapy Resulted in Re-Establishment of Chemo-Sensitivity in Adult B-Cell Acute Lymphocytic Leukemia Patients

Background: CD19-targeted chimeric antigen receptor (CAR) T cell treatment has emerged as a highly effective therapy in patients with refractory B cell acute lymphoblastic leukemia (B-ALL). Despite obvious successes, there have been documented relapses in which CAR T cells are still present but the leukemia cells have lost CD19 surface epitopes. In this study, we aim to investigate these patients and characterize their leukemic cell clones.

Methods: As part of a multi-center trial, this single-arm, phase I trial enrolled 10 patients between March 2015 and June 2016 based on the following criteria: poor physical condition, chemotherapy resistance, and lack of suitable donor. Following salvage chemotherapy, peripheral lymphocytes were collected from leukapheresis, and T cells were transduced with a 4th generation, safety-engineered, CD19scFv/CD28/CD137/CD27/ CD3zeta lentivector. The pre-infusion lymphodepleting chemotherapy conditioning is cyclophosphamide 800mg/m² x 3 days, followed by infusions at a dose of 2.13 (range from 0.42 - 5.9) x106 4SCAR19 T cells per kg body weight . Patients who achieved complete remission (CR) but later relapsed received chemotherapy of VDP (vincristine 1.4mg/m² d1,8,15,22, daunomycin 40mg/m² d1,8,15,22, prednison 40mg/m²d1-28) regimen. CD19-negative relapsed patients were assessed by surface immunophenotyping, cytoplasm immunophenotyping ,bone marrow biopsy and immunohistochemistry staining, T cell receptor(TCR) and immunoglobulin heavy chain (IgH) rearrangement analysis.

Results and Discussion: Total 10 patients received 4SCAR19 treatment and 7 achieved CR. Following CR, 5 relapsed, including 3 CD19 positive and 2 CD19 negative relapses. Surface marker analysis showed that the CD19-negative relapsed leukemia cells displayed altered surface immunophenotype different from the original CD19-positive leukemia cells. Comparing the leukemia cells of the two CD19-negative relapsed patients by immunophenotype analysis for CD19,CD10,CD34,and CD79a, the CD10 surface expression in one of the two CD19-negative leukemia cells was down-regulated from 90% to 10%, but cytoplasmic CD10 expression remained at 90% which was confirmed by immunohistochemical staining of bone marrow biopsy. The CD10 expression profile, however, did not alter in the other CD19-negative leukemia cells. The loss of surface CD19 expression was coupled with the loss of cytoplasm CD19 as determined by intracellular flow cytometry and immunohistochemical staining; however, the nuclear PAX-5 expression was positive. Another B-ALL index ,the IgH rearrangement was examined by multiplex PCR and their original clonal identify was confirmed. It is thus conceivable that the CD19 negative relapsed leukemia cells were associated with escape variants rather than selection of de novo leukemia clones following 4SCAR19 therapy. From the diverse immune phenotype of the CD19 negative relapsed leukemia cells, it should be feasible to select another surface epitope(s) as target(s) for CAR T therapy. Importantly, after CAR T therapy, the two CD19-negative relapsed patients responded to chemotherapy and achieved CR again. However, the three CD19-positive relapsed patients experienced inferior chemotherapy outcomes. These results suggest that chemo-sensitivity could be re-established after CD19-negative relapse following 4SCAR19 therapy. Other studies have found that CD19 mutations in CAR T-treated leukemic cells could result in the loss of PI3K and SFTK signaling, and further, altered signaling pathways such as p53 and PTEN mutations have been shown to contribute to chemo-resistance of leukemic cells. The forced pathway alteration in the CD19-negative leukemic clones after CAR T treatment could involve one of these pathways, and further investigation is warranted.