Loss of surface CD3 expression in allogeneic CAR T-cells.

Abstract

The use of genetically modified allogeneic chimeric antigen receptor (CAR) T cells to treat hematopoietic malignancies is steadily increasing.1 These “universal” CAR-T cells are derived from healthy donor T lymphocytes, allowing for faster manufacturing at lower cost. To reduce the risk of rejection or graft vs host disease, universal CAR T cells are designed to disrupt expression of the α/β T-cell receptor. These modifications result in improved efficacy of treatment but also have unexpected consequences in the clinical laboratory. We recently encountered two patients treated with allogeneic CAR T cells who had atypical T cell populations detectable by flow cytometry. Patient 1 was a 16-month-old girl with B-lymphoblastic leukemia with KMT2a rearrangement diagnosed originally at 3 months of age. She relapsed during maintenance therapy and, due to the presence of peripheral blasts, was referred for treatment with allogeneic CAR T cells. She was treated with good response, eventually progressing to bone marrow transplant. She remains disease-free six months later. Patient 2 was a 5-year-old boy with B-lymphoblastic leukemia with KMT2a rearrangement whose persistently high blast counts precluded collection of T-cells for autologous CAR T cell infusion. Although initially he had expansion of his CAR T cells and a decrease in circulating blasts to 5%, his disease returned one week after infusion. Peripheral blood flow cytometry with automated quantification of CD3+ T cells is commonly performed at our institution for patients receiving CAR T cells. On the sixteenth day following infusion of these cells in patient 1, our automated CD3 count by flow cytometry reported only 6.7% CD3+ cells, which suggested poor expansion of the CARs. However, peripheral blood smears showed 98.2% lymphocytes with increased activated large granular lymphocytes (Figure 1A,B). Subsequent review of the flow cytometry plots showed an unusual lymphocyte population which was CD8 positive and lacked surface CD3 (Figure 1C). These cells also expressed CD2, CD5, and CD7. Patient 2 had similar findings; manual differential cell count performed on the peripheral blood smear showed 50% lymphocytes with increased large granular lymphocytes. Flow cytometry demonstrated the same unusual sCD3-, CD8+ T cell population, consistent with allogeneic CAR T-cells (Figure 1D). In conclusion, we found that patients treated with allogeneic CAR T cells may have an abnormal T cell subset demonstrable by flow cytometry with expression of CD8 and absence of CD3. Although T-cell malignancies may lack surface CD3, this seemed highly unlikely in this clinical setting. The kinetics and morphology of cells indicated that these aberrant CD3 negative cells were allogeneic CAR T-cells. The surface TCR/CD3 complex is assembled in the endoplasmic reticulum; knockdown of TCRαβ in these genetically modified T-cells likely resulted in impaired localization of CD3. Lack of surface CD3 expression can result in incorrect CD3, CD4 and CD8 counts in the laboratory. In addition, lack of surface CD3 may also limit the subsequent use of bispecific T-cell engagers, such as blinatumomab, to target CD3 in this population.