Abstract
Introduction: While autologous anti-CD19 and anti-GD2 CAR T cell therapy has shown promise results in children with B cell acute lymphocytic leukemia (ALL) and sarcoma respectively, T cells from some patients fail to expand in culture. The anti-CD19 and anti-GD2 CAR T cell products manufactured by our center and the factors affecting CAR T cell expansion were reviewed.
Methods: The manufacturing methods for both types of CAR T cells were similar; autologous PBMC concentrates collected by apheresis were used as starting material, anti-CD3/CD28 beads were used for T cell enrichment, and IL-2 plus CD3/CD28 beads were used for T cell activation and expansion. Retroviral vector transduction of CD19 anti-GD2 CAR T cells was performed on days 2 and 3. Anti-CD19 CAR T cells from 28 patients were expanded over 7 to 11 days and anti-GD2 CAR T cells from 9 patients expanded over 10 to 11 days. The target dose of transduced cells was 1 or 3x106/Kg for both clinical protocols.
Results: Following transduction and expansion, the 28 anti-CD19 CAR T cell products contained a mean of 1,130±916 x106 transduced T cells. Four of the 28 had poor expansion defined as not yielding enough transduced cells for one dose (1x106 CAR cells/kg). When the characteristics of the PBMC concentrates collected from the 4 patients whose T cells expanded poorly was compared with that of the other 24, those collected from the poor expanders contained greater quantities of monocytes (39.8±12.9% vs 15.3±10.8%, p=0.0014) and less lymphocytes (42.3±8.4% vs 75.3±14.1%; p<0.001). The poorly expanding apheresis products also contained less CD3+ cells (27.4±10.6% vs 59.7±19.5%; p=0.0045) and more CD56+ cells (26.9±14.5% vs 6.1±5.5%; p<0.001), but there was no difference in the quantity of CD19+ cells. The sum of the proportion of granulocytes and monocytes in each PBMC concentrate was inversely associated with T cell expansion (r= -0.59). A comparison of the quantity of anti-GD2 CAR T cells produced from the 9 sarcoma patients with the quantity of anti-CD19 CAR T cells produced from the 22 ALL patients whose cells were also expanded over 10 to 11 days revealed that the anti-GD2 CAR T cell products contained less transduced cells (197±205x106 vs 1,313±917x106; p=0.0083). The PMBC concentrates collected from sarcoma patients contained more monocytes (31.4±12.4% vs 18.5±13.7%; p=0.019) than those collected from the ALL patients. Of the 9 anti-GD2 CAR T cell products, one failed to meet dose (prescribed doses ranged from 1 x 105/kg-3 x 106/kg); the apheresis product from this patient contained large quantities of myeloid cells: 37% monocytes and 55% granulocytes. However, for sarcoma patients there was no relationship between the sum of granulocytes and monocytes in the PBMC concentrate and the yield of transduced anti-GD2 CAR T cells (r=-0.11).
Among the 5 products that expanded poorly, manufacturing was repeated for 2 ALL and 1 sarcoma patient using a cryopreserved aliquot of their PBMC concentrate but incorporating an upfront step to deplete myeloid cells by plastic adherence, with or without density gradient separation. All three products expanded well following myeloid cell depletion. For the ALL patients 0 and 2.4x106 transduced anti-CD19 CAR T cells were produced initially compared to 147x106 and 160x106 transduced cells using the second manufacturing procedure. For the sarcoma patient 4.8x106 transduced anti-GD2 CAR T cells were produced by the first procedure and 5,200x106 by the second procedure.
Conclusions: These results show that the expansion of autologous T cells for CAR T cell therapy is highly variable and the variability is due at least in part to the contamination of the starting apheresis concentrate with myeloid cells. More rigorous removal of myeloid cells prior to initiating cell culture was associated with improved T cell expansion. Suppression of expansion by myeloid cells appears to more problematic in sarcoma patients than in leukemia patients.
Sabatino:Kite Pharma: Employment. Mackall:Juno: Patents & Royalties: CD22-CAR.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal