Addition of Fludarabine to Cyclophosphamide Lymphodepletion Improves In Vivo Expansion of CD19 Chimeric Antigen Receptor-Modified T Cells and Clinical Outcome in Adults with B Cell Acute Lymphoblastic Leukemia

BACKGROUND:

Chemotherapy followed by autologous T cells that are genetically modified to express a CD19-specific chimeric antigen receptor (CAR) has shown promise as a novel therapy for patients with relapsed or refractory B cell acute lymphoblastic leukemia (B-ALL); however, the risk of severe cytokine release syndrome (sCRS) and neurotoxicity has tempered enthusiasm for widespread application of this approach. The functional heterogeneity that is inherent in CAR-T cell products that are manufactured from undefined T cell subsets has hindered definition of dose-response relationships and identification of factors that may impact efficacy and toxicity.

METHODS:

We are conducting the first clinical trial that administers CD19 CAR-T cells manufactured from a defined composition of T cell subsets to adults with relapsed or refractory B-ALL. CD8⁺ and CD4⁺ T cells were enriched from each patient, transduced with a CD19 CAR lentivirus and separately expanded in vitro before formulation for infusion in a 1:1 ratio of CD8⁺:CD4⁺ CAR⁺ T cells at 2x10⁵, 2x10⁶ or 2x10⁷ CAR-T cells/kg. Prior to CAR-T cell infusion, patients underwent lymphodepletion with a high-dose cyclophosphamide (Cy)-based regimen with or without fludarabine (Flu).

RESULTS:

Twenty-nine adults with B-ALL (median age 40, range 22 - 73 years; median 17% marrow blasts, range 0 - 97%), including 10 patients who had relapsed after allogeneic transplantation, received at least one CAR-T cell infusion. Twenty-four of 26 restaged patients (92%) achieved bone marrow (BM) complete remission (CR) by flow cytometry. CD4⁺ and CD8⁺ CAR-T cells expanded in vivo after infusion and their number in blood correlated with the infused CAR-T cell dose.

Thirteen patients received lymphodepletion with Cy-based regimens without Flu. Ten of 12 restaged patients (83%) achieved BM CR by flow cytometry; however, 7 of these (70%) relapsed a median of 66 days after CAR-T cell infusion. Disease relapse correlated with a loss of CAR-T cell persistence in blood. We observed a CD8 cytotoxic T cell response to the murine scFv component of the CAR transgene that contributed to CAR-T cell rejection, and resulted in lack of CAR-T cell expansion after a second CAR-T cell infusion in 5 patients treated for persistent or relapsed disease.

To minimize immune-mediated CAR-T cell rejection 14 patients were treated with Cy followed by Flu lymphodepletion (Cy/Flu, Cy 60 mg/kg x 1 and Flu 25 mg/m² x 3-5) before CAR-T cell infusion. All patients (100%) who received Cy/Flu lymphodepletion achieved BM CR after CAR-T cell infusion. CAR-T cell expansion and persistence in blood was higher in Cy/Flu-lymphodepleted patients compared to their counterparts who received Cy alone (Day 28 after 2x10⁶ CAR-T cells/kg: CD8⁺ CAR-T cells, mean 55.8/μL vs 0.10/μL, p<0.01; CD4⁺ CAR-T cells, 2.1/μL vs 0.02/μL, p<0.01), enabling reduction in CAR-T cell dose for Cy/Flu-treated patients. Patients who received Cy/Flu lymphodepletion appear to have longer disease-free survival (DFS) than those who received Cy alone (Cy/Flu, median, not reached; Cy alone, 150 days, p=0.09).

CAR-T cell infusion was associated with sCRS, characterized by fever and hypotension requiring intensive care in 7 of 27 patients (26%) and neurotoxicity (≥ grade 3 CTCAE v4.03) in 13 of 27 patients (48%). Two patients died following complications of sCRS. Patients with sCRS or neurotoxicity had higher peak serum levels of IL-6, IFN-γ, ferritin and C-reactive protein compared to those without serious toxicity. Importantly IL-6, IFN-γ and TNF-α levels in serum collected on day 1 after CAR-T cell infusion from those who subsequently developed neurotoxicity were higher than those collected from their counterparts who did not develop neurotoxicity (IL-6, p<0.01; IFN-γ, p=0.05; TNF-α, p=0.04), providing potential biomarkers to test early intervention strategies to prevent neurotoxicity. The risks of sCRS and neurotoxicity correlated with higher leukemic marrow infiltration and increasing CAR-T cell dose. We have now adopted a risk-stratified approach to CAR-T cell dosing in which the CAR-T cell dose inversely correlates to the patient's bone marrow tumor burden.

CONCLUSION:

Risk-stratified dosing of CD19 CAR-T cells of defined subset composition is feasible and safe in a majority of patients with refractory B-ALL, and results in a CR rate of 92%. Addition of Flu to Cy-based lymphodepletion improves CAR-T cell expansion, persistence and DFS.