Cyclophosphamide and Fludarabine Conditioning Chemotherapy Induces a Key Homeostatic Cytokine Profile in Patients Prior to CAR T Cell Therapy

This study is supported in part by funding from the Cooperative Research and Development Agreement (CRADA) between the National Cancer Institute and Kite Pharma

Introduction: CAR-engineered autologous T-cell therapy has shown promising activity in relapsed/refractory B-cell malignancies in an ongoing phase 1 study (Kochenderfer et al. J Clin Oncol 2014). Lymphodepleting conditioning chemotherapy is critical for optimal CAR T-cell activity in animal models. We evaluated the effects of conditioning chemotherapy on cytokine and chemokine levels in patients dosed with anti-CD19 CAR T cells.

Methods: In this National Cancer Institute clinical trial (NCT00924326), patients with relapsed/refractory B-cell malignancies received conditioning with cyclophosphamide and fludarabine daily for 3 days starting on day -5; followed by anti-CD19 CAR T cells engineered with a CAR comprising CD28 and CD3-zeta signaling domains. Forty one cytokines, chemokines and immune response related markers were measured in the blood of patients pre (day -5) and post conditioning (day 0) by using EMD Millipore Luminex® xMAP® multiplex assays. Data acquisition and analysis were performed using a Luminex 200™ instrument and xPONENT® 3.1 data analysis software. Increases in cytokine and chemokine levels were analyzed pre- and post- conditioning, and the fold-changes in cytokine and chemokine levels were analyzed relative to clinical outcome subsequent to infusion with anti-CD19 CAR T cells. Analyses were performed with the Wilcoxon rank sum test adjusted for multiplicity with a Bonferroni correction, using a nominal level of 0.006 for significance.

Results: Samples from 15patients have been evaluated. There were significant increases pre- to post-conditioning in the levels of interleukin 15 (IL-15; p=0.001), interleukin 7 (IL-7; p=0.0002), and monocyte chemoattractant protein-1 (MCP-1; p<0.0025) in blood, five days after the initiation of conditioning chemotherapy. Levels of interferon-gamma induced protein 10 (IP-10) were elevated post-conditioning, but did not meet the threshold for significance (p=0.048). Compared with baseline, levels of IL-15 increased on average 13 fold and levels of IL-7, IP-10 and MCP-1, about 2 fold. Comparison of the fold-increases in IL-15 upon conditioning between responders and non-responders approached significance (p=0.01), but did not meet the threshold after multiplicity adjustment. Larger fold-change increases for responders versus non-responders were also observed with placental growth factor (PLGF) (median fold increase 2.6 v. 1.6, average fold increase 32 v 4.2), C-reactive protein (CRP) (median fold increase 3.5 v 2.4, average fold increase 6.6 v. 2.0), IP-10 (median fold increase 2.1 v. 0.7, average fold increase 2.6 v. 2.8), and interleukin 10 (IL-10) (median fold increase 1.8 v. 0.4, average fold increase 3.1 v. 2.0), but did not meet the threshold for significance. In addition to ongoing analysis of conditioning-mediated cytokine induction and clinical response, we are evaluating the impact of conditioning chemotherapy dose on cytokine levels, as well as the relationship between conditioning-related cytokines and CAR T-cell expansion and persistence.

Conclusions: The data obtained to date support the hypothesis that cytokines such as IL-15 play a key role in the clinical outcomes to anti-CD19 CAR T-cell therapy. Our results demonstrate that conditioning chemotherapy significantly increases the levels of homeostatic cytokines known to regulate T-cell expansion, as well as specific pro-inflammatory cytokines and chemokines. Optimization of conditioning chemotherapy is critical to the activity of CAR T-cell therapies.