Patient Characteristics and Pre-Infusion Biomarkers of Inflammation Correlate with Clinical Outcomes after Treatment with the Defined Composition, CD19-Targeted CAR T Cell Product, JCAR017

Introduction

JCAR017 is a CD19-directed 4-1BB CAR T cell product candidate administered in a defined composition at a precise dose of CD8 and CD4 CAR T cells. TRANSCEND NHL 001 is the first multicenter phase 1 trial of JCAR017 in R/R B cell NHL (NCT02631044). As previously reported (Abramson, 14-ICML 2017), JCAR017 demonstrated high CR rates and low incidence of cytokine release syndrome (CRS) and neurotoxicity (NT). Data from this trial were assessed to examine potential relationships between patient (pt) characteristics, translational variables and response and safety endpoints to help develop approaches to better risk stratify pts for CAR T cell therapy.

Methods

Pts with R/R DLBCL NOS (de novo or transformed from indolent lymphoma), PMBCL, FL grade 3B, or MCL and adequate organ function are eligible for TRANSCEND NHL 001. Pts received lymphodepletion, with fludarabine and cyclophosphamide, followed by a single dose of JCAR017 at one of two dose levels (DL1, 5 × 10⁷ cells; DL2, 1 × 10⁸ cells); a small cohort received a 2-dose schedule at DL1. Blood samples were collected per study protocol for cytokine, PK and clinical lab evaluation at protocol-defined time points. PK (CAR T cell expansion and persistence) was measured at protocol-defined time points using flow cytometry. Cytokines were measured on a Luminex platform. Cytokines and additional plasma analytes are being evaluated on additional samples and will be presented. Potential correlations of clinical endpoints with pt characteristics, baseline clinical lab values and cytokine levels, and expansion and exposure to CAR T cells was investigated. Clinical endpoints of interest include best overall response, durable response at month 3, any grade NT, any grade CRS, and Gr 3-4 NT. Data were pooled from DL1 and DL2 for analyses when appropriate. Univariate nonparametric tests were used for statistical analysis. All reported p-values are 2-sided without multiplicity adjustment.

Results

As of May 7, 2017, 55 pts in the DLBCL cohort were evaluable for safety and 54 pts for efficacy. Four MCL pts were also evaluated for safety. Blood samples were collected at various time points per study protocol and evaluated for PK, PD, and cytokine levels. In the 59 pts evaluable for safety, CRS developed in 32% (30% Gr 1-2, 0% Gr 3, 2% Gr 4); NT was observed in 20% (5% Gr 1-2, 10% Gr 3, 5% Gr 4). Dose level did not correlate with CRS or NT (p=0.565 and p=1.00, respectively). Pt factors that correlate with any grade CRS and NT were poorer performance status [e.g. ECOG PS2] (p=0.03) and higher disease burden (p <0.05) as measured by the sum of the products of diameters [SPD] on imaging. Pre-CAR T cell clinical labs and cytokines associated with the occurrence of any grade NT were higher serum LDH, ferritin, and CRP, and higher plasma IL-6, IL-8, IL-10, TNF-α, IFN-α2, MCP-1, and MIP-1β (p <0.05 for each). Higher pre-CAR T infusion plasma IL-8, IL-10, and CXCL10 were also associated with Gr 3-4 NT (p <0.05 for each).

Of the 54 pts in the DLBCL cohort evaluable for efficacy, 76% achieved best overall response of CR/PR; 51% remained in CR/PR at month 3. Higher ECOG scores and DLBCL transformed from CLL or MZL correlated with lower durable response at month 3 (p=0.02 for both). Pre-CAR T cell analytes associated with best ORR included lower values of ferritin, LDH, CXCL10, G-CSF, and IL-10, and those associated with durable response at 3 months were lower ferritin, CRP, LDH, CXCL10, IL-8, IL-10, IL-15, MCP-1, MIP-1β, TNF-α, and higher pre-CAR T cell hemoglobin and albumin (p <0.05 for each).

Characterization of apheresis and drug product composition is underway; correlations with clinical attributes will be presented. Preliminary analyses suggest T cell memory subsets and T cell functionality may correlate with clinical outcomes.

Conclusion

Preliminary analysis has identified baseline patient characteristics, including inflammatory state and high tumor burden prior to treatment, as a means to identify patients at risk for increased toxicity following administration of JCAR017. Low tumor burden and low inflammatory state appear associated with improved toxicity profile and better durability of response. These preliminary data suggest that treating pts earlier in their course of therapy, or using a panel of clinical and laboratory biomarkers to risk stratify patients for potential early intervention, is feasible and may mitigate risk of toxicity and potentially improve durability of response.