Phase I/II Study of Pembrolizumab for Progressive Diffuse Large B Cell Lymphoma after Anti-CD19 Directed Chimeric Antigen Receptor Modified T Cell Therapy

Introduction: Chimeric antigen receptor modified T cells directed against CD19 (CART19) has demonstrated efficacy in relapsed or refractory (r/r) diffuse large B-cell lymphoma (DLBCL) with durable complete remissions (CR) [Schuster et al. Blood 2016 (abstract 3026); Kochenderfer et al. Blood 2010; Turtle et al. Sci Transl Med 2016 ]. However, non-responding patients require novel therapeutic approaches. Pembrolizumab, an anti-PD1 checkpoint inhibitor, is an attractive next therapeutic option given its T cell-mediated mechanism of action. Anecdotally, we observed responses in patients treated with PD1 blockade after CART19 (Chong et al. Blood 2017). Based on these observations, we are conducting a clinical trial of pembrolizumab in patients with B cell non-Hodgkin lymphomas expressing CD19 who are r/r after CART19.

Methods: This prospective single-institution trial enrolled patients with progressive B-NHL after treatment with CART19 expressing either a murine (CTL019) or a human (CTL119) CD19-specific single chain antibody fragment fused in tandem with 4-1BB and TCRζ costimulatory / activation domains (NCT02650999). Patients who received intervening therapy between CTL019/119 infusion and enrollment were excluded. Patients received a fixed dose of pembrolizumab 200mg IV every 3 weeks until progression of disease, therapy limiting-toxicity, or elective protocol discontinuation. The primary endpoint was safety. Secondary endpoints included overall response rate (ORR) and progression-free survival (PFS) from initiation of pembrolizumab. Whole blood and peripheral blood mononuclear cells (PBMC) were collected from subjects at protocol-specified times for correlative studies. DNA extracted from whole blood was analyzed for the appropriate CAR transgene copy number using qPCR (sensitivity 25 copies/μg DNA). Flow cytometry for immunophenotyping was performed on single cell suspensions prepared from PBMC.

Results: From June 1, 2016 to June 1, 2017, 8 patients with DLBCL were enrolled [3 germinal center (2 "double hit"), 3 non-germinal center and 1 T-cell rich DLBCL, and 1 transformed follicular lymphoma). Median age was 56.7 years. Median prior therapies was 5 (range 3-7). Median PFS after CTL019/119 was 2.2 months (range 0.4-3.2); median time to first pembrolizumab dose was 2.5 months (range 0.4-3.7). Adverse events (CTCAE 4.0) per patient at least possibly related to pembrolizumab were: neutropenia (grade 4; N=3), cytokine release syndrome (grade 3 on Penn scale; N=1), delirium (grade 3; N=1), , infusion reaction (grade 2; N=1), fever (grade 1 or 2; N=3), dyspnea (grade 1; N=1), pleural effusion (grade 1, N=1), and arthralgia (grade 1; N=1). One patient developed CMV infection (unrelated; grade 4) resulting in discontinuation of protocol therapy. ORR after pembrolizumab was 25% [1 complete response (CR); 1 partial response (PR)]; 6 patients had progressive disease (PD). One patient with DLBCL continues in CR at 280 days; one patient with T-cell rich DLBCL in PR was removed from study because of CMV pneumonitis. Immunohistochemistry for PD-1 expression in the tumor microenvironment was performed for 5 patients; 4 patients had detectable PD-1 expression, including one patient with PR and 3 patients with PD. All patients, responders and non-responders, had a single initial peak in CAR transgene copy number after the first pembrolizumab dose with fold increases of 1.2-2.0 (median 1.6) between days 2-13 (median 2.5 days). Maximum CAR transgene copy number during therapy did not correlate with response, nor did the AUC over days 1-7; however, responding patients had more than one peak during therapy while non-responding patients had a single peak following the initial dose of pembrolizumab. Flow cytometry showed increases in both %CD4 CAR T cells and %CD8 CAR T cells in both responding and nonresponding patients.

Conclusions: PD1 blockade with pembrolizumab appears safe and results in clinical responses in some patients with progression of DLBCL after anti-CD19 directed chimeric antigen receptor modified T cell therapy. Analysis of the pharmacokinetics of CAR T-cells in patients treated with pembrolizumab appears to identify responding patients, and supports the hypothesis that CAR T cells may be rejuvenated by infusion of PD1 antagonists. Additional immunophenotypic and epigenetic studies are in progress to distinguish the mechanism of action in responding patients.

Figure

View largeDownload slide

Figure

View largeDownload slide

Close modal