Pembrolizumab for B-cell lymphomas relapsing after or refractory to CD19-directed CAR T-cell therapy

Elise A Chong, Cécile Alanio, Jakub Svoboda, Sunita D Nasta, Daniel J Landsburg, Simon F Lacey, Marco Ruella, Siddharth Bhattacharyya, E John Wherry, Stephen J Schuster, Elise A Chong, Cécile Alanio, Jakub Svoboda, Sunita D Nasta, Daniel J Landsburg, Simon F Lacey, Marco Ruella, Siddharth Bhattacharyya, E John Wherry, Stephen J Schuster

Abstract

CD19-directed chimeric antigen receptor-modified (CAR T) T cells achieve durable remissions in about 30% to 40% of relapsed/refractory large B-cell lymphomas. T-cell exhaustion and/or an immunosuppressive tumor microenvironment may contribute to CAR T-cell failure. Pembrolizumab, an anti-PD1 immune checkpoint inhibitor, may reverse T-cell exhaustion after CAR T-cell therapy. We treated 12 patients with B-cell lymphomas who were either refractory to (n = 9) or relapsed after (n = 3) CD19-directed CAR T-cell (4-1BB-costimulated) therapy with pembrolizumab 200 mg IV every 3 weeks. Median time from CAR T-cell infusion to first pembrolizumab dose was 3.3 months (range, 0.4-42.8 months). Pembrolizumab was well tolerated, and the only grade ≥3 adverse events related to pembrolizumab were neutropenia (n = 3; 25%). Best overall response rate after pembrolizumab was 25% (3 of 12 patients; 1 complete response; 2 partial responses). One (8%) patient had stable disease; thus, 4 of 12 (33%) patients had clinical benefit. After pembrolizumab, 4 patients with clinical benefit had an increase in percentage of CAR T cells by mass cytometry by time of flight (CyTOF); 3 of 4 of these patients also had increases in CAR19 transgene levels by quantitative polymerase chain reaction. Deep immune profiling using CyTOF revealed increased CAR T-cell activation and proliferation and less T-cell exhaustion in clinical responders. Together, PD1 blockade with pembrolizumab after CD19-directed CAR T-cell therapy appears safe and may achieve clinical responses in some patients with B-cell lymphomas refractory to or relapsed after CAR T-cell therapy. This trial was registered at www.clinicaltrials.gove as #NCT02650999.

© 2022 by The American Society of Hematology.

Figures

Graphical abstract
Graphical abstract
Figure 1.
Figure 1.
Swimmer plot depicting clinical outcomes after pembrolizumab. Bars are color coded based on 3-month response. Best ORR and 3-month ORR was 3 of 12 (25%). One patient had a CR (light blue bar), and 2 patients had PRs (light green bars). Two patients had SD with symptomatic improvement as best response, but only 1 of these patients continued to maintain SD at 3 months (yellow bar) and the other patient had PD at 3 months (red bar). The remaining 7 patients had progression of lymphoma (red bars). Orange triangles, onset of CR; green triangles, onset of PR; lavender triangles, onset of SD; black arrows, patient is alive at last follow-up; steel blue circles, patient is deceased. Response to CAR T cells (eg, relapsed or refractory) is indicated after patient ID number.
Figure 2.
Figure 2.
Changes in CAR19 copy number and percentage of CAR T cells after pembrolizumab. (A) CAR19 transgene copies/ug DNA for each patient after administration of pembrolizumab (n = 11). Pembrolizumab was started at day 0. Each diamond indicates a subsequent dose of pembrolizumab. Data are classified based on 3-month ORR. One patient with a PR did not have longitudinal data for CAR19 transgene copies available. (B) CAR19 transgene copies per μg DNA for patient 4, who had a complete response to pembrolizumab. (C) Gating strategy for identifying CAR T cells. Live cells were identified as CD45+Cisplatin−. T cells are gated as CD3+ live cells. CAR T cells are identified using a CAR19 antibody and gating the CAR T cells as CAR19+CD3+ live cells. (D) Percentage of CAR T cells among total T cells as a function of days from pembrolizumab infusion, separated by clinical response (n = 11). Patients with progressive lymphoma after pembrolizumab are indicated in red (“nonresponders,” which includes PD). Patients with clinical benefit after pembrolizumab are indicated in blue (“responders,” which includes SD, PR, and CR). (E) Cumulative dot plot showing the change (difference) in CAR T-cell percentage between the first measurement after pembrolizumab and pre-pembrolizumab baseline in patients with PD, SD, PR, and CR (n = 10).
Figure 3.
Figure 3.
Immune profiling of CAR T cells and non–CAR T cells in peripheral blood. (A) t-SNE plots of all T cells, CAR T cells, and nonCAR T cells using phenograph analysis. (B) Heatmap showing the composition of the 29 clusters of T cells most different in proportions in CAR T cells compared with nonCAR T cells. Arrows indicate clusters significantly different. Each cluster is a column. Clusters increased in CAR T cells are topped by a purple bar. Clusters decreased in CAR T cells are topped by a pink bar. (C) Composition of cluster 15 and 24 T cells after marker normalization. (D) Stacked bar plot of relative contribution of each group in each cluster (x-axis) vs fraction of cells in that cluster from each group (y-axis). All the CD3 T cells from all patients within a group were combined, and the percentage of total cells contributed to each cluster by each group was calculated and displayed on the y-axis (ie, 50% of cells in cluster X are from the nonresponders group; sum of contributions from the 2 groups for each cluster equals 100%, and each group is displayed in a different color). The clusters (listed on the x-axis) were ordered by relative abundance of responders CD3 T cells. Arrows highlight key clusters that are significantly different (further shown in panel B). (E) Heatmap showing the composition of the 6 clusters of CAR T cells significantly different in proportions in responders compared with nonresponders. Each cluster is a column. Clusters increased in responders are topped by a purple bar. Clusters decreased in responders are topped by a pink bar. (F) Heatmap showing the composition of the 3 clusters of nonCAR T cells significantly different in proportions in responders compared with nonresponders. Patients with clinical benefit (CR, PR, SD) at 3 months are grouped as responders, and patients with PD at 3 months are grouped as nonresponders (n = 11).
Figure 4.
Figure 4.
CAR T cells and non–CAR T cells in samples from baseline and after pembrolizumab, analyzed separately. (A) t-SNE plots of CAR T cells in responders and nonresponders, before and after pembrolizumab therapy, using phenograph analysis. (B) Boxplot showing frequencies of each of the defined clusters in panel A. (C) Composition of cluster 17 CAR T cells after marker normalization. (D) t-SNE plots of nonCAR T cells in responders and nonresponders, before and after pembrolizumab therapy, using Phenograph analysis. (E) Boxplot showing frequencies of each of the defined clusters in panel D. (C) Composition of cluster 17 and 18 nonCAR T cells after marker normalization. Patients with clinical benefit at 3 months are grouped as responders, and patients with PD at 3 months are grouped as nonresponders.

Source: PubMed

Подписаться