The response to lymphodepletion impacts PFS in patients with aggressive non-Hodgkin lymphoma treated with CD19 CAR T cells

Abstract

Factors associated with durable remission after CD19 chimeric antigen receptor (CAR)-modified T-cell immunotherapy for aggressive B-cell non-Hodgkin lymphoma (NHL) have not been identified. We report multivariable analyses of factors affecting response and progression-free survival (PFS) in patients with aggressive NHL treated with cyclophosphamide and fludarabine lymphodepletion followed by 2 × 106 CD19-directed CAR T cells/kg. The best overall response rate was 51%, with 40% of patients achieving complete remission. The median PFS of patients with aggressive NHL who achieved complete remission was 20.0 months (median follow-up, 26.9 months). Multivariable analysis of clinical and treatment characteristics, serum biomarkers, and CAR T-cell manufacturing and pharmacokinetic data showed that a lower pre-lymphodepletion serum lactate dehydrogenase (LDH) level and a favorable cytokine profile, defined as serum day 0 monocyte chemoattractant protein-1 (MCP-1) and peak interleukin-7 (IL-7) concentrations above the median, were associated with better PFS. MCP-1 and IL-7 concentrations increased after lymphodepletion, and higher intensity of cyclophosphamide and fludarabine lymphodepletion was associated with higher probability of a favorable cytokine profile. PFS was superior in patients who received high-intensity lymphodepletion and achieved a favorable cytokine profile compared with those who received the same intensity of lymphodepletion without achieving a favorable cytokine profile. Even in high-risk patients with pre-lymphodepletion serum LDH levels above normal, a favorable cytokine profile after lymphodepletion was associated with a low risk of a PFS event. Strategies to augment the cytokine response to lymphodepletion could be tested in future studies of CD19 CAR T-cell immunotherapy for aggressive B-cell NHL. This trial was registered at www.clinicaltrials.gov as #NCT01865617.

Conflict of interest statement

Conflict-of-interest disclosure: K.A.H. has served on advisory boards for Celgene. D.L. is an employee of and has equity ownership in Juno Therapeutics, a Celgene company. U.H.A. received research funding from Juno Therapeutics, a Celgene company. R.D.C. received research funding from Amgen, Incyte, Kite, a Gilead Company, Merck, Pfizer, and Seattle Genetics; and has served on advisory boards for Amgen, Jazz Pharmaceuticals, and Pfizer. H.-P.K. received research funding from Calimmune and Rocket Pharma; and has served on advisory boards for CSL, Homology, and Rocket Pharma. R.C.L. received research funding from Incyte, Juno Therapeutics, a Celgene company, Rhyzen Pharmaceuticals, Takeda, and TG Therapeutics. J.R. is an employee of and has equity ownership in Seattle Genetics. M.S. received research funding from Acerta Pharma, Beigene, Celgene, Genentech, Gilead Sciences, Mustang Bio, Pharmacyclics, and TG Therapeutics; and has served on advisory boards for AbbVie, AstraZeneca, Genentech, Qilu Puget Sound Biotherapeutics, and Verastem. B.G.T. received research funding from and has patents licensed to Mustang Bio. S.R.R. received research funding from and has patents licensed to Juno Therapeutics, a Celgene company; has equity ownership in Celgene; and has served on advisory boards for Adaptive Biotechnologies, Cell Medica, Juno Therapeutics, a Celgene company, and Nohla Therapeutics. D.G.M. received research funding from GlaxoSmithKline and Juno Therapeutics, a Celgene company. C.J.T. received research funding from Juno Therapeutics, a Celgene company, and Nektar Therapeutics; has patents licensed to Juno Therapeutics, a Celgene company; has served on advisory boards and has equity ownership in Caribou Biosciences, Eureka Therapeutics, and Precision Biosciences; and has served on advisory boards for Aptevo, Juno Therapeutics, a Celgene company, Kite, a Gilead Company, Nektar Therapeutics, and Novartis. The remaining authors declare no competing financial interests.

© 2019 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

PFS and OS in patients with aggressive NHL after CD19 CAR T-cell immunotherapy. Kaplan-Meier estimates of PFS (A) and OS (B) in patients with aggressive NHL who achieved CR (red) and in all patients (blue). The numbers of patients at risk at 6-month intervals are indicated.

Figure 2.

Serum biomarkers associated with PFS in patients with aggressive NHL. (A) The association of pre-lymphodepletion (pre-LD) serum LDH level with the hazard of a PFS event, adjusting for day 0 MCP-1 and peak IL-7 concentrations. The vertical dashed line shows the upper limit of normal LDH. (B) The association of serum day 0 MCP-1 with the hazard of a PFS event, adjusting for pre-LD LDH level and peak IL-7 concentration. (C) The association of serum peak IL-7 concentration with the hazard of a PFS event, adjusting for pre-LD LDH level and day 0 MCP-1 concentration. (D) Estimated combined effect of pre-LD LDH level and day 0 MCP-1 concentration on hazard of a PFS event. (E) Estimated combined effect of pre-LD LDH level and peak IL-7 concentration on hazard of a PFS event. The HRs are shown in the colored bars to the right of each figure in panels D-E. Serum biomarkers were modeled as restricted cubic splines with 3 knots.

Figure 3.

Superior PFS and OS in patients with aggressive NHL who achieved CR and had high peak IL-7 concentrations. Kaplan-Meier estimates of PFS (A) and OS (B) in patients with aggressive NHL who achieved CR and had serum peak IL-7 concentration above the median (red) compared with those who had serum peak IL-7 concentration below or equal to the median (blue). The numbers of patients at risk at 6-month intervals are indicated. Log-rank tests were used to compare between-group differences in survival probabilities. (C) Serum IL-7 concentrations in patients who remained in CR (red) and in those who subsequently relapsed (blue; P = .05 for the difference in the IL-7 area under the curve from peak to day 28 between the groups). Adjusted P values were calculated by using the Wilcoxon rank-sum test.

Figure 4.

Higher serum MCP-1 and IL-7 concentrations in patients with aggressive NHL who received more intensive lymphodepletion. Serum day 0 MCP-1 (A) and day 0 IL-7 (B) increase after lymphodepletion. Serum cytokine concentrations before lymphodepletion and on day 0 before CAR T-cell infusion for each patient are connected. P values comparing cytokine concentrations before lymphodepletion and on day 0 before CAR T-cell infusion were calculated by using the Wilcoxon rank-sum test. (C) The change in serum MCP-1 concentration between the pre-lymphodepletion and day 0 time points in patients who received high-intensity Cy/Flu lymphodepletion (red) and in those who received low-intensity Cy/Flu lymphodepletion (blue) are shown. Serum day 0 MCP-1 (D) and peak IL-7 (E) concentrations in patients who received high-intensity Cy/Flu lymphodepletion (red) and in those who received low-intensity Cy/Flu lymphodepletion (blue). Each point in panels C-E represents data from a single patient. Box and whisker plots show the median (bar) and interquartile range (box). Adjusted P values were calculated by using the Wilcoxon rank-sum test. (F) Serum day 0 MCP-1 and peak IL-7 concentrations according to lymphodepletion intensity (Spearman correlation r = 0.52; P = .0001). Each point represents data from a single patient.

Figure 5.

Higher CAR T-cell counts in patients with aggressive NHL who received high-intensity lymphodepletion or had high serum MCP-1 concentration. CAR T-cell copies/μg of DNA by qPCR in patients who received high-intensity Cy/Flu lymphodepletion (red) and in those who received low-intensity Cy/Flu lymphodepletion (blue; P = .0005 for the difference in the peak of CAR T cells between the groups) (A); and in patients with serum day 0 MCP-1 concentration above the median (red) or below or equal to the median (blue; P = .03 for the difference in the peak of CAR T cells between the groups) (B). Each thin line in panels A-B represents a single patient; bold lines represent the averaged data using local polynomial regression curve fitting approximation, with the standard error in gray. Adjusted P values were calculated by using the Wilcoxon rank-sum test. (C) Association of day 28 CAR T-cell counts by qPCR (log10 copies/μg DNA) and the hazard of PFS event using a restricted cubic spline with 3 knots. Tick marks represent data from individual patients.

Figure 6.

Impact of serum MCP-1 and IL-7 concentrations and lymphodepletion intensity on PFS. (A) Kaplan-Meier estimates of PFS according to patients with a favorable cytokine profile (serum day 0 MCP-1 and peak IL-7 concentrations above the median; red) compared with those with an unfavorable cytokine profile (serum day 0 MCP-1 and/or peak IL-7 concentrations below or equal to the median; blue). (B) Kaplan-Meier estimates of PFS in patients who received high-intensity lymphodepletion (LD; red) compared with those who received low-intensity lymphodepletion (blue). (C) Kaplan-Meier estimates of PFS in patients who received high-intensity lymphodepletion and did or did not achieve a favorable cytokine profile (day 0 MCP-1 and peak IL-7 concentrations above the median), and in those who received low-intensity lymphodepletion. (A-C) The numbers of patients at risk at 6-month intervals are indicated. Log-rank tests were used to compare between-group differences in survival probabilities. (D) Association of favorable (red) or unfavorable (blue) cytokine profile and the hazard of a PFS event according to pre-lymphodepletion serum LDH level, with 95% CI depicted in gray. Cytokines were modeled as restricted cubic splines with 3 knots. The horizontal line indicates the HR of a PFS event in the whole cohort. (E) Association of high-intensity (red) or low-intensity (blue) lymphodepletion and the hazard of a PFS event according to pre-lymphodepletion serum LDH level, with 95% CIs depicted in gray. The horizontal line indicates the HR of a PFS event in the whole cohort.