PD-1 and TIGIT downregulation distinctly affect the effector and early memory phenotypes of CD19-targeting CAR T cells

Abstract

CD19-targeting chimeric antigen receptor (CAR) T cells have become an important therapeutic option for patients with relapsed and refractory B cell malignancies. However, a significant portion of patients still do not benefit from the therapy owing to various resistance mechanisms, including high expression of multiple inhibitory immune checkpoint receptors. Here, we report a lentiviral two-in-one CAR T approach in which two checkpoint receptors are downregulated simultaneously by a dual short hairpin RNA cassette integrated into a CAR vector. Using this system, we evaluated CD19-targeting CAR T cells in the context of four different checkpoint combinations-PD-1/TIM-3, PD-1/LAG-3, PD-1/CTLA-4, and PD-1/TIGIT-and found that CAR T cells with PD-1/TIGIT downregulation uniquely exerted synergistic antitumor effects. Importantly, functional and phenotypic analyses suggested that downregulation of PD-1 enhances short-term effector function, whereas downregulation of TIGIT is primarily responsible for maintaining a less differentiated/exhausted state, providing a potential mechanism for the observed synergy. The PD-1/TIGIT-downregulated CAR T cells generated from diffuse large B cell lymphoma patient-derived T cells also showed robust antitumor activity and significantly improved persistence in vivo. The efficacy and safety of PD-1/TIGIT-downregulated CD19-targeting CAR T cells are currently being evaluated in adult patients with relapsed or refractory large B cell lymphoma (ClinicalTrials.gov: NCT04836507).

Keywords: CAR T; CD19; PD-1; TIGIT; clinical trial; immune checkpoints; shRNA.

Conflict of interest statement

Declaration of interests C.H.K., Y.-H.L., H.J.L., and Y.L. are inventors on a patent that was filed based on these works. C.H.K. is a co-founder of and holds equity in Curocell Inc. Y.K. is a consultant at Curocell Inc. H.C.K., Y.-H.L., and H.J.L. are employees at Curocell Inc. The remaining authors declare no competing interests.

Copyright © 2021 The American Society of Gene and Cell Therapy. Published by Elsevier Inc. All rights reserved.

Figures

Graphical abstract

Figure 1

PD-1 downregulation enhances the antitumor function of CAR T cells (A) Schematic representation of the lentiviral two-in-one vector carrying a CD19-CAR and shRNA expressing module. (B) PD-1 expression levels of CAR T cells with different shPD-1 candidates as determined by flow cytometry on day 2 after stimulation with gamma-irradiated Nalm-6-GL cells. Gray denotes the isotype control. Data are the pooled mean ± SD from three independent experiments, each performed in triplicate. (C) Cell counts from the homeostatic expansion of LNGFR+ CAR T cells with PD-1 downregulation candidates on days 3 and 6 after cell seeding. Data are the pooled mean ± SD from three independent experiments performed in triplicates. (D) The effects of hH1-, hU6-, and mU6-shPD1 on PD-1 expression was analyzed by flow cytometry 2 days after stimulation with gamma-irradiated K562-CD19 cells. Data are the pooled mean ± SD from two independent experiments performed in triplicates. (E) Cell counts from the homeostatic expansion of CAR T cells with each RNA Pol III promoter after LNGFR+ isolation on days 3 and 6 after cell seeding. Data are the pooled mean ± SD from two independent experiments performed in triplicates. (F) CAR T cells were incubated with GFP-expressing Nalm-6-GL or Nalm-6-GL-PD-L1 cells at a 1:1, 0.3:1, and 0.1:1 effector-to-target (E:T) ratio. GFP intensity was measured every 2 h using the IncuCyte S3 live cell imaging system. The relative percentage of total integrated GFP intensity was calculated as (GFP intensity at each time point/GFP intensity at 0 h) × 100. Representative mean ± SD from two independent experiments performed in triplicates. (G) CAR T cells were incubated with gamma-irradiated K562-CD19 or K562-CD19-PD-L1 cells at a 1:3 E:T ratio and counted on day 7. Data are the pooled mean ± SD from two independent experiments performed in triplicates. (H) 3 × 105 19GBBz or 19PBBz T cells were incubated with 9 × 106 gamma-irradiated Nalm-6-GL-PD-L1-CD80 cells every 6 days (first and second stimulation) and supernatants were collected on day 2 after each respective stimulation. Data are the pooled mean ± SD from four independent experiments. (I) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 leukemia cells. 5 days later, 1 × 106 CAR T cells were injected intravenously. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Data are from n = 3 mice (mock) and n = 5 mice (19BBz, 19GBBz, and 19PBBz). (J) PD-1 expression levels of CAR T cells from Nalm-6-GL-PD-L1-bearing mice at day 43. Gray denotes isotype control. Data are the mean ± SD from three mice per group. Statistical analysis was done by one-way ANOVA for (C), (E), (F), and (H) and an unpaired two-tailed t test for (G) and (J). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant.

Figure 2

CAR costimulatory domains differentially affect the in vivo persistence of PD-1-downregulated CAR T cells (A) PD-1 expression level of CAR T cells with CD28 or 41BB costimulatory domains 2 days after stimulation with gamma-irradiated K562-CD19 cells. Numbers denote the geometric mean fluorescence intensity (gMFI) of PD-1. (B) Unstimulated and (C) stimulated CAR T cells were incubated with Nalm-6-GL-PD-L1 cells at a 1:1, 0.3:1, and 0.1:1 E:T ratio and analyzed using the IncuCyte S3 system. Stimulated CAR T cells were generated by coincubation with Nalm-6-PD-L1-CD80 cells for 6 days prior to cytotoxicity assay. Data are the representative mean ± SD from two independent experiments performed in triplicates. (D) CAR T cells were incubated with gamma-irradiated Nalm-6-GL-PD-L1-CD80 or K562-CD19-PD-L1 cells at a 1:3 E:T ratio and counted on day 6 after each stimulation. Data are the mean ± SD from two independent experiments performed in triplicates. (E) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 leukemia cells. 5 days later, 1 × 106 CAR T cells were injected intravenously. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Data are from n = 3 mice (mock), n = 5 mice (19G28z and 19GBBz), and n = 4 mice (19P28z and 19PBBz). (F) The number of CAR T cells in mouse blood was determined on days 20 and 43 after CAR T cell injection. Data are mean ± SD from three mice per group. Statistical analysis for (A)–(D) and (F) was done by one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant.

Figure 3

Simultaneous downregulation of PD-1 and TIGIT further enhances the in vivo functionality of CD19-targeting CAR T cells (A) Schematic representation of the engineered two-in-one vector system carrying dual shRNA cassettes for two ICRs. (B) Dual downregulation efficiency of each ICR in CAR T cells stimulated for 48 h with gamma-irradiated K562-CD19 cells. FACS plots are representative data from two independent experiments performed in duplicates, and the bar graphs are the pooled mean ± SD. (C) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 leukemia cells. 5 days later, 1 × 106 CAR T cells with each dual downregulation (shGFP, shPD-1/shGFP, shPD-1/shTIM-3, shPD-1/shLAG-3, shPD-1/shTIGIT, shPD-1/shCTLA-4) were injected intravenously. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Data are from n = 3 mock, shPD-1/shTIM-3, and shPD-1/shLAG-3, n = 4 shGFP, shPD-1/shGFP, and shPD-1/shCTLA-4, and n = 5 shPD-1/shTIGIT. (D) Kaplan-Meier survival analysis with the log-rank (Mantel-Cox) test comparing each CAR T treated mice from (C). (E) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 leukemia cells. 5 days later, 0.5 × 106 or 0.25 × 106 CAR T cells with PD-1 (shPD-1/shGFP) or PD-1/TIGIT (shPD-1/shTIGIT) downregulation were injected intravenously. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Data are from n = 7 mice for the 0.5 × 106 dose groups and n = 6 mice for the 0.25 × 106 dose groups. (F) Kaplan-Meier survival analysis with log-rank (Mantel-Cox) test comparing CAR T treated mice from (E). Statistical analysis for (B) was done by one-way ANOVA. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant.

Figure 4

Downregulation of PD-1 and TIGIT distinctly affects the in vitro function of CAR T cells (A) 1 × 106 CAR T cells with single or dual downregulation were incubated with 3 × 106 gamma-irradiated Nalm-6-GL-PD-L1-CD80 cells every 6 days (first and second stimulation) and counted on day 6 after each respective stimulation. Data are the pooled mean ± SD from two independent experiments performed in triplicates. (B) 1 × 106 first-stimulated day 6 mice (19PBBz or 19PTBBz) were incubated with 3 × 106 gamma-irradiated Nalm-6-GL-PD-L1-CD80 cells with or without 10 μg/mL CD226 blockade antibody for 6 days and counted. Data are the mean ± SD from one experiment performed in triplicates. (C) The expression of CD45RO and CCR7 was measured to distinguish the differentiation state of CD4+ and CD8+ 19GBBz, 19PBBz, 19TBBz, or 19PTBBz cells on day 16 (10 days after the second stimulation). Data are the mean ± SD from 3 donors. TCM, CD45RO+CCR7+; TEM, CD45RO+CCR7−. (D) IncuCyte-based cytotoxicity kinetics of first stimulation CAR T cells against Nalm-6-GL-PD-L1 cells on day 6 at the indicated ratios. Data are the representative mean ± SD from two independent experiments performed in triplicates. Filled black dots represent co-culture with untransduced T cells. Statistical analysis for (A)–(C) was done by one-way ANOVA and for (D) by an unpaired two-tailed t test. ∗∗p < 0.01, ∗∗∗p < 0.0001, ∗∗∗∗p < 0.0001; ns, not significant.

Figure 5

RNA-seq analysis uncovers the distinct roles of PD-1 and TIGIT downregulation Second-stimulated 19GBBz, 19PBBz, 19TBBz, and 19PTBBz cells were prepared as shown in Figure S16 (A) for RNA-seq. (A) Pearson’s correlation analysis of the transcriptomic profiles and (B) hierarchical clustering of differentially expressed genes from second-stimulated 19GBBz, 19PBBz, 19TBBz, and 19PTBBz cells derived from two donors (FDR q ≤ 0.1). (C) Heatmap of selected genes associated with T cell function in second-stimulated 19GBBz, 19PBBz, 19TBBz, and 19PTBBz cells. Asterisks represent the statistical significance as measured by q value. (D) Normalized enrichment scores (NESs) of significantly enriched gene sets associated with phenotypic and functional T cell signatures in second-stimulated 19GBBz, 19PBBz, 19TBBz, and 19PTBBz cells as determined by GSEA analysis. For all gene sets, FDR q ≤ 0.03 unless otherwise indicated. ∗q < 0.05, ∗∗q < 0.01, ∗∗∗q < 0.001, ∗∗∗∗q < 0.0001.

Figure 6

Clinical-scale manufactured CAR T cells with PD-1/TIGIT downregulation showed a superior in vivo functionality against leukemia and lymphoma tumor models (A and B) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 cells. 5 days after tumor growth, mock, 19BBz-nt, and 19PTBBz-nt cells were intravenously injected at the indicated doses. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Data are from n = 4 mock mice and n = 6 mice for all CAR T cell treatment groups. (C) Kaplan-Meier survival analysis with log-rank test comparing the CAR T cell-treated mice at 0.25 × 106 dose from (A) and (B). (D) The number of CAR T cells in the blood and spleen was determined 43 days after CAR T cell injection at a 0.5 × 106 dose. Data are the mean ± SD from six mice per group. (E) Differentiation state of T cells in the spleen of mice 14 days after 0.5 × 106 CAR T infusion. The differentiation state of CD4+ and CD8+ CAR T cells was determined by flow cytometry for CCR7 and CD45RA expression. Data are the mean ± SD from 19BBz-nt-treated mice (n = 6) and 19PTBBz-nt-treated mice (n = 6). (F) IncuCyte-based cytotoxicity kinetics of CAR T cells in the spleen of mice 43 days after CAR T cell injection against Nalm-6-GL-PD-L1 cells at a 1:3 E:T ratio. Data are the mean ± SD from two mice per group performed in duplicate. (G) NSG mice were injected intravenously with 1 × 106 Nalm-6-GL-PD-L1 leukemia cells. 5 days later, 1 × 106 19BBz-nt or 19PTBBz-nt cells were injected intravenously. 3 days after CAR T cell infusion, a plasma was obtained from blood and used to measure GM-CSF levels. Data are from n = 2 (mock) and n = 4 (19BBz-nt and 19PTBBz) mice. Data are mean ± SD from each group. (H and I) NOG mice were injected subcutaneously on the right flank with 5 × 106 Raji-GL-PD-L1 lymphoma cells. When the mean tumor volume reached approximately 100 mm3, CAR T cells were injected intravenously at the indicated doses. Tumor burden was monitored based on the bioluminescence intensity from the IVIS imaging system. Statistical analysis for (D), (E), and (F) by an unpaired two-tailed t test and for (G) by a one-way ANOVA. ∗∗p < 0.01, ∗∗∗p < 0.0001, ∗∗∗∗p < 0.0001; ns, not significant.