Armored human CAR Treg cells with PD1 promoter-driven IL-10 have enhanced suppressive function

Abstract

Regulatory T cell (T_reg cell) therapy has been transformed through the use of chimeric antigen receptors (CARs). We previously found that human T_reg cells minimally produce IL-10 and have a limited capacity to control innate immunity compared to type 1 regulatory T cells (T_r1 cells). To create "hybrid" CAR T_reg cells with T_r1 cell-like properties, we examined whether the PDCD1 locus could be exploited to endow T_reg cells with CAR-regulated IL-10 expression. CRISPR-mediated PD1 deletion increased CAR T_reg cell activation, and knock-in of IL10 under control of the PD1 promoter resulted in CAR-induced IL-10 secretion. IL10 knock-in improved CAR T_reg cell function, as determined by increased suppression of dendritic cells and alloantigen- and islet autoantigen-specific T cells. In vivo, IL10 knock-in CAR T_reg cells were stable, safe, and suppressed dendritic cells and xenogeneic graft-versus-host disease. CRISPR-mediated engineering to simultaneously remove an inhibitory signal and enhance a suppressive mechanism is a previously unexplored approach to improve CAR T_reg cell potency.

Fig. 1.. Manufacturing PD1^KO and IL-10^KI A2-CAR T_reg cells.

(A) Schematic diagram of the second-generation HLA-A2–specific CAR (A2-CAR) used throughout the study. (B) Schematic diagram of the PDCD1 locus and homology-directed repair templates (HDRTs). (C) Timeline for manufacturing gene-edited T_reg cells. (D) T_reg cell expansion during manufacture, with time points of CRISPR editing and selection of tCD19⁺ cells indicated. (E) Representative (top) and average (bottom) data from phenotypic characterization of the different types of A2-CAR T_reg cells as well as unedited T_conv cells, gated on live CD4⁺ cells. (F) PD1 expression in A2-CAR T_reg cells following stimulation with HLA-A2⁺PD-L1⁻ K562 cells. Gated on live CD4⁺cMyc⁺tNGFR⁺ and tCD19⁺ cells, where appropriate. Averaged data are means ± SEM with lines connecting data points from individual subjects (n = 3 to 33). FITC, fluorescein isothiocyanate. AF647, Alexa Fluor 647; PE-Cy7, Phycoerythrin-Cyanine 7; BUV737, Brilliant UltraViolet 737.

Fig. 2.. IL-10^KI T_reg cells secrete IL-10 following CAR stimulation.

(A to C) Different types of A2-CAR T_reg cells and A2-CAR T_r1 cells were stimulated with blank (Unstim.) or HLA-A2⁺PD-L1⁻ (CAR-stim.) K562 cells. After 72 hours, amounts of (A) IL-10 and (B) other cytokines were measured. In (C), cytokine data are expressed as doughnut plots to show the relative amount of IL-10 as a proportion of the 11 cytokines analyzed. The summed ng/ml of all 11 cytokines measured is provided below each doughnut plot. (D) A2-CAR T_reg cells and T_r1 cells were stimulated with HLA-A2⁺PD-L1⁻ K562 cells. Every 24 hours, cell supernatants were collected, the cells were washed twice with PBS, and replated in fresh medium with IL-2. (E) Quantitative polymerase chain reaction (qPCR) analysis of T_reg cell IL10 expression following coculture with HLA-A2⁺PD-L1⁻ K562 cells. qPCRs were performed using primers that amplified mRNA corresponding to total IL10 (left), endogenous IL10 (middle), and exogenous IL10 (right). Schematic diagrams showing primer recognition sites are provided. CDS, coding sequence. Averaged data are means ± SEM with connecting data points from individual subjects (n = 3 to 5). Statistical significance was determined using mixed-effects analysis tests with P values shown. ND, not detected.

Fig. 3.. IL-10^KI T_reg cells retain their characteristic transcriptome, which is distinct from T_r1 cells.

A2-CAR T_reg cell types, A2-CAR T_r1 cells and A2-CAR T_conv cells were stimulated with HLA-A2⁺ beads for 16 hours, then RNA was extracted and analyzed. (A) Principal component analysis. PC, principal component. (B) Venn diagram showing the total number of significantly up- and down-regulated genes in each comparison. (C) Volcano plots showing differential gene expression of Cas9 versus IL-10^KI T_reg cells (top) and IL-10^KI T_reg cells versus T_r1 cells (bottom). [(B) and (C)] Significant DEGs were identified on the basis of P < 0.05 and a log₂ fold change of 1. (D) Normalized read counts of endogenous, exogenous, and total (endogenous + exogenous) IL10 transcripts. Data are means ± SEM with data points from individual subjects connected (n = 3). (E to G) Heatmaps showing the relative expression of (E) select secreted factors, (F) T cell/T_reg cell–associated genes, and (G) the top 50 DEGs in an IL-10^KI T_reg cell versus T_r1 cell comparison. [(E) to (G)] Statistical significance was determined using a Wald test with Benjamini-Hochberg correction. No DEGs with P < 0.05 for Cas9 versus PD1^KO T_reg cells. †P < 0.05 for Cas9 versus IL-10^KI T_reg cells. §P < 0.05 for PD1^KO versus IL-10^KI T_reg cells. n.s., not significant.

Fig. 4.. IL-10^KI T_reg cells promote differentiation of DCs with a DC10-like phenotype.

(A) Schematic diagram of experiment. CD14⁺HLA-A2⁺ cells were differentiated into dendritic cells (DCs) with GM-CSF (100 ng/ml) and IL-4 (10 ng/ml), in the absence (iDC) or in the presence of the indicated types of A2-CAR T_reg cells (10:1 monocyte:T_reg cell ratio), or with rIL-10 (DC10). After 7 days, the DC phenotype was analyzed. (B) Representative and (C) averaged data showing expression of CD163 and CD141 in live CD11c⁺CD4⁻ DCs. (D) Expression of tolerogenic markers in live CD11c⁺CD4⁻ DCs, with data from control iDC or DC10 represented with purple and black dotted lines, respectively. (E) In some experiments, CD11c⁺CD4⁻ DCs were sorted at the end of the 7-day coculture and stimulated with LPS and IFN-γ for 48 hours, and supernatants were analyzed for the indicated cytokines. Cytokine secretion data are shown relative to DCs treated with Cas9 A2-CAR T_reg cells. Averaged data are means ± SEM with connecting data points from individual subjects (n = 5 to 7). Statistical significance was determined using mixed-effects analysis tests with P values shown. n.s., not significant.

Fig. 5.. IL-10^KI T_reg cells induce a tolerogenic phenotype in mDCs.

(A) HLA-A2⁺ mDCs were cultured with varying amounts of rIL-10, in the presence or absence of A2-CAR T_reg cells (not CRISPR modified), and, after 72 hours, CD80 and CD86 expression was measured. Statistical significance indicates mixed-effects analysis only of the CAR T_reg cell–treated DCs condition (B) Schematic diagram of DC suppression assay. mDCs were cocultured with different types of A2-CAR T_reg cells for 72 hours, after which expression of co-stimulatory/inhibitory molecules on live CD11c⁺CD4⁻ cells was measured. (C) Representative and (D) pooled data showing relative MFIs compared to mDCs that were cultured without T_reg cells. Averaged data are means ± SEM with connecting data points from individual subjects (n = 3 to 7). Statistical significance was determined using mixed-effects analysis tests without (A) or with (D) Fisher’s multiple comparisons test, with P values shown. n.s., not significant.

Fig. 6.. IL-10^KI CAR T_reg cells suppress alloantigen- and autoantigen-specific T cell proliferation.

(A) Schematic diagram of alloantigen-based suppression assay shown in (B) and (C). HLA-A2⁺ DCs were cocultured with HLA-A3⁻ T_reg cells for 72 hours. Allogeneic CPD eFluor 450–labeled HLA-A3⁺CD3⁺ responder T cells were then added, and cocultures were maintained for an additional 96 hours. (B) Representative CD4⁺ responder T cell proliferation, gated on live HLA-A3⁺CPD eFluor 670⁻CPD eFluor 450⁺CD4⁺ cells. (C) % Suppression of CD4⁺ responder T cell proliferation, relative to responder T cells cultured without T_reg cells. (D) Schematic diagram of islet antigen–based suppression assay shown in (F) to (H). HLA-A2⁺HLA-DR4⁺ DCs were cocultured with T_reg cells for 48 hours. CD4⁺4.13-TCR⁺ responder T cells and 1 nM GAD65 peptide were then added, and cocultures were maintained for an additional 48 hours. (E) CD4⁺4.13-TCR⁺ T cell proliferation following 96-hour coculture with immature HLA-DR4⁺ DCs and varying GAD65 peptide concentrations (without T_reg cells). Control cells were pulsed with an irrelevant hemagglutinin peptide (100 nM). Division indices provided. (F) Representative CD4⁺ responder T cell proliferation, gated on live CD4⁺CPD eFluor 670⁻mTCRβ⁺ cells. (G) % Suppression of responder T cell proliferation, relative to responder T cells stimulated without T_reg cells. (H) Relative cytokine analysis from 1:128 T_reg cell:responder T cell (T_resp cell) ratio. TNFα, IL-17A & IL-17F were not detected. Averaged data are means ± SEM with non-linear regression lines (n = 8 to 10). Statistical significance was determined using mixed-effects analysis with P values shown. n.s., not significant.

Fig. 7.. PD1^KO and IL-10^KI T_reg cells maintain their stability following chronic, in vivo CAR stimulation.

(A) Schematic diagram of model. 1.5 to 3.2 million A2-CAR T_reg cells were coadministered with an equal number of autologous PBMCs into preconditioned A2-NSG mice. Mice were euthanized after 21 days. (B) FOXP3 and Helios expression in hCD45⁺mCD45⁻hCD4⁺cMyc⁺ T_reg cells in blood or spleen. FOXP3 and Helios gates were set on hCD4⁺ cells in mice reconstituted with PBMCs alone. (C) Upon euthanasia, hCD45⁺mCD45⁻hCD4⁺cMyc⁺ T_reg cells were isolated from splenocytes by cell sorting. Data show the methylation status of the TSDR locus in the sorted T_reg cells. Averaged data are means ± SEM with connecting data points from individual subjects (n = 3 to 5). Statistical significance was determined using one-way analysis of variance (ANOVA) with P values shown. n.s., not significant.

Fig. 8.. IL-10^KI T_reg cells secrete IL-10 in vivo and suppress DCs and xenoGVHD.

(A) Schematic diagram of model. Preconditioned NSG mice were intravenously administered HLA-A2^neg A2-CAR T_reg cell types and HLA-A2^posCD4⁺ T_conv cells at T_reg cell:T_conv cell ratios of (B to D) 1:1, 1:2, or (E and F) 1:4. Mice were monitored and euthanized at 7 or 10 weeks after cell transfer, or when predefined endpoint criteria were met. (B) % Survival (top) and clinical scores (bottom) following cell transfer. (C) Absolute numbers of HLA-A2^pos T_conv cells (top) and HLA-A2^neg T_reg cells (bottom) in peripheral blood at days 14 and 28. (D) Quantities of human IL-10 (top) and IFN-γ (bottom) in mouse plasma over time. Other cytokines (IL-2, IL-4, IL-5, IL-6, IL-9, IL-13, IL-17A, IL-17F, IL-22, and TNFα) were not detected. (E and F) (E) Representative and (F) averaged data showing the phenotype of mouse CD45⁺CD11c⁺ cells isolated from the intestine or ear skin, 7 weeks after cell injection. Representative plots show the MFI of NOD mouse MHC class I (H-2Kd), MHC class II (I-Ag7), and CD86. Averaged data are means ± SEM with each point representing one mouse (n = 3 to 4). For each group, mice were administered T_reg cells that were isolated from a different human subject; connecting data points represent each individual subject. Statistical significance was determined using a log-rank Mantel-Cox test (B), paired two-tailed Student’s t tests [(C) and (F)], or two-way ANOVA (D) with P values shown. n.s., not significant.