CD39 delineates chimeric antigen receptor regulatory T cell subsets with distinct cytotoxic & regulatory functions against human islets

Abstract

Human regulatory T cells (Treg) suppress other immune cells. Their dysfunction contributes to the pathophysiology of autoimmune diseases, including type 1 diabetes (T1D). Infusion of Tregs is being clinically evaluated as a novel way to prevent or treat T1D. Genetic modification of Tregs, most notably through the introduction of a chimeric antigen receptor (CAR) targeting Tregs to pancreatic islets, may improve their efficacy. We evaluated CAR targeting of human Tregs to monocytes, a human β cell line and human islet β cells in vitro. Targeting of HLA-A2-CAR (A2-CAR) bulk Tregs to HLA-A2⁺ cells resulted in dichotomous cytotoxic killing of human monocytes and islet β cells. In exploring subsets and mechanisms that may explain this pattern, we found that CD39 expression segregated CAR Treg cytotoxicity. CAR Tregs from individuals with more CD39^low/- Tregs and from individuals with genetic polymorphism associated with lower CD39 expression (rs10748643) had more cytotoxicity. Isolated CD39^- CAR Tregs had elevated granzyme B expression and cytotoxicity compared to the CD39⁺ CAR Treg subset. Genetic overexpression of CD39 in CD39^low CAR Tregs reduced their cytotoxicity. Importantly, β cells upregulated protein surface expression of PD-L1 and PD-L2 in response to A2-CAR Tregs. Blockade of PD-L1/PD-L2 increased β cell death in A2-CAR Treg co-cultures suggesting that the PD-1/PD-L1 pathway is important in protecting islet β cells in the setting of CAR immunotherapy. In summary, introduction of CAR can enhance biological differences in subsets of Tregs. CD39⁺ Tregs represent a safer choice for CAR Treg therapies targeting tissues for tolerance induction.

Keywords: Treg-regulatory T cell; chimeric antigen receptor; cytotoxicity; immunoregulation; type 1 diabetes.

Figure 1

CAR Treg cytotoxicity against monocytes is dichotomous with Tregs from CD39^low donors being more cytotoxic. (A) Validation of HLA-A2-CAR expression and specificity toward HLA-A2 antigens. Tregs were mock transfected or transfected with HLA-A2 or control (1x9Q, 9Q) CARs and co-cultured with HLA-A2 (A2) positive monocytes. Activation of Tregs by A2-CARs were evaluated by surface markers, CD25, CD69, PD-1 and LAG3. (B) An experimental workflow to examine the cytotoxicity of A2-CAR Tregs towards A2 positive monocytes. (C) Viability of monocytes targeted by A2-CAR Tregs from different donors showed a bimodal distribution. The graph represents the percent of live (7-AAD⁻ Annexin V⁻) monocytes co-cultured with 9Q- or A2-CAR Tregs (n=12). A2-CAR Tregs from donors reduced monocyte viability were defined as cytotoxic (magenta), while the other donors were defined as non-cytotoxic (blue). (D) The viability difference between monocytes co-cultured with 9Q- and A2-CAR Tregs is calculated to determine the cytotoxicity. Percent Δ alive mono (Δ viability) is the % live monocyte co-cultured with 9Q-CAR Tregs subtracted by the % live monocyte co-cultured with A2-CAR Tregs. Blue indicates Δ viability <10%; magenta indicates increased cytotoxicity with a Δ viability ≥ 10%. (E) Cytotoxic Treg cohort showed lower percentage of CD39⁺ cells. Percentage of CD39⁺ cells within bulk Tregs was compared between cytotoxic (magenta) and non-cytotoxic donor cohorts (blue). (D, E) Data are presented as mean ± SEM, from n= 8-10 per group. ****P < 0.0001; by Student’s t test.

Figure 2

Characterization of CD39⁺ and CD39⁻ Treg subsets showed that CD39⁺ Tregs expressed higher levels of FOXP3 and lower Granzyme B. (A) The gating hierarchy to sort Teff, bulk Tregs and CD39⁺ and CD39⁻ Treg subsets. Adjacent graph showed the proportion of CD39⁺ cells in each subset during expansion. (mean ± SEM, n = 4). (B) MLR suppression assay among bulk, CD39⁺ and CD39⁻ Tregs. Each expanded Treg subset and CD4⁺ Teff cells were labeled and co-cultured as described in Materials and Methods, and the proliferation of Teff was analyzed by flow cytometry with the percent of Teff inhibition shown. (mean ± SEM, n = 3–4). (C) Percentage of CD25⁺FOXP3⁺ cells among bulk, CD39⁺, CD39⁻ Tregs and Teff cells on day 0 and 12 of cell expansion (mean ± SEM, n =12–20). (D–I) Percentage of CD45RO⁺ CD62L⁺ central memory(cm), HELIOS⁺, Tim3⁺, PD-1⁺, PD-L1⁺ or PD-L2⁺ cells in CD39⁺ and CD39⁻ Tregs on day 0 and 12 of cell expansion. (mean ± SEM, n =5–20). (J) Percentage of Granzyme B⁺ cells among CD39⁺, CD39⁻ and bulk Tregs on day 0 and 12 of cell expansion. (mean ± SEM, n =3–19). (K) Culture expanded CD39⁺ and CD39⁻ Tregs do not show increased cytotoxicity against allogeneic monocytes after co-culture. (L) A2-CAR transduced CD39⁻ Tregs showed increased cytotoxicity against allogeneic HLA-A2⁺ monocytes after co-culture. Cells were analyzed for early apoptotic (AnnexinV⁺7-AAD⁻), late apoptotic (AnnexinV⁺7-AAD⁺), and alive (AnnexinV⁻7-AAD⁻) cell populations. (M) Percentage of Granzyme B⁺ (GzmB⁺) and/or granulysin⁺ (GNLY⁺) cells in co-cultured CAR Tregs in L. Stacked bar plots represent mean ± SEM, n = 3–7. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant by two-way ANOVA, Tukey’s multiple comparisons test.

Figure 3

HLA-A2-CAR CD39⁻ Tregs can exert cytotoxic killing of human pancreatic β cell-line, βlox5. (A) Representative confocal microscopic images of cytotoxic killing. HLA-A2⁺ βlox5 cells were plated on coverslips overnight, and then 9Q- or A2-CAR transfected Tregs were added to co-culture for 24 hours. Cells were immuno-stained for HLA-A2 (red, βlox5 cells) and DAPI (blue), and then examined for apoptosis using TUNEL (green). % TUNEL positive cells of total cells per image are shown. Scale bar: 50 μm. (mean ± SEM, n= 2–7) (B) Representative microscopic images showing differences in βlox5 cell clustering with CAR Treg co-culture. 9Q- or A2-CAR transfected Tregs were added to βlox5 plated on a non-tissue culture treated 24-well plate for 48 hours and then imaged. Scale Bar: 200 μm. (C) CD39⁺ and CD39⁻ Tregs transfected with 9Q- or A2-CARs exert cytotoxic killing of βlox5 cells. Co-cultures of βlox5 cells and 9Q- or A2-CAR Tregs were set up as described in Materials and Methods. Viability was assessed by 7-AAD/Annexin V staining and flow cytometry. Percent of viable cells (AnnexinV⁻7-AAD⁻) were normalized and presented as fold relative to untreated βlox5 cells. (mean ± SEM, n= 15–18) (D) CD39 percentage gated on CD3⁺CD4⁺CD25^highCD127^low Treg cells by flow cytometry in PBMCs from healthy donors (HD) and T1D patients. (mean ± SEM, n= 38 HD and 23 T1D) (E) Expanded CAR Tregs derived from both healthy donors (HD) and T1D patients show enhanced cytotoxic killing of βlox5 cells. βlox5 and HD or T1D CAR Treg co-cultures were set up and analyzed as in (C) (mean ± SEM, n= 9 HD and 5 T1D). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001; ns, not significant by one-way ANOVA in (A), two-way ANOVA in (C, E) Tukey’s multiple comparisons test and by Mann-Whitney test in (D).

Figure 4

CAR Tregs induce PD-L1 and PD-L2 expression in βlox5 cells, which prevents CD39⁺ CAR Treg killing. (A) PD-L1 and PD-L2 expression in βlox5 cells co-cultured with CAR Tregs. 9Q- or A2-CAR transfected CD39⁺ and CD39⁻ Tregs were co-cultured with βlox5 cells for 24 hours. PD-L1 and PD-L2 expression were analyzed by flow cytometry in βlox5 cells (mean ± SEM, n=8–13). (B) PD-1/PD-L1 blockade increased cytotoxic killing of CD39⁺ CAR Tregs. CAR Treg and βlox5 cell co-cultures were set up as in (A) with or without PD-1 (50 µg/mL), PD–L1 (75 µg/mL) and/or PD–L2 (10 µg/mL) blocking antibodies. Viability was assessed by 7–AAD/Annexin V staining and flow cytometry. (mean ± SEM, n=5–14). (C) Cytokine production of bulk, CD39⁺ and CD39⁻ Tregs. Bulk, CD39⁺ and CD39⁻ Tregs were treated with PMA–Ionomycin for 4 hours, then proceeded with intracellular staining for IFN-γ, TNF-α and IL–10 (mean ± SEM, n=7–8). (D) βlox5 cells were treated with IFN-γ (20 ng/mL), TNF-α (20 ng/mL) or IL–10 (20 ng/mL) for 24 hours, and viability was assessed as in (B) (mean ± SEM, n=3). (E) βlox5 treated as in (D) were assessed for PD–L1 and PD–L2 expression (mean ± SEM, n=3). (F) CAR Treg and βlox5 cell co–cultures were set up as in (A) with or without blocking antibodies of IFN-γ (10 µg/mL) and/or TNFα (10 µg/mL), and viability was assessed. (mean ± SEM, n=3–14). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns, not significant by one–way ANOVA, Tukey’s multiple comparisons test.

Figure 5

CD39 expression and cytotoxicity varies by donor genotype at rs10748643 and manipulation of CD39 levels alters cytotoxicity and suppression activity. (A) Genotype of healthy and T1D donors at rs10748643 obtained via SNP microarray was correlated with CD39 expression and CAR Treg cytotoxicity. (B) Cytotoxicity with CAR Tregs was specifically enriched in donors with the AA genotype at rs10748643. P value by Fisher Exact Test. (C) Knockout of CD39 increases cytotoxic potential. CD39⁺ Tregs was knocked out of CD39 by CRISPR–Cas9 and tested for knockout efficiency 4–5 days post–transfection. (D) CD39⁺, CD39KO (KO) and CD39⁻ Tregs were transfected with 9Q– or A2–CARs as indicated and co–cultured with βlox5 cells. Cell viability was assessed by flow cytometry. (mean ± SEM, n=5). (E) Mixed lymphocyte suppression assays with A2–CAR transfected CD39⁺ and CD39KO (KO) Tregs were set up as described in Materials and Methods at a Treg: Teff ratio of 1:2 and analyzed. (mean ± SEM, n=4). (F) Bulk Tregs from a CD39^low donor were lentiviral transduced to over–express CD39 and GFP. FACS–enriched GFP⁺ Tregs were expanded and analyzed for stable expression of CD39 (CD39OE). (G) Overexpression of CD39 reduces cytotoxic killing of CAR Tregs against βlox5. CD39^low bulk, CD39OE, and CD39⁻ Tregs were transfected with 9Q–, or A2–CARs as indicated, co–cultured and analyzed as in (D) (mean ± SEM, n=5). (H) Suppression activities of A2–CAR transfected CD39^low and CD39OE bulk Tregs were set up and analyzed as in (E) *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant by one–way ANOVA, Tukey’s multiple comparisons test in (D, G) and by Student’s t test in (G, H).

Figure 6

CD39⁺ A2–CAR Tregs are less cytotoxic to primary human islet β cells, which depends on the PD–L1 pathway to prevent cytotoxicity. (A) Microscopic evaluation of HLA–A2⁺ primary islets’ response to CAR Tregs. Representative images of intact human islets co–cultured with indicated CAR Tregs for 48 hours. Live/dead islet cells were stained with calcein–AM (green)/ethidium homodimer–1(magenta) and imaged by fluorescent microscope. Average size of dead cell foci per image per condition was analyzed using ImageJ and graphed in adjacent scatter plot. Scale bar: 50 μm. (mean ± SEM, n=4–12 images) (B) Insulin (INS) gene expression. Intact human islet–Treg co–cultures set up as in A for 24 hours. Post co–cultured islets were collected. RNA was extracted and assessed for insulin gene expression using quantitative RT–PCR. (mean ± SEM, n=4) (C) Flow cytometric evaluation of primary islet viability. HLA–A2⁺ human islets were dispersed into single cells and co–cultured with bulk CAR Tregs for two days and stained with HPi2, HPx1, 7–AAD and Annexin V. Percent of viable β cells (HPi2⁺ HPx1⁻ AnnexinV⁻7–AAD⁻) were normalized and presented as fold relative to untreated β cells. (mean ± SEM, n=8–13) (D) Evaluation of CD39⁺ and CD39⁻ CAR Treg cytotoxicity to primary islet β cells. Islet cells were co–cultured with 9Q– or A2–CAR transduced bulk, CD39⁺ and CD39⁻ Tregs for 48 hours and cell viability was assessed as in C (mean ± SEM, n=6–8). (E, F) Evaluation of the PD1/PD–L1/2 pathway. Islet cell co–cultures were set up with indicated CAR Tregs as above and analyzed for PD–1 expression in Tregs and PD–L1 and PD–L2 expression in β cells. (mean ± SEM, n=3–6). (G) CAR–Treg and islet cell co–cultures were set up as in D with or without PD–1 (10 µg/mL), PD–L1 (10 µg/mL) and/or PD–L2 (10 µg/mL) blocking antibodies and analyzed for islet cell viability (mean ± SEM, n=4–5). *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant by one–way ANOVA, Tukey’s multiple comparisons test in (A, C, E–G) and by Mann–Whitney test in (B, D).

Figure 7

Alternative CAR Tregs targeting NTPDase3 (NTP–CAR) were less cytotoxic to human islet β cells. (A) Representative flow staining of 1x9Q–CAR (top row) or NTP–CAR Tregs (bottom row) co–cultured with HEK–293 cells either expressing NTPDase3 (Blue histograms) or not (Red histograms). Relative expression levels of CAR molecule, along with activation markers CD25 and CD69, as well as CD62L, PD–1, CTLA4, and LAG–3 are shown. (B) Representative flow staining of 1x9Q–CAR, A2–CAR, or NTP–CAR Tregs cultured with dispersed HLA–A2⁺ human islet cells. A2–CAR and NTP–CAR Tregs showed elevated levels of activation and suppressive markers compared to 9Q–CAR Tregs. (C) Flow cytometric evaluation of primary islet β cell viability after co–culturing with CAR Tregs. HLA–A2⁺ human islets were dispersed to single cells and co–cultured with bulk 9Q–CAR, A2–CAR, or NTP–CAR Tregs for 48 hours and stained for HPi2, HPx1, 7–AAD and Annexin V. Percent of viable β cells (HPi2⁺ HPx1⁻AnnexinV⁻7–AAD⁻) were normalized and presented as fold relative to untreated β cells. (mean ± SEM, n=9–14) (D) Evaluation of the PD1/PD–L1/2 pathway. NTP–CAR Treg and islet cell co–cultures were set up as in (C) with or without PD–1 (10 g/mL), PD–L1 (10 g/mL) and/or PD–L2 (10 g/mL) blocking antibodies and analyzed for β cell viability. (mean ± SEM, n=3–8) *P < 0.05, **P < 0.01, ***P < 0.001, ns, not significant by one–way ANOVA, Tukey’s multiple comparisons test in (C, D).