High Antigenicity for Treg Cells Confers Resistance to PD-1 Blockade Therapy via High PD-1 Expression in Treg Cells

Abstract

Regulatory T (T_reg) cells have an immunosuppressive function, and programmed death-1 (PD-1)-expressing T_reg cells reportedly induce resistance to PD-1 blockade therapies through their reactivation. However, the effects of antigenicity on PD-1 expression in T_reg cells and the resistance to PD-1 blockade therapy remain unclear. Here, we show that T_reg cells gain high PD-1 expression through an antigen with high antigenicity. Additionally, tumors with high antigenicity for T_reg cells were resistant to PD-1 blockade in vivo due to PD-1⁺ T_reg-cell infiltration. Because such PD-1⁺ T_reg cells have high cytotoxic T lymphocyte antigen (CTLA)-4 expression, resistance could be overcome by combination with an anti-CTLA-4 monoclonal antibody (mAb). Patients who responded to combination therapy with anti-PD-1 and anti-CTLA-4 mAbs sequentially after primary resistance to PD-1 blockade monotherapy showed high T_reg cell infiltration. We propose that the high antigenicity of T_reg cells confers resistance to PD-1 blockade therapy via high PD-1 expression in T_reg cells, which can be overcome by combination therapy with an anti-CTLA-4 mAb.

Keywords: CTLA‐4; PD‐1; antigenicity; cancer immunotherapy; regulatory T cell.

FIGURE 1

Relationship between TCR stimulation and PD‐1 expression in human eT_reg cells. (A) Gating strategy for eT_reg cells. To analyze eT_reg cells, we used CD45RA and Foxp3, and defined CD45RA⁻Foxp3^highCD4⁺ T cells as eT_reg cells. (B–E) PD‐1 (B), CTLA‐4 (C), ICOS (D), and GITR (E) expression in eT_reg cells. Peripheral blood mononuclear cells (PBMCs) were cultured with anti‐CD3 and anti‐CD28 mAbs, and IL‐2 for 48 h and subjected to flow cytometry. Fold changes were calculated by normalizing the mean fluorescence intensity (MFI) of the no‐antibody group (0 μg/mL) to 1. Representative flow cytometry staining (left) and summaries (right) are shown (n = 3). (F) CTLA‐4, ICOS, and GITR in eT_reg cells according to PD‐1 expression in the condition with 0.5 μg/mL anti‐CD3 mAb. In vitro experiments were performed as described in (B–E). Gating strategy for PD‐1 expression (top) and summaries (bottom/left, CTLA‐4; bottom/middle, ICOS; and bottom/right, GITR) are shown (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; bars, mean; error bars, SEM.

FIGURE 2

Relationship between TCR stimulation and PD‐1 expression in mouse T_reg cells, and their suppressive function according to PD‐1 expression. (A–D) PD‐1 (A), CTLA‐4 (B), ICOS (C), and GITR (D) expression in T_reg cells. Splenocytes from C57BL/6J mice were cultured with anti‐CD3 and anti‐CD28 mAbs, and IL‐2 for 48 h and subjected to flow cytometry. Fold changes were calculated by normalizing the mean fluorescence intensity (MFI) of the no‐antibody group (0 μg/mL) to 1. Representative flow cytometry staining (left) and summaries (right) are shown (n = 3). (E) CTLA‐4, ICOS, and GITR in T_reg cells according to PD‐1 expression in the condition with 0.5 μg/mL anti‐CD3 mAb. In vitro experiments were performed as described in (A–D). Gating strategy for PD‐1 expression (leftmost) and summaries (second from the left, CTLA‐4; second from the right, ICOS; and rightmost, GITR) are shown (n = 3). (F) Suppression assay of T_reg cells. CTV‐labeled responder CD8⁺ T cells (T_resp cells) from mouse splenocytes were cocultured with or without unlabeled sorted PD‐1⁻ or PD‐1⁺ T_reg cells. Anti‐PD‐1 mAb was added to some wells. The proliferation of T_resp cells was assessed 3 days later by dilution of CTV‐labeled cells using flow cytometry. The fold change in CTV dilution of T_resp cells was calculated. Summary is shown (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; bars, mean; error bars, SEM.

FIGURE 3

Relationship between antigenicity and PD‐1 expression in T_reg cells in vitro. PD‐1 (A), CTLA‐4 (B), ICOS (C), and GITR (D) expression in OT‐II T_reg cells. CD25⁺CD4⁺ T_reg cells were purified from OT‐II splenocytes and stimulated with gamma‐irradiated antigen‐presenting cells pulsed with various peptides (OVA‐I, SIINFEKL; OVA‐II, ISQAVHAAHAEINEAGR; OVA‐II (F), ISQAVHAAFAEINEAGR; and OVA‐II (R), ISQAVHAARAEINEAGR). After 48 h, presensitized T_reg cells were analyzed with flow cytometry. Fold changes were calculated by normalizing the mean fluorescence intensity of DMSO to 1. Representative flow cytometry staining (left) and summaries (right) are shown (n = 3). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant; bars, mean; error bars, SEM.

FIGURE 4

Relationship between antigenicity and PD‐1 expression in tumor‐infiltrating T_reg cells in vivo. (A) Experimental scheme. 5 × 10⁵ parental B16F10, B16F10/OVA‐I, B16F10/OVA‐II (F), or B16F10/OVA‐II cells were injected subcutaneously into B6 SCID mice on day 0. Subsequently, 5 × 10⁵ OT‐II T_reg cells were injected intravenously on day 4. Tumors were harvested on day 8 for tumor‐infiltrating lymphocyte (TIL) analysis with flow cytometry. (B–E) PD‐1 (B), CTLA‐4 (C), ICOS (D), and GITR (E) expression in tumor‐infiltrating OT‐II T_reg cells. Representative flow cytometry staining (left) and summaries (right) are shown (n = 5). *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns, not significant; bars, mean; error bars, SEM.

FIGURE 5

Analyses of tumor‐infiltrating WT and OT‐II T_reg cells in adoptive T‐cell transfer models. (A) Experimental scheme. 5 × 10⁵ B16F10/OVA or 1 × 10⁶ LL2/OVA cells were injected subcutaneously into B6 SCID mice on day 0. Subsequently, 5 × 10⁶ WT CD8⁺ T cells and/or 5 × 10⁵ WT or OT‐II T_reg cells were injected intravenously on day 4. Anti‐PD‐1, anti‐CTLA‐4 IgG2a, or control mAbs were administered intraperitoneally from day 5 onwards three times every 3 days. B16F10 tumors were harvested on day 12 for TIL analysis with flow cytometry. (B–E) PD‐1 (B), CTLA‐4 (C), ICOS (D), and GITR (E) expression in tumor‐infiltrating WT and OT‐II T_reg cells. Representative flow cytometry staining (left) and summaries (right) are shown (n = 5). (F) CTLA‐4, ICOS, and GITR in tumor‐infiltrating OT‐II T_reg cells according to PD‐1 expression. Summaries (left, CTLA‐4; middle, ICOS; and right, GITR) are shown (n = 5). MFI, mean fluorescence intensity; **p < 0.01; ***p < 0.001; ****p < 0.0001; bars or lines, mean; error bars, SEM.

FIGURE 6

Tumor growth and analyses of tumor‐infiltrating OT‐II T_reg cells in adoptive T‐cell transfer models treated with PD‐1 and/or CTLA‐4 blockade. In vivo experiments were performed as described in Figure 5A. (A) B16F10/OVA tumor growth curves in B6 SCID mice treated with adoptive T‐cell transfer combined with PD‐1 and/or CTLA‐4 blockade. (B–F) CTLA‐4 (B), ICOS (C), GITR (D), OX‐40 (E), and Ki67 (F) expression in tumor‐infiltrating OT‐II T_reg cells treated with or without PD‐1 blockade. Representative flow cytometry staining (left) and summaries (right) are shown (n = 5). (G) T_reg/CD8⁺ T‐cell ratios in TILs treated with or without PD‐1 and/or CTLA‐4 blockade. Summary is shown (n = 5). *p < 0.05; **p < 0.01; ***p < 0.001; bars, mean; error bars, SEM.

FIGURE 7

Clinical courses and IHC. (A) Computed tomography imaging of case 1 who responded to combination with anti‐PD‐1 and anti‐CTLA‐4 mAbs sequentially after primary resistance to an anti‐PD‐1 mAb. She experienced progressive disease in lung metastases 3 months after first line nivolumab monotherapy at the first evaluation. Subsequently, she received combination therapy with ipilimumab as second line treatment. Four months later, tumors dramatically responded to the combination therapy. She achieved complete response for more than 8 years even without any therapies. (B) Computed tomography imaging of case 2 who responded to combination with anti‐PD‐1 and anti‐CTLA‐4 mAbs sequentially after primary resistance to an anti‐PD‐1 mAb. He experienced progressive disease in lung metastases 4 months after first line nivolumab monotherapy at the first evaluation, but continued monotherapy for 9 months without any tumor shrinkage. However, lung metastases further exacerbated. Subsequently, he received combination therapy with ipilimumab as second line treatment. Three months later, tumors dramatically responded to the combination therapy. (C) Computed tomography imaging of case 6 who failed to respond to combination with anti‐PD‐1 and anti‐CTLA‐4 mAbs sequentially after primary resistance to an anti‐PD‐1 mAb. She experienced progressive disease in lung metastases 3 months after 1st line pembrolizumab monotherapy at the first evaluation. Subsequently, she received combination therapy of nivolumab and ipilimumab as second line treatment. Three months later, lung tumor size increased and new liver metastases appeared (an arrow) and the treatment was discontinued. She died 5 months after the discontinuation. (D, E) immunohistochemistry (IHC) of cases 1, 2 (D) and 3–6 (E). Formalin fixed paraffin embedded (FFPE) samples before the initiation of anti‐PD‐1 mAb monotherapy were stained with an anti‐Foxp3 mAb and representative staining are shown. Scale bars, 20 μm.