Inhibition of dendritic cell differentiation and accumulation of myeloid-derived suppressor cells in cancer is regulated by S100A9 protein

Abstract

Accumulation of myeloid-derived suppressor cells (MDSCs) associated with inhibition of dendritic cell (DC) differentiation is one of the major immunological abnormalities in cancer and leads to suppression of antitumor immune responses. The molecular mechanism of this phenomenon remains unclear. We report here that STAT3-inducible up-regulation of the myeloid-related protein S100A9 enhances MDSC production in cancer. Mice lacking this protein mounted potent antitumor immune responses and rejected implanted tumors. This effect was reversed by administration of wild-type MDSCs from tumor-bearing mice to S100A9-null mice. Overexpression of S100A9 in cultured embryonic stem cells or transgenic mice inhibited the differentiation of DCs and macrophages and induced accumulation of MDSCs. This study demonstrates that tumor-induced up-regulation of S100A9 protein is critically important for accumulation of MDSCs and reveals a novel molecular mechanism of immunological abnormalities in cancer.

Figure 1.

Expression of S100A9 and A8 in myeloid cells in tumor-bearing mice. (A) HPCs from bone marrow of naive mice were cultured with GM-CSF and IL-4 for 7 d in the presence of 25% (vol/vol) conditioned media from NIH 3T3 fibroblasts (3T3) or CT26 colon cancer cells (CT26). RNA was extracted on days 0, 1, 3, and 7, and qRT-PCR was performed. The level of expression of S100A9 was normalized to 18S rRNA. Two experiments yielded the same results. (B) HPCs from naive C57BL/6 mice were cultured for 5 d with GM-CSF and IL-4 in complete medium alone (control) or conditioned media from EL-4 cells (TCM). Gr-1⁺ cells were isolated at the end of 5-d culture and added at 1:4 ratio to splenocytes from OT-1 mice. The number of IFN-γ–producing cells in response to stimulation by the specific or control peptides was evaluated in an ELISPOT assay and calculated per 2 × 10⁵ splenocytes. S.P. represents the number of spots in cells stimulated with control peptide subtracted from the number of spots in cells stimulated with specific peptide. In parallel, Gr-1⁺ cells were added at 1:4 ratio to splenocytes from naive C57BL/6 mice and stimulated with 0.5 μg/ml anti-CD3 and 5 μg/ml anti-CD28 antibodies. The number of IFN-γ cells was evaluated in ELISPOT assay. Two experiments were performed. * represents the statistically significant (P < 0.05) differences between control and TCM groups. (C) HPCs from C57BL/6 mice were cultured for 5 d with GM-CSF and IL-4 in the presence of conditioned media (25% vol/vol) from NIH 3T3 fibroblasts (NIH 3T3) or C3 fibrosarcoma cells (C3). Gr-1⁺ or Gr-1⁻ cells were isolated using magnetic beads by two rounds of isolation. Proteins extracted from cells were subjected to Western blotting using indicated antibodies. Two experiments with the same results were performed. (D) Expression of S100A8 and A9 proteins in whole cell lysates of splenocytes from CT26 tumor-bearing mice. N, naive tumor-free mice. (E) Gr-1⁺ and Gr-1⁻ cells were isolated from spleens of naive mice (control) or mice bearing 3-wk-old CT26 tumor (tumor) as described above. The levels of proteins were evaluated by Western blotting. Two experiments with the same results were performed.

Figure 2.

Lack of S100A9 affects generation of MDSCs under pathological conditions. (A) Enriched HPCs were isolated from bone marrow of S100A9 KO mice (KO) and their wild-type littermates (WT). Cells were cultured with GM-CSF and IL-4 for 5 d in complete culture medium or in the presence of EL-4 TCM. The cell phenotypes were evaluated by flow cytometry. Mean ± SD of the proportions of indicated cell populations from three mice are shown. (B) EL-4 cells (3 × 10⁵) were injected s.c. into wild-type (WT) or S100A9 knockout (KO) mice and tumor size was measured. Anti-CD8 antibody (200 μg) was injected i.p. into KO mice 3 and 1 d before injection of tumor cells. The S100A9KO group included 12 mice, WT group 15 mice, and KO mice treated with anti-CD8 antibody 5 mice. Tumor size for each mouse is shown. (C) Phenotypes of splenocytes from wild-type (S100A9^+/+) and knockout (S100A9^−/−) mice 15 d after EL-4 cell injection were evaluated using multicolor flow cytometry. * represents statistically significant differences between the groups (P < 0.05). (D) Wild-type and S100A9 knockout mice were injected s.c. with 200 μl of CFA. The presence of Gr-1⁺CD11b⁺ cells was monitored in peripheral blood. Each group included 4 mice. Mean ± SD are shown. (E) Gr-1⁺CD11b⁺ cells in spleens of wild-type and S100A9 knockout mice on day 12 after CFA injection. Four mice per genotype were analyzed. * represents statistically significant differences between the groups (P < 0.05). (F) S100A9^−/− mice were injected s.c. with 2 × 10⁵ EL-4 tumor cells and then split into 2 groups of 5 mice. One group was left untreated (no transfer) and the other group was treated i.v. with 3 × 10⁶ Gr-1⁺CD11b⁺ cells isolated from tumor-bearing wild-type mice (MDSC) on days 1, 3, 5, and 7 after tumor inoculation. Tumor size for each mouse is shown.

Figure 3.

Effect of S100A9 overexpression on DC differentiation from ES cells. ES cells were transfected with empty vector (vector), S100A1, S100A8, S100A9, or a combination of S100A8 and 9 (S100A8/9). (A) Expression of S100A8 and A9 was evaluated in transfected ES cells using qRT-PCR. (B) Western blot assay for S100A9 and A8 in ES cells transfected with pcDNA-S100A9 and pcDNA-S100A8. Lane 1, empty vector; lane 2, S100A8; lane 3, S100A9; and lane 4, S100A8/9. (C) Phenotypes of cells generated from ES cells cultured for 35 d using the combination of cytokines described in Materials and methods. Expression of different surface molecules was evaluated by flow cytometry. Mean and SD from four different experiments are shown. (D) Effector T cells were isolated from allogeneic BALB/c mice and 10⁵ cells per well were mixed in triplicates of U-bottomed 96-well plates with different numbers of irradiated cells generated from ES cells. Vector, ES cells transfected with empty vector; S100A8/9, ES cells transfected with S100A8 and A9. Two different clones (#6 and #7) of ES cells are shown. HPC-DC, DCs generated from bone marrow progenitor cells by 5-d culture with GM-CSF and IL-4. T cell proliferation was measured by ³[H]thymidine uptake and expressed as counts per minute (CPM). Three experiments were performed. Mean ± SD is shown. T cells alone showed <500 CPM count (not depicted). (E and F) Cells were generated from ES cells and either used for analysis of phenotype by flow cytometry (E) or for colony formation in semisolid medium supporting the growth of myeloid colonies (StemCell Technologies). (F) Colony formation was evaluated in duplicates, and the number of colonies was calculated per 10⁵ cells.

Figure 4.

Effect of S100A9 overexpression on myeloid cell differentiation in transgenic mice. (A) Vector used for generation of transgenic mice. (B, left) GFP⁺ cells in peripheral blood of wild-type (WT) or transgenic mice (Tg) that express S100A9. (right) The level of S100A9 protein in spleens of wild-type (WT) and S100A9 transgenic mice (Tg). (C) Bone marrow progenitor cells were enriched by lineage depletion kit (MACS), and then stained with a cocktail of lineage-specific antibodies. Lin⁻c-kit⁺ GFP⁻ or GFP⁺ populations were sorted by flow cytometry. (D) Sorted cells were cultured for 5 d with cocktail of cytokines as described in Materials and methods. After that time, media was replaced and 0.5 × 10⁶ cells were plated in 24-well plates and cultured with GM-CSF alone for additional 7 d. LPS was added for the last 24 h of culture. The phenotype of cells was evaluated using multicolor flow cytometry. WT, wild-type FVB/N mice; GFP⁻ and GFP⁺, S100A9 Tg mice. Three experiments were performed. (E) Sorted cells were used as stimulators of allogeneic (BALB/c) T cells. Cell proliferation was measured in triplicate 4-d cultures by ³[H]thymidine uptake. Mean ± SD from two experiments are shown. T cells alone showed count <500 CPM.

Figure 5.

Phenotype and functional activity of myeloid cells in S100A9 transgenic mice. (A) S100A9Tg mice and their wild-type littermates were killed at indicated times after birth. The number of Gr-1⁺CD11b⁺ cells in spleens of wild-type (control) and transgenic mice were evaluated. GFP⁻ and GFP⁺ splenocytes were counted separately. Each group included three mice. Mean ± SD is shown. (B) The proportion CD11c⁺ DCs and F4/80⁺ macrophages in spleens of 3-wk-old wild-type and transgenic mice. (C) The proportion of indicated populations of cells in bone marrow of wild-type and S100A9Tg mice. In transgenic mice, the proportion of cells was calculated separately for GFP⁺ and GFP⁻ cells. Each group included three mice. Mean ± SD. * represents statistically significant differences between GFP⁺ and GFP⁻ cells (P < 0.05). (D) Colony formation assay of splenocytes and bone marrow cells isolated from wild-type or S100A9Tg mice. Total myeloid colonies were calculated per 2 × 10⁵ cells in spleen and per 2 × 10⁴ cells in bone marrow. * represents statistically significant differences between WT and Tg mice (P < 0.05). (E) Gr-1⁺CD11b⁺ IMC were sorted from spleens of wild-type (WT) or S100A9Tg mice (Tg) and added at a 1:3 or 1:6 ratio to splenocytes from FVB/N mice immunized with specific MHC class I–restricted PDSLRDLSVF peptide. Cells were stimulated with control (RAHYNIVTF) or specific peptides, and the number of IFN-γ–producing cells was evaluated in quadruplicates in ELISPOT assay. The numbers of spots in cells stimulated with control peptide were subtracted from values in cells stimulated with specific peptide. Mean ± SD from three experiments are shown. * represents statistically significant differences from splenocytes cultured without IMC (P < 0.05). (F) Gr-1⁺CD11b⁺ IMCs were sorted from WT or S100A9Tg mice (Tg). In transgenic mice, these cells were separated into GFP⁺ and GFP⁻ cells. T cells were isolated from naive FVB/N mice and DCs were generated from bone marrow of naive BALB/c mice. DCs were mixed with T cells at 1:50 ratio and IMC were added at the indicated ratio. Cells were cultured in triplicates in round-bottomed 96-well plates for 4 d. ³[H]thymidine uptake was measured. Mean ± SD from three experiments are shown. * represents statistically significant differences from splenocytes cultured without IMC (P < 0.05). (G) AVN cells (5 × 10⁵) were injected s.c. into wild-type or S100A9Tg mice and tumor size was monitored. Each group included six mice.

Figure 6.

Regulation of S100A9 expression by STAT3 (A) ChIP assay. Nuclear extracts from 32D cells were sonicated and ChIP with either anti-STAT3 antibody (STAT3) or control rabbit IgG (IgG) was performed. PCR was performed with primers specific for promoter regions of the S100A9 (A9), S100A8 (A8), or β-actin (C) genes. Input, PCR reaction performed with DNA isolated from nuclear extract without immunoprecipitation. (B and C) R1 ES cells were transfected with either control plasmid (R1-C) or Stat3C plasmid (R1-Stat3C) (56) and cultured with complete medium containing LIF. (B) The level of Stat3 protein was determined by Western blotting prior (0 h) and 48 h (48 h) after LIF withdrawal from ES cells. (C) The expression of S100A9 was measured by Southern blotting of RT-PCR products. (D) Deletion of STAT3 was achieved in STAT3^loxP/loxP x Mx-Cre transgenic mice by two successive i.p. injections of 100 μg poly(I:C) 7 d apart. Undeleted STAT3^loxP/loxP mice were used as a control. Gr-1⁺ and Gr-1⁻ cells were isolated from spleens of control (STAT3 +) and STAT3 deficient mice (STAT3 -), and the levels of STAT3, S100A8, and S100A9 proteins were evaluated by Western blotting. Two experiments with the same results were performed. (E and F) 32D cells were transfected with STAT3 siRNA. As a control, nontargeting siRNA pool was used (siGenome). (E) Expression of S100A8 and A9 mRNA was measured in quadruplicates by qPCR. Mean ± SD are shown. * represents statistically significant difference from control (P < 0.05). (F) The level of indicated proteins was evaluated in Western blotting. Two experiments with the same results were performed.

Figure 7.

S100A9 affects myeloid cell differentiation via up-regulation of ROS. (A) ROS levels in S100A9Tg mice. Splenocytes and bone marrow cells were collected from 3-wk-old wild-type and S100A9Tg FVB/N mice. Cells were stimulated with PMA, labeled with APC-conjugated anti–Gr-1 antibody, and loaded with the oxidation-sensitive dye hydroethidine. The level of ROS production was evaluated in GFP⁺ or GFP⁻Gr-1⁺ cells. (left) fluorescence histograms from one typical experiment. (right) Graphs representing mean fluorescence intensity (MFI) in all performed experiments. Red, Gr-1⁺ cells from wild-type mice; green, GFP⁻ Gr-1⁺ cells from S100A9 Tg mice; blue, GFP⁺Gr-1⁺ cells from S100A9 Tg mice. Three experiments with the same results were performed. (B) EL-4 tumors were established in wild-type and S100A9KO mice as described in Fig. 3. Splenocytes were collected 12 d after tumor inoculation. Cells were stimulated with PMA and labeled with APC-conjugated anti–Gr-1 antibody and PE-conjugated anti-CD11b antibody. ROS were measured in Gr-1⁺CD11b⁺ cells using DCFDA. Two experiments with similar results were performed. (C) EL-4 tumors were established in wild-type C57BL/6 or gp91^−/− mice. 3 wk later, when tumors reached 1.5 cm in diameter, MDSCs were isolated from spleens using magnetic beads and cultured for 7 d in vitro, as described in Fig. 3 B. Proportions of different cell populations were evaluated. Cumulative results from three experiments are shown. (inset) The level of S100A9 protein in freshly isolated MDSCs. (D) Phenotypes of cells in F1 offspring from crosses between S100A9Tg and gp91^phox KO mice or wild-type C57BL/6 mice. Lin⁻c-kit⁺ progenitor cells were sorted and cultured with cocktail of cytokines to generate myeloid cells. Proportions of different cell populations were evaluated by flow cytometry. Mean ± SD of cumulative results of two experiments is shown.