Critical Role of Myeloid-Derived Suppressor Cells in Tumor-Induced Liver Immune Suppression through Inhibition of NKT Cell Function

Abstract

Metastasis followed by the tumor development is the primary cause of death for cancer patients. However, the underlying molecular mechanisms of how the growth of tumor resulted in the immune suppression, especially at the blood-enriched organ such as liver, were largely unknown. In this report, we studied the liver immune response of tumor-bearing (TB) mice using concanavalin A (Con A)-induced hepatitis model. We demonstrated that TB mice displayed an immune suppression phenotype, with attenuated alanine aminotransferase levels and liver damage upon Con A treatment. We also elucidated that large amounts of myeloid-derived suppressor cells (MDSCs) being influx into the liver in TB mice and these MDSCs were essential for liver immune suppression through both depletion and reconstitution approaches. We further determined that these MDSCs selectively suppressed the IFN-γ production deriving from NKT cells through membrane-bound transforming growth factor β (TGF-β). Finally, we defined a tumor-derived TGF-β-triggered CXCL1/2/5- and CXCR2-dependent recruitment of MDSC into the liver. In summary, our results defined a novel mechanism of liver immune suppression triggered by growing living tumor and provided possible therapeutic targets against these MDSCs.

Keywords: CXCR2; MDSC; TGF-β; liver immune suppression; remote tumor.

Figure 1

Alleviated concanavalin A (Con A)-induced hepatitis in tumor-bearing (TB) mice. Sex- and age-matched C57BL/6 wild-type mice were either untreated or subcutaneously injected with B16 or EL4 tumor cells (1 × 10⁶ cells/mouse) to prepare TB mice (TB-B16 or TB-EL4), respectively. On 15 days posttumor inoculations, these mice were challenged with Con A (10 mg/kg body weight). (A) Serum samples were collected at different time points post-Con A treatments and used for analysis of alanine aminotransferase (ALT) levels (n = 8). (B) Results from one representative experiment at the time point of 12 h posttreatment are shown. (C) Liver tissues collected at 12 h post-Con A treatments were fixed for hematoxylin and eosin staining, and one representative tissue staining is shown. Scale bars, 200 µm. (D) The relationship of serum ALT levels and tumor sizes was analyzed by SPSS, and results from one representative experiment are shown (R² = −0.5863, p = 0.0170).

Figure 2

Induction of immune suppression in tumor-bearing (TB) (B16) mice. (A) Sex- and age-matched wild-type (WT) mice were either untreated or inoculated with B16 tumor cells as described above (named TB-B16), followed by concanavalin A (Con A) injections. Serum samples were collected at 0, 2, 4, 6, 8, 12, and 24 h post-Con A treatments and used for determining cytokine levels by ELISA (n = 6). Results shown are one of three independent experiments. (B) Livers of WT and TB mice were harvested before and 2 h after Con A challenge. Fas mRNA levels were determined via quantitative real-time PCR (n = 3). (C) A total of 1 mg/kg anti-Fas (clone Jo2) were injected intravenously into WT and TB mice. At 12 h after injections, serum alanine aminotransferase levels were measured (n = 6). (D) Liver tissues were fixed for hematoxylin and eosin staining, and one representative tissue staining is shown. Scale bars, 200 µm.

Figure 3

Increased liver immune-suppressive cells in the liver of tumor-bearing (TB) mice. Sex- and age-matched wild-type (WT) mice were either untreated or inoculated with 1 × 10⁶ B16 cells. Fifteen days postinoculation, liver mononuclear cells were prepared and analyzed through flow cytometry (n = 6). (A) The percentage and absolute number of CD11b⁺Gr1⁺ myeloid-derived suppressor cells (MDSCs) (mean ± SEM) are shown. (B) One representative FACS staining in panel (A) is shown. (C) The percentage and absolute number of CD4⁺Foxp3⁺ T regulatory cells (Tregs) (mean ± SEM) are shown. (D) One representative FACS staining in panel (C) is shown. (E) Depletion of MDSC, but not Tregs, restored concanavalin A (Con A)-induced hepatitis in TB mice. TB mice were either untreated or intraperitoneally injected with 0.2 mg anti-Gr1 (clone RB68C5) or anti-CD25 (clone PC-61.5.3) prior to Con A treatment as described in Section “Materials and Methods.” Serum samples collected at 12 h post-Con A treatment were used for alanine aminotransferase (ALT) detection, and results from one typical experiment are shown (mean ± SEM, n = 7). (F) Adoptive transfer of MDSCs, but not Tregs, from TB mice protected WT mice against Con A-induced hepatitis. Sex- and age-matched WT mice were either untreated or adoptively transferred with 5 × 10⁵ CD11b⁺Gr1⁺ cells or CD4⁺CD25⁺ Treg cells from TB mice via intravenous injection prior to Con A injection. Serum samples collected at 12 h post-Con A treatment were used for analyzing ALT levels. Results from one representative experiment are shown (mean ± SEM, n = 5). Data represent at least three independent experiments with similar results.

Figure 4

IFN-γ production of NKT, not CD4+ T cells, was inhibited by myeloid-derived suppressor cells (MDSCs). (A) Wild-type (WT) and tumor-bearing (TB) mice were challenged by concanavalin A (Con A) (10 mg/kg body weight) 15 days post-B16 injections; 12 h later, mice were sacrificed and liver mononuclear cells were isolated. Afterward, intracellular IFN-γ intracellular staining was performed as described in Section “Materials and Methods,” and cells were analyzed by flow cytometry (n = 5). (B) One typical flow cytometry plot in panel (A) is shown. (C) MDSC depletion restored IFN-γ production of NKT cells. WT and TB mice were challenged by Con A 24 h after anti-Gr1 treatment. Then, 12 h later, mice were sacrificed and IFN-γ intracellular staining was performed, and cells were analyzed by flow cytometry (n = 5). (D) One typical flow cytometry plot in panel (C) is shown. (E) MDSC transfer inhibited IFN-γ production of NKT cells. 5 × 10⁵ TB MDSCs were intravenously transferred into WT mice, followed by Con A challenge (10 mg/kg), and 12 h later, livers were isolated. IFN-γ intracellular staining was performed, and cells were analyzed by flow cytometry (n = 5). (F) One typical flow cytometry plot in panel (E) is shown.

Figure 5

Myeloid-derived suppressor cells (MDSCs) inhibited IFN-γ production of NKT cells via membrane-bounded transforming growth factor β (TGF-β) in a cell contact-dependent manner. (A) Membrane-bound TGF-β is the main factor for suppression of NKT cell IFN-γ secretion. NKT cells (CD3⁺NK1.1⁺) were isolated from liver tissues of wild-type (WT) mice, co-cultured with MDSCs sorted from livers of tumor-bearing (TB) mice at the ratio of 1:1 under various conditions, including either transwell (0.4 µm) separation or in the presence of 10 ng/ml anti-TGF-β1 mAb or 0.5 ng/ml rmTGF-β1 for 6 h, and then cells were then stimulated with PMA (25 ng/ml) plus ionomycin (1 µg/ml). Eighteen hours later, supernatants were harvested for determination of IFN-γ levels using ELISA kit. Results (mean ± SEM) of triplicated wells from one representative experiment are shown. (B) Expression levels of arginase 1 (Arg-1) and inducible nitric oxide synthase (iNOS) were not changed in liver tissues of TB mice. Livers were isolated from either WT or TB mice (day 15 posttumor inoculation), and mRNA levels of Arg-1 and iNOS were determined via quantitative real-time PCR (n = 3). Results from one representative experiment are shown. (C) Reactive oxygen species (ROS) expression was increased in the liver of TB mice. Livers were isolated from either WT or TB mice 15 days posttumor inoculation. ROS levels were determined via ROS assay kit as described in Section “Materials and Methods” (n = 3). Results from one representative experiment are shown. (D) Arg-1, iNOS, and ROS inhibitors failed to restore IFN production of NKT cells in vitro. NKT cells were co-cultured with MDSCs isolated from the liver of TB mice at the ratio of 1:1 as described above in the absence or presence of various inhibitors, including N-hydroxy-nor-arginine (1 mM), catalase (1,000 U/ml), or l-NIL (0.5 µM) for 6 h. Then cells were stimulated with PMA plus ionomycin, and IFN-γ levels at the supernatant were determined using ELISA kit. Results from triplicated wells (mean ± SEM) are shown.

Figure 6

Recruitment of myeloid-derived suppressor cells (MDSCs) into tumor-bearing (TB) (B16) mice livers was CXCR2 mediated. (A) Liver mononuclear cells (MNCs) were isolated from either wild-type (WT) or TB mice 15 days posttumor inoculation, and mRNA levels of CXCRs were determined via quantitative real-time PCR (n = 3). (B) Liver tissues were isolated from either WT or TB mice 15 days posttumor inoculation, and mRNA levels of CXCRs were determined via quantitative real-time PCR (n = 3). (C) CXCR2-dependent liver recruitment of MDSC in TB mice. Sex- and age-matched WT mice and CXCR2^−/− mice were either untreated or inoculated with B16 tumor cells to prepare TB mice as described above, and liver MNCs was prepared from the livers and then used for analyzing the percentage of CD11b⁺Gr1⁺ through flow cytometry (n = 6). (D) One representative FACS plot in panel (C) is shown. (E) Deficiency of CXCR2 rendered susceptibility to Con A-induced liver damage in TB mice. Sex- and age-matched WT or CXCR2^−/− mice were injected with B16 tumor cells, and on day 15 posttumor inoculation, mice were challenged with concanavalin A (Con A). Serum samples collected at 12 h post-Con A injections were used for analyzing the levels of alanine aminotransferase (n = 6). (F) Liver tissues were fixed for hematoxylin and eosin staining, and one representative tissue staining is shown. Scale bars, 200 µm.

Figure 7

Recruitment of myeloid-derived suppressor cells (MDSCs) into livers was regulated by tumor-derived transforming growth factor β (TGF-β). (A) Tumor-derived TGF-β contributed to tumor-induced immune suppression in livers. Sex- and age-matched WT mice were either untreated or inoculated with B16 cells or sh RNA-transfected stable B16 cells prepared in our laboratory (1 × 10⁶ cells/mouse) and named TB-B16 or TB-sh B16, respectively. After 15 days posttumor inoculations, mice were challenged with concanavalin A; 12 h later, serum alanine aminotransferase levels were analyzed. Results from one of three independent experiments are shown (n = 8). (B) Recruitment of MDSCs into livers was dependent on tumor-derived TGF-β. Three groups of mice, as described above after day 15 posttumor inoculations, were used for percentage analysis of CD11b⁺Gr1⁺ MDSCs through flow cytometry (n = 6). (C) One representative FACS plot in panel (B) is shown. (D) Tumor-derived TGF-β elevated expression of CXCLs in the liver. Liver tissues collected from the three groups of mice as shown above were used for analyzing mRNA levels of CXCLs via quantitative real-time PCR (n = 3). (E) WT mice were intraperitoneally injected 1 µg rmTGF-β 3 h prior to sacrifice. Then WT, TB (B16), TB (sh B16), and WT injected rmTGF-β were sacrificed for analyzing the percentage of CD11b⁺Gr1⁺ MDSCs through flow cytometry (n = 3). (F) One representative FACS plot in panel (E) is shown. (G) Liver tissues collected from the four groups of mice in panel (E) were used for analyzing mRNA levels of CXCLs via quantitative real-time PCR (n = 3).

Figure 7

Recruitment of myeloid-derived suppressor cells (MDSCs) into livers was regulated by tumor-derived transforming growth factor β (TGF-β). (A) Tumor-derived TGF-β contributed to tumor-induced immune suppression in livers. Sex- and age-matched WT mice were either untreated or inoculated with B16 cells or sh RNA-transfected stable B16 cells prepared in our laboratory (1 × 10⁶ cells/mouse) and named TB-B16 or TB-sh B16, respectively. After 15 days posttumor inoculations, mice were challenged with concanavalin A; 12 h later, serum alanine aminotransferase levels were analyzed. Results from one of three independent experiments are shown (n = 8). (B) Recruitment of MDSCs into livers was dependent on tumor-derived TGF-β. Three groups of mice, as described above after day 15 posttumor inoculations, were used for percentage analysis of CD11b⁺Gr1⁺ MDSCs through flow cytometry (n = 6). (C) One representative FACS plot in panel (B) is shown. (D) Tumor-derived TGF-β elevated expression of CXCLs in the liver. Liver tissues collected from the three groups of mice as shown above were used for analyzing mRNA levels of CXCLs via quantitative real-time PCR (n = 3). (E) WT mice were intraperitoneally injected 1 µg rmTGF-β 3 h prior to sacrifice. Then WT, TB (B16), TB (sh B16), and WT injected rmTGF-β were sacrificed for analyzing the percentage of CD11b⁺Gr1⁺ MDSCs through flow cytometry (n = 3). (F) One representative FACS plot in panel (E) is shown. (G) Liver tissues collected from the four groups of mice in panel (E) were used for analyzing mRNA levels of CXCLs via quantitative real-time PCR (n = 3).