Tumor-Induced Myeloid-derived Suppressor Cell Function is Independent of IFNγ and IL-4Rα

Summary

Myeloid-derived suppressor cells (MDSC) are present in most cancer patients and experimental animals where they exert a profound immune suppression and are a significant obstacle to immunotherapy. IFNγ and IL-4Rα have been implicated as essential molecules for MDSC development and immunosuppressive function. If IFNγ and IL-4Rα are critical regulators of MDSC, then they are potential targets for preventing MDSC accumulation or inhibiting MDSC function. Because data supporting a role for IFNγ and IL-4Rα are not definitive, we have examined MDSC induced in IFNγ-deficient, IFNγR-deficient, and IL-4Rα-deficient mice carrying three C57BL/6-derived tumors (B16 melanoma, MC38 colon carcinoma, and 3LL lung adenocarcinoma), and three BALB/c-derived tumors (4T1 and TS/A mammary carcinomas, and CT26 colon carcinoma). We report that although MDSC express functional IFNγR and IL-4Rα, and have the potential to signal through the STAT1 and STAT6 pathways, respectively, neither IFNγ nor IL-4Rα impacts the phenotype, accumulation, or T cell suppressive potency of MDSC, although IFNγ and IL-4Rα modestly alter MDSC-macrophage IL-10 cross-talk. Therefore, neither IFNγ nor IL-4Rα is a key regulator of MDSC and thus targeting these molecules is unlikely to significantly alter MDSC accumulation or function.

Introduction

Individuals with advanced cancer are frequently immunosuppressed, lack effective innate and adaptive anti-tumor immunity, and are poorly responsive to active immunotherapy. Assorted tumor-secreted factors drive the accumulation of multiple immune suppressive mechanisms [1]. Tumor-secreted factors act directly to activate suppressive mechanisms, or they act indirectly by inducing host cells that reduce immunocompetence [2]. Different cancers stimulate diverse inhibitory mechanisms; however, myeloid-derived suppressor cells (MDSC) are induced by virtually all cancer cells and are a major obstacle to anti-tumor immunity [3]. Mouse MDSC are a heterogeneous cell population, consisting of CD11b⁺Gr1⁺ cells. Two major subpopulations have been defined based on the differential expression of Ly6C and Ly6G, the components of Gr1. Monocytic MDSC (MO-MDSC) are mononuclear and CD11b⁺Ly6G⁻Ly6C^hi, while granulocytic MDSC (PMN-MDSC) are polymorphonuclear and CD11b⁺Ly6G⁺Ly6C^low/− [4, 5]. Gr1 levels roughly correlate with Ly6G levels, so that CD11b⁺Gr1^hi/med cells tend to be CD11b⁺Ly6G⁺Ly6C^−/low PMN-MDSC [6]. Both subpopulations suppress by the production of arginase, while MO-MDSC also produce nitric oxide (NO)[4, 5]. Although not as well characterized, comparable subpopulations exist in cancer patients [7-9].

Various tumor-produced factors, including GM-CSF [6, 8, 10-13], IL-1β [14, 15], IL-6 [16], cyclooxygenase-2 and prostaglandin E2 [17, 18], S100A8/A9 [19, 20] and vascular endothelial growth factor [21] facilitate MDSC development and/or enhance MDSC suppressive activity. Because MDSC are induced by any one of these factors, no single molecule is essential for generating MDSC. In contrast, IFNγ [10, 22] and IL-4Rα [9, 23] have been reported as essential for MDSC development and/or suppressive activity. Two of these studies used MDSC “cell lines” [22, 23], so the applicability of the results to primary MDSC is unclear. The requirement for IFNγ [4] and IL-4Rα [9, 16] have been attributed to the development and suppressive activity of MO-MDSC and PMN-MDSC, respectively. IL-4Rα is also considered a marker for human MDSC [9]. However, other studies demonstrated that IL-4Rα [5, 24] and IFNγ [25]are not essential for murine MDSC accumulation or suppressive function. If IFNγ and/or IL-4Rα are critical for MDSC development and function, then manipulation of these molecules could impact MDSC-mediated immune suppression. Therefore, it is important to clarify the role of IFNγ and IL-4Rα in MDSC biology. Given the inconsistencies in the literature we evaluated the role of these molecules using IFNγ-deficient, IFNγR-deficient, and IL-4Rα-deficient mice using three C57BL/6-derived and three BALB/c-derived murine tumors that induce monocytic and granulocytic MDSC. We now report that neither IFNγ nor IL-4Rα is essential for the development or suppressive activity of mouse MDSC.

Results and Discussion

MDSC from wild type, IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice have similar phenotypes

To determine if IFNγ or IL-4Rα impacts MDSC development, wild type BALB/c, IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice were inoculated with syngeneic TS/A, 4T1, or CT26 tumor cells, and wild type C57BL/6, IFNγ^−/−, and IFNγR^−/− mice were inoculated with syngeneic MC38, 3LL, or B16 tumor cells. MDSC were harvested from the blood of all mice when primary tumors within each group of wild type and knockout mice carrying the same tumor were approximately equal in size, and analyzed for by flow cytometry for phenotype (Fig. 1). Microscopy images were obtained to confirm morphology (Supplemental Fig. 1). Percentages of MO-MDSC and PMN-MDSC did not significantly differ between wild type, IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice with the same tumor (Fig. 1A and 1B). As reported previously, MO-MDSC (CD11b⁺Ly6G⁻ Ly6C^hi) express more CD115, F4/80, and iNOS compared to PMN-MDSC (CD11b⁺Ly6G⁺Ly6C^low/−), while all MDSC populations contain similar quantities of IL-4Rα and arginase [4, 5] (representative flow profiles for individual mice are in Fig. 1C; average mean channel fluorescence pooled from three mice per group are in Fig. 1D). Likewise, MDSC induced by the six tumors in their respective syngeneic wild type, IFNγ^−/−, and IFNγR^−/− hosts do not substantially differ in their expression of CD11b, Gr1, Ly6C, Ly6G, IL-4Rα, CD115, F4/80, arginase, iNOS, or ROS. Similarly, MDSC induced by the three tumors in BALB/c and IL-4Rα^−/− mice express similar levels of CD11b, Gr1, Ly6C, Ly6G, CD115, F4/80, arginase, iNOS, and ROS. Therefore, IFNγ and IL-4Rα do not alter the phenotype of either MO-MDSC or PMN-MDSC with respect to the markers that define these cells, or impact the accumulation of MDSC.

Figure 1. MO-MDSC and PMN-MDSC from wild type, IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice have similar phenotypes.

Wild type, IFNγ^−/−, and IFNγR^−/− mice on a C57BL/6 background or wild type, IFNγ^−/−, IFNγR^−/−, and IL4Rα^−/− mice on a BALB/c background were inoculated with MC38, 3LL, or B16 tumor cells (C57BL/6) or TS/A, 4T1, or CT26 tumor cells (BALB/c). Mice with 4T1 and B16 tumors were bled on day 21 - 28 when primary tumors for each group of wild type and knockout mice with the same tumor were 11.38 ± 0.94 and 6.29 ± 0.68 mm in diameter, respectively. Mice with MC38, 3LL, TS/A, and CT26 tumors were bled on day 25-26 when primary tumors for each group of wild type and knockout mice with the same tumor were 13.29 ± 2.1, 8.2 ± 0.77, 12.4 ± 1.00, 13.55 ± 1.03 mm in diameter, respectively. RBCs were removed by lysis and blood leukocytes were stained with fluorescently labeled mAbs for CD11b and Gr1, or CD11b, Ly6C and Ly6G. (A) Gr1 and CD11b labeled cells (top row dot plot of each tumor); CD11b, Ly6G, and Ly6C labeled cells gated for CD11b⁺ cells and analyzed for Ly6C and Ly6G expression (bottom row dot plot of each tumor). Right hand graphs show percentages of total MDSC (Gr1⁺CD11b⁺ - top panel), and MO-MDSC (CD11b⁺Ly6G⁻Ly6C^hi) and PMN-MDSC (CD11b⁺Ly6G⁺Ly6C^low/− - bottom panel). Each flow profile shows the results of individual mice and the bar graphs represent average + SD of three mice per group. There is no statistically significant difference between the levels of total MDSC, MO-MDSC, and PMN-MDSC for wild type and knockout mice with the same tumor (ANOVA single factor; p > 0.05 for all comparisons). Data are from one of three independent experiments. (B) BALB/c and C57BL/6 mice and their associated knockout strains with 4T1 or B16 tumors were bled when tumor diameters were at the indicated sizes and leukocytes stained as in panel A for CD11b and Gr1. Each symbol represents an individual mouse. There is no statistically significant difference between wild type and knockout groups with the same tumor. n= 4 mice/group/strain/tumor diameter (Anova: single factor analysis p > 0.05). Data were pooled from four experiments. (C) C57BL/6 and BALB/c mice and their associated knockout strains with B16 or 4T1 tumors were bled and leukocytes stained as in panel A for CD11b and Gr1. Gated CD11b⁺ Gr1⁺cells were analyzed for IL-4Rα, CD115, F4/80, arginase, iNOS, and ROS. Data are from one of three independent experiments. (D) Gated CD11b⁺ Gr1⁺cells of panel A were analyzed for IL-4Rα, CD115, F4/80, arginase, iNOS, and ROS. Data shown are average + SD of mean channel fluorescence (MCF) (Experimental MCF – control MCF) of three mice per group. There is no statistically significant difference between the levels of IL-4Rα, CD115, F4/80, arginase, iNOS, or ROS in wild type and knockout groups with the same tumor (ANOVA single factor; p > 0.05 for all comparisons).

MDSC-mediated T cell suppression is independent of IFNγ and IL-4Rα

To determine if IFNγ or IL-4Rα is essential for T cell suppression by MDSC, MDSC were harvested from tumor-bearing wild type and knockout mice, and tested for their ability to inhibit the activation of antigen-specific transgenic T cells. MDSC induced by the same tumor were similarly suppressive for CD8⁺ and CD4⁺ T cells regardless of whether they were generated in wild type, IFNγ^−/−, IFNγR^−/− or IL-4Rα^−/− mice (Fig. 2A). Therefore, the T cell suppressive function of MDSC is not affected by IFNγ or IL-4Rα.

Figure 2.

IFNγ and IL-4Rα do not affect T cell suppressive activity of MDSC. MDSC were generated in B16 tumor-bearing wild type, IFNγ^−/−, and IFNγR^−/− C57BL/6 mice, and in 4T1, TS/A, and CT26 tumor-bearing wild type, IFNγR^−/−, and IL4Rα^−/− BALB/c mice. Magnetic bead purified blood MDSC were >90% Gr1⁺CD11b⁺ cells. (A) MDSC were incubated at the indicated ratios with 1×10⁵ splenocytes from OT-I (upper left-hand panel), Clone 4 (bottom four left-hand panels), or DO11.10 (right-hand panels) transgenic mice and cognate peptide [24]. T cell activation was measured by incorporation of ³H-thymidine. Data are the average of triplicate points + SD from one of two independent experiments with MDSC pooled from 3-4 mice/group. There is no statistically significant difference between the suppressive activity of MDSC induced by the same tumor in wild type vs. knockout mice (Student’s two tailed t-Test for paired two samples for means, p>0.05), with the exception of 4T1 induced IFNγR^−/− MDSC which are more suppressive than BALB/c MDSC for CD8⁺T cells (p<0.05). (B) MDSC production of IL-10 and macrophage-induced MDSC production of IL-10 is modestly affected by IFNγ and IL-4Rα; MDSC-mediated down-regulation of macrophage-produced IL-12 is not affected by IFNγ or IL-4Rα. Peritoneal macrophages (>90% F4/80⁺) from wild type BALB/c mice were untreated or activated with LPS and IFNγ and co-cultured overnight with MDSC from 4T1 tumor-bearing wild type, IFNγ^−/−, IFNγR^−/−, or IL-4Rα^−/− BALB/c mice at 0.75×10⁶ macrophages/well. Culture supernatants were assayed by ELISA for IL-10 and IL-12 [24]. Data are the average + SD of quadruplicates from three mice/group and are representative of three independent experiments. (Student’s two tailed t-Test for paired two sample for means). Top panel: Macrophages increase wild type and knockout MDSC production of IL-10; knockout MDSC produce less IL-10 than wild type MDSC. Bottom Panel: Wild type and knockout MDSC decrease macrophage production of IL-12 by >87%. (C) IFNγ and IL-4 activate STAT1 and STAT6 signaling pathways, respectively, in MDSC. (Left panel) Blood MDSC from tumor-free and tumor-bearing wild type, IFNγ^−/−, and IFNγR^−/− C57BL/6 (B16) and BALB/c (4T1) mice were unstimulated or stimulated in vitro for 15 min with IFNγ and then stained for Gr1, CD11b, and pSTAT1. Gated Gr1⁺CD11b⁺ cells were analyzed for phosphorylated STAT1 (pSTAT1). (Right panel) MDSC from tumor-free wild type, or tumor-bearing wild type and IL-4Rα^−/− mice were stimulated in vitro for 15 min with IL-4 and then stained for Gr1, CD11b, and pSTAT6. Gated Gr1⁺CD11b⁺ cells were analyzed for pSTAT6. Each panel depicts MDSC pooled from three or more mice. Data are from one of three experiments.

MDSC-macrophage cross-talk is independent of IFNγ and IL-4Rα

MDSC also promote tumor progression by polarizing immunity towards a type 2 response through their cross-talk with macrophages which reduces macrophage production of IL-12 and increases MDSC production of IL-10 [24]. MDSC from 4T1 tumor-bearing wild type, IFNγ^−/−, IFNγR^−/−, or IL-4Rα^−/− BALB/c mice co-cultured with wild type BALB/c macrophages produced more IL-10 as compared to MDSC cultured in the absence of macrophages. However, MDSC from IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice produced less IL-10 than MDSC from wild type mice (Fig. 2B). Therefore, MDSC production of IL-10 and macrophage-induced MDSC production of IL-10 is modestly affected by IFNγ and IL-4Rα. Macrophage production of IL-12 was reduced >87% by MDSC from wild type, IFNγ^−/−, IFNγR^−/−, and IL-4Rα^−/− mice. Since MDSC production of IL-10 down-regulates macrophage production of IL-12 [24], and macrophage production of IL-12 is not impacted, it is likely that the modest decrease in IL-10, although statistically significant, is not physiologically relevant.

MDSC have functional receptors for IFNγ and IL-4

Synthesis of iNOS and nitric oxide (NO) by MO-MDSC are attributed to STAT1 activation of IFNγ[4]. To determine if this pathway is active, B16- and 4T1-induced MDSC were examined for signaling through STAT1. CD11b⁺Gr1⁺ MDSC from wild type, but not from IFNγR^−/− mice, expressed IFNγR and IFNγ-deficiency did not affect expression of IFNγR (Supplemental Fig. 2). IFNγ-treated MDSC from wild type and IFNγ^−/− mice, but not from control IFNγR^−/− mice, contained activated STAT1 (phosphorylated STAT1 (pSTAT1) (Fig. 2C) indicating that MDSC have the potential to respond to IFNγ.

Production of arginase, the predominant effecter molecule of PMN-MDSC, has been attributed to IL-4 and IL-13 signaling through their receptor which consists of the common γ and IL-4Rα chains [9, 26]. Stimulation of MDSC from wild type, but not from IL-4Rα^−/− mice with IL-4, activated STAT6 (pSTAT6) (Fig 2C), demonstrating that MDSC have the potential to respond to IL-4 through IL-4Rα.

Concluding remarks

These studies demonstrate that although MDSC can respond to IFNγ and IL-4, IFNγ and IL-4Rα do not regulate MDSC accumulation, phenotype, or suppression of T cell activation. Therefore, targeting IFNγ and/or IL-4Rα will not reduce the quantity of MDSC, alter MDSC phenotype, or restore T cell activation. MDSC production of IL-10 and macrophage-induced MDSC production of IL-10 are partially controlled by IFNγ and IL-4Rα. However, targeting these molecules is unlikely to facilitate polarization towards a type 1 response because the minimal reduction in MDSC production of IL-10 will not restore macrophage production of IL-12. Therefore, treatments aimed at down-regulating IFNγ and/or IL-4Rα are unlikely to be therapeutically effective.

Materials and methods

Mice

Breeding stock for BALB/c, transgenic D011.10 (TcR is I-A^d-restricted, ovalbumin (OVA) peptide 323-339-specific), transgenic OT-1 (TcR is H-2K^b-restricted, OVA peptide SINNFEKL-specific), IFNγR-deficient C57BL/6, IFNγ-deficient C57BL/6, IFNγ-deficient BALB/c and IL-4Rα-deficient BALB/c was obtained from The Jackson Laboratory (Bar Harbor, ME). Initial stocks of transgenic BALB/c Clone 4 (TcR is H-2K^d-restricted, influenza hemagglutinin (HA) peptide 518–526-specific) mice were provided by Dr. E. Fuchs (Johns Hopkins). IFNγR-deficient BALB/c mice were generated from 129-IFNγR^−/− mice (The Jackson Laboratory) by backcrossing to BALB/c for 12 generations. PCR screening was performed as described on The Jackson Laboratory website (http://jaxmice.jax.org/protocolsdb/f?p=116:2:723076526485070::NO:2:P2_MASTER_PROTOCOL_ID,P2_JRS_CODE:7034,002702). Pups from the F12 generation were intercrossed and PCR screened to identify homozygous BALB/c IFNγR^−/− mice. Mice were bred in the University of Maryland Baltimore County (UMBC) animal facility and all animal procedures were approved by the UMBC Institutional Animal Care and Use Committee.

Antibodies and flow cytometry

Fluorescently-coupled Gr1 (clone RB68C5), CD11b, Ly6C (clone AL-21), Ly6G (clone 1A8), IL4Rα, IFNγR, CD115, F4/80, CD3, CD4, CD8, DO11.10 TCR (clone KJ1-26), Vβ_8.1&8.2 mAbs, uncoupled mAbs to arginase I and iNOs, rat anti-mouse IgG1-FITC (A85-1), and isotype control mAbs were from BD Pharmingen (San Diego, CA). Goat anti-mouse IgG2a-FITC was from Southern Biotech (Birmingham, AL). Staining for flow cytometry was performed as described [25]. Samples were run on a Beckman/Coulter XL or CyAn ADP flow cytometer and analyzed using FCS-Express or Summit software.

Tumor cells, tumor cell inoculation, and MDSC collection

BALB/c-derived 4T1 mouse mammary carcinoma cells were maintained as described [27]. C57BL/6-derived B78H1-GM-CSF cells (B16 variant called B16 in the present study) were provided by Dr. H. Levitsky (Johns Hopkins) and were maintained as described [11]. C57BL/6-derived 3LL lung carcinoma and the BALB/c-derived CT26 and MC38 colon carcinomas (latter two lines provided by Drs. D. Klinman (NIH) and D. Gabrilovich (Moffit), respectively), were maintained as described [5]. BALB/c-derived TS/A mammary carcinoma was maintained as described [28]. Mice were inoculated in the abdominal mammary gland with 7000 4T1 or 1×10⁶ TS/A cells, or in the abdominal flank with 1×10⁶ B16, 3LL, MC38, or CT26 cells. Mice were observed three times per week and sacrificed when they became moribund. Mice were bled from the tail, retroorbital sinus, or submandibular vein into 500μl of a 0.008% heparin solution, RBCs were removed by lysis, and the remaining leukocytes were characterized by flow cytometry [14, 24, 25].

In vitro T cell suppression

Splenocytes from DO11.10, Clone 4, or OT-I mice were co-cultured with cognate peptide and varying quantities of irradiated blood MDSC (> 90% Gr1⁺CD11b⁺ cells) isolated by magnetic bead sorting of Gr1⁺ cells using Miltenyi Biotec magenetic beads from tumor-bearing mice as described [19].

Macrophage and MDSC co-culture

Thioglycolate-induced BALB/c peritoneal macrophages were generated and co-cultured with blood-derived MDSC as described [24]. Briefly, macrophages were plated at 7.5×10⁵ cells/well in 500 μl macrophage medium (DMEM supplemented with 10% FBS) in 24-well plates and incubated at 37°C for 3 h. Non-adherent cells were removed, and 500 μl 5% FBS in DMEM was added to the adherent cells. MDSC (1.5×10⁶) in 500 μl 5% FBS in DMEM were added to some wells. Macrophages and MDSC co-cultures were then stimulated with LPS (100 ng/ml) and/or IFNγ (2 ng/ml), and cultures were incubated at 37°C for 24 h. Culture supernatants were collected and stored at −80°C until analyzed for IL-10 and IL-12 using duo set ELISA kits (R&D Systems, Minneapolis, MN, USA), according to the manufacturer’s protocols.

IFNγ and IL-4 stimulation

Blood leukocytes from tumor-free and tumor-bearing mice were either untreated or incubated for 15 min at 37°C with 2ng/ml IFNγ (Pierce Endogen, Rockford, IL), or 10ng/ml IL-4 and subsequently stained according to the manufacturer’s protocol (BD Biosciences) with mAb to phosphorylated STAT1 or phosphorylated STAT6, respectively, and mAbs to CD11b and Gr1.

Statistical analysis

Anova and Students’ t test were performed using Microsoft Excel 2007. p values <0.05 were considered significant.

Supplementary Material

Supplementary_Figure_1

Supplemental Figure 1: MDSC from wild type, IFNγR^−/−, IFNγ^−/−, IL4Rα^−/− mice have similar nuclear morphology. MACS sorted MDSC were adhered to glass slides using Cell-Tak (BD Biosciences), fixed with formaldehyde and stained with DAPI in Prolong Gold antifade reagent (Invitrogen). The cells were visualized with a Zeiss Axiolmager fluorescent microscope and Apotome (Zeiss, Germany) fitted with 40xl0NA, and analyzed with AxioVision LE software.

Supplementary_Figure_2

Supplemental Figure 2: MDSC have receptors for IFNγ. MDSC were obtained from tumor free and 4T1 or B16 tumor-bearing wild type, IFNγ−/−, and IFNγR−/− mice and labeled for Grl, CD lib, and IFNγR. Gated Grl⁺CDllb⁺ cells were analyzed for IFNyR expression.

List of Abbreviations

IFNγR

Interferon gamma receptor

IL-4Rα

IL-4 receptor alpha

MDSC

myeloid-derived suppressor cells

MO-MDSC

mononuclear MDSC

PMN-MDSC

polymorphonuclear MDSC

pSTAT1

phosphorylated STAT1

pSTAT6

phosphorylated STAT6