Myeloid-derived suppressor cells (MDSC) include immature monocytic (M-MDSC) and granulocytic (PMN-MDSC) cells that share the ability to suppress adaptive immunity and to hinder the effectiveness of anticancer treatments. Of note, in response to IFNγ, M-MDSCs release the tumor-promoting and immunosuppressive molecule nitric oxide (NO), whereas macrophages largely express antitumor properties. Investigating these opposing activities, we found that tumor-derived prostaglandin E2 (PGE2) induces nuclear accumulation of p50 NF-κB in M-MDSCs, diverting their response to IFNγ toward NO-mediated immunosuppression and reducing TNFα expression. At the genome level, p50 NF-κB promoted binding of STAT1 to regulatory regions of selected IFNγ-dependent genes, including inducible nitric oxide synthase (Nos2). In agreement, ablation of p50 as well as pharmacologic inhibition of either the PGE2 receptor EP2 or NO production reprogrammed M-MDSCs toward a NOS2low/TNFαhigh phenotype, restoring the in vivo antitumor activity of IFNγ. Our results indicate that inhibition of the PGE2/p50/NO axis prevents MDSC-suppressive functions and restores the efficacy of anticancer immunotherapy.

Significance:

Tumor-derived PGE2-mediated induction of nuclear p50 NF-κB epigenetically reprograms the response of monocytic cells to IFNγ toward an immunosuppressive phenotype, thus retrieving the anticancer properties of IFNγ.

Microenvironmental signals are sensed by myeloid cells through cytokine and/or innate immune receptors, whose differential engagement leads different polarized programs (1, 2). This functional plasticity is exemplified in the M1 versus M2 extremes of macrophage polarization (1–3) and has major impacts in the orchestration of cancer-related inflammation, immunosuppression (4–6), and clinical responses (2, 7). Of note, a time-dependent M1 to M2 transcriptional reprogramming of myeloid cells was reported in response to prolonged exposure to Toll-like receptors (TLR) ligands (e.g., lipopolisaccharide; i.e., LPS tolerance; refs. 8, 9), which requires the nuclear accumulation of p50 NF-κB and results in altered responsiveness to polarizing cytokines, such as IFNγ and IL4 (9). Cancer fuels myeloid cells heterogeneity also by sustaining altered myelopoiesis (10–13) that supports resistance to anticancer immunotherapy (14). Of relevance, divergent outcomes have been reported in response to immunotherapy, either with cytokines (15–17) or checkpoint inhibitors (18, 19). In particular, IFNγ, originally termed “macrophage activating factor” (20), was paradoxically shown to be equally necessary for melanoma development and rejection (21). Indeed, IFNγ has pleiotropic and contrasting effects in the tumor microenvironment. On one hand it exerts antiangiogenic activities, suppression of protumorigenic properties, enhancement of tumoricidal activity of macrophages, and of processing and presentation of tumor antigens to T lymphocytes (17, 22). On the other hand, IFNγ promotes immunosuppressive functions in myeloid cells (13) inducing expression of the immunosuppressive enzymes indoleamine 2,3 dioxygenase (Ido) and inducible nitric oxide synthase (Nos2), involved in the catabolism of l-tryptophan (23) and l-arginine (12), as well as the ligand programmed-death receptor-ligand 1 (PD-L1, B7-H1; ref. 24), whose interaction with the coinhibitory receptor programmed death-1 (PD-1) can be blocked to restore antitumor immunity (25). Mixed responses to IFNγ were also reported in different human malignancies (16, 17, 26). We have previously reported that accumulation of nuclear p50 NF-κB plays an essential role in the M2 orientation of tumor-associated macrophages (TAM; refs. 9, 10), which share common myeloid precursors and a similar M2-like gene signature with M-MDSCs (10). Here, we investigated how tumor-derived signals affect transcriptional activities and myeloid cell functions in response to immune-stimulatory cytokines and the impact of these events on cytokine-mediated cancer immunotherapy.

Mice and ethics statement

The study was designed in compliance with Italian Governing Law (Legislative Decree 116 of January 27, 1992); EU directives and guidelines (EEC Council Directive 86/609, OJ L 358, 12/12/1986); Legislative Decree September 19, 1994, no. 626 (89/391/CEE, 89/654/CEE, 89/655/CEE, 89/656/CEE, 90/269/CEE, 90/270/CEE, 90/394/CEE, 90/679/CEE); the NIH Guide for the Care and Use of Laboratory Animals (1996 edition). All experiments involving animals described in this study were approved by the Ministry of Health (authorization numbers 160/2012-B and 25/2018-PR). The study was approved by the scientific board of Humanitas Clinical and Research Center. Mice were monitored daily and euthanized when they displayed excessive discomfort. p50 NF-κB-deficient mice on the C57BL/6J background were available in the laboratory (27). The NF-κB1flox/flox (p50flox) mice was recently generated (28). p50flox mice were crossed with B6.Cg-Tg(Tek-Cre)1Ywa mice (Jackson Laboratories) to generate p50flox; Tie2Cre mice (p50Tie2 mice). OT-I mice were obtained from Jackson Laboratories.

Tumor models

Eight-week-old mice were injected intramuscularly in the left leg with 105 cells of murine fibrosarcoma (MN/MCA1) or subcutaneously with 5 × 105 cells of murine melanoma (B16). Tumor growth was monitored three times a week with a caliper. Bone marrow (BM) transfer and schedule of treatments are described in the Supplementary Materials and Methods

MDSC purification

MDSC purification from the spleens of tumor-bearing mice has been performed by magnetic separation (MACS, Miltenyi Biotec; ref. 11), as described in the Supplementary Materials and Methods.

Cell culture and reagents for mouse studies

BM-MDSCs were derived from BM cells isolated from C57BL/6 mice as described previously (29) and reported in the Supplementary Materials and Methods. Peritoneal exudate macrophages (PEC) were obtained as described previously (10). Preparation of MN/MCA1 tumor supernatant (TSN) and reagents used for mouse studies are described in the Supplementary Materials and Methods.

Cell culture and reagents for human studies

Human monocytes were isolated from peripheral blood as described previously (9) and cultured in RPMI1640 containing 10% FBS, 2 mmol/L glutamine, 100 U/mL penicillin–streptomycin. Culture of human pancreatic carcinoma cell line PANC1 and preparation of tumor-conditioned supernatant (TSN) are described in the Supplementary Materials and Methods.

The concentration for the different treatments were as follows: PANC1 supernatant (30%), human IFNγ (PeproTech) 20 ng/mL, prostaglandin E2 (PGE2) receptor (EP2/EP1) antagonist AH6809 (Tocris) 10−5 mol/L.

Nitrite production

A total of 2 × 105 cells were plated and stimulated with IFNγ (PeproTech, 200 U/mL) for indicated time points. Nitric oxide production was evaluated in culture supernatant using the Griess Reagent System (Promega) as described in Supplementary Materials and Methods.

Suppression assay

M-MDSCs (CD11b+Ly6C+Ly6G cells) and PMN-MDSCs (CD11b+Ly6Clow/−Ly6G+ cells) were isolated from spleen and tumor of wild-type (WT) and p50−/− MN-MCA1 tumor-bearing mice. A total of 2 × 105 splenocytes from naive C57Bl6 mice were labelled with 1 μmol/L CFSE and then cocultured with 1 × 105, 5 × 104 or 2.5 × 104 WT M-MDSCs, in the presence of anti-CD3 (3 μg/mL, 2C11, BioLegend) and anti-CD28 (2 μg/mL, 37.5; BD Biosciences). Similarly, M- and PMN-MDSC from WT and p50−/− mice were stimulated with IFNγ (200 U/mL) for 3 days and then cocultured with activated CFSE-labeled splenocytes. After 3 days of coculture, cells were stained and CFSE signal of gated lymphocytes was used to analyze cell proliferation. For antigen-specific suppression assay, mixed-lymphocyte reaction was performed as reported previously (11). M-MDSCs isolated from the spleen of tumor-bearing mice were then stimulated with IFNγ (200 U/mL), in the presence or absence of 500 μmol/L of L-NG-monomethylarginine (L-NMMA; Calbiochem). At day 3, 50 μL of supernatant was tested for nitric oxide (NO) production (as control) and 2 × 105 splenocytes from OT-I mice were added for additional 72 hours in the presence of 250 μg/mL of ovalbumin peptide (OVA257–264; Sigma). [3H] thymidine was added for the last 16 hours of culture and its incorporation was analyzed by MicroBeta plate counter (Perkin Elmer). As controls, OT-1 splenocytes alone were pulsed with OVA peptide (250 μg/mL) or kept in culture media. All conditions were evaluated in triplicates.

Isolation of MDSCs from patients with colorectal cancer

The study was conducted in accordance with recognized ethical guidelines (Declaration of Helsinki) was approved by the Institute Ethical Committee, and written informed consent was obtained from 20 patients. Twenty milliliters of peripheral blood were collected from healthy donors and patients with T2 or T3 colorectal cancer. Patients with colorectal cancer did not receive radiation or chemotherapy before sample collection. Blood was stratified on Ficoll gradient to separate peripheral blood mononuclear cells (PBMC). All blood samples were analyzed within 3 hours after collection by FACS analysis. Briefly, 1 × 106 cells were resuspended in Hank's Balanced Salt Solution (HBSS; Lonza) supplemented with 0.5% BSA (Sigma). Staining was performed at 4°C for 30 minutes, with a cocktail of mAbs to HLA-DR-Pacific Blue (clone L243); CD14-PE, -APC, -FITC (clone M5E2); CD33-PerCp-Cy5.5 (clone WM53); iNOS/NOS Type II-FITC (Clone 6/iNOS/NOS Type II), from BD Biosciences or BioLegend. Furthermore, we used unconjugated rabbit monoclonal anti-human EP2R [clone EPR8030(B); Abcam] followed by incubation with secondary goat anti-rabbit Alexa Fluor 488–conjugated antibody (Life Technologies). For intracellular staining, Foxp3/Transcription Factor Staining Buffer Set (eBioscence) were used according to the manufacturer's instructions. Cells were analyzed using the BD FACSCanto II or BD LSRFortessa and BD FACSDiva and FlowJo (9.3.2) software. When needed, cells were stained, sorted using a BD FACSAria III cell sorter, and subsequently analyzed by confocal microscopy.

BM colony formation

BM cells were isolated from the tibias and femurs of WT or p50−/− mice and the colony formation capacity was measured by detection and quantification using MethoCult GF M3434, which supports optimal growth of granulocyte–macrophage progenitors (CFU-GM, CFU-M, CFU-G), of erythroid progenitors (BFU-E) and multi-potential granulocyte, erythroid, macrophage, megakaryocyte progenitors (CFU-GEMM). Twelve days after seeded, colonies were counted independently by two separate operators.

Histopathologic analysis of mice BM

Sections of WT and p50−/− BM were routinely stained with hematoxylin and eosin and analyzed by an expert pathologist (C. Tripodo) using a Leica DM2000 optical microscope (×400 and ×630 magnification) and microphotographs were collected with a Leica DFC320 digital camera using the Leica IM50 imaging software.

Quantification of circulating granulocytes in peripheral blood smears

Cell counts were visually performed on five May–Grunwald Giemsa-stained smears on high-power microscopic fields (×400 magnification) and the average number of total and immature granulocytes was determined by averaging the counts.

Flow cytometry and sorting

Splenocytes were collected from spleen after disaggregation and filtration through Falcon strainers (70 μm). Primary tumors were cut into small pieces, disaggregated with 0.5 mg/ml collagenase IV and 150 U/ml DNase I in RPMI1640 for 30 minutes at 37°C and filtered through strainers. Cells (106) were resuspended in HBSS (Lonza) supplemented with 0.5% BSA (Sigma) and the staining was performed at 4°C for 20 minutes with specific antibodies (detailed information is provided in the Supplementary Materials and Methods). Cells were analyzed using the BD FACSCanto II or BD LSRFortessa and BD FACSDiva and FlowJo (9.3.2) software.

mRNA Sequencing

Total RNA was extracted from 1–5 × 106 M-MDSC (RNeasy kit, Qiagen), and 2–5 μg were used to generate sequencing libraries with a Truseq RNA Sample Prep Kit V2 (Illumina) according to the manufacturer's instructions. Sequencing was performed on a HiSeq2000 (Illumina).

Chromatin immunoprecipitation

Chromatin immunoprecipitation (ChIP) was carried out with a previously described high-throughput protocol (30, 31). Details are reported in the Supplementary Materials and Methods.

Computational methods

Computational analysis of data generated by RNA sequencing (RNA-seq) and ChIP-seq is described in the Supplementary Experimental Procedures.

Confocal microscopy on BM-MDSCs, spleen MDSCs, and PECs

Cells were seeded on Poly-l-lysine (Sigma-Aldrich) coated sterile rounded glasses at 2 × 105 cells/mL in medium and fixed with 4% PFA for 10 minutes at room temperature. Cell staining and analysis are described in Supplementary Material and Methods.

Statistical analysis

Statistical significance between two groups was determined by two-tailed Student t test or two-way ANOVA corrected for multiple comparison by Sidak test and among more than two groups by one-way or two-way ANOVA corrected for multiple comparison by Tukey test (Prism version 6). *, P < 0.05; **, P <0.01; ***, P <0.001.

Real-time PCR analysis, MTT assay, ELISA are described in the Supplementary Materials and Methods

p50 NF-κB controls both M-MDSC–suppressive functions and differentiation of myeloid precursors

Nuclear accumulation of p50 NF-κB, as occurring in TAMs and LPS-tolerant macrophages, impairs both M1 polarization and antitumor activities (9, 10). Similar to TAMs, confocal microscopy analysis showed selective nuclear accumulation of p50 NF-κB in splenic and tumor-infiltrating monocytic (M)-MDSCs (CD11b+Ly6GLy6C+ cells) from fibrosarcoma (MN/MCA1)-bearing mice, at advanced stages of tumor development (day 21; Fig. 1A; Supplementary Fig. S1A). As inhibition of T-cell proliferation validated the suppressive activity of these cells (Supplementary Fig. S1B), we explored whether p50 NF-κB could be actually involved in accumulation of suppressive M-MDSCs occurring during tumor development. According to our previous findings (10, 27), C57BL/6 p50−/− mice displayed both reduced tumor growth and metastasis formation, as compared with WT mice (Fig. 1B) but, paradoxically, displayed increased accumulation of splenic and tumor myeloid cells expressing the M-MDSC phenotype CD11b+Ly6GLy6C+ (Fig. 1C and D; Supplementary Fig. S1C). Of note this event coincided temporally with the nuclear accumulation of p50, as observed in WT mice (21 days; Fig. 1A). Cells expressing the PMN-MDSCs phenotype CD11b+ Ly6G+Ly6Clow/− also accumulated during disease progression (Fig. 1C and D; Supplementary Fig. S1C); however, without exhibiting nuclear accumulation of p50 NF-κB (Fig. 1A; Supplementary Fig. S1A). Hence, we analyzed whether p50 could affect the expression of genes encoding for immunosuppressive activities (12, 32). Noteworthy, in the absence of p50, magnetically sorted splenic CD11b+Ly6GLy6C+ cells (Supplementary Fig. S1D) showed a drastic reduction of both Nos2 gene expression (Fig. 1E) and NO production (Fig. 1F), in response to IFNγ treatment. Accordingly, p50 depletion impaired in vivo NO production in tumor tissues (Supplementary Fig. S1E). In contrast, both Nos2 mRNA (Fig. 1E) and NO production (Fig. 1F) were poorly induced by IFNγ treatment in both WT and p50−/− splenic CD11b+Ly6G+Ly6Clow/− cells (Supplementary Fig. S1D). Furthermore, the expression of arginase I (Arg1; ref. 13), indoleamine-2,3-dioxygenase (Ido1), and programmed cell death 1 ligand 1 (CD274) were poorly affected by the lack of p50 in both splenic and tumor CD11b+Ly6GLy6C+ and CD11b+Ly6G+Ly6Clow/− cells (Supplementary Fig. S1F and S1G) and not further investigated. To establish the actual role of p50 NF-κB in the suppressive activity of both myeloid subsets, CD11b+Ly6GLy6C+ and CD11b+Ly6G+Ly6Clow/− cells were isolated from tumor and spleen of fibrosarcoma (MN/MCA1)-bearing mice, activated with IFNγ and tested for their ability to inhibit T-cell proliferation. Consistently, both splenic and tumor M-MDSCs from WT mice strongly inhibited T-cell proliferation, whereas the p50-deficient counterparts displayed impaired suppressive ability (Fig. 1G and H).

In accordance with the lack of p50 nuclear accumulation, CD11b+Ly6G+Ly6Clow/−, both WT and p50-deficient, showed no suppressive activity (Fig. 1G and H). To further characterize the mechanisms underpinning the suppressive activity of p50 NF-κB, splenic M-MDSCs isolated from tumor-bearing mice were activated with IFNγ and tested for antigen-specific suppressive activity. In keeping with the data above, p50−/− M-MDSCs displayed reduced suppressive activity (Fig. 1I, left), and NO levels in the coculture supernatants (Fig. 1I, right). The M-MDSC–suppressive activity was NO-dependent, as addition of the NOS inhibitor L-NMMA to the coculture abolished both T-cell suppression (Fig. 1I, left) and NO production (Fig. 1I, right). These data suggested that the NO-dependent suppressive capacity of M-MDSCs in response to IFNγ, relies on p50 nuclear accumulation. We next determined whether the nuclear accumulation of p50 was also observable in blood M-MDSCs (13, 33) from patients with colorectal carcinoma. Compared with healthy donors, the frequency of M-MDSC (CD33+CD14+HLA-DRlow/−) was increased and these cells expressed higher level of NOS2 (Fig. 1J; Supplementary Fig. S2A), as well as increased nuclear levels of p50 (Fig. 1K), as compared with peripheral blood mononuclear CD14+ cells from healthy donors. To confirm the role of p50 in human myeloid cells, circulating CD14+HLA-DR+ mononuclear cells from healthy donors were transfected with a small interfering RNA against p50 (p50 siRNA). Silencing of p50 (Supplementary Fig. S2B), significantly inhibited Nos2 mRNA expression in response to IFNγ treatment (Supplementary Fig. S2C).

The predominant increase in the M-MDSCs subset within the spleen of tumor-bearing mice might results from its accelerated proliferation rate or preferential skewing of precursors. To study the implication of p50 in either normal or emergency hematopoiesis, we first evaluated the BM of naïve mice for composition in hematopoietic stem cells (HSC), along with their proliferation and differentiation potential. HSCs are immunophenotypically defined as cells lacking lineage specific markers (Lin−) but expressing Sca-1 and c-Kit (Lin-Sca-1+c-kit+, LSK), while the methylcellulose-based colony-forming unit (CFU) assay allows to quantify their derived progeny in vitro. We observed a significant enrichment of LSK progenitors (Lin-Sca-1+c-kit+) in BM of p50−/− mice (Fig. 2A and B), associated with higher clonogenic potential of p50−/− HSCs in both GM-CFU and M-CFU progenitors, but not in G-CFU (Fig. 2C). These results indicate that lack of p50 results in a preferential skewing of HSCs toward the monocytic branch at the myeloid/granulocytic bifurcation. Accordingly, histopathologic analysis showed a severe impairment of terminal granulopoiesis in the BM of p50−/− mice, which was associated with an increased number of immature myeloid precursors (Fig. 2D). Consistently, the hematopoietic parenchyma of p50−/− mice was characterized by the marked reduction of mature segmented granulocytes and by the expansion of myeloid blasts showing abnormal interstitial localization and aggregation in clusters (Fig. 2D). Monocytic differentiation (Fig. 2E) was preserved in the BM of p50−/− mice. Consistent with BM histopathology, flow cytometry revealed a neat decrease in the Gr1+c-Kit granulocytic population and a paralleled increase in Gr1+c-kit+ myeloblasts in p50−/−, as compared with control WT BM (Fig. 2F). Accordingly, blood granulocytes from p50−/− mice were enriched in immature and blast-like forms compared with circulating granulocytes from WT controls (Fig. 2G). Overall, these results confirm previous observations (34) and demonstrate that p50 NF-κB deficiency is associated with defective granulocytic differentiation in favor of the monocytic lineage. Lack of p50 resulted in a significant accumulation of LSK progenitors (Lin-Sca-1+c-kit+) also in the BM of tumor bearers (Fig. 2H). Furthermore, both histopathology (Fig. 2I) and flow cytometry (Fig. 2J) of BM confirmed the reduction of the mature Gr1+c-kit and the concomitant increase of the immature Gr1+c-kit+ granulocytic cells in p50−/− tumor bearers. Of relevance, cytofluorimetric analysis of tumor-infiltrating MDSCs revealed a significant increase of the negative regulator of MDSC expansion IRF8 (35) expression by the p50-deficient PMN-MDSC subset (CD11b+Ly6G+ cells; Fig. 2K), which also acts as promoter of terminal monocytic maturation (36). This result is in agreement with the preferential expansion of nonsuppressive Ly6C+ monocytic cells (M-MDSC–like) observed in tumor-bearing p50-deficient mice.

Hematopoietic deletion of p50 NF-κB abolishes NO production by M-MDSCs and impairs metastasis formation

To establish whether the expression of p50 NF-κB in the hematopoietic compartment was uniquely responsible for the suppressive activity of M-MDSCs, we generated chimeric mice by transplanting WT or p50−/− BM cells into sublethally irradiated WT and p50−/− mice, which were subsequently implanted with the MN/MCA1 fibrosarcoma. Splenic CD11b+Ly6GLy6C+ cells isolated from mice receiving p50−/− BM cells, and subsequently activated with IFNγ, were strongly impaired in their capacity to suppress both T-cell proliferation and NO production (Fig. 3A). To investigate further the in vivo importance of p50 in the hematopoietic compartment, we used the deleter strain B6.Cg-Tg(Tek-Cre)1Ywa to ablate p50 in all hematopoietic lineage cells (p50Tie2 mice) (37). In agreement with Fig. 1A and B, we observed inhibition of lung metastasis formation in tumor-bearing p50Tie2 mice, which was paralleled by an increased number of splenic and tumor-infiltrating CD11b+ Ly6GLy6Chigh cells, while an increase of CD11b+Ly6Clow/−Ly6G+ cells was observed at the tumor site (Fig. 3B).

Furthermore, splenic and tumor-associated CD11b+Ly6GLy6C+ cells from p50Tie2 mice (Fig. 3C) showed decreased NOS2 but higher TNFα expression, associated with increased number of both splenic and tumor-infiltrating IFNγ expressing CD4+ and CD8+ T cells (Fig. 3D), these latter also expressing high levels of the cytotoxic molecule granzyme B (GZMB). Accordingly, CD11b+Ly6GLy6C+ cells isolated from p50Tie2 mice showed a reduced ability to inhibit T cells' proliferation, whereas CD11b+Ly6G+Ly6CLow/− cells lacked suppressive activity (Fig 3E). Of note, transplantation of B16 melanoma cells (38) in p50Tie2 mice fully recapitulated the phenotype of the MN-MCA1 fibrosarcoma, both in terms of tumor growth (Supplementary Fig. S3A) and accumulation of splenic and tumor-infiltrating immune cells (Supplementary Fig. S3B–S3D).

p50 influences IFNγ-induced Stat1 recruitment to a subset of p50-dependent genes

Because the absence of p50 has strongly altered the response of CD11b+Ly6GLy6C+ cells to IFNγ, we further investigated this aspect. IFNγ-induced Stat1 phosphorylation (39) was not reduced in the absence of p50 (Fig. 4A), suggesting that downstream events occurring in p50−/− cells could be responsible for defective IFNγ-mediated gene expression. To directly assess this hypothesis, we generated mRNA-seq datasets of WT and p50−/− CD11b+Ly6GLy6C+ cells, isolated from the spleen of tumor-bearing mice, treated or not with IFNγ. p50−/− CD11b+Ly6G Ly6C+ cells showed selective gene expression defects in response to IFNγ stimulation as compared with WT controls (Fig. 4B; Supplementary Table S1). Using cutoffs, a log2 (fold change) ≥ 1 and FDR ≤ 0.05, 15 (2.4%, cluster 2) of the 628 genes induced by IFNγ in WT CD11b+Ly6GLy6C+ cells 15 (2.38%) were down-regulated in p50−/− cells, whereas 16 (2.5%) were upregulated. Analysis of the generated gene profiles confirmed that p50 deficiency results in a significant impaired induction of Nos2 (Fig. 4C) along with an increased induction of some IFN-inducible genes, such as Oas1g and Pyhin1/Ifix, (Supplementary Table S1). Collectively, our data highlight a specific function of p50 in controlling a subset of functionally relevant genes in response to IFNγ stimulation. Because transcriptional responses to IFNγ predominantly rely on the Stat1 transcription factor, we evaluated the recruitment of Stat1 on the Nos2 gene using ChIP with a validated antibody directed against Stat1 (30). Lack of p50 reduced binding of Stat1 on the Nos2 gene in CD11b+Ly6Chigh cells under both steady-state and IFNγ treatment (Fig. 4D), therefore indicating that p50 controls IFNγ-dependent responses at epigenetic level.

Along with MDSCs, macrophages produce NO in response to IFNγ, which mediate either their immunosuppressive (4, 13) or tumoricidal capacity (40). Similar to M-MDSCs, in thioglycollate elicited macrophages (PEC) lack of p50 decreased Nos2 mRNA expression and NO production in response to IFNγ (Supplementary Fig. S4A), without reducing Stat1 phosphorylation (Supplementary Fig. S4B). In further analogy with MDSCs, PECs primed with the tumor supernatant (TSN) produced higher level of NO in response to IFNγ, in a p50-dependent manner (Supplementary Fig. S4C). Moreover, transcriptional analysis (mRNA-seq) of WT and p50−/− PECs confirmed that p50−/− PECs showed selective gene expression defects in response to IFNγ (Supplementary Fig. S4D and Supplementary Table S2), including Nos2. On the other hand, in agreement with Fig. 2H, p50 deficiency resulted in an increased induction of the monopoiesis-inducing transcription factor Irf8 (36; Supplementary Table S2). We then performed ChIP coupled to next-generation sequencing (ChIP-seq). Almost all DNA binding events occurred only after Stat1 activation (Supplementary Table S3) and positively correlated with IFNγ-induced gene expression in a statistically significant manner (Supplementary Fig. S4E), highlighting the prominent role of Stat1 as a transcriptional activator. A discrete fraction of the Stat1 cistrome was selectively affected by p50 deficiency, with an abrogation or reduction of Stat1 occupancy at 2571 sites (8.3% of all inducible peaks) in p50−/− PEC relative to WT controls. For instance, we confirmed that Stat1 was not efficiently recruited to either promoters or enhancers of Nos2 gene in p50−/− macrophages, and this was associated with its reduced induction in response to IFNγ (Supplementary Fig. S4F). These observations were then validated at a genomic scale by computationally integrating our ChIP-seq and mRNA-seq datasets (Supplementary Table S4). As shown in Supplementary Fig. S4G, p50-dependent genes were located at shorter distances from p50-dependent Stat1 peaks than p50-independent genes. Conversely, p50-independent genes were closer to p50-independent Stat1 peaks.

These findings identify a role for p50 in controlling IFNγ-induced Nos2 gene expression in myeloid cells, and are consistent with a model of p50-dependent assistance of Stat1 recruitment to selected p50-dependent genes in response to IFNγ treatment.

Ablation of p50 NF-κB in M-MDSCs restores the in vivo antitumor activity of IFNγ

We next investigated whether increased nuclear p50 NF-κB in M-MDSCs limits cytokine-mediated immunotherapy in vivo. While IFNγ treatment of WT tumor-bearing mice was ineffective, the same treatment significantly reduced tumor development in p50−/− mice (Fig. 5A). Noteworthy, in vivo depletion of CD4+ and CD8+ T cells in p50 NF-κB-deficient tumor bearing mice restored tumor growth (Fig. 5A, center) and metastasis formation (Fig. 5A, right), indicating that myeloid-specific p50 NF-κB is suppressing specific antitumor immunity. Hence, to test the specific contribution of p50 to the immunosuppressive activity of M-MDSCs, p50−/− tumor-bearing mice were adoptively transferred with WT M-MDSCs (1 × 106) and treated daily with IFNγ. Such M-MDSCs transfer restored tumor growth in p50−/− mice (Fig. 5B) and decreased both IFNγ production and GZMB expression by CD8+ T cells, both in the spleen and the primary tumor (Fig. 5C and D). Of note, transfer of p50−/− CD11b+Ly6GLy6C+ cells did not affect tumor growth, as well as the phenotype of CD4+ and CD8+ T cells. These data demonstrate that accumulation of nuclear p50 in M-MDSCs promotes their suppressive functions in blunting the efficacy of cytokine-mediated immunotherapy (i.e., IFNγ). This conclusion was further strengthened observing that ablation of p50 in tumor-bearing mice improved the antitumor and antimetastatic activity of the immunostimulatory cytokine IL12 (Supplementary Fig. S5).

Tumor-derived PGE2 primes M-MDSCs for higher IFNγ-induced NO-mediated suppressive activity

In the attempt to identify tumor-derived signals controlling nuclear p50 accumulation, we used BM-MDSCs (29) to test molecules detected in the TSN (Fig. 6A) and reported to either induce nuclear accumulation of p50 NF-κB homodimers in macrophages (i.e., PGE2, IL10, TGFβ; ref. 10) or myeloid cell differentiation (GM-CSF, G-CSF, and M-CSF). Confocal microscopy analysis showed that neither M-, GM- nor G-CSF induced p50 accumulation (Supplementary Fig. S6A). In contrast, PGE2 efficiently induced nuclear p50, whereas the PGE2 receptor antagonists EP2 (AH6809) and EP4 (AH23848; ref. 41) inhibited the TSN-induced nuclear accumulation of p50 (Fig. 6B). Furthermore, BM-MDSCs primed with TSN showed higher levels of both Nos2 mRNA and NO production in response to IFNγ (Fig. 6C), which was prevented by AH6809 (Fig. 6D). In agreement, pretreatment with Butaprost, a selective agonist of EP2, enhanced IFNγ-mediated NO production in a p50-dependent manner (Fig. 6D).

We next evaluated the capacity of AH6809 to interfere with the MN/MCA1 fibrosarcoma in vivo. Because MN-MCA1 cells also express the EP2R (Supplementary Fig. S6B), we evaluated the effects of AH6809 on tumor cell viability. MN-MCA1 cells were cultured in presence of 10 μmol/L AH6809, 10 μmol/L PGE2 or the combination of AH6809 and PGE2, up to 72 hours. As shown (Supplementary Fig. S6C), AH6809 did not affect tumor cell proliferation and viability. In contrast, daily treatment with AH6809 inhibited tumor growth (Fig. 6E, left) and reduced the number of lung metastasis (Fig. 6E, right) in WT mice. Consistently, no tumor growth inhibition was observed in p50−/− mice upon treatment with AH6809 (Fig. 6E, left). In agreement with the phenotype of p50−/− mice (Fig. 3C and D), flow cytometry of tumor tissue (MN/MCA1) from tumor-bearing WT mice, showed that AH6809 treatment significantly reduced the expression of NOS2 (Fig. 6F, left) by CD11b+Ly6GLy6C+ cells, while increasing the expression of TNFα (Fig. 6F, center), as well as the IFNγ+CD4+ and IFNγ+CD8+ T cells (Fig. 6F, right). Similar results were obtained in splenic M-MDSCs (Fig. 6G). Further supporting the protumoral activity of the PGE2–p50–NO axis, in vivo treatment with AH6809 displayed an antitumor effects comparable with an antibody against the PD-1 immune checkpoint inhibitor (anti-PD-1), both in terms of primary tumor growth (MN/MCA1) and lung metastasis formation (Supplementary Fig. S6D), while the expression of NOS2 in CD11b+Ly6C+ WT M-MDSCs cells was reduced to levels comparable with p50−/− M-MDSC (Supplementary Fig S6E). In agreement, treatment of B16-bearing WT mice with AH6809 reduced tumor growth and NOS2 expression by CD11b+Ly6C+ M-MDSCs, mimicking p50 NF-κB ablation (Supplementary Fig. S6F). Despite also B16 melanoma cells expressed EP2R (Supplementary Fig. S6G), AH6809 did not elicit direct cytotoxic effects in vitro (Supplementary Fig. S6H), neither antitumor effects in p50-deficient mice (Supplementary Fig. S6F).

Finally, we investigated the effect of NO on the antitumor activity of IFNγ in vivo. WT B16-bearing mice were then treated with IFNγ and/or with L-NMMA. While treatment with IFNγ elicited poor effects on tumor growth, L-NMMA alone displayed a significant antitumor effect, which was enhanced in combination with IFNγ (Fig. 7A). Moreover, L-NMMA, and even more its combination with IFNγ, reduced the expression of NOS2 by tumor M-MDSCs (Fig. 7B, top) and enhanced the expression of TNFα (Fig. 7B, bottom), while increasing the number of tumor-infiltrating IFNγ+CD8+ T cells (Fig. 7C). These results indicate NO as a terminal mediator of suppressive myeloid cell functions and suggest that blocking the PGE2-p50-NO axis either upstream (i.e., EP2 inhibitors) or downstream (i.e., NOS2 inhibitor L-NMMA) may restore antitumor immunity.

To corroborate this finding in patients with cancer, we analyzed the expression of both the PGE2 receptor EP2 and NOS2 in PBMCs from patients with advanced colorectal cancer (stage T2/T3; n = 5). In agreement with mouse data, human M-MDSCs (HLA-DRlow/−CD14+CD33high; Supplementary Fig. S2A) expressed the PGE2 receptor EP2 (Fig. 7D). Moreover, human peripheral blood monocytes primed with TSN from the human pancreatic carcinoma cell line PANC1, containing 923 μg/mL of PGE2, and subsequently treated with IFNγ, showed higher levels of intracellular NOS2 protein, that were reduced by treatment with AH6809 (Fig. 7E). While confirming the suppressive activity of PGE2 in cancer bearers (42), our results are the first to identify the PGE2-driven nuclear accumulation of p50 NF-κB as leading event in the differentiation of suppressive M-MDSCs.

Cancers support a pathologic generation of immunosuppressive and tumor-promoting MDSC populations (12, 13), that share phenotypic traits with TAMs, including the expression of anti-inflammatory M2 polarized genes (10). Here, we demonstrate that tumor-derived PGE2 drives the suppressive phenotype of M-MDSCs through upregulation of nuclear p50 NF-κB, a key player in the resolution of the inflammatory response, that links unresolved cancer inflammation with tolerance (9, 28). Accumulation of nuclear p50 NF-κB, in both mouse (CD11b+Ly6GLy6C+) and human (patients with colorectal cancer; CD14+HLA-DRlow/−) M-MDSCs, diverts their transcriptional responses to IFNγ toward enhanced production of the immunosuppressive molecule NO. Consistently, ablation of p50 NF-κB reactivated specific antitumor immunity and reinstated the in vivo antitumor activity of IL12 and IFNγ, two immunostimulatory cytokines evaluated in clinical trials against various human tumors (e.g., colorectal cancer, soft tissue sarcoma, melanoma, and plasma cell neoplasms; ref. 16). This observation provides new insights into the ambivalent action of IFNγ (22, 43, 44), which in clinical protocols has shown either moderate or poor efficacy (17, 21, 45). This ambiguity is reminiscent of the bimodal immunologic activities of IFNγ, which promotes transcription of genes involved in the activation of immune responses (e.g. MHC class I and class II, IL12; ref. 46) and simultaneously induces immunosuppressive pathways, including B7-H1 (47, 48) and the immunosuppressive enzymes IDO (23) and iNOS/NOS2 (12, 13). The biological activity of IFNγ requires the nuclear translocation and the DNA binding of the STAT1 homodimer to gamma activated sequence (GAS) sites on the promoters of downstream target genes (49), including Nos2 (50). Accumulation of p50 NF-κB in myeloid cells does not affect IFNγ-dependent STAT1 phosphorylation, but rather controls their chromatin landscape to promote binding of STAT1 onto specific gene regulatory elements of IFNγ-responsive genes, including Nos2. We also observed that lack of p50 results in increased CD11b+Ly6GLy6C+ cells in the spleen of tumor-bearing mice, with low NO production and suppressive activity, as well as in the preferential skewing of HSCs toward the monocytic branch in the bone marrow. This hematopoietic output is in agreement with the Irf8high feature that we observed in p50-deficient PMN-MDSCs, because IRF8 is a cell fate–switching factor driving terminal differentiation of the monocytic lineage (36) ad preventing MDSC expansion (35).

COX2-stimulated production of PGE2 enhances the expansion of MDSCs and promotes suppression of adaptive immunity (51). We show that antagonists of the PGE2 receptors EP1/EP2 reprogram M-MDSC functions in response to IFNγ toward an inflammatory phenotype (i.e., NOS2low/TNFhigh), associated with increased specific antitumor immunity. Of relevance, in analogy with the results obtained both with the inhibition of PGE2 receptors and with the ablation of the NF-κB p50 gene, the NOS L-NMMA inhibitor reprogrammed the M-MDSCs toward an NOS2low/TNFαhigh phenotype, associated with an increased antitumoral effect of IFNγ. This evidence indicates the gasotransmitter NO (52) as the final effector of the protumoral reprogramming of myeloid cells, guided by the PGE2/p50 axis.

COX inhibitors (e.g., aspirin) were suggested for prevention and treatment of cancer (42). Nevertheless, due to their severe side effects, targeting specific prostanoid receptors endowed with immunosuppressive properties, such as EP2 (53–55), might provide an alternative selective and safer approach to burst specific immunity in cancer patients. Our work indicates that p50 NF-κB accumulation diverts the differentiation and activation status of myeloid cells, leading to the resolution of immune and inflammatory responses. It further demonstrates that in cancer an accumulation of p50 NF-κB, induced by PGE2, acts as an epigenetic hacker of M-MDSC functions, which establishes their suppressive phenotype, suggesting the use of PGE2 receptor antagonists as potential adjuvants for anticancer immunotherapy.

V. Bronte has ownership interest (including patents) in Xios Therapeutics, Codiak BioSciences, EMD Serono, and BioNTech AG and is a consultant/advisory board member for Xios Therapeutics, Codiak BioSciences, EMD Serono, Incyte Corporation, and Ganymed Pharmaceuticals AG. F. Balzac is a consultant at KITHER. E. Hirsch is a consultant/advisory board member for Kither Biotech. No potential conflicts of interest were disclosed by the other authors.

Conception and design: P. Larghi, V. Bronte, A. Sica

Development of methodology: S. Morlacchi, A. Bleve, M.G. Totaro, P. Larghi, M. Rimoldi, L. Strauss, M. Storto, A. Ippolito, A. Doni, R. Ostuni, V. Bronte

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C. Porta, F.M. Consonni, S. Sangaletti, A. Bleve, M. Storto, T. Pressiani, L. Rimassa, S. Tartari, A. Ippolito, A. Doni, R. Ostuni, F. Balzac, E. Turco, E. Hirsch

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Porta, C. Tripodo, L. Strauss, A. Ippolito, G. Soldà, S. Duga, V. Piccolo, R. Ostuni, G. Natoli, A. Sica

Writing, review, and/or revision of the manuscript: C. Porta, L. Rimassa, G. Soldà, S. Duga, V. Bronte, M.P. Colombo, A. Sica

Study supervision: G. Natoli, A. Sica

Other (performed the experiments): S. Morlacchi

Other (performed some experiments): S. Banfi

Other (acquisitions of confocal optical microscopy and colocalization analysis): A. Doni

This work was supported by Associazione Italiana per la Ricerca sul Cancro (AIRC) Italy (12810, to S. Sangaletti; IG numbers 10137 to M.P. Colombo; IG numbers 15585, 19885 to A. Sica; AIRC 5 per Mille, number 22757 to A. Sica; AIRC 5 per Mille, Program Innovative Tools for Cancer Risk Assessment and Diagnosis, number 12162 to C. Tripodo; Three-Year fellowship “Pierluigi Meneghelli,” project code 19682, to F.M. Consonni). Ministero dell'istruzione, dell'Università e della Ricerca (MIUR), Italy (PRIN 2015YYKPNN and 2017BA9LM5_001 to A. Sica); Ministero della Salute, Ricerca Finalizzata RF-2016-0236842 to A. Sica and GR-2013-02355637 to S. Sangaletti; European Research Council (ERC) Advanced grant ERCAdGN.692789 to G. Natoli; Fondazione Cariplo 2016/0871 to A. Sica, Associazione “Augusto per la Vita,” Novellara, Italy, to A. Sica, Associazione “Medicine Rocks,” Milan, Italy, to A. Sica, and Università del Piemonte Orientale, Novara, Italy, Ricerca locale 2019 to C. Porta.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Mosser
DM
,
Edwards
JP
. 
Exploring the full spectrum of macrophage activation
.
Nat Rev Immunol
2008
;
8
:
958
69
.
2.
Sica
A
,
Mantovani
A
. 
Macrophage plasticity and polarization: in vivo veritas
.
J Clin Invest
2012
;
122
:
787
95
.
3.
Mantovani
A
,
Sozzani
S
,
Locati
M
,
Allavena
P
,
Sica
A
. 
Macrophage polarization: tumor-associated macrophages as a paradigm for polarized M2 mononuclear phagocytes
.
Trends Immunol
2002
;
23
:
549
55
.
4.
Mantovani
A
,
Sica
A
. 
Macrophages, innate immunity and cancer: balance, tolerance, and diversity
.
Curr Opin Immunol
2010
;
22
:
231
7
.
5.
Mantovani
A
,
Allavena
P
,
Sica
A
,
Balkwill
F
. 
Cancer-related inflammation
.
Nature
2008
;
454
:
436
44
.
6.
Martinez
FO
,
Helming
L
,
Gordon
S
. 
Alternative activation of macrophages: an immunologic functional perspective
.
Annu Rev Immunol
2009
;
27
:
451
83
.
7.
Noy
R
,
Pollard
JW
. 
Tumor-associated macrophages: from mechanisms to therapy
.
Immunity
2014
;
41
:
49
61
.
8.
Ziegler-Heitbrock
HW
,
Petersmann
I
,
Frankenberger
M
. 
p50 (NF-kappa B1) is upregulated in LPS tolerant P388D1 murine macrophages
.
Immunobiology
1997
;
198
:
73
80
.
9.
Porta
C
,
Rimoldi
M
,
Raes
G
,
Brys
L
,
Ghezzi
P
,
Di Liberto
D
, et al
Tolerance and M2 (alternative) macrophage polarization are related processes orchestrated by p50 nuclear factor kappaB
.
Proc Natl Acad Sci U S A
2009
;
106
:
14978
83
.
10.
Biswas
SK
,
Gangi
L
,
Paul
S
,
Schioppa
T
,
Saccani
A
,
Sironi
M
, et al
A distinct and unique transcriptional program expressed by tumor-associated macrophages (defective NF-kappaB and enhanced IRF-3/STAT1 activation)
.
Blood
2006
;
107
:
2112
22
.
11.
Strauss
L
,
Sangaletti
S
,
Consonni
FM
,
Szebeni
G
,
Morlacchi
S
,
Totaro
MG
, et al
RORC1 regulates tumor-promoting "Emergency" granulo-monocytopoiesis
.
Cancer Cell
2015
;
28
:
253
69
.
12.
Sica
A
,
Bronte
V
. 
Altered macrophage differentiation and immune dysfunction in tumor development
.
J Clin Invest
2007
;
117
:
1155
66
.
13.
Gabrilovich
DI
,
Ostrand-Rosenberg
S
,
Bronte
V
. 
Coordinated regulation of myeloid cells by tumours
.
Nat Rev Immunol
2012
;
12
:
253
68
.
14.
Martin
F
,
Apetoh
L
,
Ghiringhelli
F
. 
Role of myeloid-derived suppressor cells in tumor immunotherapy
.
Immunotherapy
2012
;
4
:
43
57
.
15.
Zaidi
MR
,
Davis
S
,
Noonan
FP
,
Graff-Cherry
C
,
Hawley
TS
,
Walker
RL
, et al
Interferon-gamma links ultraviolet radiation to melanomagenesis in mice
.
Nature
2011
;
469
:
548
53
.
16.
Vacchelli
E
,
Galluzzi
L
,
Eggermont
A
,
Galon
J
,
Tartour
E
,
Zitvogel
L
, et al
Trial watch: immunostimulatory cytokines
.
Oncoimmunology
2012
;
1
:
493
506
.
17.
Miller
CH
,
Maher
SG
,
Young
HA
. 
Clinical use of interferon-gamma
.
Ann N Y Acad Sci
2009
;
1182
:
69
79
.
18.
Assi
HI
,
Kamphorst
AO
,
Moukalled
NM
,
Ramalingam
SS
. 
Immune checkpoint inhibitors in advanced non-small cell lung cancer
.
Cancer
2018
;
124
:
248
61
.
19.
Lo Russo
G
,
Moro
M
,
Sommariva
M
,
Cancila
V
,
Boeri
M
,
Centonze
G
, et al
Antibody-Fc/FcR interaction on macrophages as a mechanism for hyperprogressive disease in non-small cell lung cancer subsequent to PD-1/PD-L1 blockade
.
Clin Cancer Res
2019
;
25
:
989
99
.
20.
Adams
DO
,
Hamilton
TA
. 
The cell biology of macrophage activation
.
Annu Rev Immunol
1984
;
2
:
283
318
.
21.
Zaidi
MR
,
Merlino
G
. 
The two faces of interferon-gamma in cancer
.
Clin Cancer Res
2011
;
17
:
6118
24
.
22.
Dighe
AS
,
Richards
E
,
Old
LJ
,
Schreiber
RD
. 
Enhanced in vivo growth and resistance to rejection of tumor cells expressing dominant negative IFN gamma receptors
.
Immunity
1994
;
1
:
447
56
.
23.
Pallotta
MT
,
Orabona
C
,
Volpi
C
,
Vacca
C
,
Belladonna
ML
,
Bianchi
R
, et al
Indoleamine 2,3-dioxygenase is a signaling protein in long-term tolerance by dendritic cells
.
Nat Immunol
2011
;
12
:
870
8
.
24.
Kondo
A
,
Yamashita
T
,
Tamura
H
,
Zhao
W
,
Tsuji
T
,
Shimizu
M
, et al
Interferon-gamma and tumor necrosis factor-alpha induce an immunoinhibitory molecule, B7-H1, via nuclear factor-kappaB activation in blasts in myelodysplastic syndromes
.
Blood
2010
;
116
:
1124
31
.
25.
Topalian
SL
,
Drake
CG
,
Pardoll
DM
. 
Immune checkpoint blockade: a common denominator approach to cancer therapy
.
Cancer Cell
2015
;
27
:
450
61
.
26.
Chan
SR
,
Vermi
W
,
Luo
J
,
Lucini
L
,
Rickert
C
,
Fowler
AM
, et al
STAT1-deficient mice spontaneously develop estrogen receptor alpha-positive luminal mammary carcinomas
.
Breast Cancer Res
2012
;
14
:
R16
.
27.
Saccani
A
,
Schioppa
T
,
Porta
C
,
Biswas
SK
,
Nebuloni
M
,
Vago
L
, et al
p50 nuclear factor-kappaB overexpression in tumor-associated macrophages inhibits M1 inflammatory responses and antitumor resistance
.
Cancer Res
2006
;
66
:
11432
40
.
28.
Porta
C
,
Ippolito
A
,
Consonni
FM
,
Carraro
L
,
Celesti
G
,
Correale
C
, et al
Protumor steering of cancer inflammation by p50 NF-kappaB enhances colorectal cancer progression
.
Cancer Immunol Res
2018
;
6
:
578
93
.
29.
Marigo
I
,
Bosio
E
,
Solito
S
,
Mesa
C
,
Fernandez
A
,
Dolcetti
L
, et al
Tumor-induced tolerance and immune suppression depend on the C/EBPbeta transcription factor
.
Immunity
2010
;
32
:
790
802
.
30.
Ostuni
R
,
Natoli
G
. 
Lineages, cell types and functional states: a genomic view
.
Curr Opin Cell Biol
2013
;
25
:
759
64
.
31.
Garber
M
,
Yosef
N
,
Goren
A
,
Raychowdhury
R
,
Thielke
A
,
Guttman
M
, et al
A high-throughput chromatin immunoprecipitation approach reveals principles of dynamic gene regulation in mammals
.
Mol Cell
2012
;
47
:
810
22
.
32.
Movahedi
K
,
Guilliams
M
,
Van den Bossche
J
,
Van den Bergh
R
,
Gysemans
C
,
Beschin
A
, et al
Identification of discrete tumor-induced myeloid-derived suppressor cell subpopulations with distinct T cell-suppressive activity
.
Blood
2008
;
111
:
4233
44
.
33.
Mandruzzato
S
,
Solito
S
,
Falisi
E
,
Francescato
S
,
Chiarion-Sileni
V
,
Mocellin
S
, et al
IL4Ralpha+ myeloid-derived suppressor cell expansion in cancer patients
.
J Immunol
2009
;
182
:
6562
8
.
34.
Wang
D
,
Paz-Priel
I
,
Friedman
AD
. 
NF-kappa B p50 regulates C/EBP alpha expression and inflammatory cytokine-induced neutrophil production
.
J Immunol
2009
;
182
:
5757
62
.
35.
Waight
JD
,
Netherby
C
,
Hensen
ML
,
Miller
A
,
Hu
Q
,
Liu
S
, et al
Myeloid-derived suppressor cell development is regulated by a STAT/IRF-8 axis
.
J Clin Invest
2013
;
123
:
4464
78
.
36.
Friedman
AD
. 
Transcriptional control of granulocyte and monocyte development
.
Oncogene
2007
;
26
:
6816
28
.
37.
El Kasmi
KC
,
Qualls
JE
,
Pesce
JT
,
Smith
AM
,
Thompson
RW
,
Henao-Tamayo
M
, et al
Toll-like receptor-induced arginase 1 in macrophages thwarts effective immunity against intracellular pathogens
.
Nat Immunol
2008
;
9
:
1399
406
.
38.
Sica
A
,
Saccani
A
,
Bottazzi
B
,
Polentarutti
N
,
Vecchi
A
,
van Damme
J
, et al
Autocrine production of IL-10 mediates defective IL-12 production and NF-kappa B activation in tumor-associated macrophages
.
J Immunol
2000
;
164
:
762
7
.
39.
Platanias
LC
. 
Mechanisms of type-I- and type-II-interferon-mediated signalling
.
Nat Rev Immunol
2005
;
5
:
375
86
.
40.
Dinapoli
MR
,
Calderon
CL
,
Lopez
DM
. 
The altered tumoricidal capacity of macrophages isolated from tumor-bearing mice is related to reduce expression of the inducible nitric oxide synthase gene
.
J Exp Med
1996
;
183
:
1323
9
.
41.
Zhang
T
,
Gong
X
,
Hu
G
,
Wang
X
. 
EP2-PKA signaling is suppressed by triptolide in lipopolysaccharide-induced microglia activation
.
J Neuroinflammation
2015
;
12
:
50
.
42.
Zelenay
S
,
van der Veen
AG
,
Bottcher
JP
,
Snelgrove
KJ
,
Rogers
N
,
Acton
SE
, et al
Cyclooxygenase-dependent tumor growth through evasion of immunity
.
Cell
2015
;
162
:
1257
70
.
43.
Kaplan
DH
,
Shankaran
V
,
Dighe
AS
,
Stockert
E
,
Aguet
M
,
Old
LJ
, et al
Demonstration of an interferon gamma-dependent tumor surveillance system in immunocompetent mice
.
Proc Natl Acad Sci U S A
1998
;
95
:
7556
61
.
44.
Shankaran
V
,
Ikeda
H
,
Bruce
AT
,
White
JM
,
Swanson
PE
,
Old
LJ
, et al
IFNgamma and lymphocytes prevent primary tumour development and shape tumour immunogenicity
.
Nature
2001
;
410
:
1107
11
.
45.
Wilke
CM
,
Wei
S
,
Wang
L
,
Kryczek
I
,
Kao
J
,
Zou
W
. 
Dual biological effects of the cytokines interleukin-10 and interferon-gamma
.
Cancer Immunol Immunother
2011
;
60
:
1529
41
.
46.
Lawrence
T
,
Natoli
G
. 
Transcriptional regulation of macrophage polarization: enabling diversity with identity
.
Nat Rev Immunol
2011
;
11
:
750
61
.
47.
Curiel
TJ
,
Wei
S
,
Dong
H
,
Alvarez
X
,
Cheng
P
,
Mottram
P
, et al
Blockade of B7-H1 improves myeloid dendritic cell-mediated antitumor immunity
.
Nat Med
2003
;
9
:
562
7
.
48.
Wu
K
,
Kryczek
I
,
Chen
L
,
Zou
W
,
Welling
TH
. 
Kupffer cell suppression of CD8+ T cells in human hepatocellular carcinoma is mediated by B7-H1/programmed death-1 interactions
.
Cancer Res
2009
;
69
:
8067
75
.
49.
Platanias
LC
. 
Introduction: interferon signals: what is classical and what is nonclassical?
J Interferon Cytokine Res
2005
;
25
:
732
.
50.
Ganster
RW
,
Taylor
BS
,
Shao
L
,
Geller
DA
. 
Complex regulation of human inducible nitric oxide synthase gene transcription by Stat 1 and NF-kappa B
.
Proc Natl Acad Sci U S A
2001
;
98
:
8638
43
.
51.
Mao
Y
,
Sarhan
D
,
Steven
A
,
Seliger
B
,
Kiessling
R
,
Lundqvist
A
. 
Inhibition of tumor-derived prostaglandin-e2 blocks the induction of myeloid-derived suppressor cells and recovers natural killer cell activity
.
Clin Cancer Res
2014
;
20
:
4096
106
.
52.
Kobayashi
Y
. 
The regulatory role of nitric oxide in proinflammatory cytokine expression during the induction and resolution of inflammation
.
J Leukoc Biol
2010
;
88
:
1157
62
.
53.
Ganesh
T
. 
Prostanoid receptor EP2 as a therapeutic target
.
J Med Chem
2014
;
57
:
4454
65
.
54.
O'Brien
AJ
,
Fullerton
JN
,
Massey
KA
,
Auld
G
,
Sewell
G
,
James
S
, et al
Immunosuppression in acutely decompensated cirrhosis is mediated by prostaglandin E2
.
Nat Med
2014
;
20
:
518
23
.
55.
O'Callaghan
G
,
Houston
A
. 
Prostaglandin E2 and the EP receptors in malignancy: possible therapeutic targets?
Br J Pharmacol
2015
;
172
:
5239
50
.