FFAR2 expressing myeloid-derived suppressor cells drive cancer immunoevasion

Mouse Lewis lung carcinoma cell (LLC), melanoma cell (B16F10), fibroblast cell (3T3), human melanoma cell (SK-MEL-2), lung adenocarcinoma cell (A549), human normal epithelial cell (BEAS-2B) and umbilical vein endothelial cell (HUVEC) lines were purchased from the American Type Culture Collection (ATCC, USA). NCM460 and PC-9 cell line was obtained from Cell Resource Center of East China Normal University. LLC was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1 × penicillin–streptomycin. LLC stably expressing firefly luciferase (LLC-luci) was generated by our lab as described [17, 18] and cultured in complete DMEM medium with 200 ng/mL G418 (Gibco). B16F10, HUVEC, PC-9, A549 and 3T3 were cultured in RPMI 1640 Medium supplemented with 10% fetal bovine serum (FBS), 1 × penicillin–streptomycin. SK-MEL-2 was cultured in Modified Eagle’s Medium (MEM) supplemented with 10% fetal bovine serum (FBS), 1 × penicillin–streptomycin. All cell lines cultured maintained at 37 °C and 5% CO₂, and were regularly tested for mycoplasma-free.

RPMI 1640, Dulbecco’s modified Eagle’s medium (DMEM), penicillin–streptomycin, and fetal bovine serum (FBS) was purchased from Gibco. TRIzol reagent and PrimeScript RT Master Mix were acquired from Takara. SYBR Green PCR Master Mix was purchased from Yeasen. Mouse TNF-α ELISA kit was purchased from Biolegend and Mouse IL-12 p70 ELISA kit was purchased from Invitrogen. GM-CSF and IL-6 were purchased from Proteintech. InVivoMab anti-mouse Ly6G/Ly6C (Gr-1; clone RB6-8C5), InVivoMab IgG2a isotype antibody (clone LTF-2) and InVivoMab anti-mouse PD-1 (clone RMP1-14) were purchased from BioXCell. Mouse Myeloid-Derived Suppressor Cell Isolation Kit, mouse CD4⁺ T Cell Isolation Kit, mouse CD8⁺ T Cell Isolation Kit, LS Separation columns, MS Separation columns, MACS BSA Stock Solution, autoMACS Rinsing Solution, autoMACS Running Buffer and mouse T Cell Activation/Expansion Kit were purchased from Miltenyi. FFAR2 Agonist (#371725), FFAR2 inhibitor (GLPG0974, #SML2443), Sodium acetate (#S2889) and Urethane (#U2500) were purchased from Sigma-Aldrich. Gαq inhibitor (YM-254890) and FFAR2 inhibitor (CATPB) were purchased from MCE. Ca²⁺ inhibitor (2-APB) was purchased from Tocris and PPAR-γ-inhibitor (GW9662) was purchased from Selleck. Live/dead dye and antibodies used for flow cytometry were purchased from Biolegend unless indicated otherwise: Fixable Viability Dye (BD, Horizon™ Fixable Viability Stain 780), FITC-conjugated anti-mouse CD45 (clone 30-F11), APC-conjugated anti-mouse CD3 (clone 17A2), Brilliant Violet 421™-conjugated anti-mouse CD4 (clone RM4-4), PE-conjugated anti-mouse CD8a (clone 53-6.7), PE/Cyanine7-conjugated anti-mouse F4/80 (clone BM8), PE-conjugated anti-mouse CD11c (clone N418), TruStain fcX™ anti-mouse CD16/32 (clone 93), Brilliant Violet 605™-conjugated anti-mouse I-A/I-E (clone M5/114.15.2), PE-conjugated anti-mouse CD8a (BD Pharmingen, clone 53-6.7), BV421-conjugated anti-mouse LY-6G (BD Pharmingen, clone 1A8), APC-conjugated anti-mouse LY-6C (BD Pharmingen, clone AL-21), FITC-conjugated anti-mouse CD11b (BD Pharmingen, clone M1/70), FITC-conjugated anti-mouse CD8a (clone 53-6.7), PE-conjugated anti-human/mouse Arginase 1 (eBioscience™, clone A1exF5), Brilliant Violet 421™-conjugated anti-mouse/human CD11b (clone M1/70), PE-conjugated anti-mouse IFNγ (clone XMG1.2), Brilliant Violet 421™-conjugated anti-mouse/human CD11b (clone M1/70), PE-conjugated anti-mouse IFNγ (clone XMG1.2), FITC-conjugated anti-mouse CD8a (clone 53-6.7), PE-conjugated anti-mouse iNOS (clone W16030C), PE/Cyanine7-conjugated anti-mouse IL-10 (clone JES5-16E3), Alexa Fluor® 700 -conjugated anti-mouse NK-1.1 (clone PK136) and Brilliant Violet 421™-conjugated anti-mouse CD25 (clone A18246A). Antibodies used for Western blotting are as follows: PPAR-γ (#2443), Arginase-1 (#93668), STAT1 (#9172), p-STAT1 (#9171), STAT3 (#4904), p-STAT3 (#9145) and AlexaFluor® 488/555 mouse, rabbit secondary antibodies were purchased from Cell Signaling Technology (CST). C/EBPβ (H7, sc-7962) and p-C/EBPβ (Thr217, sc-16993-R) were purchased from Santa Cruz Biotechnology. Antibodies used for immunofluorescence are as follows: Anti-human FFAR2 (ab124272) and anti-human CD15 (ab17080) were purchased from abcam. Anti-mouse Gr-1 (clone RB6-8C5) was purchased from Biolegend. Anti-CD8 (GB13429), anti-CD4 (GB13064-2), anti-human ARG1 (GB11285) and anti-PPAR-γ (GB112205) were purchased from Servicebio.

C57BL/6 mice (6–8 weeks old) were purchased from Gempharmatech Co., Ltd. Lyz2-cre mice were purchased from Jackson Laboratory. Global knockout mice Ffar2^−/− and conditional knockout mice Ffar2^fl/fl were constructed using the CRISPR–Cas9 genome-editing system. Ffar2^fl/fl mice were crossed with Lyz2-cre mice to obtain mice with targeted Ffar2 deletion. All mouse breeding and mouse animal experiments are done at the specific-pathogen-free conditions Experimental Animal Center of East China Normal University. All animal experiments were approved by the Institutional Animal Ethics Committee of East China Normal University. The protocol was approved by the East China Normal University Center for Animal Research (m20230201).

Sample preparation of cell supernatants: 3T3, LLC, B16F10, BEAS-2B, NCM460, HUVEC, PC-9, A549 and SK-MEL-2 cells were cultured in recommended complete medium for a few days, and seeded then in 6-well plates (1 × 10⁶ cells per well), and cultured in RPMI 1640 supplemented with 10% FBS and 1 × penicillin–streptomycin (3 ml complete medium per well) for 16 h. Cell culture supernatants were then collected and kept in −80 °C. Preparation of urethane-induced lung cancer tissue: Urethane-induced mice lung cancer tissue and normal mice lung tissue was dissected, washed twice with precooled PBS, and kept then in −80 °C. All samples were sent to Suzhou Meixin Bioscience Co., Ltd and analyzed and quantified by GC–MS.

Bone marrow cells were harvested from tibias and femurs of C57BL/6 mice (6–8 weeks old), after red cell lysis, cell suspension was cultured RPMI 1640 supplemented with 1 × Penicillin and Streptomycin, 10% FBS, GM-CSF (40 ng/ml) and IL6 (40 ng/ml) medium for 4 days.

Mice were subcutaneously injected with LLC (1 × 10⁶ cells/mouse), and LLC tumors were excised at day 21. LLC tumors were minced into small pieces (less than 3 mm in diameter) and re-suspended into T75 culture flask with RPMI 1640 medium (without FBS) and incubated for 16–18 h. After incubation, supernatants were collected. After centrifugation and filtration (0.22 μm filters), supernatants were directly used or kept in −80 °C.

Disrupt spleen in recommended buffer and pass through 70 μm nylon cell strainer (Corning) and followed red blood cell lysis. Prepared single cell suspensions were determined cell number, and used for following cell isolation. CD4⁺ T cells and CD8⁺ T cells were isolated from naïve C57BL/6 mouse spleen via Mouse CD4⁺ T Cell Isolation Kit (Miltenyi) or Mouse CD8⁺ T Cell Isolation Kit (Miltenyi), following the manufacturer’s instructions. MDSCs were isolated from LLC tumor-bearing C57BL/6 mouse spleen via Myeloid-Derived Suppressor Cell Isolation Kit (Miltenyi), following the manufacturer’s instructions.

LLC cells were subcutaneously injected into Ffar2^fl/fl and Ffar2^fl/flLyz2-cre mice (1 × 10⁶ cells/mouse), and LLC tumors were excised at day 21. LLC tumors were minced into small pieces (less than 3 mm in diameter) and resuspended with RPMI 1640, 400 U/mL Collagenase IV (Gibco) and 30 U/mL DNase I (Gibco). Tumor small pieces were incubated at 37 °C for half an hour. Stopping tumor samples digestion used complete medium (RPMI 1640 + 10% FBS), and tumor samples were filtered through 70 μm nylon cell strainer (Corning). After red cell lysis, and CD8⁺ T cells isolated from generated single-cell suspensions via Mouse CD8a Positive Selection Kit II (STEMCELL Technologies) following the manufacturer's instructions.

For tumor infiltrating immune leukocytes, tumor single-cell suspensions created as described (CD8⁺ T cells isolation from LLC tumors). For spleen infiltrating immune leukocytes, disrupt spleen in PBS supplemented with 2% FBS and pass through a 70 μm nylon cell strainer (Corning) and followed red blood cell lysis. All samples were blocked FcγII/III with anti-CD16/32 (BD Pharmingen) at 4 °C for 30 min, and surface marker was stained at 4 °C for 30 min. Samples were then stained with indicated fluorescence-conjugated antibodies. Fixable Viability Stain 780 (BD Pharmingen) was used to gate out non-viable cells. For intracellular Arg1 and IFNγ staining, Cytofix/Cytoperm Soln Kit (BD Pharmingen) was used to fix and permeabilize cells, following the manufacturer’s instructions. All samples were run on LSRFortessa (BD Pharmingen) and analyzed by FlowJo software (Tree Star).

CD4⁺ T cells, CD8⁺ T cells isolation from naïve mice spleen and MDSCs isolation from tumor-bearing mice spleen described as above. MDSCs were plated in the 48-well plates and cocultured with 1 μM carboxyfluorescein succinimidyl ester (CFSE) labeled CD4⁺ or CD8⁺ T cells at different ratios in the complete medium (RPMI 1640 supplemented with 10% FBS). For T cells activation, T Cell Activation/Expansion beads (Miltenyi) was added to coculture of T cells and MDSCs. After 72 h, T cells proliferation and IFNγ expression were measured by flow cytometry.

For urethane-induced lung cancer, Ffar2^+/+, Ffar2^−/−, Ffar2^fl/fl and Ffar2^fl/flLyz2-cre mice were intraperitoneally injected with urethane (1 g/kg body weight in 200 μl PBS) once per week for 10 weeks, lung tissues were excised and collected at 28 weeks. Then, the lungs were soaked in 4% paraformaldehyde in a fixed shape for 2 weeks. After that, lung nodules were quantified and photographed. For mice subcutaneous tumor model, LLC, LLC-luci or B16F10 cells were injected subcutaneously into Ffar2^+/+, Ffar2^−/−, Ffar2^fl/fl and Ffar2^fl/flLyz2-cre mice (1 × 10⁶ cells/mouse), and tumor volume assessed using calipers and calculated using the formula [(small diameter)² × (large diameter) × 0.5]. For MDSCs deletion in vivo, Ffar2^+/+ or Ffar2^−/− mice were injected subcutaneously with LLC cells (1 × 10⁶ cells/mouse) and injected intraperitoneally with isotype or anti-Gr-1 antibody (200 μg/mouse, every 4 days) from day 4 to day 23. For bone marrow chimeras, bone-marrow cells after red blood cell lysis were collected from Ffar2^+/+ or Ffar2^−/− mice. Prepared Ffar2^+/+ or Ffar2^−/− mice were lethally irradiated with 8.5 Gy, and lethally irradiated mice received bone marrow transplants from Ffar2^+/+ or Ffar2^−/− mice. Ten weeks after transplantation, chimeric mice were subcutaneously injected with LLC (1 × 10⁶ cells/mouse), and tumor growth was recorded. For the combined treatment with FFAR2 inhibitor and anti-PD-1 antibody, WT mice were injected subcutaneously with LLC (1 × 10⁶ cells/mouse). LLC-tumor bearing mice treated with FFAR2 inhibitor (5 mg/kg per day, single esophageal gavage), anti-PD1 antibody (200 μg/mouse every 4 days), FFAR2 inhibitor + anti-PD1 antibody or control (PBS containing 0.5% DMSO). Tumor-bearing mice were treated starting at day 4 post-tumor injection. Tumor growth and survival curve were recorded in two independent experiments. For survival analysis, mice were euthanized when total tumor burden approached IACUC guidelines with a tumor burden exceeding 1500 mm³ in volume.

MDSCs were isolated from tumor-bearing mice spleen via Myeloid-Derived Suppressor Cell Isolation Kit (Miltenyi) as described. WT mice were then injected with tumor cells (LLC or B16F10; 5 × 10⁵ cells/mouse) or co-injected with tumor cells and Ffar2^+/+ MDSCs (5 × 10⁵:5 × 10⁵) or tumor cells and Ffar2^−/− MDSCs (5 × 10⁵:5 × 10⁵). Tumor volumes were recorded.

Urethane-induced lung cancer tissues, LLC and B16F10 tumors were dissected, and washed twice by precooled PBS. Samples were fixed then in 4% paraformaldehyde for overnight, and embedded into paraffin. All paraffin embedded samples were sent to Servicebio. Hematoxylin and eosin (H&E) and immunofluorescence assay were stained and analyzed by Servicebio.

Generated bone marrow-derived MDSCs were first resting overnight, and then stimulated by GM-CSF (40 ng/ml) + IL6 (40 ng/ml) for the indicated time. After stimulation, BM-MDSCs were harvested and lysed with radio immunoprecipitation assay (RIPA) buffer (CoWin Biosciences, China, catalog# CW2333) supplemented with complete Mini Protease and Phosphatase inhibitor Cocktail (Roche, catalog# 4693159001 and 4906837001). Cell lysates were separated by standard SDS-PAGE and analyzed by immunoblotting.

Sample preparation of cell supernatants: generated bone marrow-derived MDSCs were first resting overnight, and then stimulated by GM-CSF (40 ng/ml) + IL6 (40 ng/ml) for 48 h. Remove particulates by centrifugation and assay immediately or store samples at -80℃. Preparation of LLC tumor tissue extracts: Mice were subcutaneously injected with LLC (1 × 10⁶ cells/mouse), and LLC tumors were excised at day 21. Add appropriate amount of PBS to the tissue and mash it. Centrifuge at 3000 rpm for 10 min to take the supernatants and kept then in -80℃. The concentration of L-Arginine from cell supernatants and LLC tumor tissue extracts were quantified using mouse L-Arginine (L-Arg) ELISA Kit (Shanghai Coibo Bio Technology Co.Ltd). Consumption of L-Arginine by MDSCs was calculated according to the formula: (concentration of L-Arginine in medium-concentration of L-Arginine in culture supernatants) × volume. TNF-α and IL-12 p70 in the culture supernatants and lysates of MDSCs were determined by ELISA following the manufacturer's instructions.

Human lung adenocarcinoma tissues and adjacent normal tissues were obtained from Huzhou Central Hospital, Affiliated Hospital of Zhejiang University (Huzhou, 313000, Zhejiang, China) and statements that informed written consent of lung adenocarcinoma patients were obtained. All tissues were collected and handled according to the ethical and safety procedures approved by the Clinical Ethics Committee of the Huzhou Central Hospital, Affiliated Hospital of Zhejiang University (reference ethics number 20180701-02). The lung cancer tissue array (HLugA180Su08) was purchased from Shanghai Outdo Biotech.

Ffar2^+/+ and Ffar2^−/− bone marrow-derived MDSCs (1 × 10⁶ cells/well) were seeded in a 6-well plate and cultured in complete RPMI 1640 medium overnight. After overnight resting, BM-MDSCs were re-stimulated by combination GM-CSF (40 ng/ml) with IL-6 (40 ng/ml) for 24 h. After that, RNA was extracted using the RNA extraction kit (Magen, R4801-02) and sequenced by BGI (Beijing Genomic Institute in ShenZhen). The sequencing data were filtered with SOAPnuke (v1.5.2) by removing reads containing sequencing adapters, removing reads whose low-quality base ratio (base quality less than or equal to 5) was more than 20%, and removing reads whose unknown base (“N” base) ratio was more than 5%. After this, the clean reads were obtained and stored in a FASTQ format. The clean reads were aligned to the reference genome using HISAT2 (v2.0.4). Fusion genes and differential splicing genes (DSGs) were detected through Ericscript (v0.5.5), and rMATS (v3.2.5). Bowtie2 (v2.2.5) was used to align the clean reads to the gene set, a database for this organism was built by BGI (Beijing Genomic Institute in ShenZhen), coding transcripts were included, and the expression levels of genes were calculated using RSEM (v1.2.12). Differential expression analysis was performed using DESeq2 (v1.4.5) with a P value ≤ 0.05 and |log₂ (fold change)|≥ 0.5. Volcano map was plotted by using the EnhancedVolcano package in R (v4.5.0). GSEA of KEGG pathway was performed by clusterProfiler package and visualized by ggplot2 package and gseaplot2 package. P value ≤ 0.05 was set as the cut-off criteria.

Total RNA of cells or tissues was isolated with TRIzol (TaKaRa), and RNA concentration was measured by NanoDrop 2000 (Thermo Fisher Scientific). RNA was reverse transcribed into cDNA by PrimeScript RT Master Mix (TaKaRa). Quantitative real-time RT-PCR (RT-qPCR) was performed using the QuantStudio 3 Real Time PCR System (Applied Biosystems). The expression of each gene was normalized to the expression level of GAPDH and reported as relative mRNA expression (2^−ΔΔCt) or fold change. The sequence-specific primers are shown in Additional file 2: Table S1.

Statistical analyses were analyzed by Prism 6.0 (GraphPad Software). Statistical differences between two groups were analyzed using unpaired two-tailed Student’s t-test. The statistical differences for more than two groups were analyzed using ANOVA. Survival analysis was performed using the Log-rank (Mantel-Cox) test. All data were shown as mean ± SEM. P value ≤ 0.05 was considered to be statistically significant. (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, NS, not significant).

To assess the aberrant SCFAs accumulation in tumor microenvironment, we employed gas chromatography-mass spectrometry (GC–MS)-based targeted metabolomics. Then acetic acids were predominantly accumulated in lung adenocarcinoma patients -derived tumor tissues (Fig. 1A). Similarly, acetic acids were also significantly increased in urethane-induced mice lung cancer tissues (Fig. 1B–D). Accordingly, both human (SK-MEL-2, PC-9 and A549) and mouse (B16F10 and LLC) tumor cells produced much more SCFAs (especially acetic acids) than control human HUVEC, BEAS-2B and NCM460 or mouse 3T3 cells (Fig. 1E, F, and Additional file 1: Fig. S1A and B). However, the serum acetic acids were little changed in tumor-bearing mice (Additional file 1: Fig. S1C), suggesting the specific accumulation of acetic acids in tumor microenvironment. Accordingly, elevated expression of ACLY and FASN (enzymes involved in fatty acid synthesis) was associated with reduced OS (overall survival) (Additional file 1: Fig. S1D and E). On the contrary, the enzymes (ACSS1 and ACSS2) involved in fatty acids degradation are extremely associated with increased OS (Additional file 1: Fig. S1F and G). Furthermore, intraperitoneal (i.p.) injections of NaAc (alternative mimetics of acetic acids) significantly increased tumor volume and weight in both LLC (Fig. 1G–I) and B16F10 (Fig. 1J) tumor model. These results suggest that metabolic reprogramming of cancer cells results in the accumulation of acetic acids in tumor microenvironment which facilitate the tumor progression.

Although both FFAR2 and FFAR3 are receptors for SCFAs, only FFAR2 is highly expressed in tumor tissues (Fig. 2A, B) and positively correlated with poor prognosis of lung adenocarcinoma patients (Fig. 2C). In addition, the mRNA expression of FFAR2 was greatly increased in urethane-induced mice lung cancer tissues as compared to normal lung tissues (Fig. 2D). Furthermore, the number and size of urethane-induced tumor nodules (Fig. 2E), as well as solid adenoma lesions (Fig. 2F) were all significantly reduced in Ffar2^−/− mice. In addition, FFAR2 deficiency markedly restrained tumor growth (Fig. 2G, H) and prolonged the survival of syngeneic LLC (Fig. 2I) mouse model. Moreover, esophageal gavage of FFAR2 inhibitor GLPG0974 [19, 20] and intraperitoneal injection of another FFAR2 inhibitor CATPB significantly restrained the growth of LLC tumors in Ffar2^+/+ mice, but not Ffar2^−/− mice (Fig. 2J, K, and Additional file 1: Fig. S2A). Whereas, the tumor growth of LLC (Fig. 2L) and B16F10 (Fig. 2M) could only be notably accelerated by NaAc in Ffar2^+/+ mice. Meanwhile, NaAc and FFAR2 inhibitors could not influence the proliferation and survival of LLC and B16F10 tumor cells in vitro (Additional file 1: Fig. S2B–E). Collectively, these data demonstrate the predominant role of FFAR2 in acetic acids promoted tumor progression.

Tumor-infiltrating myeloid cells (TIMs) including MDSCs, TAMs and tumor-associated DCs are the most abundant immune-related cells in the TME, and TIMs play a key role in modulation of immunosuppressive tumor microenvironment [21]. Therefore, we investigated whether the effects of FFAR2 on tumor growth were through MDSCs, TAMs and DCs (Additional file 1: Fig. S3A). As shown in Fig. 3A–C, FFAR2 deficiency significantly attenuated Gr-1⁺ MDSC infiltration in urethane-induced mice lung tumor tissues, LLC and B16F10 tumor tissues. However, the other TIMs including CD11C⁺MHCII⁺-DCs and CD11b⁺Gr-1⁻F4/80⁺-TAMs were little changed (Fig. 3D and Additional file 1: Fig. S4A). Furthermore, CD11b⁺Gr-1^high-PMN-MDSCs, CD11b⁺Ly-6G⁺-PMN-MDSCs and CD11b⁺Ly-6C^high-M-MDSCs in the blood and spleens of Ffar2^−/− LLC tumor-bearing mice were also decreased significantly (Fig. 3E, F). And FFAR2 is also highly expressed in human lung adenocarcinoma tumor tissues (Fig. 3G) and tumor-bearing mice splenic MDSCs (Additional file 1: Fig. S4B). Both flow cytometry and immunofluorescent staining showed that the infiltration of CD4⁺ and CD8⁺ T cells was increased significantly in Ffar2^−/− LLC, B16F10 tumors and urethane-induced lung tumors (Additional file 1: Fig. S4C–E), while the infiltration of NK and Tregs was little changed (Additional file 1: Fig. S4F and G). Taken together, FFAR2 deletion reduces MDSCs accumulation, but increases CD4⁺ and CD8⁺ T cell infiltration in tumors significantly.

Especially, the growth of LLC tumors was markedly restrained in mice reconstituted with Ffar2^−/− bone marrow cells as compared to that reconstituted with Ffar2^+/+ bone marrow cells (Fig. 4A). FFAR2 deficiency restricted tumor growth was eliminated by depletion of MDSCs (Fig. 4B) and the infiltration of MDSCs, CD4, CD8, DCs, NKs and macrophages in LLC tumor-bearing mouse upon isotype and anti-Gr-1 antibody treatment was shown in Additional file 1: Fig. S5A–F, implying the dominant role of FFAR2 in the protumoral phenotypes of MDSCs. Moreover, the tumor growth was little changed between co-injection of LLC with wild-type and Ffar2^−/− macrophages (Additional file 1: Fig. S6A). Additionally, co-injection of Ffar2^−/− MDSCs markedly delayed tumor growth than Ffar2^+/+ MDSCs both in LLC (Fig. 4C) and B16F10 (Additional file 1: Fig. S6B and C) tumor models. Furthermore, the number and size of urethane-induced nodules were dramatically decreased in the Ffar2^fl/flLyz2-cre mice which FFAR2 was specifically depleted in myeloid cells (Fig. 4D, E). Moreover, the subcutaneous LLC and B16F10 tumor growth was also restrained in Ffar2^fl/flLyz2-cre mice (Fig. 4F and Additional file 1: Fig. S6D). Immunofluorescent staining and flow cytometry demonstrated that conditional knockout of Ffar2 in myeloid cells led to much lower MDSCs infiltration (Fig. 4G, H), but significantly enhanced CD4⁺ and CD8⁺ T cell infiltration (Fig. 4I–K and Additional file 1: Fig. S6E). These data suggested the great potential of FFAR2 as a specific target for MDSCs in reshaping immune suppressive TME.

Then, bone marrow derived MDSCs (BM-MDSCs) were generated [4, 22] and activated using GM-CSF plus IL-6 with or without LLC tumor explant supernatants. We found FFAR2 deficiency did not affect the maturation, proliferation and apoptosis of BM-MDSCs (Additional file 1: Fig. S6F–H). However, key suppressive effectors including Arg1 and iNOS [5] were significantly reduced while pro-inflammatory factors including IL12-p40 and TNF-α were markedly elevated in Ffar2^−/− BM-MDSCs (Fig. 5A–C). Similar outcomes as well as reduced IL-10 expression were also observed in splenetic MDSCs of LLC tumor-bearing mice (Additional file 1: Fig. S6l–K). Furthermore, flow cytometry showed that percentage of Arg1⁺ cells were markedly reduced in LLC tumor derived-MDSCs (Fig. 5D) and urethane-induced lung cancer tissues of Ffar2^−/− mice (Additional file 1: Fig. S7A). Meanwhile, the proliferation (Fig. 5E, F, and Additional file 1: Fig. S7B and C) and IFNγ production (Fig. 5G, H, Additional file 1: Fig. S7D and E) was also enhanced significantly in Ffar2^−/− MDSCs treated T cells. As a consequence, the producing of IFNγ and TNF-α by CD8⁺ T cells from tumors was notably increased in Ffar2^fl/flLyz2-cre mice (Additional file 1: Fig. S7F). Collectively, these data suggest that FFAR2 deletion impairs the immunosuppressive activity of MDSCs on T cells in the tumor microenvironment.

After analyzing the transcriptome profile of BM-MDSCs by RNA sequencing (RNA-seq), 737 down-regulated 301 up-regulated genes were identified in Ffar2^−/− BM-MDSCs. In particular, key immunosuppressive genes Arg1 and NOS2 were significantly down-regulated in Ffar2^−/− BM-MDSCs (Fig. 6A). In addition, KEGG pathway gene set enrichment analysis revealed signaling pathways associated with calcium, PPAR and arginine metabolism were significantly changed in Ffar2^−/− BM-MDSCs, which may result in enhanced L-Arginine and T cell anti-tumor responses [23] (Fig. 6B). In order to explore the correlation between FFAR2, PPAR-γ and Arg1, we treated the MDSCs with indicated agonists or inhibitors. As shown in Fig. 6C, D, and Additional file 1: Fig. S8A, both the FFAR2 ligand (NaAc) and agonist upregulated Arg1 expression, while PPAR-γ-inhibitor (GW9662), Gαq-inhibitor (YM-254890) and Ca²⁺-inhibitor (2-APB) markedly suppressed the increased Arg1 expression, indicating that FFAR2 promoted Arg1 expression depended on Gαq/Calcium/PPAR-γ signaling pathway. Accordingly, the accumulation of L-Arginine in the tumor from Ffar2^fl/flLyz2-cre mice was enhanced significantly (Fig. 6E), which is consistent with the reduced consumption of L-Arginine in Ffar2^−/− BM-MDSCs (Additional file 1: Fig. S8B). Furthermore, only the expression of PPAR-γ and Arg1 but not other key transcription factors such as STAT1, STAT3 and C/EBPβ was dramatically reduced in Ffar2^−/− MDSCs (Fig. 6F and G, and Additional file 1: Fig. S8C and D) and urethane-induced lung cancer tissues (Additional file 1: Fig. S8E). The expression of PPAR-γ and Arg1 are increased significantly by FFAR2 ligand (NaAc) in the LLC tumors from Ffar2^+/+ mice. Whereas, the tumors from Ffar2^−/− mice were little influenced (Fig. 6H). Accordingly, the NaAc reduced CD8⁺ T cell infiltration was eliminated by PPAR-γ-inhibitor (GW9662) and L-Arginine, suggesting the predominant role of L-Arginine in T cell mediated antitumor immunity (Fig. 6I). In addition, much more FFAR2^highARG1^highCD15⁺-MDSCs were observed in the tumor of lung cancer patients with short survival compared to that with longer survival (Fig. 6J, K). Collectively, these data suggest the predominant role of Gαq/Calcium/PPAR-γ/Arg1 signaling as well as L-Arginine consumption in FFAR2 mediated immune suppression, which have great potential to be targets for cancer immunotherapy.

To evaluate the potential of FFAR2 inhibition in immune checkpoint therapy, we treated tumor bearing mice with FFAR2 inhibitor (GLPG0974) with or without anti-PD-1 antibody. Although the LLC tumor growth was restrained by FFAR2 inhibitor and anti-PD-1 antibody respectively, the combination therapy dramatically enhanced this inhibition on both tumor size and weight (Fig. 7A–C), and extended the survival of tumor-bearing mice (Fig. 7D). Then, we dissected LLC tumors to analyze tumor-infiltrating leukocytes (TILs) at day 30. Flow cytometry and immunofluorescent staining showed that the infiltrating CD8⁺ T cells was significantly increased in tumors from combination therapy group mice (Fig. 7E–I). Which is consistent with in vitro data, both FFAR2 inhibitor and combination therapy (FFAR2 inhibitor plus anti-PD-1 antibody) downregulated the expression of Arg1 in LLC tumors (Fig. 7J, K). Moreover, immunofluorescent staining showed that the infiltrating Gr-1⁺-MDSC was significantly decreased in tumor from FFAR2 inhibitor or combination therapy group (Additional file 1: Fig. S9A–C). Above data suggest that FFAR2 deletion or pharmaceutical inhibition could be a new strategy to overcome resistance to PD-1/PD-L1 blockade in lung cancer immunotherapy.