MDSC suppresses T cell antitumor immunity in CAC via GPNMB in a MyD88‐dependent manner

Abstract Background Myeloid‐derived suppressor cells (MDSCs) played an essential role in tumor microenvironment to suppress host antitumor immunity and help cancer cells escape immune surveillance. However, the molecular mechanism behind tumor evasion mediated by MDSCs is not fully understood. Glycoprotein nonmetastatic melanoma protein B (GPNMB) is considered to associate with tumor initiation, metastasis and angiogenesis. Blocking GPNMB function is a potentially valuable therapy for cancer by eliminating GPNMB+MDSCs. Our previous study has proved that blockage the MyD88 signaling with the MyD88 inhibitor, TJ‐M2010‐5, may completely prevent the development of CAC in mice, accompanying with downregulation of GPNMB mRNA in the inhibitor‐treated mice of CAC. Methods We here focus on the underlying the relationship between GPNMB function and MyD88 signaling pathway activation in MDSCs' antitumor activity in CAC. Results CAC development in the mouse model is associated with expanded GPNMB+MDSCs by a MyD88‐dependent pathway. The GPNMB expression on MDSCs is associated with MyD88 signaling activation. The inhibitory effect of MDSCs on T cell proliferation, activation and antitumor cytotoxicity in CAC is mediated by GPNMB in a MyD8‐dependent manner. Conclusion MyD88 signaling pathway plays an essential role in GPNMB+MDSC‐mediated tumor immune escape during CAC development and is a promising focus for revealing the mechanisms of MDSC that facilitate immunosuppression and tumor progression.

Colorectal cancer remains the leading incidence and mortality rates around the world for decades.Inflammatory bowel disease is a crucial trigger for the occurrence of colorectal cancer. 1 Chronic inflammation of the colon causes intestinal cells damages and genetic mutations, and activates the intestinal immune system.The interaction between mutated cancer cells and host immune system in the colon, where is called tumor microenvironment (TME), promotes the development of colorectal cancer.Multiple molecules, mediators, and cells participate in the construction of TME, which contributes to immune escape, expansion, and metastasis of tumor cells. 2 Myeloid-derived suppressor cells (MDSCs) have been confirmed to be one of the key players engaged in TME to inhibit host antitumor immunity and accelerate tumor progression.Many studies have declared that MDSCs are one of potential therapy for cancer treatment.Currently, multiple treatment approaches targeting MDSCs have been examined in preclinical and clinical studies. 3In colitis-associated colorectal cancer (CAC), MDSCs massively expand and infiltrate the TME. 4 Decreasing the expanding MDSCs or attenuating their immunosuppressive function in host is an important therapeutic method to prevent the development of CAC.
The Toll-like receptor (TLR) pathway plays an important role during inflammation and cancer development and its activation through sensing of danger-associated molecular patterns (DAMP) released in TME helps host initiate anti-inflammatory and antitumor immune response. 5Myeloid differentiation factor 88 (MyD88) is a central adaptor-signaling molecule for most TLRs except TLR3.7][8][9] Blocking Myd88 signaling has been proved to induce antitumor effects by reversing the immunosuppressive effect of MDSCs. 10,11lycoprotein nonmetastatic melanoma protein B (GPNMB) is a glycosylated type I transmembrane protein.
It is expressed in a variety of cell types, such as melanoma, glioblastoma, dendritic cells (DCs), macrophages (M Φ ), and MDSCs in the immune system. 12,13There is growing evidence supporting that GPNMB is overexpressed among various cancers and it exhibits immunosuppressive function to promote tumor growth and metastasis. 14,15GPNMB is considered to be an applicable marker to identify MDSCs with suppressive activity 16 and blocking GPNMB function is a potentially valuable therapy for colorectal cancer by eliminating GPNMB + MDSCs. 17Our previous study has proved that the administration of TJ-M2010-5 (a patented MyD88 inhibitor, synthesized by Dr. Zhou's group) may completely prevent the development of CAC in mice, 18 accompanying with downregulation of GPNMB mRNA in the inhibitor-treated mice of CAC.We here focus on the underlying the relationship between GPNMB function and MyD88 signaling pathway activation in MDSCs' antitumor activity in CAC.

| Animals
Six-week old wild type (WT) and MyD88 −/− BALB/c female mice, SPF grade, were purchased from Gempharmatech Company (Nanjing, China).The experimental operation has been approved by the Animal Care and Research Committee of Huazhong University of Science and Technology.

| CAC model
The mouse model of CAC was induced as previously described. 18Briefly, each mouse was injected intraperitoneally (i.p.) with 10 mg/kg azoxymethane (AOM, Sigma-Aldrich Chemical, Germany).One week later, the mice were fed with 3 cycles of 2.5% dextran sodium sulfate (DSS, Meilunbio, China) solution for 1 week and 2 weeks of plain drinking water throughout a 10-week observation period.A group of sex-and age-matched normal control (NC) mice received plain drinking water.Samples [spleen, peripheral blood (PB), and bone marrow (BM)] were collected at 10 weeks (W10).

| MyD88 inhibitor (TJ-M2010-5) treatment
All WT mice were randomly divided into three groups: NC, CAC model (CAC), and MyD88 inhibitor-treated CAC (Inhibitor-CAC) groups.TJ-M2010-5 was i.p. administered to mice in the Inhibitor-CAC group daily at 50 mg/ kg body weight for 10 weeks beginning 2 days before the first DSS administration.

| Preparation of Splenocytes, bone marrow cells (BMCs), and peripheral blood cells (PBCs) suspensions
Splenocytes: spleens of mice were removed, diced in phosphate buffer saline (PBS) and pressed through a 70 μM strainer, and the harvested cells were washed with PBS and suspended.BMCs: femurs and tibia of mice were aseptically cut and debrided of surrounding skeletal muscle and other tissue, and BMCs were flushed from the femur and tibia using PBS and strained through a 100 μM strainer.PBCs: Blood was drawn from the inner canthus of mice.PBCs were suspended after the red blood cells were removed from single-cell suspensions with red blood cells lysis solution (Solarbio Comp., Beijing, China).

| Flow cytometry analysis
The following antibodies (Abs) were used for flow cytometry analysis: CD11b and CD4 from BD Biosciences (USA); Gr-1 from Biolegend Inc (USA); GPNMB, CD8a, and CD107a from ThermoFisher Scientific (USA).The cells were surface-labeled with anti-mouse Abs for 30 min at 4°C in the dark, and subsequently washed, fixed, and permeabilized by the Cytofix/Cytoperm Solution Kit (BD Biosciences, USA).The permeabilized cells were labeled by anti-mouse CD107a for 30 min at 4°C in the dark.The cells were read using a FACS Celesta flow cytometer, and data were analyzed by FlowJo software (Version 10.0).

| MDSC suppression assay on T cells proliferation and activation
CD11b + Gr-1 + MDSCs sorted from BMCs of mice and 5 × 10 5 CD4 + or CD8 + T cells sorted from autologous splenocytes were cocultured in 48-well plates in complete RPMI medium at ratios of 1:1.T cells activation was performed by means of anti-CD3/anti-CD28-coated microbeads (ThermoFisher Scientific, USA) activation for 4 days.The T cell proliferation was performed using carboxy-fluorescein diacetate succinimidyl ester (CFSE, ThermoFisher Scientific, USA) staining assay.Suppressor ability of MDSCs was indicated as percentage of suppression: 1 -b/a × 100%, where a is the proliferation rate of T cells only; and b is the proliferation rate of T cells in cocultures at 1:1 ratio.IFN-γ levels in the cultures were assessed using ELISA Kit (BOSTER Biological Technology Co., China) for activation assay.

T cells
Sorted splenic CD8 + T cells were cocultured with target tumor cells (CT26.WT, BOSTER Biological Technology Co., China, 3rd to 10th passage) labeled with CFSE (Biolegend Inc, USA) at 5:1 ratio w/o sorted CD11b + Gr-1 + MDSCs from BMCs of mice with CAC, in RPMI 1640 complete medium (Gibco, USA).After 72 h, CD8 + T cell cytotoxicity was measured by the expression of effector molecules, such as CD107a, on activated CD8 + T cells and mortality of target tumor cells by Zombie Aqua (Biolegend Inc, USA) staining by FACS.

BM-derived immature myeloid cells in vitro
CD11b − cells sorted from BMCs were cultured in the presence of 10 ng/mL of GM-CSF (Peprotech, USA) and 1 μg/mL LPS (Sigma-Aldrich Chemical, Germany) in RPMI 1640 complete medium supplemented with sodium-pyruvate (1 mM, Meilunbio, China) and βmercaptoethanol (50 μM, Solarbio, China) w/o 20 μM MyD88 inhibitor administration.At 8 days of cultivation, the population of CD11b + Gr-1 + MDSC was calculated to express the differentiation of MDSC from immature myeloid cells.GPNMB expression on MDSC were detected by flow cytometry, and the mRNA expression of GPNMB on induced MDSCs were detected by RT-qPCR.

| Statistical analyses
Comparisons of means ± standard deviation (SD) between groups were performed using Student's t-test and one-way analysis of variance with GraphPad Prism softwar (version 9.0; La Jolla, USA).p-value <0.05 were considered statistically significance.

GPNMB-expressing MDSCs on T cell proliferation and activation depends on the MyD88 pathway
It is reported that GPNMB blockade restores the T-cell response to antitumor by suppressing MDSC function. 17e here also detected the role of GPNMB pathway in the suppression effect of MDSCs on T cell function in the mice model of CAC.CD11b + Gr-1 + MDSCs sorted from BMCs of mice with CAC were cocultured with autologous CD4 + or CD8 + T cells.CD4 + and CD8 + T cell proliferation rates were analyzed by CFSE assay in Figure 2A, which showed significant suppressive effects on CD4 + T cell proliferation rate [16.0 ± 5.3% vs. 62.2 ± 1.1% in positive control (PC), p = 0.0000] and CD8 + T cell proliferation rate (13.0 ± 4.5% vs. 67.0± 11.5% in PC, p = 0.0003).T cell activation was measured by IFN-γ production in the cultures (Figure 2B), with significantly reduced concentration in the coculture groups (23.36 ± 2.64 pg/mL vs. 104.80± 15.32 pg/mL in PC of CD4 + T cells, p = 0.0000; 11.85 ± 2.00 pg/mL vs. 258.63± 33.05 pg/mL in PC of CD8 + T cells, p = 0.0000).The suppressor ability assessed by the percentages of suppression of CD4 + and CD8 + T cell proliferation rates.For CAC, GPNMB expression correlated positively with higher suppressor activity (R 2 = 0.7934/p = 0.0030 for CD4 + T cell proliferation and R 2 = 0.9265/p = 0.0001 for CD8 + T cell proliferation, Figure 2C).

| MyD88 inhibitor suppressed MDSCs differentiation and GPNMB expression in vitro
We have proved that blocking the MyD88 signaling pathway with the MyD88 inhibitor could significantly suppressed the differentiation of CD11b − BMC induced to CD11b + Gr1 + MDSCs in vitro. 11Here, we further confirmed that the suppression was related to the expression of GPNMB on MDSCs.As shown in Figure 4B, the expression of GPNMB on MDSCs induced differentiation was significantly decreased after MyD88 inhibitor treatment (25.7%), compared with 70.4% in no-inhibitor-treated group, companied by a decreased in CD11b + Gr1 + MDSC population (1.52% vs. 19.5% in the no-inhibitor-treated group, Figure 4A).The mRNA expression of GPNMB in induced MDSCs was also significantly declined after the MyD88 inhibitor administration (Figure 4C).Thus, the GPNMB expression on MDSCs is associated with MyD88 signaling pathway activation.

| DISCUSSION
TME is composed of malignant cells, various immuneinfiltrating cells (neutrophils, M Φ s, DCs, MDSCs, B and T lymphocytes, and natural killer cells), fibroblasts, endothelial cells, as well as the surrounding stroma. 19teractions between TME and tumor cells could mostly influence tumor progression and metastasis. 20During tumorigenesis, there is a balance between protumor and antitumor immunity.In the process of tumor development, tumor cells evolve immunosuppressive mechanisms to escape immune surveillance, including expressing immunosuppressive molecules.One of these molecules is GPNMB whose role in enhancing immunosuppression has been increasingly confirmed in malignancy, 14,21 such as melanoma, 22 breast cancer, 23 glioma, 24 and gastric cancer, 25 which may offer an attractive target for cancer immunotherapy. 26An antibody anti-GPNMB (glembatumumab vedotin) 27 has been used in phase II clinical trials in the treatments of melanoma 28,29 and breast cancer. 30PNMB is a highly glycosylated type 1 transmembrane protein that was first described in 1995 in melanoma cell lines. 31It is widely expressed in various tissues, such as the long bones, calvaria, bone marrow, skeletal muscle, skin, heart, lung, liver, pancreas, and placenta, and its expression is increased in cancer cells. 27It is encoded by the GPNMB gene located on chromosome 7p15 12 .GPNMB protein contains an extracellular domain, a single-pass transmembrane domain, and a 53 amino acid cytoplasmic tail. 32It can be cleaved and release a soluble fragment that can bind to many partners and receptors, including integrins, heparin, Syndecan-4, CD44, Na + -K + -ATPase, epidermal growth factor receptor, tyrosine kinase receptors, vascular endothelial growth factor receptor, and others, 21,33 and trigger multiple cellular responses. 34uman GPNMB + MDSCs (CD14 + HLA − DR no/low ) and murine GPNMB + MDSCs (CD11b + Gr1 + ) both play a considerable role in tumorigenesis, where they have been described as the significant suppressors of T cell function. 14GPNMB + CD11b + Gr1 + cells were massively expanded in melanoma-bearing animals, and showed prominent suppressive effects on T cell activation both in vivo and in vitro.Administration of an anti-GPNMB monoclonal Ab (mAb) or completely GPNMB knockout on CD11b + Gr1 + cells, could abrogate CD11b + Gr1 + cells expansion, impact their suppressor function on T cells, and diminish melanoma development in vivo. 35n clinical study, CD14 + HLA − DR no/low cells from metastatic melanoma patients showed increased GPNMB expression.These CD14 + HLA − DR no/low cells isolated from melanoma patients significantly reduced the IFN-γ secretion of T cell in the cocultures of GPNMB + MDSCs and autologous activated T cells. 36Similarly, GPNMB + MDSCs isolated from CAC and pancreatic cancer patients exhibited significant suppressive effect on T cell function.However, administration of anti-GPNMB mAb in the MDSC and T cell cocultures inhibited MDSC activity and restored the IFN-γ production by T cells. 17ur results also dedicated that GPNMB-expressing  MDSCs are widely expanded in the mouse model of CAC (Figure 1), and those from mice of CAC have significant suppressive effects on T cells proliferation and IFN-γ production (Figure 2).Additionally, our findings expressed the suppressive effect of MDSC on antitumor cytotoxicity of CD8 + T cell (Figure 3) for the first time, and illustrated the positive correlation between the suppressor activity of MDSC and GPNMB expression level on it.Thus, here we provide valuable information that GPNMB regulation on MDSC may serve as a tumortargeting therapy for CAC.
There is limited research on the regulatory and effect molecular mechanisms of GPNMB expressing on MDSCs.It is reported that microphthalmia transcription factor (MITF) binds and trans-activates the Gpnmb promoter, and drives GPNMB expression. 37Through enhancing PI3K/AKT/mTOR pathway signaling and βcatenin activity, GPNMB promotes breast tumor initiation and growth mediated by Wnt-1 signaling. 38GPNMB mechanistically augments IL-33-mediated tumorigenesis by activating CD44 receptor. 39Binding of MDSC-derived GPNMB to SD-4 on activated T cells greatly blocked CD4 + and CD8 + T cell responses. 13TLR activation by agonists induces amplified GPNMB/SD-4 pathway, which is a direct result of NF-κB binding to the promoter and induction of de novo transcription. 40In our study, blockage of TLR/ MyD88 signaling pathway results in less of GPNMBexpressing MDSCs expanding in CAC (Figure 1), less of suppression effect of GPNMB-expressing MDSCs on T cell proliferation, activation (Figure 2), and antitumor cytotoxicity (Figure 3), and less of MDSCs differentiation and GPNMB expression in vitro (Figure 4).Our results firstly proved the relationship between TLR/MyD88 signaling pathway activation in innate immunity and MDSC's suppressive function on antitumor T-cell adoptive immunity.Therefore, our studies provide strong support for the suppressive effect of MDSC on T cell antitumor immunity in CAC via GPNMB in a MyD88-dependent manner.

| CONCLUSIONS
In summary, we have uncovered that the CAC's immune silence derive from GPNMB-expressing MDSC-mediated T cell suppression, and the suppressive effect of MDSCs is positively correlated with GPNMB expression level.TLR/ MyD88 signaling pathway is proved to play an essential role in MDSC-mediated immune escape in CAC development and is a promising focus for revealing the mechanisms of MDSC that facilitate immunosuppression and tumor progression.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations and guidelines issued by the Ethics Committee of Huazhong University of Science and Technology (Wuhan, China).Written informed consent for publication of this paper was obtained from the Tongji Hospital, Huazhong University of Science and Technology (No. TJH-201803001) and all authors.ORCID Lin Xie https://orcid.org/0000-0002-7058-0851

F
I G U R E 1 GPNMB + MDSCs are expanded in the mouse model of CAC by a MyD88-dependent pathway.(A) Population of CD11b + Gr-1 + MDSCs in the spleen (SPL), BM, and PB by flow cytometry 10 weeks after induction in the normal control (NC), the CAC group (CAC) and the MyD88 inhibitor-treated CAC (Inhibitor-CAC) groups.(B) Population of GPNMB + MDSCs (gated on CD11b + Gr-1 + cells) in the SPL, BM, and PB by flow cytometry 10 weeks after induction in the NC, CAC, and Inhibitor-CAC groups.(C) Relative levels of GPNMB mRNA transcripts in BMCs from mice with CAC examined by RT-qPCR.Data of RT-qPCR were normalized to β-Actin expression in each sample, and the relative expression levels were compared with that in NC group.**p < 0.01; ***p < 0.001.N = 5 per group.Data are expressed as the mean ± SD of each group.coculturedwith CD4 + T cells, p = 0.0004; 24.4 ± 1.9% vs. 30.5 ± 3.1% on CAC-MDSCs cocultured with CD8 + T cells, p = 0.0003, Figure2D).These results supported that GPNMB pathway mediated the suppression effect of MDSCs on T cell function in CAC and this process was dependent on the MyD88 pathway.

F I G U R E 2 3 . 3 |
Suppressive effect of GPNMB + MDSCs on T cell proliferation and IFN-γ secretion depends on the MyD88 pathway.CD11b + Gr-1 + MDSCs sorted from BM of mice were cocultured with CFSE-labeled autologous CD4 + or CD8 + T cells in proportion (1:1) in the present of anti-CD3/anti-CD28-antibody-coated microbeads.(A) Percentage of CFSE-labeled CD4 + T (upper panels) and CD8 + T (lower panels) cells after 72 h in culture, with no stimulation (negative control, Neg C), stimulation (positive control, Pos C) and cocultured MDSCs from mice with CAC (CAC-MDSC) or MyD88 inhibitor-treated mice with CAC (Inhibitor-CAC-MDSC). Data are expressed as the mean of each group.(B) Concentrations of IFN-γ in the cultures were measured by ELISA.Data are expressed as the mean ± SD of each group.(C) Correlation of percentage of GPNMB + cells among MDSCs from mice with CAC and their ability to suppress CD4 + or CD8 + T cell proliferation rates.Values (percentages) were plotted and analyzed for correlation coefficient R 2 .(D) Population of GPNMB + MDSCs (gated on CD11b + Gr-1 + cells) sorted from the NC, CAC, Inhibitor-CAC groups by flow cytometry cocultured with autologous CD4 + or CD8 + T cells.N = 8 per group.Data are expressed as the mean of each group.**p < 0.01; ***p < 0.001.Suppressive effect of

F I G U R E 3
Suppressive effect of GPNMB + MDSCs on antitumor cytotoxicity of CD8 + T cells depends on the MyD88 pathway.(A) Cytotoxicity of CT26.WT cells labeled by CFSE was detected by Zombie Aqua staining by flow cytometry cocultured with autologous CD8 + T cells w/o MDSCs.(B) Population of CD107a + cells (gated on CD8 + T cells) in the cocultures by flow cytometry.N = 5 per group.Data are expressed as the mean of each group.