MAGE-C3 promotes cancer metastasis by inducing epithelial-mesenchymal transition and immunosuppression in esophageal squamous cell carcinoma

2.1 Cell culture

The ESCC cell lines KYSE30, KYSE410, and KYSE140 were gifts from Yoshikazu Shimada of Kyoto University (Sakyo-ku, Kyoto-shi, Kyoto, JAPAN) and cultured in RPMI 1640 medium with 10% fetal bovine serum (FBS) (Gibco, Thermo Fisher, Waltham, MA, USA). The Jurkat cell line was bought from national infrastructure of cell line resource and cultured in RPMI 1640 medium with 10% FBS. The B16f10 cell line was bought from national infrastructure of cell line resource and cultured in Dulbecco's modified eagle medium (DMEM) with 10% FBS. These were authenticated using short tandem repeat profiling. The stable overexpression cell lines KYSE30, KYSE140, and KYSE410 were cultured in RPMI 1640 with 10% FBS and 0.5μg/mL puromycin (Sigma, St Louis, MO, USA). The cells were culture condition was maintained at 37°C with 5% CO₂. For IFN-γ treatment, cells were incubated with 50ng/mL IFN-γ (Sigma,) for 24 h. For WP1066 treatment (S2796, Selleck Chemicals, Houston, Texas, USA), cells were incubated with 5μM WP1066 for 20 h. For Ruxolitinib treatment (HY-50856, MedChemExpress, NJ, Princeton, USA) cells were incubated with 2μM ruxolitinib for 20 h.

2.2 Plasmid construction and infection

2.3 Immunohistochemical analysis

A tissue microarray constructed using tumor tissues from 93 patients with primary ESCC diagnosed between January 2019 and December 2010 was purchased from Shanghai Outdo Biotech Co. Ltd (Shanghai, China). Of the 93 patients, 87 had matched adjacent normal tissues (77 males and 16 females at the age from 49 to 85, median age 65.8±8.7 years). All patients were not treated with chemotherapy or radiotherapy before surgery. In parallel, we collected demographic data, tumor pathological grades, clinical stages, survival data, and outcomes, including local invasion, LNM, and distant metastasis from the medical records. The follow-up period was 78 months or until death. Informed consent from the patients and ethics approval from the Institutional Research Ethics Committee was obtained.

For immunohistochemistry, tissue samples were deparaffinized in xylene and rehydrated in graded ethanol. Samples were immersed in ethylenediaminetetraacetic acid (EDTA) antigenic retrieval solution (pH8.0) and heated for antigenic retrieval for 15 min. The sections were incubated in normal goat serum blocking buffer at room temperature (25°C) for 1h. The sections were then incubated with a 1:100 dilution of anti-MAGE-C3 primary antibody (1:200; PA5-62546; Invitrogen, Thermo Fisher, Waltham, MA, USA,) or anti-CD8 alpha (1:100; ab217344; Abcam, Cambridge, UK,) at 4°C overnight, and then incubated with peroxidase (HRP)-conjugated secondary antibody (DS-0004, ZSGB-BIO, Beijing, China) for 1 h at room temperature. Cell nuclei were stained with hematoxylin staining solution (ZLI-9610, ZSGB-BIO). Immunostaining intensity of MAGE-C3 was evaluated, -: none, +: light yellow, ++: yellow brown, +++: brown. All patients were divided into two groups as follows: low expression group, -, +; high expression group, ++, +++.

2.4 Transwell migration/invasion assays

For the migration assay, 2 × 10⁵ cells suspended in 100 μL of serum-free medium were added to the upper chamber of the transwell cell culture chambers (3422, Corning, NY, USA). Then, 600 μL of medium with 20% FBS was added to the lower chamber and incubated for 10-20 h. Finally, the migrated cells on the lower surface of the membrane were fixed with methanol before staining with crystal violet (C0121, Beyotime, Beijing, China). In the invasion assay, 100 μL of medium containing 2% Matrigel matrix (354234, Corning) was pre-coated to the membrane of the upper chamber.

For cell migration and invasion detected via the xCELLigence DP system (Roche, Basel, Switzerland), we seeded 1 × 10⁴ cells in 100 μL of serum-free medium in for migration and invasion assays. The E16-plate was placed in an RTCA-DP device (Roche) at 37°C with 5% CO₂. Cell indexes were read every 15 min and the recorded curve was shown as a cell index.

2.5 Cell proliferation

Cell proliferation was detected using the xCELLigence MP system (Roche), which is a real-time cell growth analysis system. 50 μL of culture medium was added to each well of E-Plate 96 (Roche) to set the impedance baseline. Next, 1,000 cells in 150 μL of culture medium were seeded in an E-Plate 96 and was placed in an RTCA-MP device (Roche) at 37°C with 5% CO₂; and cell indexes were read every 15 min, and the recorded curve was shown as a cell index.

2.6 Animal experiments

All animal care procedures and experiments were conducted according to the animal welfare guidelines and based on the “3R” principle (reduction, replacement, refinement). Mouse experiments were approved by the ethical committee of Peking University Cancer Hospital & Institute (Beijing, China). Three independent experiments were conducted.

For xenograft studies, 5-week-old female BALB/c nude mice or C57/BL6 mice (Vital River Laboratories, Beijing, China) were used. Further, 1 × 10⁶ MAGE-C3 stably overexpressing KYSE30 cells or 5 × 10⁵ MAGE-C3 stably overexpressing B16F10 cells and control cells were injected subcutaneously into the bilateral dorsal flank, respectively. The mice were euthanized after 1 month, tumor weights were measured, and tumor volumes were calculated using the formula: π/6 × larger diameter × smaller diameter square.

For metastasis assay, 2 × 10⁵ cells were injected into 5-week-old female BALB/c nude mice or C57/BL6 mice via the tail the vein. The mice were euthanized after 1 month, and the lungs were enucleated and paraffin-embedded. Tissue sections were collected and stained with hematoxylin and eosin.

2.7 Western blotting and immunoprecipitation assays

Cells were collected and lysed in RIPA buffer with a protease inhibitor cocktail (04693132001, Roche) for 40 min on ice. The supernatant was collected after centrifugation at 12,000 × g for 15 min. Approximately, 30 μg of protein was resolved by SDS-PAGE and transferred to a ployvinylidene difluoride (PVDF) membrane (1620177, Bio-rad, Hercules, USA). The membrane was incubated with primary antibodies at 4°C overnight and then incubated with secondary antibodies for 1h at room temperature. The following antibodies were used: anti-PD-L1 (1:500; ab21093; Abcam), anti-STAT1 (phospho Y701) (1:1000; ab29045; Abcam), anti-STAT1 (phospho S727) (1:1000; ab109461; Abcam,, anti-MAGE-C3 (1:500; PA5-68045; Thermo Fisher), anti-IFNGR2 (1:1000; 10266-1-AP; Proteintech), anti-Vimentin (1:1000; 10366-1-AP; Proteintech), anti-interferon regulatory factor 1 (1:1000; 11335-1-AP; Proteintech), anti-Signal transducer and activator Of transcription 3 (1:1000; 9139; Cell Signaling Technology, Boston, MA, USA), anti-phospho-STAT3 (Tyr705) (1:500; 9145; Cell Signaling Technology), anti-phospho-Jak2 (Tyr1007/1008) (1:500; 3771; Cell Signaling Technology), anti-N-Cadherin (1:500; 13116; Cell Signaling Technology), anti-flag (1:500; F3165; Sigma), anti-rabbit IgG, HRP-linked antibody (1:5000; 7074; Cell Signaling Technology), anti-mouse IgG, HRP-linked Antibody (1:5000; 7076; Cell Signaling Technology).

For immunoprecipitation, cells were lysed RIPA buffer with a protease inhibitor cocktail (04693132001, Roche) for 4 × 10 min on ice. The 2 mg cell lysate was incubated with 2μg primary antibodies and protein A-agarose magnetic beads (HY-K0202, MedChemExpress). Immunoprecipitates were washed 5 times with lysis buffer and collected using magnetic stand. The immunoprecipitates were degenerated and loaded into each well of a 10-15% SDS-PAGE, and then transferred onto a PVDF membrane. Finally, the membrane was incubated with anti-STAT1 (1:1000; 11335-1-AP; Proteintech) or anti-MAGE-C3 (1:1000; 11335-1-AP; Proteintech) overnight at 4°C and then incubated with secondary antibodies anti-rabbit IgG, HRP-linked antibody (1:5000;7074; Cell Signaling Technology), anti-mouse IgG, HRP-linked antibody (1:5000; 7076; Cell Signaling Technology). The target bands would be exposed by Amersham Imager 600 (GE, Boston, MA, USA).

2.8 T-lymphocyte-mediated tumor killing assay

Peripheral blood mononuclear cells (PBMCs) from 10 healthy donors in Peking Cancer Hospital were isolated using lymphocyte separation medium (Tbdscience, Tianjin, China). PBMCs were activated with Dynabeads Human T-Activator CD3/CD8 (Gibco, Thermo Fisher) for 3 days. The tumor cells were pre-stained with 1 μM 5,6- carboxyfluorescein diacetate,succinimidyl ester (CFSE) (C0051, Beyotime) for 10 min at 37°C and then washed three times with PBS. The tumor cells were co-cultured with stimulated lymphocytes at ratio of tumor cells/T cells of 1:20 for 2 days in 24-well plates, then the cells were collected and labeled with 1 mg/mL propidium iodide (PI) (ST511, Beyotime) for 5 min at room temperature, then subjected to flow cytometry (BD Bioscience, East Rutherford, NJ, USA), CFSE⁺PI⁺ cells represented dead tumor cells that killed by T cells.

2.9 Flow cytometry analysis

The cells were digested by pancreatin (C125C1, NCM Biotech, Suzhou, Jiangsu, China,) and incubated with PD-L1-Alexa Fluor 647 (1:100; ab209960; Abcam, Cambridge, UK,) and the control were incubated with isotype (1:100; ab199093; Abcam) at 4°C for 30 min. After incubation, the cells were washed twice and suspended in phosphate-buffered saline (PBS). The cells were analyzed by Flow Cytometry (BD Bioscience).

The mouse lung tissues were digested with collagenase I (A004194, Sangon Biotech, Shanghai, China). Cells were stained with fluorescently labeled antibodies. Then cells were analyzed by flow cytometry (BD Bioscience), and the data were analyzed using the FlowJo10 software (BD Bioscienc). The following antibodies were used: anti-CD45 PE/Cy7-conjugated (1;100;103113;BioLegend, San Diego, CA, USA), anti-CD3ε PerCP/Cy5.5-conjugated (1:100;100327; BioLegend), anti-PD-1-APC-conjugated (1:100; 135209; BioLegend).

2.10 Cytokine measurements

Cytokine in supernatants were measured using commercial ELISA kits from NeoBioscience (Shenzhen, Guangdong, China): human interleukin-10 ELISA Kit (EHC009H), human tumor necrosis factor -α ELISA Kit (EHC103a.96), human interleukin-1β ELISA Kit (EHC002b), and human interferon-γ ELISA Kit (EHC102g.96). Briefly, the cell culture supernatant was centrifugated for 15 min at 1,000 × g to remove particulates. Further, 100 μL of sample was added per well and incubated for 2 h at 37°C to remove the liquid. Add Biotin-antibody was added into the well and incubated for 1 h at 37°C. Then aspirate each well was aspirate and washed with wash buffer from the kit for three times. HRP-acidin was added to each well and incubated for 1 h at 37°C. The aspiration was repeated for five times. Furthermore, 3,3',5,5'-Tetramethylbenzidine (TMB) substrate was added to each well and incubated for 30 min at 37°C. Finally, add the stop solution was added and determine the optical density at 450 nm was determined using by infinite M200PRO (TECAN, Thermo Fisher).

2.11 GAS luciferase reporter assay

GAS luciferase reporter plasmids (336841, Qiagen, Dusseldorf, German) and MAGE-C3-flag plasmid were co-transfected into cells using Lipofectamine 2000 (11668019, Thermo Fisher) according to the manufacturer's instruction. Luciferase activity was measured 48 h after transfection using the Dual-Luciferase Reporter Assay System 10-Pack (N1610, Promega, Madison, WI, USA). Briefly, the growth medium of cultured cells was removed and the cells were rinsed in PBS. Passive lysis buffer was then added, and the culture vessel was gently rocked for 15 min at room temperature. The lysate was centrifugated for 30 s at 3,000 × g to remove insoluble substances. The lysate was then transferred to a 96-well plate. Next, 100 μL of LAR II and Stop & Glo reagent from kit was added, and the samples were measured using for measurements by infinite M200PRO (TECAN, Thermo Fishe). Photinus luciferase activity was measured relative to Renilla luciferase activity was measured.

2.12 Nuclear and cytoplasmic extraction assay

Nuclear and cytoplasmic extraction assay was performed using NE-PER® Nuclear and Cytoplasmic Extraction Reagents (78835, Thermo Fisher) according to the manual. For 3 × 10⁶ cells (about 50 μL), 500 μL of ice-cold cytosol extraction reagent I from the kit was added in a tube and vortexed vigorously for 15 s to suspend the cell pellet. The tube was incubated on ice for 10 min. Next, ice-cold cytosol extraction reagent II from the kit was added to the tube and vortexed for 5 s and then incubated on ice for 1 min followed by vortexing for another 5 s and centrifugation for 5 min at 12,000 × g. After centrifugation, the supernatant (cytoplasmic extract) was transferred to a clean tube, and the insoluble fraction was suspended in ice-cold buffer nuclear extraction buffer and vortexed for 15 s. The sample was then placed on ice and vortexed for 15 s every 10 min for four times. Finally, the tube was centrifuged at 13,000 × g for 10 min, and the supernatant fraction (nuclear extract) was immediately transferred to a clean pre-chilled tube.

2.13 Quantitative real-time quantitative polymerase chain reaction (PCR) analysis

2.14 Immunofluorescence

Cells were seeded on glass bottom dished (150680, Thermo Fisher) for 24 h and fixed in 4% paraformaldehyde for 15 min. The cells were incubated with the primary antibodies anti-flag (1:200; F3165; Sigma) and anti-STAT1 (1:200;14994; cell signaling technology) overnight at 4°C, and then stained with followed by Alexa Fluor 488 and Alexa Fluor 594-secondary antibody (1:100; ZF-0511; ZF-0513; ZSGB-BIO). Nuclei were detected by DAPI (D9542; Sigam). Immunofluorescence was analyzed using a confocal laser scanning confocal microscope (Leica Microsystems Heiderg GmbH).

2.15 RNA-seq analysis

RNA extraction and RNA-seq analysis were performed by CapitalBio Technology (Beijing, China). A total amount of 2-μg RNA per sample was used for the RNA sample preparations. Sequencing libraries were generated using Collibri Stranded RNA Library Prep Kit for Illumina Systems (Thermo Fisher) following the manufacturer's instructions. The library was sequenced on the HiSeq 2500 sequencers (Illumina, San Diego, CA). Raw reads were filtered by NGSQC toolkit (version 2.3.3). Clean reads were mapped to the reference genome by HISAT (version 2.0) with default parameters. Gene expression values of the transcripts were analyzed by Cuffquant and Cuffnorm (version 2.2.0). Cuffdiff (version 2.2.0) was used to determine differentially expressed genes (DEGs) between two samples. Genes were considered as significantly differentially expressed if q value < =  0.05 and |fold change| > 1.0. Gene ontology analysis was performed using DAVID.

2.16 Statistics

All analyses from experiments were performed at least three times using SPSS 11.5 software (IBM Japan, Tokyo, Japan). Pearson chi-square test was used to analyze the relationship between MAGE-C3 expression and pathological grade, clinical stage, and LNM and Student t-test was used to compare the means. Kaplan-Meier method and log-rank test were used to analyze the overall survival (OS). The starting point for OS was the time that patient was diagnosed with ESCC. Significance was set at P < 0.05, and all tests were 2-tailed.

3.1 MAGE-C3 expression was associated correlates with prognosis and LNM in ESCC

In our previous study, we conducted whole-genome analysis in ESCC patients and found that MAGE-C3 exhibited an 8.0% mutation rate and 8.6% copy number amplification rate in ESCC patients [28], suggesting an important role of MAGE-C3 during tumor progression. To explore the role of MAGE-C3 in ESCC, in this study, a tissue microarray which containing 93 cancer tissue and 87 matched adjacent normal tissue sample was used for immunohistochemistry. We found that MAGE-C3 was highly expressed in 74.7% (65/87) of cancer tissues, whereas it was only expressed in 12.6% (11/87) of adjacent normal tissue tissues (P < 0.001; Figure 1A and Supplementary Table S1). Based on the expression levels of MAGE-C3 in cancer tissues, we divided all ESCC patients into high and low expression groups. Out of 93 cancer samples, 71 (76.3%) displayed high expression, whereas 22 (23.7%) showed low expression. We further analyzed the relationship between MAGE-C3 expression and clinical parameters (Supplementary Table S2). We did not observe a significant difference in MAGE-C3 expression based on sex (P = 0.53) or age (P = 0.311). Interestingly, 93.8% (45/48) of patients with LNM displayed high MAGE-C3 expression, whereas only 57.8% (26/45) of patients without LNM displayed high MAGE-C3 expression, and a positive correlation between MAGE-C3 expression and LNM was identified (P < 0.001; Figure 1B). Additionally, MAGE-C3 expression was significantly associated with pathological grade (P < 0.001; Figure 1C) and clinical stage (P < 0.001; Figure 1D). Therefore, we further examined the role of MAGE-C3 in ESCC prognosis. During the follow-up period (average: 31 months, range: 1-78 months), there were 60 cancer-related deaths, including 8 cases in the low expression group and 52 cases in the high expression group. ESCC patients with low MAGE-C3 expression showed longer OS whereas those with high MAGE-C3 expression revealed shorter OS, as determined by Kaplan-Meier analysis (P = 0.003; Figure 1E).

3.2 MAGE-C3 enhanced tumor cell mobility and promoted metastasis

A high MAGE-C3 expression was associated with LNM in ESCC patients; we investigated whether MAGE-C3 influenced the metastatic capacity of ESCC cells. We measured the endogenous MAGE-C3 in 9 ESCC cell lines (Supplementary Figure S1A). The KYSE410 and KYSE140, which display the lowest and highest MAGE-C3 level, were used to introduce MAGE-C3-flag plasmids and MAGE-C3-targeted siRNAs respectively. And the KYSE30 cells, which displayed medium MAGE-C3 level, were used for both overexpressing and knocking down (Supplementary Figure S1B). We found that MAGE-C3 overexpression in KYSE30 cells increased cellular migration and invasion compared with that of an empty vector via transwell assays (Figure 2A); these results were confirmed in KYSE410 cells using the xCELLigence Real-Time Cell Analyzer (Figure 2B). In contrast, MAGE-C3 knockdown attenuated the migration and invasion of KYSE30 (Figure 2C) and KYSE140 cells (Figure 2D). These data imply that MAGE-C3 promoted cell mobility in vitro.

Next, to determine the role of MAGE-C3 in metastasis in vivo, KYSE30 cells stably expressing MAGE-C3 (pLVX-MAGE-C3) or an empty vector (pLVX-nc) were established and injected into BALB/c nude mice intravenously. One month later, almost all the mice injected with MAGE-C3-overexpressing cells developed lung metastases, whereas only 30.0% (3/10) of control mice developed lung metastases (Figure 2E). The average number of metastatic tumor nodules in the MAGE-C3-overexpressing group was significantly higher than that in the control group (Figure 2E).

Because metastasis is closely associated with increased proliferation, we performed immunohistochemical with ki67 in metastatic nodules. The ki67 staining in metastatic nodules of MAGE-C3 overexpression group was nearly same as that in control group (Supplementary Figure S1C). Moreover, we showed that MAGE-C3 did not significantly alter the proliferation of ESCC cells in vitro (Supplementary Figure S1D-E). Similarly, overexpression of MAGE-C3 did not enhance tumor formation and tumor volume in the nude mouse xenograft formation (Supplementary Figure S1F). Collectively, these observations showed that MAGE-C3 promoted tumor metastasis, but was not involved in tumor cell proliferation in vitro and tumor formation in vivo.

3.3 MAGE-C3 enhanced tumor cell invasion and migration via STAT3-mediated EMT

Activation of EMT is important for driving cancer cell metastasis, which is usually accompanied by a spindle-like morphological change [29]. Compared with the control cells, we observed a more spindle-like phenotype in MAGE-C3 overexpressing cells (Figure 3A). Thus, we hypothesized that MAGE-C3 promotes ESCC cell invasion and migration via EMT progression. To test this hypothesis, we detected changes in the EMT-related markers, E-cadherin, N-cadherin, and Vimentin, in relation to MAGE-C3 levels. To reduce endogenous MAGE-C3 as much as possible, we stably knocked down MAGE-C3 in KYSE30 and KYSE140 cells using two combined lentiviral short hairpin RNAs (Shc3) and a negative control (Shnc). As shown in Figure. 3B, a reduction in E-cadherin level and upregulation of Vimentin and N-cadherin were observed in MAGE-C3 overexpression KYSE30 and KYSE410 cells, whereas E-cadherin was upregulated and vimentin and N-cadherin levels were reduced in MAGE-C3-knockdown KYSE30 and KYSE140 cells (Figure 3C). Several cell-intrinsic signaling pathways cooperate to induce the EMT, such as the TGF-β, Wnt, and Notch pathways, as well as mitogenic growth factor receptor-induced pathways [9]. To verify the mechanisms involved in MAGE-C3-regulated EMT, we analyzed some important transcription factors that induce EMT, such as STAT3, Nuclear Factor Kappa Beta (NF-κB), and Smad2/3 [9]. We found that STAT3 was activated in MAGE-C3 overexpressing KYSE30 and KYSE410 cells with increased pSTAT3^Tyr705 levels (Figure 3D), whereas NF-κB and Smad2/3 levels remained nearly unchanged (Supplementary Figure S2). Hence, we hypothesized that MAGE-C3-induced EMT occurs via STAT3. WP1066, a STAT3-specific inhibitor, was employed to confirm the role of STAT3 in MAGE-C3-mediated EMT. When MAGE-C3 overexpressing cells were treated with WP1066, the levels of pSTAT3^Tyr705, E-cadherin, and Vimentin were almost unchanged, whereas pSTAT3^Tyr705 and Vimentin levels were increased and E-cadherin levels were reduced in MAGE-C3 overexpressing cells without WP1066 treatment as we described above (Figure 3E). Meanwhile, the effects of MAGE-C3 on migration and invasion were attenuated after WP1066 treatment (Figure 3F). Taken together, our findings reveal that MAGE-C3-mediated promotion of tumor migration and invasion is partially dependent on STAT3.

3.4 MAGE-C3 repressed antitumor immunity responses and regulated cytokine secretion by T cells

Successful metastatic spread requires cellular mobility and immune escape, which suggests a potential link between tumor metastasis and immune resistance [11]. To examine whether MAGE-C3 assists tumor cell evasion from immune monitoring and destruction, we performed we performed lymphocyte-mediated cytotoxicity assays to visualize the function of MAGE-C3 in tumor immune suppression. After co-cultured with T cells, MAGE-C3 knockdown KYSE30 or KYSE140 cells exhibited a significantly higher percentage of PI⁺ CFSE-labeled tumor cells (Figures 4A and 4B), reflecting high T-cell-mediated tumor cell killing. We also measured the levels of T cell-related cytokines and found that T cells that were co-cultured with MAGE-C3 knockdown cells had increased TNF-α and IFN-γ secretion in the culture medium, while they exhibited decreased IL-1β and IL-10 release (Figures 4C and 4D). These findings demonstrate that MAGE-C3-mediated suppression of antitumor immune responses might affect tumor-reactive T cell functions in ESCC.

3.5 MAGE-C3 facilitated IFN-γ signaling

To investigate the mechanism of MAGE-C3-mediated resistance to T cell-induced cytotoxicity, we next employed RNA-seq in MAGE-C3 overexpressing KYSE30 cells and empty vector cells (Supplementary Figure S3A). Gene ontology (GO) enrichment analyses showed that the most altered pathways in MAGE-C3 overexpressing cells were IFN response pathways (Figure 5A). We also performed RNA-seq in MAGE-C3 knockdown KYSE30 cells and control cells and found that differentially expressed genes were also enriched in terms of canonical IFN-γ signaling (Supplementary Figure S3B and C).

Next, we treated KYSE30, KYSE140 and KYSE410 cells with various concentrations of IFN-γ and measured phosphorylated STAT1 (pSTAT1^Tyr701 and pSTAT1^Tyr727) levels to evaluate the activation of IFN-γ pathway. As shown in Supplementary Figure S4, KYSE30 and KYSE410 cells, but not KYSE140 cells, were sensitive to IFN-γ treatment. Notably, MAGE-C3 levels were unaltered in response to IFN-γ treatment. Consistently, enhanced IFN-γ signaling was confirmed in MAGE-C3 overexpressing cells, as indicated by increased levels of phosphorylated JAK2, pSTAT1^Tyr701 and IRF1 (Figure 5B). In contrast, the levels of phosphorylated JAK2, pSTAT1^Tyr701 and IRF1 were reduced in MAGE-C3 knockdown cells (Figure 5C). However, the pSTAT1^Tyr727 level remained unchanged. Subsequently, we performed nuclear and cytoplasmic extraction assays to determine the STAT1 subcellular distribution in MAGE-C3 overexpressing cells. Nuclear accumulation of STAT1 was increased in MAGE-C3 overexpressing KYSE30 and KYSE410 cells with or without IFN-γ treatment (Figure 5D). Immunofluorescence assays showed increased nuclear accumulation of STAT1 in MAGE-C3 overexpressing cells treated with IFN-γ (Supplementary Figure S5). Furthermore, we found that MAGE-C3 overexpression promoted IFN-γ-inducible gene transcription, as determined by a GAS luciferase reporter assay (Figure 5E). These results suggest that MAGE-C3 facilitates the activation of IFN-γ signaling.

PD-L1, which plays a crucial role in the immune escape, is a downstream element of IFN-γ signaling [30]. We found that CD274 (encoding PD-L1) was among the differentially expressed genes identified via RNA-seq (data not shown). Interestingly, the mRNA, protein, and membrane levels of PD-L1 were positively associated with MAGE-C3 expression (Figures 5G and 5H). These findings suggest that MAGE-C3 can activate IFN-γ signaling pathway and upregulate PD-L1 expression.

3.6 MAGE-C3 strengthened the interaction between IFNGR1 and STAT1 to activate IFN-γ signaling

To determine which component of the IFN-γ pathway is directly regulated by MAGE-C3, we treated MAGE-C3 overexpressing KYSE30 and KYS410 cells with the JAK1/2 inhibitor Ruxolitinib before adding IFN-γ. We found that pSTAT1^Tyr701 expression levels were unchanged following cell exposure to Ruxolitinib (Figure 6A and Supplementary Figure S6A). These data suggest that MAGE-C3 might act on the upstream elements of JAKs. We then used IFNGR1- and IFNGR2-targeted siRNAs to investigate whether IFN-γ receptors were involved. As shown in Figure 6B and Supplementary Figure S6B, the upregulation of pSTAT1^Tyr701 in MAGE-C3 overexpressing cells were attenuated when IFNGR1 levels were depleted in the presence of IFN-γ, whereas the function of MAGE-C3 was not affected after IFNGR2 knockdown (Figure 6C and Supplementary Figure S6C), suggesting that the effect of MAGE-C3 on STAT1 was mediated by IFNGR1. Next, we performed immunoprecipitation assays and found that MAGE-C3 was associated with IFNGR1 either in the presence or absence of IFN-γ stimulation (Figure 6D and Supplementary Figure S6D). Previous study suggested that IFNGR1 together with IFN-γ and STAT1 can translocate to the nucleus as a macromolecular complex, following IFN-γ activation [31]. Therefore, co-immunoprecipitation assays were used to test the interaction between IFNGR1 and STAT1 after overexpressing MAGE-C3. The results showed that MAGE-C3 greatly enhanced the interactions between IFNGR1 and STAT1 either in the presence or absence of IFN-γ (Figure 6E and Supplementary Figure S6E). Taken together, our findings demonstrate that MAGE-C3 binds to IFNGR1 and strengthens the interaction between IFNGR1 and STAT1 to activate IFN-γ signaling.

3.7 MAGE-C3 promoted tumor progression through PD-L1-involved immunosuppression in vivo

Melanoma is one of the most suitable models for studying immune therapy. Since the murine ESCC cell line was not available at present, we established stable MAGE-C3 overexpressing murine melanoma cells B16F10 to confirm the function of MAGE-C3 in vivo (Figure 7A). Consistent with our above results, MAGE-C3 overexpressing did not significantly alter cellular proliferation (Figure. 7A). Subsequently, the tumorigenicity of MAGE-C3 overexpressing B16F10 cells was examined in immunodeficient BALB/c nude mice and immunocompetent C57/BL6 mice. There was no significant effect of MAGE-C3 on tumor growth in nude mice after subcutaneous injection (Figure 7B). However, in C57/BL6 mice, tumor weight and volume dramatically increased in the MAGE-C3 overexpression group compared with those in the control group in C57/BL6 mice (Figure 7C). Mice bearing MAGE-C3 overexpressing tumors displayed shorter survival (Figure 7D). These results suggest that the integrated immune microenvironment is crucial for MAGE-C3 to promote tumor progression. Additionally, C57/BL6 mice injected with MAGE-C3 overexpressing cells developed severe lung metastasis and had much heavier lung weights than those of the control group (Figure 7E). Furthermore, immunohistochemistry and flow cytometry analysis showed that the infiltrated CD8⁺ T cell population was decreased in lung metastasis of MAGE-C3 overexpressing mice (Figure 7F). Because MAGE-C3 activated the IFN-γ pathway and upregulated PD-L1 levels, we further detected PD-1⁺CD8⁺ T cells among tumor-infiltrating CD8⁺ T cells in lung metastasis. Interestingly, the percentage of PD-1⁺CD8⁺ T cells was increased in MAGE-C3 overexpressing mice (Figure. 7G). Collectively, these findings suggest that MAGE-C3 promotes tumor progression through PD-L1-involved immunosuppression.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

The study protocol was approved by the ethics committee of Peking University Cancer Hospital & Institute (permit number: 2020KT132). The tissue samples were obtained with written informed consent from each patient. The animal study was carried out in compliance with the guidance suggestion of Animal Care Committee of Peking University Cancer Hospital & Institute (permit number: EAEC2020-16).

Human ESCC samples were provided by Shanghai National Engineering Research Center of Biochip. Written informed consent was obtained from all patients prior to the study. The use of the clinical specimens for research purposes was approved by the Institutional Research Ethics Committee of Peking University Cancer Hospital.

CONSENT FOR PUBLICATION

Not applicable.

AVAILABILITY OF DATA AND MATERIALS

Data of this study are available within the article and its supplementary information files.

COMPETING INTERESTS

The authors declare that they have no competing interests.

AUTHORS’ CONTRIBUTIONS

Conceptualization: Q.Z., Q.W., W.Z.

Experiments design: Q.W., W.Z.

Methodology: Q.W., Y.W.

Formal analysis and investigation: Q.W., W.Z. Q.M.

Writing original draft preparation: Q.W., W.Z.

Review and editing: Q.Z. Q.W.

Material preparation and data collection: Q.W., Y.W., H.Z., D.Z.

Supervision: Q.Z.

All authors read and approved the final manuscript.