A pan-cancer analysis of CMTM6 was conducted by the GEPIA2 (Tang et al., 2019). The CMTM6 mRNA expression level was determined in OV and their corresponding normal tissues in Oncomine, and the cutoff of p value and fold change was as following: p value:0.05; fold change:2; and gene rank:10% (Rhodes et al., 2007). Besides, we investigated the CMTM6 expression level based on The Cancer Genome Atlas (TCGA) database and the Genotype-Tissue Expression (GTEx) database in 427 tumor samples and 88 normal tissues by R package “ggplot2” visualization and log2 (TPM+1) for log-scale. CMTM6 expression in 55 OV cell lines was identified in the Cancer Cell Line Encyclopedia (CCLE) datasets (Barretina et al., 2012). Immunohistochemistry (IHC) images from the Human Protein Atlas (HPA) database were used to manifest the CMTM6 protein expressed level. Eventually, the UALCAN platform was queried to examine the correlation between CMTM6 gene expression and clinical features. These pre-defined subgroups include individual cancer stages, patient’s race, patient’s age, tumor grade, and TP53 mutation status. The difference was made by comparing median values.

Normal ovary cell line (IOSE80) and ovarian cancer cell lines (A2780, ES2, Hey, and SKOV3) obtained from the American Type Culture Collection (ATCC, Manassas, VA, United States) were cultured in Roswell Park Memorial Institute (RPMI)1,640 medium (HyClone, United States) supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin. All cell lines were maintained in a humidified incubator in an atmosphere of 5% CO₂ at 37°C.

Total RNA of cell lines was extracted using RNAiso Plus reagent (TaKaRa, Shanghai, China) and transcribed into cDNA with the reverse transcription kit (TaKaRa) according to the manufacturer’s protocol. Quantitative real-time PCR was performed using cDNA primers specific for mRNA. All the real-time PCR reactions were performed using the QuantStudio^TM Design and Analysis Software1.3.1 PCR System. The 2^−ΔΔCt method was used for quantification, and fold change for targeted genes was normalized by internal control. The PCR reaction conditions were as follows: 95°C for 30 s, followed by 95°C for 5 s, and 60°C for 30 s: 40 cycles and then 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s again. The expression levels were normalized against those of the internal reference gene GAPDH. The primer sequences are listed as follows: CMTM6 forward 5′-ATG​AAG​GCC​AGC​AGA​GAC​AG-3′, reverse 5′-GTG​TAC​AGC​CCC​ACT​ACG​GA-3′; GAPDH forward 5′-CGC​TCT​CTG​CTC​CTC​CTG​TTC-3′; reverse: 5′-ATC​CGT​TGA​CTC​CGA​CCT​TCA​C-3′.

The cells were lysed for 30 min with RIPA buffer and then centrifuged at 12,000 g for 30 min at 4°C to collect the supernatants. After mixing with SDS sample buffer, the proteins were heated to 99°C for 10 min, separated on 12.5% SDS-polyacrylamide gels, transferred to PVDF membranes, and probed with antibodies against CMTM6 (1:1,000, CST, United States) and actin (1:1,000, CST, United States) at 4°C overnight. A 1:2000 dilution of the HRP-linked anti-IgG (CST, United States) was used as secondary antibody, and blots were visualized by ECL (Millipore, United States), photographed using a cooled Tanon chemiluminescence gel imaging system (Thermo Fisher, United States), and band intensities were quantified using ImageJ software. The results were expressed as a ratio of band density to actin.

To uncover the prognostic value of the CMTM6 gene in OV patients, survival analysis, for instance overall survival (OS) and disease-free survival (DFS), was conducted. Relevant data of OV patients were downloaded from TCGA datasets, which were visualized with Kaplan–Meier curves. The analyses were accomplished through the R packages “survival” and “survminer” (Liu et al., 2018). The p-value was calculated by the log-rank test. A statistically significant variation was observed when the p-value was <0.05. The time-dependent receiver operating characteristic (ROC) curve was used to judge the predictive accuracy of the prognostic significance by using the pROC package (Vivian et al., 2017).

The Kaplan–Meier plotter database was further wielded to estimate the prognostic value between gene expression and survival in cancers (Lanczky et al., 2016). The relations between CMTM6 expression and survival in OV were validated by the PrognoScan database (Mizuno et al., 2009). Log-rank p-values and the hazard ratio (HR) with 95% confidence intervals (95% CIs) were calculated.

In some situations, GEPIA2 was also used to evaluate the prognostic value of CMTM6.

GSEA is a powerful analytical method to predict whether a gene set is enriched in a specific biological state. The clusterProfiler package in R was used for GSEA to reveal the critical biological process and relevant signaling pathways affected by CMTM6 expression. The cutoff criteria were p < 0.05 and normalized enrichment score (NES) > 1.

The top 100 coexpressed genes with CMTM6 were extracted from coexpression analysis datasets in the Oncomine database. Fold change≥2 and P-value≤0.001 were selected as thresholds.

The GeneMANIA database is a resource-rich website containing gene information, analyzes gene lists, and prioritizes genes for functional assays with a high accuracy of the prediction algorithm (Warde-Farley et al., 2010). A figure where nodes symbolize genes and links represent protein–protein network (PPI) of coexpressed genes was used to display interactions. Cytoscape (version 3.8.2), an open bioinformatics software platform, was used to visualize the molecular interaction network.

The Gene Ontology (GO) analysis comprising cellular component (CC), molecular function (MF), and biological process (BP), as well as the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of coexpressed genes were implemented via the R package clusterProfiler (3.14.3) (Yu et al., 2012). P < 0.05 was considered statistically significant for functional enrichment analysis.

TIMER2.0 (tumor immune estimation resource, version 2) is a comprehensive resource for systematical analysis of immune cell infiltrate across diverse cancer types (Li et al., 2020). Spearman’s analysis was executed to illuminate the correlation between CMTM6 expression and the degree of immune infiltrates by the “Gene” module, and log2 (TPM) was used for log-scale.

TISIDB, well known as a web portal for tumor and immune system interaction, integrates multiple heterogenous data types (Ru et al., 2019). Here, TISIDB was used to explore the links between CMTM6 and multiple immune regulatory factors.

Meanwhile, the relationships between the expression of CMTM6 and gene markers of immune cells by the “correlation analysis” module of GEPIA2 were analyzed.

UCSC Xena is available to inspect gene expression, copy number, methylation, and somatic mutation in TCGA OV patient cohort (Goldman et al., 2020).

The genomic alteration types, alteration frequency, and CMTM6 mRNA level vs. methylation in OV were analyzed through the “OncoPrint” and “Plots” module, and the survival situation of the altered group or not were analyzed by the “Comparison/Survival” module in cBioPortal (Gao et al., 2013).

Partial statistical analyses were performed by default by website recourse. The Wilcoxon rank sum test was adopted to compare the CMTM6 expression levels in tumor tissues and normal tissues based on TCGA and GTEx databases. All R packages were chosen by R software version 3.6.3. The box plots were generated with GraphPad Prism 9 software, data are presented as mean ± standard deviation, and one-way analysis of variance (ANOVA) was used to determine the differences of CMTM6 expression in cell lines. All cell experiments had been repeated at least three times. Only P value < 0.05, adjusted p < 0.05, and log-rank p value < 0.05 were thought to be statistical significant (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).

According to the GEPIA2 database query, we found that CMTM6 expression was upregulated in COAD (colon adenocarcinoma), GBM (glioblastoma multiforme), LAML (acute myeloid leukemia), LGG (brain lower grade glioma), OV (ovarian serous cystadenocarcinoma), PAAD (pancreatic adenocarcinoma), READ (rectum adenocarcinoma), STAD (stomach adenocarcinoma), THCA (thyroid carcinoma), and UCEC (uterine corpus endometrial carcinoma), especially in OV, compared with the corresponding normal tissues (Figure 1A). Consistently, higher mRNA expression of CMTM6 by the Oncomine database was observed in OV patients than that in normal samples (Figures 1B–F). The detailed results of CMTM6 expression in the Oncomine database are summarized in Table 1. Meanwhile, CMTM6 was also apparently increased when combining TCGA (427 tumor tissues) with GTEx (88 non-tumor tissues) (Figure 1G). In the end, representative IHC images from the THPA database further investigate that the CMTM6 protein level was obviously elevated in OV tissue compared with normal ovary tissue (Figure 1H). Therefore, we chose CMTM6 for further exploration.

Then, we dug the CMTM6 expression among groups of patients in line with different clinical parameters by using the UALCAN online tool, such as individual cancer stages, race, age, tumor grade, and TP53-mutation status. In terms of individual cancer stages, without increasing trend of CMTM6 level was witnessed as the aggravation of stages (Supplementary Figure S1A). Regrettably, regarding tumor grade, we could not know the effect of CMTM6 expression on tumor progression due to lack of clinical patient data for grade I and IV (Supplementary Figure S1B). Additionally, there were no differences in race, age groups, or TP53-mutation status (Supplementary Figures S1C–E).

The data of cancer cell lines in the CCLE provided a large body of support about the expression of genes on many more cancer subtypes of various tissues of origin. With the use of the CCLE database, we discovered that CMTM6 was overexpressed in all cell lines of ovarian cancer (Figure 2A). Afterward, we examine the expression of CMTM6 mRNA and protein in four OV cell lines (A2780, ES2, Hey, and SKOV3) and human normal ovarian epithelial cell (IOSE80) by means of qRT-PCR and Western blot separately. The results confirmed that CMTM6 expression level was upregulated in OV cell lines compared to normal ovarian epithelial cells (Figures 2B–D). These results implied that high expression of CMTM6 may be closely correlated with the biological features of OV.

Based on the abovementioned consequences, we hypothesized that CMTM6 expression could well predict the survival of OV patients. To evaluate the role of CMTM6 in the OV patient survival, we classified the cancer cases from TCGA cohort into high-expression and low-expression groups. An unexpected observation in this regard is that the influence of CMTM6 overexpression in OV is related to preferable OS (Figure 3A, log-rank p = 0.032), and DFS (Figure 3B, log-rank p = 0.043), compared with low expression of CMTM6.

After that, we took advantage of the PrognoScan database to further verify that the elevated expression of CMTM6 has better prognosis in OV (Table 2). Although the outcomes in GSE17260 and GSE14764 were not statistically discrepant, we observed the similar tendency, which we attributed to insufficient samples. Moreover, ROC analysis was conducted on data from TCGA and GTEx to measure the discrimination value of CMTM6. The area under the curve (AUC) was 0.979 (Figure 3C). These outcomes not only strongly indicate a relevant role that high CMTM6 expression has with increased OS and DFS in OV but it is also a marker for poor prognosis in at least four solid tumors (Supplementary Figure S2). These puzzling findings caused us to further research the function of CMTM6 in OV.

To better comprehend the prognostic role and assess the effect of CMTM6, we attempted to probe the connection between CMTM6 mRNA expression and clinical features applying the Kaplan–Meier database. We noticed that high CMTM6 expression in ovarian serous adenocarcinoma was linked to preferable OS and progression-free survival (PFS), but in ovarian endometrioid adenocarcinoma, high-expressed CMTM6 only exhibited decent OS (Figure 3D). Regarding different tumor stages, higher expression of CMTM6 only appeared beneficial to OS and PFS in the first stage of OV patients (Figure 3D). Similarly, a significant association between CMTM6 expression and superior OS in grade I and II was observed; however, better PFS can be found only in pathologic grade I (Figure 3D). Furthermore, CMTM6 upregulation was associated with positive OS in OV patients without TP53 mutation (Figure 3D). Hence, these analyses illustrate the prognostic value of the expression of CMTM6 for patients with OV.

As CMTM6 expression is positively correlated with survival of OV, we tried to identify signaling pathways modulated by CMTM6 expression level. The GSEA was conducted, and biological pathways were enriched in the natural killer cell–mediated cytotoxicity, chemokine signaling pathway, B cell receptor signaling pathway, and T cell receptor signaling pathway (Figures 4A–D), which are immune-related pathways and in favor of antitumor effect. That might partially explain the connection of CMTM6 with beneficial prognosis; at the same time, it is found that GSEA results are also enriched in apoptosis (Figure 4E) and cell cycle (Figure 4F). Surprisingly, some metabolism pathways, such as adipocytokine signaling pathway, alanine aspartate and glutamate metabolism, fructose and mannose metabolism, and ubiquitin-mediated proteolysis were equally enriched (Supplementary Figure S3A). These certified that CMTM6 is involved in the immune response and metabolic control in OV.

To explore the biological significance of CMTM6 in OV, we performed coexpression data mining to obtain the top 100 coexpressed genes of CMTM6 from a large cluster of 19,574 genes across Tothill Ovarian 295 samples in the Oncomine database (Figure 5). A PPI network analysis of coexpressing genes was generated by utilizing the GeneMANIA database (Figure 6A) and moved to the bioinformatics software platform Cytoscape (Version 3.8.2) to visualize the interaction network (Figure 6B). Afterward, to obtain a more extensive and in-depth cognition of the selected gene, we attempted to perform GO and KEGG pathway enrichment analyses of coexpressed genes. GO analysis showed that CMTM6 coexpressed genes were remarkedly enriched in the BP and CC. The top 10 significant terms of BP and CC are presented in Figures 7A,B. More specifically, coexpressed genes of CMTM6 were enriched in BPs principally involving establishment of protein localization to membrane, protein targeting to membrane, viral gene expression, translational initiation, viral transcription, nuclear-transcribed mRNA catabolic process, nonsense-mediated decay, establishment of protein localization to the endoplasmic reticulum, protein targeting to the ER, cotranslational protein targeting to membrane, SRP-dependent cotranslational protein targeting to membrane, and so on (Figure 7A; Supplementary Table S1). Regarding CCs, the genes were involved in cell–substrate junction, focal adhesion, ribosome, ribosomal subunit, cytosolic ribosome, large and small ribosomal subunit, cytosolic large ribosomal subunit, chromosome, centromeric region, kinetochore, and so on (Figure 7B; Supplementary Table S2). KEGG pathway analysis showed that CMTM6 coexpressed genes were remarkedly enriched in the ribosome, neurotrophin signaling pathway, and pathways in prostate cancer and glioma (Figure 7C; Supplementary Table S3).

Previous studies showed tumor-infiltrating lymphocytes (TILs) as independent predictors of prognosis in cancers (Azimi et al., 2012). Further analysis was performed to probe the immunomodulatory effect of CMTM6. We first detected the correlation of CMTM6 and immune cells in OV using the TIMER, EPIC, and XCELL algorithms of the TIMER2 database. The results suggested associations between tumor purity (r = 0.164, p = 9.38e-03), CD4⁺ T cell (r = 0.263, p = 2.64e-05), neutrophils (r = 0.199, p = 1.60e-03), and myeloid dendritic cells (r = 0.162, p = 1.05e-02) by CMTM6 expression (Supplementary Figure S3B). Moreover, we performed the association between CMTM6 expression and the immune-related signatures via the TISIDB portal, including the abundance of 28 TIL types and immunomodulators (Figures 7A–D). The results exhibited the correlations between CMTM6 and TILs in OV, such as Act_B_ abundance, Act_DC_ abundance, neutrophil_ abundance, Th17_abundance (Figure 7A). Immunomodulators can be further divided into immunoinhibitors, immunostimulators, and MHC molecules. As for immunoinhibitors, CMTM6 expression showed a positive correlation with BTLA_exp and IL10RB_exp and negative correlation with ADORA2A_exp, CD160_exp, KDR_exp, TGFBR1_exp, and TGFB1_exp (Figure 7B). With respect to immunostimulators, CMTM6 expression was positively linked to CD40_exp, HHAL2_exp, IL6_exp, IL6R_exp, KLRC1_exp, TMEM173_exp, and TNFSF13_exp (Figure 7C). Regarding MHC molecules, CMTM6 expression was related to HLA-DMA_exp, HLA-DMB_exp, TAP2_exp, TAPBP_exp, and TNFSF13_exp (Figure 7D). To elucidate the association of CMTM6 with immune cell migration, finally, we comprehensively analyzed the connection with chemokines and chemokine receptors. Figures 7E–J displayed positive correlations of CMTM6 expression with CCL20_exp, CXCL5_exp, CXCL8-exp, CXCL11_exp, CXCL16_exp, and CXCR4_exp. The microarray data sets GSE26913 (n = 107) of the Kaplan–Meier plotter disclose that the upregulation of the CCL20, CXCL5, CXCL8, and CXCL16 cohorts had a better prognosis in OV (Figures 8A–C,E), and there was no obvious significance of CXCL11 and CXCR4 expression in the survival of OV (Figures 8D,F). Following that, we searched the prognostic significance of CMTM6 in OV built on GSE26913. Similarly, higher expression of CMTM6 suggests a favorable prognosis (Supplementary Figures S3C,D). Overall, it is seen that CMTM6 possibly played an important role in mediating the OV immune microenvironment and may become a potential biomarker which has crucial impact on prognosis of patients through cross-talk with immune-related factors.

To enhance our understanding of CMTM6 cross-talk with immune regulation, we analyzed the relationships between CMTM6 expression and varied immune signatures using the correlation pattern of GEPIA2, including monocytes, TAMs, M1 macrophages, M2 macrophages, B cells, CD8+T cells, T cells (general), NK cells, neutrophils, and dendritic cells. Simultaneously, many T cell subgroups, such as Th1, Th2, Th17, Tfh, Treg, resting Tregs, effector Tregs, effector T cells, naïve T cells, effector memory T cells, resident memory T cells, and exhausted T cells, were analyzed as well (Tables 3, 4). We discovered a prominent positive association of CMTM6 with monocyte markers, natural killer cell markers, neutrophil markers, and dendritic cells markers (Table 3). We also noticed that CMTM6 showed a connection with markers of M1 macrophages and M2 macrophages and appeared to be more correlated with M1 than M2 (Table 3), reminding that CMTM6 is more likely to lead to M1-polarized macrophages in OV, and M1 macrophages have been known to enhance antitumor activity by activating macrophage-mediated inflammation (Ma et al., 2017). In addition to this, CMTM6 was markedly related with 11 of 13 T cell subgroups, and the p-value was all less than 0.01 (Table 4). The information ulteriorly demonstrated that CMTM6 has an influence on immune infiltration cells. Of course, longer-term studies need to be implemented to explore whether CMTM6 is a crucial factor that affects immune infiltration in the tumor microenvironment.

CNV refers to the insertion, deletion, duplication, translocation, and derived chromosomal structural variation of DNA fragments, which has been a pivotal factor in genetic variation and biodiversity (Iafrate et al., 2004). DNA methylation is a vital mode of regulation of gene expression, and hyper-methylation of promoter regions reduces gene expression (Saghafinia et al., 2018). Previous gene resequencing efforts have acknowledged that gene mutation is associated with human cancers (Citron et al., 2002). Because of CMTM6 expression being obviously increased in OV than control tissues, we tried to analyze the links between CMTM6 level with CNV, DNA methylation, and gene mutation by the UCSC Xena database. The heatmap revealed that CMTM6 expression was related with CNV and DNA methylation, but not with somatic mutation in OV (Figure 9A). Next, we analyzed the genetic alterations of CMTM6 in the TCGA-OV cohorts through the cBioPortal online tool. Two kinds of alterations (amplification and deletion) were detected in TCGA sources of OV (Figure 9B). CMTM6 was altered in four samples of 1,680 samples with OV, occupying 2.3% (Figure 9C). Interestingly, there was no relationship between the genetic alteration of CMTM6 and clinical prognosis of OV (Supplementary Figures S3E,F). After that, the DNA methylation status of CMTM6 was analyzed; there was a negative link between CMTM6 expression and the methylation level for OV (Figure 9D p = 3.28e-7, r < 0, Spearman’s analysis; p = 3.406e-4, r < 0, Pearson’s analysis). Considering all this, CNV and DNA methylation might contribute to the elevated level of CMTM6 in OV.