This study explored the potential role of human MECOM (NM_001105077.4 for mRNA or NP_001098547.3 for protein, Figure S1A) in multiple cancer types. The phylogenetic tree illustrated the evolutionary relationship of the MECOM protein among species (Figure S1B), suggesting its potential involvement in crucial biological processes.

To form an overview of MECOM expression in humans, we initially analyzed MECOM expression in normal tissues using the HPA, GTEx, and Function Annotation of the Mammalian Genome 5 (FANTOM5) datasets. Although MECOM exhibited the highest mRNA expression in the stomach, followed by the kidneys and lungs, it was discernible in nearly all tissues, indicating low tissue specificity (Figure S2A). Single cell type expression analysis revealed that alveolar cell type 2 exhibited the highest MECOM expression, followed by collecting duct cells (Figure S2B).

In order to investigate the expression of MECOM in tumors, we then used TIMER2 to analyze the expression status of MECOM across cancer types in TCGA database. Compared to the corresponding control tissues, tumor tissues of BRCA, head and neck squamous cell carcinoma (HNSC), kidney chromophobe (KICH), KIRC, kidney renal papillary cell carcinoma (KIRP), LUAD, lung squamous cell carcinoma (LUSC), and PRAD showed lower MECOM expression levels. Conversely, cholangiocarcinoma (CHOL), glioblastoma multiforme (GBM), liver hepatocellular carcinoma (LIHC), pancreatic adenocarcinoma (PAAD), pheochromocytoma and paraganglioma (PCPG), stomach adenocarcinoma (STAD), thyroid carcinoma (THCA), and UCEC showed higher MECOM expression levels (Fig. 1A).

MECOM mRNA and protein expression in various human tumor types. A MECOM differential expression in tumor and adjacent tissues as analyzed by TIMER2. *P < 0.05; **P < 0.01; ***P < 0.001. B Box plot data of MECOM expression in several tumor types in TCGA and corresponding normal tissues in the GTEx database. *P < 0.05. C Comparison of MECOM protein expression in breast cancer, ovarian cancer, clear cell RCC, and UCEC in the CPTAC dataset. ***P < 0.001. D Violin plots of MECOM mRNA level in patients with various tumor types in TCGA database based on their TNM stage. Log2 (TPM + 1) was used for log-scale

The GTEx dataset was then impacted to assess MECOM expression in tumors for which no normal tissue data were available in TCGA database. The results showed that the expression levels of BRCA, KIRC, KIRP, LAML, PRAD, and testicular germ cell tumors (TGCT) tumor tissues were lower than in corresponding normal tissues. However, the opposite result was seen in tissues of CHOL, colon adenocarcinoma (COAD), OV, PCPG, rectum adenocarcinoma (READ), thymoma (THYM), UCEC, and uterine carcinosarcoma (UCS) (Fig. 1B,  P < 0.05). Notably, differences for other tumors, such as bladder urothelial carcinoma (BLCA), cervical squamous cell carcinoma (CESC), and endocervical adenocarcinoma, were found to be insignificant (Figure S3).

We proceeded to investigate the protein expression of MECOM across various tumors. Total MECOM protein expression in the CPTAC dataset exhibited significant differences between tumor and normal tissues in several cancer types, including breast cancer, ovarian cancer, clear cell renal cell carcinoma (RCC), and UCEC (Fig. 1C,   P < 0.001).

We used GEPIA2 to get the MECOM expression data in specific stages across the cancers in TCGA, aiming to elucidate the relationship between MECOM expression and tumor stage. MECOM expression was significantly related to the pathological stage in certain tumors, such as BRCA and KIRC (Fig. 1D,   P < 0.05), while no clear association was found in other cases (Figure S4). Notably, a declining trend in MECOM expression was evident with the progression of the tumor stage in KIRC. Therefore, we used clinical specimens from our center to further evaluate MECOM expression in KIRC. We found that MECOM mRNA and protein levels were downregulated in tumor tissues compared to normal tissues (Fig. 2A, B, E). Similar results were observed in KIRC cell lines at the mRNA (Figure S5). Furthermore, its IHC staining level decreased with the progression of the T stage (Fig. 2C, D). IHC scores in the early stage (I + II) of KIRC were significantly higher than those in the late stage (III + IV) (Fig. 2D).

Validation of MECOM expression levels in clinical KIRC samples. A Relative MECOM mRNA levels in tumor and adjacent normal tissues of 24 patients with KIRC. B MECOM mRNA differential expression in 24 patients with KIRC between tumor and adjacent normal tissues. C Representative images of MECOM IHC staining in tumor and adjacent normal tissues in patients with various T stages (n = 147). D IHC scores of MECOM expression in tumor tissues of patients with KIRC at various clinical stages. E MECOM protein levels in tumor and adjacent normal tissues of eight patients of KIRC

Next, we compared the MECOM expression data in TCGA with IHC results in the HPA database. Our analysis revealed stronger MECOM staining in IHC images of normal tissues compared to tumor tissues in the prostate, kidney, lung, and breast but not in normal uterine tissues, consistent with the analysis results of the mRNA expression data in TCGA (Figure S6A–E).

We categorized patients based on high or low MECOM expression in various cancers, mainly in TCGA and GEO datasets, to determine the association between MECOM expression and patient prognosis. In comparison to those with low expression levels, a worse OS was seen in patients with high MECOM expression in cancers such as CHOL (P = 0.044), KIRP (P = 0.015), brain lower grade glioma (LGG) (P = 0.00044) and PAAD (P = 0.042). Conversely, worse OS was related to low MECOM expression in cancers like LUAD (P = 0.041) and OV (P = 0.046) (Fig. 3A). The high MECOM expression group seemed to have a shorter DFS in KIRP (P = 0.013), LGG (P = 0.04), and UCS (P = 0.046) (Fig. 3B).

The correlation between MECOM expression level and survival prognosis of patients with various cancers in TCGA. GEPIA2 was applied to assess the overall survival A and disease-free survival B using survival maps and Kaplan–Meier curves in all TCGA tumors based on the expression of the MECOM gene. The positive results are listed

Next, we used the Kaplan–Meier plotter tool to analyze the survival data in gastric, lung, ovarian, breast, and liver cancers. Notably, a significant correlation was observed between low MECOM expression and poor OS, first progression (FP), and post-progression survival (PPS) in lung cancer. Additionally, poor progression-free survival (PFS) in ovarian cancer, relapse-free survival (RFS) in breast cancer. However, we did not identify a correlation between MECOM expression and the prognosis of patients with gastric and liver cancers (Figure S7). A meta-analysis of the above data further confirmed the significant differences in prognosis between the low and high MECOM expression groups in lung (P < 0.00001) and ovarian (P = 0.005) cancers, but not in gastric and liver cancers (Figure S8).

Next, we directed our focus towards the MECOM genetic alterations in several human tumor types in TCGA. We noticed that the primary alteration type in LUSC (~ 40%), OV (~ 30%), and esophageal carcinoma (ESCA) (~ 20%) was ‘amplification’. In contrast, the ‘mutation’ type (~ 20%) was the predominant alteration in skin cutaneous melanoma (SKCM) (Fig. 4A). We further depicted the case number, locations, and MECOM genetic alteration types. Among the most frequent alterations was truncation mutation G614Efs*30, detected in 10 cancer cases, including COAD/READ, STAD, BRCA, CESC and ESCA (Fig. 4B). This mutation site was further observed in the 3D structure of the MECOM protein (Fig. 4C). Subsequently, we performed a survival analysis of patients with various MECOM alteration statuses in several cancer types to explore the potential relationship between the clinical survival prognosis and genetic alterations in MECOM. Our findings indicated that patients with MECOM alterations tended to exhibit worse OS (P = 0.0117) and DFS (P = 0.0210) but not PFS (P = 0.0619) and disease-specific survival (DSS) (P = 0.0651) (Fig. 4D).

Mutation features of MECOM in TCGA tumors. The cBioPortal tool was used to analyze the mutation status of MECOM in TCGA tumors. The alteration frequency with mutation type A and mutation site B are shown. C The G614Efs*30 mutation site is displayed in the 3D structure of MECOM. D Correlation analysis between MECOM mutation status and OS (overall survival), PFS (progression-free survival), DSS (disease-specific survival), and DFS (disease-free survival) in patients with PAAD, as performed using the cBioPortal tool

We performed analyses of tumor mutational burden (TMB) and microsatellite instability (MSI) based on MECOM expression across all the tumors in TCGA. The results revealed a positive association between MECOM expression and TMB in ACC (P = 0.043) and READ (P = 0.0022), and a negative association with LUAD (P = 0.00015) and KIRP (P = 3.8e−05) (Figure S9). MECOM MSI was positively associated with READ (P = 0.0082), STAD (P = 0.02), and TGCT (P = 0.019) and negatively associated with KICH (P = 0.00012), PRAD (P = 0.00011), SKCM (P = 0.04), and UCEC (P = 5.2e−05) (Figure S10). These results deserve further investigation.

We then analyzed the MECOM protein phosphorylation levels and sites in UCEC, ovarian cancer, breast cancer, clear cell RCC, and LUAD using the CPTAC dataset (Fig. 5A). Increased phosphorylation levels were respectively found at S128, S427, S501, S603, and S925 in UCEC (Fig. 5B), S526, T530, and S1048 in ovarian cancer (Fig. 5C), Y64, S726, and S1048 in colon cancer (Fig. 5F), and S427 in LUAD (Fig. 5G). Decreased phosphorylation levels were identified at S603 in breast cancer (Fig. 5D) and S501 and S603 in clear cell RCC (Fig. 5E).

MECOM protein phosphorylation analysis in various tumor types. A The schematic diagram of phosphoprotein sites in MECOM detected by the CPTAC dataset. B–G Box plots of MECOM phosphoprotein levels based on the UALCAN tool in UCEC (B), ovarian cancer (C), breast cancer (D), clear cell RCC (E), colon cancer (F), and LUAD (G)

Tumor immune cell infiltration is closely related to cancer initiation and progression [15, 16]. Moreover, cancer-associated fibroblasts, endothelial cells, and T cells in the tumor microenvironment exert significant influence on various tumor functions [17, 18]. Therefore, we investigated the potential relationship between MECOM expression levels and the estimated infiltration values of immune cells in multiple cancer types in TCGA using the EPIC, MCPCOUNTER, XCELL, TIDE, CIBERSORT, and CIBERSORT-ABS algorithms. A positive correlation was found between MECOM expression and infiltration of cancer-associated fibroblasts in BRCA, BRCA-LumA, HNSC, HNSC-HPV-, LIHC, LUSC, mesothelioma (MESO), SKCM, SKCM-Metastasis, TGCT, and THYM, while a negative correlation was found in BLCA and STAD (Fig. 6A). We also noted a positive correlation between MECOM expression and the estimated infiltration values of endothelial cells in tumors such as ACC, BRCA, and KIRC (Fig. 6B) and a negative correlation for T cell infiltration in LUSC and TGCT (Fig. 6C). The potential association between multiple checkpoint markers and MECOM expression in different cancer types is presented in Figure S11.

Correlation analysis between MECOM expression and immune infiltration of cancer-associated fibroblasts, endothelial cells, and T cells. A–C The expression level of MECOM was analyzed using the EPIC, MCPCOUNTER, XCELL, TIDE, CIBERSORT, and CIBERSORT-ABS algorithms and correlated with the level of cancer-associated fibroblasts (A), endothelial cells (B), and T cells (C) across all tumors in TCGA

Pathway enrichment analysis screened out the MECOM-interacting proteins and the MECOM expression-correlated genes using the STRING and GEPIA2 tools. We first used STRING to construct an interaction network comprising 50 experimentally detected proteins that could bind to MECOM (Fig. 7A). Subsequently, we integrated expression tumor data from TCGA to identify the top 100 genes closely associated with MECOM expression by using GEPIA2. Figure 7B displays the positive correlation between MECOM expression and the top five genes, including SOX17 (R = 0.42), C1orf106 (R = 0.42), GALNT4 (R = 0.49), MCU (R = 0.47), and PRKCI (R = 0.43) (P < 0.001 for all). The heatmap data also revealed the positive correlations mentioned above in most cancer types (Fig. 7C). We conducted additional analysis using cell lines of KIRC, PRAD, and BLCA to validate the correlations at the mRNA level (Figure S12).

MECOM-related gene enrichment analysis. A The STRING tool was used to obtain the experimentally determined MECOM-binding proteins. B GEPIA2 was used to determine the correlation between MECOM and five representative genes (SOX17, C1orf106, GALNT4, MCU, and PRKCI) of the top 100 MECOM-correlated genes in TCGA project. C Heatmap of the expression correlation between MECOM and SOX17, C1orf106, GALNT4, MCU, and PRKCI across the tumors in TCGA. D GO/KEGG pathway analysis was performed according to the MECOM-binding and interacted genes

Finally, we conducted Gene Ontology (GO) and KEGG analyses following the results reported above. Figure 7D presents the top 20 pathways, shedding light on the potential impact of MECOM on tumor pathogenesis.