Multidimensional comprehensive and integrated analysis of the potential function of TMEM25 in renal clear cell carcinoma with low expression status

Differential expression of TMEM25 in ccRCC tissues

This study presents a comprehensive examination of TMEM25 expression across various cancers, accomplished through a pancancer analysis. The outcomes of this pan-cancer TMEM25 analysis, derived from the TIMER online platform. Specifically, TMEM25 exhibited high expression levels in breast invasive carcinoma (BRCA) and lung adenocarcinoma (LUAD). However, a noticeable reduction in expression was evident in several other cancers, including cervical squamous cell carcinoma and endocervical adenocarcinoma (CESC), colon adenocarcinoma (COAD), glioblastoma multiforme (GBM), kidney chromophobe (KICH), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), liver hepatocellular carcinoma (LIHC), rectum adenocarcinoma (READ), stomach adenocarcinoma (STAD), and uterine corpus endometrial carcinoma (UCEC) tissues when compared to their respective normal tissues (Figure 1A). Within the TCGA-KIRC cohort, we conducted an in-depth analysis of TMEM25 expression, which showed a significant decrease in TMEM25 expression levels in ccRCC compared to normal tissue. This finding was confirmed in paired analyses of ccRCC and normal tissue samples (Figure 1B). Concurrently, our investigations extended to additional cohorts such as GSE46699 and GSE40435, where TMEM25 exhibited lower expression in ccRCC in contrast to normal tissues. This observation was further affirmed through qPCR analysis of renal cancer cell lines and renal clear cell carcinoma tissues (Figure 1C, 1D). Meanwhile, immunohistochemical assessment from the Human Protein Atlas (HPA) database confirmed low expression of TMEM25 in ccRCC (Figure 1E).

Correlation analysis of TMEM25 expression and clinicopathological features

Based on the levels of TMEM25 RNA expression within the TCGA-KIRC cohort, patients were categorized into distinct groups: those with high and low expression levels. By combining these results with clinically relevant features, we revealed noteworthy associations between TMEM25 and several factors. These encompassed age (p=0.011), gender (p=0.040), T-stage (p<0.001), N-stage (p=0.009), M-stage (p<0.001), Pathologic stage (p<0.001), and Histologic grade (p<0.001) (Table 1). In addition, our study highlights large differences in key survival factors. Specifically, overall survival (OS), disease-specific survival (DSS) and progression-free interval (PFI) exhibited significant differences between the two groups of ccRCC patients (all p-values < 0.001) (Table 1). In this study, we found that for OS, DSS, and PFI the expression of TMEM25 was lower in the presence of adverse events (for example death events). Similarly, the expression of TMEM25 was significantly lower in later T-stages, N-stages, M-stages, pathological stages and histological stages than in corresponding earlier stages (Figure 2A–2H).

Evaluation of TMEM25 in ccRCC for diagnosis and determination of prognostic ability, and construction and validation of nomogram

We designed a Receiver Operating Characteristic (ROC) curve to evaluate the capability of TMEM25 expression in effectively distinguishing between ccRCC and the adjacent normal tissue. The outcomes showed a remarkable area under the curve (AUC) value of 0.903, strongly indicating the substantial potential of TMEM25 as a promising diagnostic marker (Figure 2I). Taking these findings further, our study proceeded to validate the predictive power of TMEM25 in determining the prognosis of ccRCC patients. Specifically, ccRCC patients exhibiting lower TMEM25 gene expression consistently experienced shorter OS, DSS, and PFI (all p-values <0.001) (Figure 2J–2L). These results have highlighted the important impact of TMEM25 expression on the prognosis of ccRCC patients.

In our univariate COX regression analysis, we observed that a range of factors, including T stage, N stage, M stage, pathologic stage, histologic stage, and TMEM25 expression had a significant effect on OS, DSS, and PFI in ccRCC patients. On the basis of these preliminary results, our subsequent multivariate COX regression analyses consistently highlighted the potential significance of TMEM25 expression. It emerged as a substantial and independent risk factor, holding promise in predicting these outcomes (Tables 2–4). This further underscores the clinical relevance of TMEM25 expression in offering valuable prognostic insights for patients with ccRCC.

These findings underscore the potential significance of TMEM25 as a prognostic factor within the context of ccRCC. Expanding on the diagnostic and prognostic capabilities of TMEM25 in ccRCC, we developed a comprehensive nomogram that integrates some variables including age, pathologic grade, histologic grade and TMEM25 expression. ROC curves calculated from the nomogram in the training group (Figure 3A) yielded impressive AUC values of 0.864, 0.84, and 0.808 at 1, 3, and 5 years respectively (Figure 3B). These robust AUC values reflect the nomogram’s strong predictive power. Additionally, calibration curves and clinical decision curves further affirmed the nomogram’s effectiveness, showing that the predictions of the nomograms were in perfect agreement with the actual survival trends of the patients (Figure 3C, 3D). Furthermore, the AUC values for the ROC curves in the test group at 1, 3, and 5 years were 0.823, 0.749, and 0.71 respectively (Figure 3E), highlighting a consistent correlation between the calibration curves and the clinical decision curve (Figure 3F, 3G). In summary, the inclusion of TMEM25 in the nomogram offers a valuable and practical tool for predicting the survival probabilities of patients with ccRCC.

Discovery of potential functions of TMEM25 in ccRCC

By using the capabilities of the STRING database, we were able to construct a protein-protein interaction network centered around TMEM25. This notable set of genes includes ANKRD13B, TMEM91, LPPR3, PCDH20, PIP5KL1, PNKD, TMEM39A, TMEM30B, FAM19A4, and TMEM207 (Figure 4A). Subsequent to this network construction, we embarked on a Spearman analysis to unravel the correlation between these ten genes and TMEM25. This analysis was conducted using TCGA-KIRC tumor and normal tissues, with the help of the GEPIA2 website. The outcomes of this study yielded valuable results. We observed a distinctive pattern where the expression levels of TMEM25 exhibited a negative correlation with ANKRD13B, TMEM91, and LPPR3. In contrast, the expression levels of PCDH20, PIP5KL1, PNKD, TMEM39A, TMEM30B, FAM19A4, and TMEM207 displayed a positive correlation with TMEM25 expression (Figure 4B). This complex web of correlations deepens our understanding of TMEM25’s potential functional roles within the context of ccRCC.

We carried out functional enrichment analyses based on 178 up-regulated genes and 4365 down-regulated genes identified in the TCGA-KIRC dataset from the differential analysis of the TMEM25 high and low expression groups. Our exploration revealed intriguing insights into the potential roles of TMEM25 within ccRCC. Regarding the Gene Ontology (GO) results, it was apparent that TMEM25 could be intricately associated with an array of biological processes (BP). These processes encompassed “humoral immune response,” “protein activation cascade,” “complement activation,” and “humoral immune response mediated by circulating immunoglobulin.” In terms of cell components (CC), TMEM25’s influence extended to structures like “immunoglobulin complex,” “external side of plasma membrane,” “immunoglobulin complex, circulating,” “blood microparticle,” and “nucleosome.” Moving on to molecular functions (MF), the enrichment was evident in areas such as “antigen binding,” “immunoglobulin receptor binding,” “serine hydrolase activity,” “peptidase inhibitor activity,” and “serine-type endopeptidase activity.” Parallel to these results, the KEGG enrichment results highlighted significant pathways. These pathways encompassed “neuroactive ligand-receptor interaction,” “systemic lupus erythematosus,” “alcoholism,” “complement and coagulation cascades,” and “fat digestion and absorption” (Figure 4C).

Furthermore, through Gene Set Enrichment Analysis (GSEA), we uncovered a series of enriched gene sets linked with TMEM25 in ccRCC. These sets included “complement cascade,” “creation of C4 and C2 activators,” “initial triggering of complement,” “CD22-mediated BCR regulation,” “FCGAMMA receptor FCGR-dependent phagocytosis,” “FCGR activation,” “FCGR3A-mediated IL10 synthesis,” “immunoregulatory interactions between a lymphoid and a non-lymphoid cell,” “antigen activates B cell receptor BCR leading to generation of second messengers,” and “signaling by the B cell receptor BCR” (Figure 4D). These comprehensive findings collectively offer valuable insights into the potential functional roles of TMEM25 within complexities of ccRCC.

The relationship between TMEM25 expression and ccRCC immune infiltration

In the context of functional enrichment analysis, it was revealed that TMEM25 might possess immunological significance in ccRCC. Subsequent ssGSEA calculations demonstrated that the expression of TMEM25 correlated both positively and negatively with immune infiltration. Specifically, it was positively correlated with Treg (regulatory T cells) infiltration and negatively correlated with resting NK (natural killer) cell infiltration, as well as with NK cell infiltration overall (Figure 5A). Survival analysis conducted using data from the TIMER2.0 website suggested that ccRCC patients exhibiting low TMEM25 expression, along with reduced infiltration of activated and resting mast cells, as well as resting NK cells, and elevated infiltration of Tregs, were associated with poorer clinical outcomes (Figure 5B–5G). Further comprehensive analysis carried out using the TISIDB platform indicated a strong association between TMEM25 expression in ccRCC and a range of immune checkpoints, immunosuppressive agents, and major histocompatibility complex (MHC) molecules (Figure 5H–5J).

In addition, the TISIDB analysis also explored the relationship between TMEM25 expression in ccRCC and distinct immune subtypes. This analysis revealed that TMEM25 expression was most statistically significant in ccRCC subtype C5 (characterized as immunologically quiet) and lowest in ccRCC subtype C6 (dominated by TGF-b signaling). This implies that TMEM25 expression in the ccRCC immunomicroenvironment has a significant impact on the immune environment to some extent (Figure 5K). Another aspect of investigation involved examining the connection between TMEM25 copy numbers and the presence of immune infiltrating cells. The results indicated that higher TMEM25 levels associated with arm-level gain status were inversely correlated with the presence of B cells, CD8+ T cells, CD4+ T cells, macrophages, neutrophils, and dendritic cells, in contrast to the diploid/normal state. This suggests a significant relationship between TMEM25 copy number and the extent of immune infiltration in ccRCC (Figure 5L).

DNA methylation analysis of TMEM25 in ccRCC patients

Upon conducting a thorough analysis of TMEM25 DNA methylation data using the UCSC Xena and UALCAN platforms, compelling findings have come to explore. The investigation has unveiled a significant increase in DNA methylation levels within the TMEM25 gene in ccRCC tissues compared to their normal tissues. This elevated methylation trend was consistently observed not only in cases with nodal metastasis but also across varying cancer stages and diverse tumor grades (Figure 6A–6E). Further exploration into the DNMIVD database has yielded deeper insights. The analysis has revealed distinct associations between specific CpG loci within the TMEM25 gene and both the diagnosis and prognosis of ccRCC patients. Notably, four particular CpG loci (cg19715094, cg10260050, cg20001829, cg15694715) have emerged as especially noteworthy. Among these, cg19715094 has been assigned the highest importance score, whereas cg15694715 has received the lowest importance score (Figure 6F, 6G).

The resulting diagnostic model, formed by using the information from these four identified loci, has demonstrated remarkable efficacy in discerning ccRCC tissue from normal tissue. This is evident from the ROC curves, where the diagnostic model’s ROC curve boasts a substantial AUC value of 0.923, showing its significant diagnostic capability (Figure 6H).

Analysis of genetic alterations of TMEM25 in ccRCC

The cBioPortal website provides a comprehensive dataset encompassing 538 ccRCC patients from the TCGA, particularly through the Firehose Legacy project. This dataset includes essential information about genetic alterations and prognostic survival outcomes. Within this cohort, a relatively small percentage (0.7%) of ccRCC patients exhibit mutations in the TMEM25 gene. These mutations primarily manifest as missense mutations and amplifications. Notably, the patients with TMEM25-altered genomes exhibit significantly lower overall survival rates in comparison to those with unaltered genomes, a difference that holds statistical significance (p=0.0253). These trends are effectively visualized in Figure 7A–7D. COSMIC website shows the types of mutations in TMEM25 in different cancers. Missense substitutions and synonymous substitutions are main mutation types, with C>T and G>A mutations emerging as the most common substitution alterations (Figure 7E, 7F). Employing the muTarget website for mutation status and TMEM25 expression analysis has uncovered an intriguing correlation. It’s evident that TMEM25 expression experiences a notable reduction in the presence of mutant phenotypes linked to genes like BAP1, EYS, SETD2, UNC80, and XIRP2 (Figure 7G–7K). This analysis highlights the intricate interactions between TMEM25 mutations and the expression profiles of these related genes.

In summary, these careful analyses performed on different platforms have combined to greatly clarify the genetic landscape shaped by mutations in the TMEM25 gene in ccRCC. Moreover, it underscores their impact on patient survival rates and their interconnectedness with genes of significance.

Drug sensitivity analysis of TMEM25 in ccRCC

Based on an assessment of the median expression levels of TMEM25 in ccRCC, the TCGA-KIRC cohort was divided into two distinct groups: those exhibiting high TMEM25 expression and those with low TMEM25 expression. Subsequent calculations involved determining the IC50 values for a variety of drugs. A compelling trend is that ccRCC cases with lower levels of TMEM25 expression have correspondingly lower IC50 values across the range of drugs tested. Notably, this sensitivity was observed for drugs like sunitinib, pazopanib, gemcitabine, crizotinib, bryostatin, epothilone, and 5-fluorouracil. This intriguing correlation suggests that these medications tend to be more effective against ccRCC instances characterized by diminished TMEM25 expression.

However, an interesting divergence from this pattern is evident in the case of erlotinib. Here, ccRCC patients featuring low TMEM25 expression exhibited a lowered sensitivity to erlotinib compared to those with elevated TMEM25 expression (Figure 8A–8P). This nuanced observation highlights the multifaceted role that TMEM25 expression plays in influencing ccRCC’s response to a range of therapeutic agents. This analysis radically reveals the complex interactions between TMEM25 expression levels and drug sensitivity in ccRCC. The findings garnered from this exploration potentially carry implications for the development of personalized treatment strategies tailored to individual patients.

Patient dataset acquisition

We leveraged the TCGA database (https://portal.gdc.cancer.gov/), an extensive genome sequencing repository encompassing genetic data from 33 different cancer types. From this resource, we extracted RNA expression profiles and relevant clinical data originating from 539 ccRCC samples, alongside 72 normal tissues adjacent to cancer. Furthermore, we turned to the GEO database [47] (https://www.ncbi.nlm.nih.gov/geo/) for validation purposes, utilizing GSE40435 [48] (101 tumor tissues, 101 normal tissues, platform GPL10558) and GSE46699 [49] (65 tumor tissues, 65 normal tissues, platform GPL570). The GEO database compiles gene expression data contributed by research entities globally, providing a reliable source for verifying gene expression disparities. To enhance the robustness of our findings, we procured an additional set of 15 ccRCC tumor tissue samples along with corresponding adjacent normal tissue samples at our research center. We adhered to strict ethical guidelines and all samples were ethically approved by the ethics committee of the research centre and patient approval was obtained for sample collection. Further corroborating our observations, we accessed the HPA database [50] (https://www.proteinatlas.org/), which contains comprehensive information on the tissue and cells of various human proteins. This database validated the divergent distribution patterns of TMEM25 in tumor tissues versus normal tissues of ccRCC patients.

Cell lines and cell culture

The qPCR experiments involved normal renal epithelial cells (HK-2), the human embryonic kidney cell line (293-T), and a renal carcinoma cell line (A498, 769-P, 786-O and Caki-1) procured from the Chinese Academy of Sciences (Shanghai, China). All cell types were cultured under suitable conditions, utilizing a medium supplemented with 10% fetal bovine serum and 1% streptomycin and penicillin. The incubation temperature was maintained at 37° C with 5% CO2 to create an optimal growth environment.

Analysis of the correlation between TMEM25 and clinicopathological features and its diagnostic and prognostic power for ccRCC

Using the RNA expression data extracted from the TCGA database and clinically relevant information, we classified the TCGA samples into different groups based on clinicopathological characteristics. Subsequently, we employed the “pROC” package within the R programming language to assess the diagnostic potential of TMEM25 in ccRCC. Furthermore, we utilized the “ggplot2”, “survival”, and “survminer” R packages to scrutinize the differential expression of TMEM25 in ccRCC across various clinical contexts. This comprehensive analysis also enabled us to determine whether such variations in expression were correlated with patient prognoses.

Quantitative real-time polymerase chain reaction(qPCR)

Utilizing TRIzol reagent (Cwbio, Taizhou, China), we extracted RNA from both cells and tissues. The extracted RNA was then quantified using NanoDrop2000 software. Subsequently, we performed reverse transcription using reagents from TransGen Biotech ( Beijing, China). For the PCR analysis, we employed SYBR Real-Time PCR reagent from the same source (TransGen Biotech, Beijing, China), and the examination was conducted utilizing the 2^-ΔΔCt method, with β-actin serving as the internal reference.

The forward primer sequence for TMEM25 was 5′-CTTGGCACACAACCTCTCGGTG-3′, and the reverse sequence was 5′-AAGGCAAGTCCTCCAGCCACAA-3’. The reference gene was β-actin, in which the forward sequence was 5′-TCTCCCAAGTCCACACAGG-3′ and the reverse sequence was 5′-GGCACGAAGGCTCATCA-3’.

Univariate and multivariate COX regression analyses and construction and validation of nomograms

To further validate the independent prognostic significance of TMEM25 in ccRCC, we partitioned patients within the KIRC dataset into two groups: TMEM25 high expression and TMEM25 low expression, based on the median TMEM25 expression values. Our initial step involved selecting OS, DSS, and PFI as dependent variables. Subsequently, we conducted univariate and multivariate COX regression analyses, incorporating TMEM25 and significant clinical characteristics. This aimed to ascertain whether TMEM25 stood as an independent prognostic factor, calculated by determining the corresponding Hazard Ratios (HR) and the two-sided p-values within a 95% confidence interval. Proceeding to the second step, we constructed a nomogram using the R package “rms”. This nomogram allowed us to predict the overall survival at 1, 3, and 5 years for ccRCC patients, utilizing the data derived from the previous analysis. The final step encompassed the validation of the nomogram. To achieve this, we randomly partitioned the TCGA-KIRC cohort into training and test groups using the “Caret” package. Subsequently, we evaluated the discrimination performance of the nomograms and their consistency with actual observations by means of the C-index and the calibration graphs, respectively. Additionally, we plotted DCA curves for both the training and test groups to determine the clinical utility of the nomogram.

Enrichment analysis of protein-protein interaction network and functional pathway of TMEM25 in ccRCC

A protein-protein interaction network was established using the STRING database, wherein the combined interaction score for TMEM25 exceeded 0.4. The GEPIA2 website was employed to analyze the correlation between TMEM25 and these identified genes using the pearson correlation method. We used the “DESeq2” R package to detect gene expression differences between high and low TMEM25 expression samples in ccRCC and screened for differential genes based on p-values as well as logFC values. To comprehensively understand the functions of TMEM25 in ccRCC, we conducted an in-depth assessment. This involved analyzing GO, KEGG, and GSEA data, with BP, CC and MF corrections applied. The R packages “clusterProfiler” and “org.Hs.eg.db” were pivotal in these analyses. For GSEA, the reference gene set utilized was c2.cp.v7.2.symbols.gmt, with a seed number of 2020 and 1000 permutations for calculations.

Analysis of immune infiltration correlation of TMEM25 in ccRCC

By employing the “GSVA” R package, we investigated the distribution and relationship of TMEM25 with immune infiltration enrichment across diverse immune cell types in ccRCC patients. This exploration encompassed the utilization of both the ssGSEA and Spearman methods. Extending our inquiry, we investigated how TMEM25 expression in ccRCC correlated with specific immune cell states. Employing the CIBERSORT-ABS method available on the TIMER2.0 website, coupled with a prognostic analysis, we unveiled potential associations between TMEM25 expression and factors such as mast cell activation, NK cell resting, and T cell regulation. Furthermore, our investigations unveiled associations between TMEM25 and various pivotal immune checkpoints, immunosuppressive agents, and MHC molecules. These results are based on the TISIDB website, which is a multifunctional platform integrating gene and tumour immune system interactions. Additionally, we leveraged the SCNA section of TIMER2.0 to dissect the impact of somatic copy number alterations of the TMEM25 gene on tumor infiltration levels.

Genetic alterations, methylation analysis and drug sensitivity analysis of TMEM25 in ccRCC

The genetic alterations associated with TMEM25 in ccRCC were carefully analysed using ccRCC data accessed through the cBioPortal website, in particular the Firehose Legacy data from TCGA. The analysis allowed us to identify the main types of TMEM25 mutations and their distribution patterns. To establish the relations between TMEM25 expression levels and gene mutations, we leveraged the capabilities of the muTarget platform, which yielded meaningful insights into these relationships.

In our pursuit of understanding the role of TMEM25 methylation in ccRCC, we turned to platforms like UCSC Xena and UALCAN. These resources illuminated the distinct methylation patterns within ccRCCs and their correlation with various clinical characteristics. In addition, findings from the DNMIVD database suggest that DNA methylation levels of TMEM25 can be used as a predictive marker for survival and diagnosis of ccRCC patients.

Intriguingly, we also aimed to measure the differential sensitivity of certain drugs between groups characterized by high and low TMEM25 expression. To achieve this, we harnessed the “pRRophetic” R package, enabling us to compute the respective IC50 values based on the “cgp2016ExprRma” dataset. This analysis facilitated a comparative assessment of drug sensitivity.

Statistical analysis

The statistical analysis for this study was executed using the R programming language software, version 4.1.2. To examine variations in gene expression between ccRCC tissues and their corresponding normal tissues adjacent to cancer, we employed the Wilcoxon test. In adherence to conventional practice, significance was attributed to results where p<0.05, indicating statistically significant differences.

Consent

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.