The unified and standardized pan-cancer dataset was downloaded from the UCSC (https://xenabrowser.net/) database, Therapeutically Applicable Research to Generate Effective Treatments (TARGET; includes 7 pediatric cancers), and Genotype-Tissue Expression (GTEx; covers 54 normal tissues). Those samples with fewer than three tumor types were excluded from the analysis. Single-cell sequencing datasets were obtained from the Gene Expression Omnibus (GEO) database. They included breast invasive carcinoma (BRCA, GSE114727), colon adenocarcinoma (COAD, GSE146771), skin cutaneous melanoma (SKCM, GSE48190), liver hepatocellular carcinoma (LIHC, GSE98638), nasopharyngeal carcinoma (NPC, GSE150430), and kidney renal clear cell carcinoma (KIRC, GSE 111360).

Diseases associated with CHMP7 were analyzed with the Open Target website (https://www.opentargets.org/) and represented in bubble charts. The sample information obtained from UCSC was analyzed with R software (version 3.6.4) to calculate the difference in CHMP7 expression between normal and tumor samples in each tumor, and the significance of differences was performed using the unpaired Wilcoxon rank sum and signed rank tests. Protein level data were obtained from the UALCAN database (http://ualcan.path.uab.edu/) and further utilized to reveal the differential expression of CHMP7 in normal and tumors (Patra et al., 2021). The UALCAN database also facilitated further analysis of the correlation between CHMP7 expression levels and tumor stage (Patra et al., 2023). The SURVIVAL package from the R software was applied to analyze the association between CHMP7 expression and pan-cancer survival with risk tables.

IHC staining analysis was performed on fifty specimens of postoperative paraffin slides that underwent surgical treatment at the Department of Colorectal Surgery, Second Hospital of Jilin University, from January 2017 to December 2017, and whose postoperative pathology was confirmed as colorectal cancer. The experiment was approved by the ethics committee. The antibody was obtained from Proteintech under item number 16424-1-AP for IHC staining. Paraffin slides were stained by the Servicebio company. The slides were dried and observed with a light microscope, and two professional pathologists were invited to determine the intensity of the staining. Relevant clinical information of patients was collected and followed up to determine whether there was any significant difference in the effect of different CHMP7 expression intensities on the prognosis of patients.

cBioPortal (https://www.cbioportal.org/) is a website that enables researchers to explore, visualize, and analyze multidimensional cancer genomic data. Using this website, we can easily explore genetic alterations in different tumor types, genes, and pathways. We queried the mutation characteristics of CHMP7 in different tumor types and further explored the mutation types of the genes. The prognostic impact of CHMP7 mutations on patients with tumors was evaluated through the “Comparison” module of the website.

To further analyze the correlation between CHMP7 and genomic stability, the R software was used to compute the correlation between tumor mutation burden (TMB), microsatellite instability (MSI), homologous recombination deficiency (HRD), and neoantigen (NEO) data for each tumor and CHMP7 expression levels. A heat map demonstrated the correlation between CHMP7 expression levels and five mismatch repair genes (MLH1, MSH2, MSH6, PMS2, and EPCAM) in each tumor type (Levy et al., 2017).

The Stemness index is an indicator to assess the similarity of tumor cells to stem cells, which is associated with active biological processes in stem cells and a more advanced degree of tumor dedifferentiation. Wiznerowicz et al. constructed predictive models for multipotential stem cell samples from the PCBC dataset with the one-class logistic regression (OCLR) machine learning algorithm. The predictive model was subsequently applied to a training set such as TCGA to calculate the stemness score for each sample (Malta et al., 2018). The tumor stemness index calculated by mRNA expression and methylation signature was obtained from previous studies, and the stemness index and gene expression data of the samples and their correlation performed were further integrated (Kocaturk et al., 2019).

Homologous recombination repair (HRR), as one of the core DNA damage repair methods, is a DNA repair mechanism that maintains genome integrity to ensure high-fidelity transmission of genetic information. Mutation of related HRR genes or methylation of gene promoters can trigger HRR dysfunction, leading to genomic instability. Tumor cells tend to exploit HRR to save cells from apoptosis. HRR is a complex signaling pathway involving multiple steps, in which the most critical genes are BRCA1 and BRCA2, and other related genes include MLH1, MSH2, ATM, and TP53. The HRR-related genes analyzed in the present study were cited in the research of Liang et al. (Liang et al., 2022). The correlation between CHMP7 and HRR signature can be analyzed by applying the GEPIA2.0 online website (Ledermann et al., 2016). Heat maps were applied to visualize the correlation between CHMP7 and 44 RNA-modified genes.

AS refers to the process of mRNA precursor to mature mRNA, in which various splicing methods can allow the same gene to produce multiple different mature mRNAs, resulting in the translation of different proteins (Ule and Blencowe, 2019). AS is a major mechanism for maintaining protein diversity (Li et al., 2017). It produces specific shear isoforms in certain tissues or conditions at different stages of development. The OncoSplicing website (http://www.oncosplicing.com/) was employed to explore SplAdder and SliceSeq projects that contain AS events for CHMP7. Differences in percent spliced-in (PSI) and AS events were further compared between TCGA tumor tissues and GTEx normal tissues. We also explored the impact of AS events of CHMP7 on patient prognosis in diverse tumors and confirmed shear isoforms of CHMP7 in pan-cancer.

The protein-protein interaction (PPI) network of CHMP7 was explored using the STRING website (https://www.string-db.org/). Tumor development involves the aberrant activation of multiple critical pathways. Genes in the corresponding pathways were collected from TCGA and analyzed using the GSVA package of R software. The correlation between CHMP7 and pathway scores was analyzed using Spearman correlation analysis. The similar gene detection function in the GEPIA2.0 website helped us obtain genes that were the top 100 co-expressed with CHMP7 in tumor tissues and performed gene ontology (GO) functional analysis of the top 100 genes with the clusterProfiler package. Gene set enrichment analysis (GSEA) further explored the functional enrichment-related pathways CHMP7 may affect.

The role of immune cells in TME has attracted much attention in recent years (Padma et al., 2023). Different immune cells perform various roles in tumorigenesis. The immune cells of tumor types are frequently quantified in studies of tumorigenesis, treatment, and other mechanisms. The correlation scores between CHMP7 and tumor tissue immune infiltration levels were calculated with the ESTIMATE package of R software, and the bar graphs present the stromal, immune, and ESTIMATE scores for each tumor (Yoshihara et al., 2013). The correlation between CHMP7 and immune checkpoints is also shown in a heat map.

The TISDB online website (http://cis.hku.hk/TISIDB/) was utilized to analyze the differential expression of CHMP7 in distinct immune subtypes of pan-cancerous tissues (Thorsson et al., 2018). The site was also exploited to construct heat maps exhibiting the correlation between CHMP7 and chemokines, chemokine receptors, immunostimulatory factors, and immunoinhibitory factors.

M2 macrophages are regarded to serve in pathogen clearance, anti-inflammatory response, and tumor progression (Biswas and Mantovani, 2010). The TIMER2.0 website was applied to analyze the relationship between CHMP7 and M2 macrophages according to different algorithms. Additionally, spatial transcripts in SpatialDB (https://www.spatialomics.org/SpatialDB/) were subjected to analysis of the overlapping levels of CHMP7 expression and spatial expression of M2 macrophage markers (CD68 and CD163). As for the TISCH website (http://tisch.comp-genomics.org/), single-cell data from multiple tumors were presented to compare the expression of CHMP7 in various cell subtypes of tumor tissues.

CTLs are one of the critical immune surveillance cells. A high abundance of CTL with a killing function in tumor tissues is a promising prognostic indicator, and increasing the proportion of CTL in patients’ tumor tissues can help inhibit tumor progression and eventual elimination (Mami-Chouaib et al., 2018). Multiple algorithms are available on the TIMER2.0 website to analyze the correlation between CHMP7 and CTL and the impact of CHMP7 expression levels and CTL infiltration levels on patient prognosis. Tumor immune dysfunction and exclusion (TIDE, a web tool) also enabled us to investigate the role of CHMP7 in T cell dysfunction and CTL-related prognosis in tumor tissues.

To elucidate the predictive value of CHMP7 in tumor immunotherapy, we explored it from public databases. The predictive role of CHMP7 as a new biomarker was compared with classical markers such as TMB at the TIDE website to calculate the predictive role of CHMP7. The TIDE website further enabled us to predict the therapeutic response of CHMP7 in the core dataset, immunotherapy dataset, CRISPR screening dataset, and mechanistic follow-up experiments of immunosuppressive cell types. And the differential expression of CHMP7 was compared between the responding and non-responding groups in 30 immunotherapy cohorts such as IMvigor210.

The ROC Plotter dataset (http://www.rocplot.org/site/index) was adopted to analyze the gene expression levels in the response and non-response groups of multiple chemotherapeutic agents. The RNAactDrug database (hrbmu.edu.cn) can facilitate our queries on the association between CHMP7 and drug sensitivity (Xie et al., 2022).

The overall survival (OS) differences between the high and low gene expression groups were explored using the log-rank test. The correlation coefficients were quantified using Spearman or Pearson. All data analyses included in this article, such as differential expression, gene interaction, immune infiltration, and drug sensitivity analyses, were considered significant only at p < 0.05.

The Open Target website presents diseases associated with CHMP7, and the bubble chart illustrates that CHMP7 may be associated with multiple tumors (Figure 2A). The differential expression of CHMP7 in tumor tissues and corresponding normal tissues were analyzed with TCGA, TARGET, and GTEx databases. The results indicated that CHMP7 expression levels were significantly upregulated in GBM, LGG, BRCA, KIPAN, STAD, HNSC, SKCM, PAAD, LAML, and CHOL tumor tissues, while markedly downregulated in ESCA, STES, KIRP, COAD, PRAD, LUSC, BLCA, THCA, READ, OV, TGCT, UCS, ACC, and KICH tumor tissues (Figure 2B). The UALCAN website analyzed the differences in CHMP7 protein levels, revealing that CHMP7 protein was downregulated in BRCA, COAD, and UCEC tumor tissues and upregulated in LUAD tumor tissues (Figure 2C). The expression level of CHMP7 is related to the stage of many tumors. For example, the more advanced the stage in COAD, the lower the CHMP7 expression level (Figure 2D). The TCGA data were employed to analyze the predictive value of CHMP7 on prognosis and demonstrated that low CHMP7 expression was associated with poorer OS in BRCA, ESCA, HNSC, KIRC, SARC, and SKCM, while in LAML, patients in the high CHMP7 expression group had a worse prognosis (Figure 2E).

The IHC staining was performed on the collected 50 colorectal cancer paraffin slides. According to the staining intensity, specimens were divided into CHMP7 low expression group (Figures 2F1–3) and CHMP7 high expression group (Figures 2F4–6). The relevant information was collected into a baseline clinical information table (Table 1), and the collected data were processed and analyzed using R software. The results indicated that CHMP7 expression level was not significantly correlated with patients’ age, weight, gender, concomitant disease status (hypertension and diabetes), and T-stage, and the group with low CHMP7 expression had advanced N-stage and TNM-stage. The prognostic impact of CHMP7 on patients was analyzed by Kaplan-Meier survival analysis, which revealed that patients with low CHMP7 expression had a relatively poor prognosis (Figure 2G).

To explore the possible role of CHMP7 in tumor development in detail, the enrichment pathways of CHMP7 and its related genes were analyzed. The PPI network for CHMP7 was constructed with the STRING website (Figure 3A). The tumor data from the TCGA database were exploited to analyze the correlations between CHMP7 and multiple pathways, and the results were represented as scatter plots. CHMP7 and angiogenesis, apoptosis, inflammatory response, EMT markers, tumor proliferation signature, and tumor inflammation signature all demonstrated a significant positive correlation (Figure 3B).

The top 100 genes tightly associated with CHMP7 were explored via the GEPIA website, and heat maps were drawn to present the top 5 genes of relevance (Figure 3C; Supplementary Material S1). The genes of CHMP7 and its related top hundred were subjected to GO and KEGG enrichment analysis, revealing a strong association with RNA splicing, ESCRT complex, and DNA damage checkpoint (Figures 3D, E).

Mutation analysis of CHMP7 in the TCGA database was performed on the cBioPortal website. The results reveal that the top five tumors with alterations are PRAD, OV, LIHC, BLCA, and COAD, with the main mutation type being deletion. The main loci of mutations are shown in Figure 4A. To further analyze the effect of CHMP7 mutations on patient prognosis, COAD and LUSC patients in the CHMP7-altered group had poorer disease-free survival (DFS; Figure 4B).

TMB and MSI are important clinical biomarkers that can be effectively predicted for tumor treatment (Baretti and Le, 2018; Zhang et al., 2022). Radar plots present the correlation between CHMP7 and TMB, and MSI in different tumor tissues (Figure 4C). CHMP7 is positively correlated with TMB in five tumors (LGG, COAD, STES, STAD, and UCEC) and significantly negatively correlated with TMB in three tumors (LUAD, KIRP, and THCA). As for MSI, CHMP7 is significantly positively correlated in CESC, COAD, ESCA, STES, STAD, UCEC, READ, and UVM and negatively correlated in PRAD, THCA, and DLBC.

Ploidy is tightly associated with chromosomal instability in tumor development (Müller et al., 2021). The bar chart demonstrates the correlation of CHMP7 with ploidy, which is significantly negatively correlated in nine types of tumors (BRCA, STES, KIPAN, PRAD, UCEC, KIRC, LUSC, THYM, and BLCA). CHMP7 is significantly negatively correlated with HRD in CESC, LUAD, COAD, BRCA, ESCA, STES, KIPAN, STAD, UCEC, KIRC, LUSC, READ, TGCT, SKCM, UCS, BLCA, and CHOL. It presents a positive correlation in LGG (Figure 4D). NEO is abundantly expressed in tumor cells with strong immunogenicity and tumor heterogeneity, making it an attractive target for tumor immunotherapy (Blass and Ott, 2021). Scatter plots indicate a significant positive correlation between CHMP7 and NEO in COAD, LGG, UCEC, and CESC, while a negative correlation in BLCA and PCPG (Figure 4E). All analyses suggest that CHMP7 is altered in various tumor tissues and associated with genomic instability.

The stability of the genome relies on the combined action of multiple repair mechanisms, such as MMR and HRR. The diagonal heat map demonstrates the correlation between CHMP7 and MMR-related genes (MLH1, MSH2, MSH6, PMS2, and EPCAM), exhibiting positive correlation in various tumor tissues, including ACC, BRCA, and KIRC (Figure 5A). The correlation between CHMP7 and cancer stemness was analyzed based on six tumor stemness indices calculated by mRNA expression and methylation signature, which were significantly positively correlated in HNSC and negatively correlated in LUAD, COAD, COADREAD, BRCA, THYM, PCPG, and BLCA (Figure 5B; Supplementary Material S2). A positive correlation between CHMP7 and HRR was found in multiple tumor tissues, indicating that CHMP7 may be associated with DNA repair (Figure 5C).

DNAss is a cancer stemness score based on the methylation profile calculated for each tumor and further analyzed for CHMP7 DNA methylation in the TCGA database (Malta et al., 2018). A significant negative correlation was observed between CHMP7 and DNA methylation probe cg00140501 in several tumor tissues of TCGA, including ACC, BRCA, BLCA, COAD, LIHC, and LUAD (Figure 5D). In addition to methylation analysis, the correlation between CHMP7 and RNA-modified genes was analyzed, revealing a heat map showing a positive correlation between CHMP7 and major RNA-modified genes (Figure 5E).

AS is an essential mechanism for regulating gene expression and generating proteomic diversity, which has the potential to serve as a new biomarker in oncology and provide many new targets for drug development (Blencowe, 2006). The AS events for CHMP7 were evaluated at the OncoSplicing website, and alt_5prime_195490 is presented in Figure 6A and the rest in Supplementary Material S3. A higher PSI was observed in BRCA, CHOL, KIRP, and READ tumors. The difference in PSI between tumor tissues and adjacent/normal tissues was compared, and CHMP7_alt_5prime_195490 exhibited higher PSI in various tumor tissues, such as BRCA and STAD (Figure 6B). The predictive value of PSI values on the prognosis of patients with tumors was analyzed. The results suggested that high PSI values were associated with poorer OS in ESCA, COAD, HNSC, and UVM and poorer DFI in LIHC, SKCM, and TGCT (Figure 6C). Shear isoforms of CHMP7 in pan-cancer were also demonstrated (Figure 6D). All these results indicate that AS events of CHMP7 are critical in tumor research.

To investigate the relationship between CHMP7 and immune infiltration in the tumor TME, we calculated the stromal and immune scores of tumor samples based on CHMP7 expression data using the ESTIMATE package of R software. The two scores were summed to obtain the ESTIMATE score, which can be used to estimate tumor purity. Significant correlation between CHMP7 expression and immune infiltration was observed in 20 cancer species, 12 of which were significantly positively correlated, including TCGA-BRCA, TCGA-STES, TCGA-KIPAN, TCGA-COAD, TCGA-COADREAD, TCGA-PRAD, TCGA-STAD, TCGA-HNSC, TCGA -KIRC, TCGA-BLCA, TCGA-PAAD, and TCGA-LAML. There were eight significant negative correlations, such as TCGA-GBM, TCGA-GBMLGG, TCGA-LGG, TCGA-SARC, TARGET-WT, TCGA-THCA, TARGET-NB, and TCGA-ACC (Figure 7A). CHMP7 was significantly positively correlated with most immune checkpoints and immunoregulatory genes in pan-cancerous tissues. Notably, CHMP7 was negatively correlated with most immune checkpoints and immunoregulatory genes in THYM (Figure 7B; Supplementary Material S4).

The TISIDB enabled us to investigate the differential expression of CHMP7 in different immune subtypes. The results revealed that CHMP7 expression was significantly elevated in the C4 subtype in BRCA, KIRP, and LGG, indicating that CHMP7 may be associated with lymphocyte function (Figures 7C, D). Heat maps demonstrate the correlation of CHMP7 with cytokines, cytokine receptors, immunostimulatory factors, and immunosuppressive factors (Figure 7E). The TISMO website was utilized to compare the changes in CHMP7 expression levels in tumor cell lines in both pre- and post-cytokine treatment (Figure 7F). CHMP7 expression levels increased in several cell lines after IFN-β treatment, and the same was observed in several cell lines treated with IFN-γ. The results suggest that the downregulation of our CHMP7 expression levels may lead to the suppression of immune checkpoint function, which is associated with suppressive TME.

The relationship between CHMP7 and immune cell infiltration in TME was analyzed using the EPIC algorithm (Sturm et al., 2019), and seven immune cells associated with CHMP7 expression were obtained (Figure 8A). The results indicate that CHMP7 negatively correlates with CD8⁺ T cells in CESC, KICH, KIRC, PRAD, THCA, THYM, and UVM. The negative correlation of CHMP7 with tumor-associated macrophages has been observed in BLCA, COAD, HNSC, LAML, PAAD, and PRAD, while the opposite has been observed in CHOL, GBM, KIRP, LIHC, and THYM. The relationship between CHMP7 and immune cells in COAD is shown in scatter plots (Figure 8B).

M2 macrophages can secrete suppressive cytokines such as TGF-β and IL-10, producing an immunosuppressive TME that promotes tumor progression (Genin et al., 2015). The TIMER2.0 website facilitated us to analyze the correlation between CHMP7 and M2 macrophage infiltration levels with multiple algorithms, and consistent correlations were observed in KIRP, THYM, and UCEC, suggesting that downregulation of CHMP7 expression may be associated with M2 macrophage infiltration (Figure 8C). The correlation between CHMP7 and M2 macrophage markers (CD68, CD163) was analyzed at the spatial transcriptional level utilizing the SpatialDB, and CHMP7 and M2 macrophage biomarkers presented a similar spatial distribution (Figure 8D).

Subsequently, we investigated the correlation between CHMP7 and immune cell infiltration from the single-cell level. The relationship between CHMP7 and immune cell infiltration was investigated with the help of the TISCH (Figure 8E). The correlation between CHMP7 expression and immune cells in BRCA, COAD, SKCM, NPC, LIHC, and KIRC was analyzed, and there were differences in the expression levels of CHMP7 in various infiltrating immune cells (Figure 8F).

In addition to M2 macrophages, the relevance of CHMP7 to CTL has also attracted our attention. Similar to M2 macrophages, we analyzed the correlation between CHMP7 and CTL with multiple immune infiltration algorithms in TIMER2.0 and expressed consistent significant positive correlations in BLCA, BRCA-Basal, BRCA-LumA, PAAD, PRAD, SARC, SKCM, STAD, THCA, and UVM (Figure 9A). Combined CHMP7 expression and CTL infiltration levels were analyzed for their predictive value for patient prognosis. The results reveal a poorer prognosis in the CHMP7 low expression and CTL low infiltration groups in BRCA, CSC, HNSC, LIHC, KIRC, and SKCM (Figure 9B).

The correlation between CHMP7 and CTL dysfunction was observed in SKCM, UCEC, BRCA, and LAML through the TIDE website. A significant positive correlation between CHMP7 and CTL was demonstrated in SKCM and LAML (Figure 9C). The results suggest that CHMP7 is associated with many immune cells, with the possible involvement of killing tumor cells by targeting CTLs.

To investigate the possible role of CHMP7 in tumor therapy, we investigated the predictive value of CHMP7 compared to classical biomarkers for immunotherapeutic response through the TIDE website (Figure 10A). The bar chart indicates that CHMP7 presents an AUC above 0.5 in 11 of the 25 immunotherapy cohorts, suggesting a predictive significance. In 11 cohorts, the predictive value of CHMP7 was superior to the MSI score of CHMP7. Higher predictive value of CHMP7 was observed in 14, seven, 18, 18, 16, nine, and six immunotherapy cohorts compared with TIDE, TMB, CD274, CD8, IFN, T. Clonality, T. Clonality, and Merck18, respectively. It was observed in the antiPD-L1 treatment group of the EMT6 mouse breast cancer cell line that CHMP7 expression was higher in the response group (Figure 10B). The TIDE website also assisted us in predicting the therapeutic response of CHMP7 in the core dataset, immunotherapy dataset, CRISPR screening dataset, and mechanistic follow-up experiments with immunosuppressive cell types (Figure 10C).

In addition to the predictive value of CHMP7 for immunotherapy efficacy, CHMP7-sensitive small molecule drugs were also analyzed using the CTRP and GDSC datasets (Figure 10D). It was further compared that CHMP7 was differentially expressed in the responding and non-responding groups of multiple immunotherapy cohorts. The results revealed that CHMP7 was differentially expressed in response and non-response groups in multiple immunotherapy cohorts including SRP094781, GSE67501, SRP230414, and IMvigor210 (Supplementary Material S5). The RNAact Drug website also facilitated our predictions (Supplementary Material S6). The results suggest that the top 5 sensitive drugs associated with CHMP7 expression based on the GDSC dataset are GSK429286A, KU-55933, BX-912, CCT007093, and Tretinoin (p < 0.05). Performing further exploration of the differential expression of CHMP7 in the response and non-response groups of multiple chemotherapeutic agents, we discovered that high CHMP7 expression may be associated with drug sensitivity to chemotherapeutic agents such as paclitaxel, oxaliplatin, gemcitabine, and methotrexate, which are now commonly recommended (Figure 10E).