A total of 33 types of cancer RNA-seq data were downloaded from the The Cancer Genome Atlas (TCGA) database. ENSG00000023892 (DEF6) was the target gene for data processing and extract. Normal tissue RNA-seq data were acquired from the Genotype-Tissue Expression (GTEx) database (https://commonfund.nih.gov/GTEx) (10). We transformed the RNA-seq data to TPM, and a log₂(1+TPM) transformation was performed.

A total of 1,406 cancer cell lines from 33 types of disease were downloaded from the Cancer Cell Line Encyclopedia (CCLE) database (https://sites.broadinstitute.org/ccle) (11, 12). The downloaded raw data were transformed into TPM and a log₂(1+TPM) transformation was performed.

Curated clinical data (n=12,591), including high-quality survival outcomes (overall survival, progression-free interval, disease-specific survival, and disease-free interval), were extracted from a published paper (13).

Expression data for 44 marker genes for three classes of RNA modifiers [i.e., of N1-methyladenosine (m1A) (10), 5-methylcytosine (m5C) (13), and N6-methyladenosine (m6A) (14)] in each sample were extracted from downloaded RNA-seq data. In addition, Spearman correlation was used for the correlation between DEF6 and RNA-modified genes.

We extracted expression data for three classes of immune pathways [i.e., chemokine, major histocompatibility complex (MHC), and receptor] from the downloaded TCGA dataset and analyzed the Spearman’s correlation of DEF6 expression with these genes.

The “ESTIMATE” R package (v1.0.13) calculates stromal, immune, and ESTIMATE scores for each patient based on gene expression (15). We evaluated the scores of each patient’s immune cells using the Timer method via the “IOBR” R package (16).

MuTect2 software processed the level 4 simple nucleotide variation dataset downloaded from TCGA. The tumor mutational burden (TMB) and mutant-allele tumor heterogeneity (MATH) for each tumor were calculated using the TMB and inner heterogeneity functions of the “maftools” R package (ver. 2.8.05), and the TMB and MATH scores were combined with gene expression data (17). We obtained tumor purity and microsatellite instability (MSI) data for each sample from previous studies and merged these with expression data, followed by Spearman’s correlation analysis (18, 19).

We analyzed genetic alterations using the cBio Cancer Genomics Portal (http://cbioportal.org), an open-access resource for interactive exploration of multidimensional cancer genomic datasets (20). In addition, the “Cancer Types Summary” submenu was used to analyze and visualize genetic alteration frequencies. To assess the relationship between DNA methylation and the copy number alteration (CNA) profile of SUSd4, the “mutation” module in the Gene Set Cancer Analyses (GSCA) (http://bioinfo.life.hust.edu.cn/GSCA/#/mutation) was utilized (21).

To construct the protein–protein interaction network (PPI), we used the STRING database (https://cn.string-db.org/) (14, 22–24). As for STRING parameters, the minimum interaction score was 0.15, and the top 50 relative proteins were obtained. The “clusterProfiler” (v4.4.4) R package was used to perform gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis.

For gene set enrichment analysis (GSEA), the GSEA software (v 3.0) was utilized, and samples were grouped by DEF6 expression level (cut-off value 50%) (25). The KEGG symbol matrix was acquired from the Molecular Signatures Database (http://www.gsea-msigdb.org/gsea/downloads.jsp) to evaluate the potential pathways and mechanisms based on gene expression profiles and groupings, in which the minimum and maximum gene set function were set to 5 and 5,000 re-samplings, respectively (26).

Tissue samples were harvested from colorectal cancer patients who underwent surgical treatment at the Department of Colorectal Surgery of the Second Hospital of Harbin Medical University, and a total 20 pairs of samples were obtained. Ethics approval for this study was obtained from the Second Affiliated Hospital of Harbin Medical University. The method of immunohistochemistry was consistent with previous studies (27, 28). We calculated the immunohistochemical (IHC) score as the ratio of positively stained cells to staining intensity. The IHC score can help us to assess the level of DEF6 expression. The DEF6 antibody was purchased from Abcam (ab279395). All procedures performed in this study were in accordance with the Declaration of Helsinki (1964) and its later amendments.

Unpaired Wilcoxon tests were used to calculate the differential expression in normal and tumor samples. The Cox proportional hazards regression model was subsequently established using the coxph module of the “survival’” (v5.6-2) and “survminer”(v0.4.9) R packages to analyze the inner link between gene expression and prognosis. The “forestplot” R package (v2.0.1) was used to map the forest plot. We also calculated the optimal cut-off value of risk score using the “maxstat” R package (version 0.7-25), in which the minimum sample size was set to greater than 30%. We used the Spearman’s correlation method for correlation analysis and the Benjamin–Hochberg (BH) method for p-value adjustment.

Statistical analysis was carried out using R 4.2.0 software and the above-mentioned R packages.

High levels of expression of the gene in cancerous tissues imply a possible association with tumorigenesis or progression. We first analyzed the RNA expression of DEF6 in cancer tissues compared with normal tissues using TCGA data ( Figure 1A ). However, the number of normal tissue RNA-seq data is insufficient; we extracted normal tissue data from the GTEx dataset to pair with TCGA cancers for comparability; as shown in Figure 1B , we found that DEF6 was significantly aberrantly expressed in multiple cancers after supplemental normal tissue data. Next, we analyzed the expression of DEF6 in normal tissues using the GTEx database, and concluded that DEF6 showed a different distribution trend in normal tissues, as seen in Figure 1C , which shows that levels of DEF6 expression were lower in muscle and heart tissues and higher in blood and spleen tissues. Finally, we analyzed trends in the levels of DEF6 expression in cancer cell lines, which were lowest in kidney cancer, neuroblastoma, and liposarcoma cell lines and highest in leukemia, myeloma, and lymphoma cell lines ( Figure 1D ). Among solid tumors, the level of DEF6 expression was highest in head and neck cancer, esophageal cancer, and pancreatic cancer cell lines.

To validate DEF6 as a predictor of cancer prognosis, we calculated the relationship between DEF6 expression and survival by univariate Cox survival analysis. Figure 2 , a forest plot of DEF6 mRNA expression versus overall survival (OS), shows the outcomes of progression-free interval (PFI), disease-free interval (DFI), and disease-specific survival (DSS) outcomes.

Figure 2A shows that a high level of DEF6 mRNA expression is related to poor prognosis in KIRC, LAML, COREAD, and LGG and better prognosis in BRCA, SKCM, CESC, LUAD, BLCA, HNSC, and PCPG. As shown in Figure 2B–D , we found that a high level of DEF6 expression in COREAD was associated with all four of the poorer prognostic outcomes, whereas DEF6 expression in BRCA and CECS was associated with a better prognosis. This implies that DEF6 may have different functions in different cancers, and that in COREAD DEF6 may mainly play the role of oncogene. Next, we utilized algorithms to define an optimal cut-off point to group the samples and performed log-rank survival analysis (29). As shown in Figure 3 , by using the best cut-off value method, it was possible to distinguish survival differences in 31 cancers but not in TGCT. A high level of expression of DEF6 in ACC, BLCA, BRCA, CESC, SKCM, CHOL, DLBC, ESCA, STAD, HNSC, KICH, KIRP, LIHC, LUAD, LUSC, MESO, and UVM was associated with a better prognosis. High DEF6 transcription levels were associated with a worse prognosis in COREAD, GBM, KIRC, LAML, LGG, THCA, OV, PAAD, and PRAD.

We performed an IHC analysis to verify the relationship between DEF6 and colorectal cancer. Supplementary Figure 1A shows an IHC colorectal cancer pathology section and B shows IHC results of paired normal tissues. We found significantly higher IHC scores for colorectal cancer tissue than for normal tissue ( Supplementary Figure 1C ). The IHC score of most cancerous tissues was 4 or 5, whereas the IHC score of most paired normal tissues was 2. We divided the patients into two groups—high and low level of expression—based on IHC scores. As shown in Table 1 , we found that the high level of expression group had a higher positive rate of lymph node metastasis (p=0.031); however, there was no statistical difference at the TNM stage (p=0.081). We also found that higher levels of expression of DEF6 mean higher levels of CA199 and larger tumor sizes. This suggests that DEF6 may be associated with tumor development.

Indicators associated with tumor heterogeneity include MSI, TMB, purity, and MATH. We calculated the correlation between these indicators and DEF6 expression to reveal the possible effects of DEF6 on tumors. Microsatellite instability (MSI) is caused by defects in the mismatch repair system, resulting in hypermutation patterns. The correlation between DEF6 and MSI is demonstrated in Figure 4A . DEF6 expression was significantly negatively correlated with MSI in USC, TGCT, ESCA, and KIRP, but positively correlated with MSI in LUSC, LUAD, COREAD, HNSC, PRAD, STAD, THCA, and DLBC (p < 0.05). As shown in Figure 4B , we found that the association between DEF6 expression and TMB was not strong, showing negative correlations in LAML, GBM, ESCA, and LIHC, and positive correlations only in STAD (p < 0.05). Figure 4C demonstrates the association between DEF6 expression and tumor purity. Tumor purity was associated with immune infiltration, and we found that tumor purity decreased significantly with increased expression of DEF6, which partly proves that DEF6 expression is associated with immune infiltration. MATH is an algorithm to calculate tumor intratumor genetic heterogeneity, and greater heterogeneity is associated with a worse prognosis (30, 31). We evaluated the correlation between DEF6 expression and MATH in Figure 4D . In TGCT, ESCA, THYM, STAD, BLCA, LUAD, THCA, PRAD, and LUSC, higher levels of expression of DEF6 correlated with low MATH scores, whereas, in LIHC, MATH scores increased with higher levels of DEF6 expression.

Gene modifications play a critical role in cancer development, and we evaluated the correlation between DEF6 expression and methylation-modified genes, including m6A, m5C, and m1A.

Figure 5A shows the correlation heatmap between m1A-modified genes and DEF6 expression, and we can conclude that DEF6 expression showed a positive correlation with m1A-modified genes in LIHC, COREAD, HNSC, STAD, KICH, TCGT, ACC, CESC, and ESCA, and a negative correlation with m1A-modified genes in OV, SKCM, THYM, and PCPG. The correlation between m5C- and m6A-modified genes and DEF6 expression was similar to the correlation between DEF6 expression and m1A-modified genes, being mainly positive in LIHC, COREAD, HNSC, STAD, KICH, TGCT, and ACC, and negative in SKCM and THYM ( Figures 5B, C ).

We analyzed the association between DEF6 expression and gene mutations. First, we obtained the mutation data and extracted the copy number variation data of DEF6. We found that DEF6 expression was predominantly associated with heterozygous deletion in KICH, with a very high mutation rate (77.27%), whereas in SARC (26.46%), LUSC (23.95%), BLCA (21.57%), PAAD (33.70%), and KIRC (18.56%), heterozygous deletion also accounted for the majority of mutation types. In other cancers, heterozygous amplification was the predominant form of mutation; interestingly, DEF6 mutations were not found in THCA and LAML ( Figure 6A ). Previously, we analyzed the correlation between RNA-modified genes and DEF6 expression; subsequently, we analyzed the correlation between DEF6 expression and the degree of methylation. Surprisingly, as can be seen in Figure 6B , DEF6 expression showed a significant negative correlation with methylation. By contrast, the correlation between DEF6 expression and copy number variation was mainly positive in HNSC, LUSC, BRCA, BLCA, KICH, OV, PAAD, PRAD, ESCA, and CESC, but negative in LGG and SKCM ( Figure 6C ). We explored the genetic alterations of DEF6 in the TCGA pan-cancer datasets through the cBioPortal online resource. The results revealed that the overall frequency of DEF6 was relatively high in cancers and dominated by “copy number mutations”, with the highest value being found in melanoma, at over 5% ( Figure 6D ).

Since DEF6 deficiency is linked to several immunological disorders, we surmised that DEF6 may also be relevant to genes that regulate the immune system in cancer. Therefore, we analyzed the correlation between DEF6 expression and chemokine, MHC, and receptor genes. As shown in Figure 7A , DEF6 expression showed a strong positive correlation with chemokines by correlation analysis in TGCT, KICH, LIHC, KIRC, KIRP, ACC, SKCM, CHOL, MESO, PCPG, SARC, THCA, GBM, and LGG. CCL5, XCL1, XCL2, CCL17, CCL22, and CCL19 showed a positive correlation with DEF6 in most cancers, and DEF6 may function in regulating these chemokines in cancer. The main functions of the MHC include participation in antigen presentation and processing, which plays an important function in cancer immunity. The results, as shown in Figure 7B , indicate that there was a significant positive correlation between DEF6 expression and MHC regulatory genes. With the exception of ESCA, UCEC, READ, and LUSC, there was a general trend toward a larger number receptor-regulated genes as DEF6 expression increased ( Figure 7C ).

Malignant tumor tissues include not only tumor cells, but also normal epithelial and stromal cells, immune cells, and vascular cells associated with the tumor. Stromal cells are closely associated with tumor growth, disease progression, and tumor resistance. We used the ESTIMATE algorithm, including stromal score, immune score, and ESTIMATE score, to estimate the relationship between SUSD4 expression level and tumor microenvironment (TME) ( Figure 8 ). Figure 8A–C shows the top eight cancers with the strongest correlation between DEF6 expression and ESTIMATE score, immune score, and stromal score. We discovered that the three immune infiltration scores significantly increased in GBM, LGG, SARC, KIRP, KIRC, LIHC, and KICH with an increase in level of expression of DEF6, suggesting that DEF6 expression may be crucial for immune infiltration in these malignancies. Finally, Figure 8D provides an overview of the correlation between DEF6 expression and immune infiltration in 32 cancers.

Previous findings suggest that DEF6 expression may play a key role in cancer, and using enrichment analysis we expect to elucidate the pathways and activities. First, we performed the analysis of the PPI network with DEF6 by STRING in Figure 9A . Using genes related to DEF6 as input for KEGG enrichment analysis, we found that, consistent with predictions, multiple immune-related pathways could be enriched ( Figure 9B ).

Figures 9C, D shows the top 10 terms of GO enriched by PPI genes. We found that DEF6 expression may play a function in regulating the cell cycle, and, as shown in Figure 9C , we enriched a variety of cell cycle-related functions. In COREAD, the level of DEF6 expression was high in cancerous tissues, and a high level of expression was associated with worse prognosis. We grouped DEF6 according to median DEF6 expression and performed GSEA, and the results are shown in Figure 9E–G . We enriched many immune-related pathways, as shown in Figure 9E . DEF6 expression may be involved in the tumor immune process, affecting the development and progression of COREAD through chemokines, immune cell infiltration, etc. Figure 9F shows the enriched cancer-related pathways, and these results suggest that DEF6 expression may indeed be associated with cancer development. We found that the MAPK pathway, the TGF-β pathway, the WNT pathway, and the VEGF pathway were significantly enriched. These pathways are closely related to cancer, and DEF6 expression may interact with or regulate these pathways ( Figure 9G ).