Transcription profiles from high-throughput sequencing fragments per kilobase per million (HTSeq-FPKM) and corresponding clinical information were obtained from the TCGA website (https://www.cancer.gov/about-nci/organization/ccg/research/structural-genomics/tcga), which includes 539 ccRCC samples and 72 normal samples. Nine microarray datasets were also downloaded from the GEO website (https://www.ncbi.nlm.nih.gov/gds/). The data from the GEO databases were translated into log2 values for sequencing analysis. According to a previously reported method, if standardized data were not available, the raw data were downloaded. If a gene was found to have multiple probes in the same chip, the average value of all probes was taken as the expression value of the gene (24). Samples which were neither ccRCC, nor adjacent kidney tissue, were excluded from the present study. The ‘sva’ R package (version 3.6.1) was used to remove the batch effect (25). There were more than 700 specimens obtained from the following GEO datasets: GSE4282, GSE46699, GSE53757, GSE15641, GSE68417, GSE14994, GSE40435, GSE71963 and GSE76351.

A total of 146 paired samples from patients diagnosed with ccRCC were included in the resent study. The patients ranged in age from 20 to 83 years, including 97 males and 51 females. All patients underwent renal resection at Peking University First Hospital between June 2008 and January 2011. Clinical data of the recruited patients was obtained from medical records, such as Fuhrman score and body mass index (BMI). The present study was supported by the Ethics Committee of Peking University First Hospital and written informed consent was obtained from all patients. All procedures were performed according to the World Medical Association Declaration of Helsinki.

Firstly, the tissue samples were fixed in 10% formalin for 24 h at room temperature. Subsequently, the 4-µm paraffin-embedded tissue sections were prepared at room temperature. Immunostaining was performed using a two-step detection kit (cat. no. PV-9000; OriGene Technologies, Inc.) according to the manufacturer's protocol. The sections were deparaffinized in xylene at room temperature, rehydrated in a graded alcohol series and then boiled in citrate buffer (pH 6.0) for 30 min in an autoclave. Endogenous peroxidases were blocked by incubation in 3% H₂O₂ for 30 min at room temperature. The sections were washed in PBS, blocked with 10% goat serum (OriGene Technologies, Inc.) at room temperature for 1 h and incubated with anti-DEF6 (cat. no. ab247011; 1:20,000; Abcam) at 4°C overnight. The sections were washed in PBS solution three times and incubated with a reaction enhancer kit (cat. no. PV-9000; OriGene Technologies, Inc.) for 20 min at room temperature, then washed in PBS solution three times. The sections were then incubated with peroxidase-conjugated secondary antibodies (cat. no. PV-9000; 1:1,000; OriGene Technologies, Inc.) for 30 min at room temperature. All slides were counterstained with DAB solution for 3 min and 20% hematoxylin for 1 min at room temperature and dehydrated. The primary antibody diluent was used as a negative control.

Two experienced independent investigators (ZZ and CX) examined all tumor slides by examining five random fields of view and observing 100 cells per view at ×400 magnification using a light microscope (Olympus Corporation). The staining intensity was classified as 0 (no staining), 1 (weak), 2 (moderate) or 3 (strong); The proportion of stained tumor cells was scored as 0 (0–5%), 1 (6–25%), 2 (26–50%), 3 (51–75%), 4 (>75%). The product of these two variables was used to calculate a final score based on a previous study (26), as follows: 0 (product of 0–3); 1 (product of 4–6); 2 (product of 7–9); 3 (product of 10–12).

HK-2, 293, 786-O, 769-P, Caki-1, ACHN, A498 and OSCR2 were acquired from the American Type Culture Collection. Cells were cultured in DMEM (Invitrogen; Thermo Fisher Scientific, Inc.) or 1640 media (Invitrogen; Thermo Fisher Scientific) containing 10% FCS (Invitrogen; Thermo Fisher Scientific, Inc.) and 1% penicillin-streptomycin (Gibco; Thermo Fisher Scientific, Inc.). Cells were cultured in 10 mm culture dishes in a 5% humidified atmosphere at 37°C.

Total RNA from 20 paired clinical samples was obtained from 20 patients diagnosed with ccRCC by Peking University First Hospital and extracted using TRIzol reagent^® according to the manufacturer's protocol (Invitrogen; Thermo Fisher Scientific). Subsequently, cDNA was synthesized from 5–10 µg of total RNA using the Super Master Mix synthesis kit (Takara Bio, Inc), and the used condition of synthesizing cDNA were as follow: 15 min at 37°C, 5 sec at 85°C and 5 min at 4°C. Then, quantification of all gene transcripts was performed by RT-qPCR using the SYBR Premix ExTaq kit (Takara Bio, Inc.) and α-tubulin was used as a normalizing control. The primer pairs used were as follows: DEF6 forward primer, 5′-TACATGCCCTACCTCAACAAGT-3′ and reverse primer, 5′-TGTTCCCGTTGCTATCTGCC-3′; α-tubulin forward primer, 5′-ACCTTAACCGCCTTATTAGCCA-3′ and reverse primer, 5′-CACCACGGTACAACAGGCA-3′. Each reaction was performed four times and the qPCR conditions used were as follows: 10 min at 95°C, 40 cycles of 15 sec at 95°C and 1 min at 60°C. The 2^−ΔΔCq method was used to calculate the relative gene expression level (27). The relative mRNA expression levels were further normalized using the following formula: Z=(x-μ)/σ, where ‘x’ represents the observation of the sample expression, ‘μ’ represents the sample mean expression, and ‘σ’ represents the sample SD (28).

Total proteins from cells were extracted using the NP-40 lysis buffer (cat. no. P0013F; Beyotime Institute of Biotechnology) and quantified using the BCA method. The supernatant (20 µg of protein) was denatured and separated on 10% SDS-PAGE. Samples were then transferred to 0.22-µm PVDF membranes and blocked in skimmed milk for 1 h at room temperature. After that, samples were incubated overnight at 4°C with the antibodies against DEF6 (cat. no. ab247011; 1:1,000; Abcam) and β-actin (cat. no. sc47778; 1:1,000; Santa Cruz Biotechnology, Inc.). After incubation with peroxidase-coupled anti-rabbit IgG (cat. no. 7074; 1:1,000, Cell Signaling Technology, Inc.) at 37°C for 2 h, bound proteins were visualized using an ECL kit (Pierce; Thermo Fisher Scientific, Inc.) and detected using a DNR Bioimaging System (DNR Bio-Imaging Systems, Ltd.). Relative protein levels were quantified using β-actin as the loading control.

According to the DEF6 median expression in ccRCC samples, the analyzed groups were divided into a low DEF6 expression group (n=269) and a high DEF6 expression group (n=270) (29). Subsequently, the ‘DESeq2’ package and ‘edgeR’ R package (version 3.6.1) were used to screen DEGs between samples. Those with an absolute log2 fold-change >1 and a false discovery rate (FDR) <0.05 in the TCGA or GEO database were considered statistically significant.

STRING version 11.0 (https://string-db.org/) was used to evaluate the PPI information of all DEGs related to DEF6 expression. The PPI network included 108 nodes and 324 edges. Subsequently, Cytoscape (version 3.7.1; http://cytoscape.org/) was used to analyze the PPI network, where a low degree value is correlated with a small node size, and a low co-expression value is related to a small edge size. Furthermore, genes with degrees ≥20 were selected as hub genes, and interactions (combined score >0.4) were considered significant. The interaction network of these proteins was visualized using Cytoscape 3.7.1 and the molecular complex detection (MCODE) plug selected necessary modules were applied (both MCODE score and node number >4) (30).

To further explore the mechanism of action DEF6 in regulating ccRCC, gene ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed using the ‘ggplot2’ R package (version 3.6.1; http://www.rstudio.com/). A P-value of <0.05 was considered statistically significant (31).

Moreover, gene set enrichment analysis (GSEA) was performed to identify the potential biological pathways. GSEA software (version 4.0.3; http://software.broadinstitute.org/gsea/index.jsp) was conducted on the JAVA 8.0 platform. By using the TCGA database, the high group and low group were classified according to the average mRNA expression levels of DEF6. For each analysis, gene set permutations were implemented 1,000 times. Significantly enriched gene sets were identified, which produced FDR q-values <0.05 (32).

The TISIDB database (http://cis.hku.hk/TISIDB) integrated high-throughput data from 988 reported immune-related anti-tumor genes. The database enabled the analysis of correlations for the expression of selected genes with the expression of lymphocytes, immunomodulators and chemokines (33). In the present study, the relationships between the expression levels of DEF6 and lymphocytes as well as immunomodulators across various types human cancers, especially ccRCC, were investigated.

In the present study, unpaired t-tests were used to compare the mRNA expression levels in ccRCC tissues and normal renal tissues from the TCGA and GEO database, using SPSS 26.0 (IBM Corp.). One-way ANOVAs were conducted to calculate the difference among multiple groups, and Bonferroni's corrections were used for the post-hoc tests. Patients with unclear values (NX, MX, GX) or missing values were excluded from subsequent analyses. Paired t-tests were performed to calculate the differences between paired samples and the Wilcoxon signed rank test was used to calculate the differences among IHC scores of clinical samples. The Pearson's correlation analysis was used to calculate the correlation between DEF6 expression and clinical parameters. The Kaplan-Meier method and Cox regression were used to compare the impact of DEF6 expression on the overall survival (OS) of patients alongside other clinical variables in both the TCGA database and the present clinical database. All P-values are based on a two-sided statistical analysis, and P<0.05 was considered to indicate a statistically significant difference.

the DEF6 mRNA expression levels were assessed in ccRCC using the GEO database, the TCGA database, clinical samples collected from the present study and cell lines. Firstly, compared with normal samples, in the ccRCC samples, the DEF6 mRNA expression levels were upregulated in the 9 GEO datasets (Fig. 1A). Furthermore, a forest plot, based on the standardized mean difference, showed the meta-analysis of 9 GEO datasets (Fig. 1E). The outcomes similarly indicated that DEF6 mRNA expression levels in ccRCC samples were upregulated in the 9 GEO datasets compared with the normal samples. In the TCGA database, the DEF6 mRNA expression levels were also increased in ccRCC (Fig. 1B). Furthermore, the DEF6 mRNA expression levels were elevated in 20 paired clinical samples (Fig. 1C). The relative DEF6 mRNA expression levels between the ccRCC tissues and paired normal tissues in each sample is also shown (Fig. 1F). In most samples, the mRNA expression levels of DEF6 in the ccRCC tissues was higher than that of the normal controls. In addition, the outcome of RT-qPCR in RCC cell lines showed that the DEF6 mRNA expression levels were significantly upregulated in the 786-O, ACHN, OSRC-2, and A498 cells compared to the HK-2 cells (Fig. 1D).

Representative staining of tissues (Fig. 2A) and the representative scores (Fig. 2C) are shown. A higher DEF6 expression level was observed in the ccRCC tissue according to the IHC scores of 146 paired ccRCC tissues and corresponding adjacent normal renal tissues. The mean IHC scores of ccRCC tissues and adjacent normal renal tissues were 1.73 and 1.20, which showed significant difference. The frequency distribution of these scores is demonstrated in Fig. 2D. Western bolt analysis also showed that the DEF6 protein expression levels were increased in the ccRCC cell lines compared with the healthy cell lines (Fig. 2B).

To evaluate the prognostic value of DEF6 in ccRCC, the patient OS in both the TCGA database and the present study's clinical database were analyzed using the Kaplan-Meier log-rank test. It was found that high DEF6 expression was correlated with a poor prognosis for patients with ccRCC in both the TCGA database and the present database (Fig. 3A and B). Furthermore, the correlation between DEF6 expression and clinicopathological characteristics in patients with ccRCC was explored in the TCGA database. DEF6 expression was significantly associated with the pathologic grade (Fig. 3C), pathologic T (Fig. 3D), pathologic M (Fig. 3E) and pathologic stage (Fig. 3F).

Whether DEF6 could be an independent prognostic factor for patients with ccRCC was subsequently explored. Cox regression analysis was conducted using both the TCGA database and the present clinical database. Detailed patient information from both the TCGA database (Table SI) and the present clinical database (Table SII) is shown. As shown in Table SI, the clinical characteristics (size, pathologic T, pathologic M and pathologic N) showed a significant difference between the high- and low-DEF6 expression groups in the TCGA database. Meanwhile, as shown in Table SII, the clinical characteristics (size) showed a significant difference between the high- and low-DEF6 expression groups in the present clinical database.

In the TCGA database, univariate Cox regression analysis showed that DEF6 expression, age at diagnosis, tumor size, histologic grade, pathologic T and pathologic M stages were correlated with the OS of patients with ccRCC (Table I). Moreover, multivariate Cox regression analysis indicated that DEF6 expression, age at diagnosis and pathologic M stage were independent prognostic factors for OS (Table I).

Moreover, in the present clinical database, DEF6 expression, tumor size and pathologic T were correlated with the OS of patients with ccRCC (Table II). In addition, multivariate Cox regression analysis indicated that DEF6 expression and pathologic T stage were independent prognostic factors for OS (Table II).

The analysis contained 539 ccRCC samples in the TCGA database, 101 ccRCC samples in GSE40435 and 71 ccRCC samples in GSE53757. A total of 188 overlapping DEGs related to DEF6 expression were identified from 3 datasets, including 180 upregulated genes and 8 downregulated genes (Fig. 4A).

Moreover, GO, KEGG and GSEA analyses were conducted to explore the biological functions of these DEGs. The GO analysis showed that the DEGs were mainly enriched in the activation and regulation of the immune microenvironment (Fig. 4C). Moreover, the KEGG pathway enrichment analysis indicated that the primary functions of the identified DEGs were the regulation of the immune microenvironment. Interestingly, programmed cell death ligand 1 expression and programmed cell death 1 checkpoint pathways in cancer were also enriched in the KEGG pathway (Fig. 4D). Furthermore, GSEA analysis showed that 7 gene sets were upregulated and 4 gene sets were downregulated, with significant enrichment at both the NOM P-value<0.05 and FDR q-value<0.05 (Table III). These gene sets were mainly correlated with the immune response and tumor metabolism (Fig. 4B).

To identify the hub genes of these DEGs related to DEF6 expression, a PPI network was constructed using STRING 11.0 (https://string-db.org/), including 108 nodes and 324 edges. Subsequently, Cytoscape (version 3.7.1) was used to analyze the PPI network, where a low degree value is correlated with a small node size and a low co-expression value is related to a small edge size. The co-expression network showed that protein tyrosine phosphatase receptor type c (PTPRC), integrin subunit beta 2 (ITGB2), lymphocyte-specific protein tyrosine kinase (LCK) and cytotoxic T-lymphocyte associated protein 4 (CTLA4) were hub genes with a degree >20 (Fig. 5A). The MCODE plugin was then used to explore the critical modules of target genes and two imperative modules were identified with MCODE score and node number >4 (Fig. 5B and C). Two imperative modules might act as the core regulatory network for the biological functions.

The aforementioned biological analyses showed that DEF6 was associated with the immune microenvironment. As such, the TISIDB database was used to explore the correlations between DEF6 expression and lymphocytes and immunomodulators, using Spearman's correlation tests. Figs. S1A and 6A showed the correlation between DEF6 expression and tumor-infiltrating lymphocytes, and the 4 types of tumor-infiltrating lymphocytes that significant correlated with DEF6 expression, including myeloid-derived suppressor cells (MDSCs), activate CD8 T cells (Act_CD8), activate B cells (Act_B) and effector memory CD8 T cells (Tem_CD8). Moreover, Figs. S1B and 6B showed the correlation between DEF6 expression and immunoinhibitors, and the 4 significant immunoinhibitors included lymphocyte activating 3 (LAG3), programmed cell death 1 (PDCD1), T Cell Immunoreceptor With Ig And ITIM Domains (TIGIT) and CD96. Furthermore, Figs. S1C and 6C showed the correlation between DEF6 expression and immunostimulators, and the 4 significant immunostimulators included CD27, Lymphotoxin Alpha (LTA), Killer Cell Lectin Like Receptor K1 (KLRK1) and CD48. Figs. S1D and 6D showed the correlation between DEF6 expression and major histocompatibility complexes (MHC) molecules, and the 4 significant MHC molecules included 4 major histocompatibility complexes (HLAs), which are HLA-DOB, HLA-DPB1, HLA-DMA and HLA-DRA. Hence, DEF6 may influence the immune microenvironment by regulating the aforementioned immune molecules.