UALCAN (http://ualcan.path.uab.edu) is a comprehensive cancer data website based on The Cancer Genome Atlas (TCGA) database (14,15). UALCAN was used to obtain data (TCGA-BLCA) on CBX mRNA expression and promoter methylation levels in BLCA tissues and normal tissues (from healthy controls), as well as the association between CBX expression and clinicopathological parameters. The beta value (percentage of methylation signal strength) is the most commonly used quantitative method for methylation levels. Differences in CBX mRNA expression and promoter methylation levels compared with the corresponding controls (from healthy controls) were analyzed by Welch's T-test, and P<0.05 was considered to indicate a statistically significant difference. Differences in CBX expression were compared among individual cancer stages by one-way ANOVA followed by the Dunnett's post hoc multiple comparisons test, and P<0.05 was considered to indicate a statistically significant difference.

Expression analyses were conducted using ONCOMINE (www.oncomine.org). An unpaired t-test was used to examine differential expression between the cancer and normal groups (from healthy controls), and the error detection rate was used as a correction measure of significance (16). Fold change and cut-off P-values were set as 1.5 and 0.01, respectively.

The SV-HUC-1 human ureteral epithelial immortalized cell line was obtained from the American Type Culture Collection and cultured in f12k (iCell Bioscience, Inc.) media with 10% fetal bovine serum (Gibco; Thermo Fisher Scientific, Inc.) at 37°C with 5% CO₂ in humidified air. The 5637 bladder cancer cell line was obtained from the American Type Culture Collection and cultured in 1640 media (Gibco; Thermo Fisher Scientific, Inc.) with 10% fetal bovine serum at 37°C with 5% CO₂ in humidified air. The UMUC3 bladder cancer cell line was obtained from the American Type Culture Collection and cultured in 1640 media with 10% fetal bovine serum at 37°C with 5% CO₂ in humidified air. Media were not supplemented with antibiotics. The culture medium was replaced every 48 h.

Total RNA (from SV-HUC-1, 5637 and UMUC3 cells) was isolated using the total RNA extraction kit (Takara Bio, Inc.). The Bestar™ qPCR RT Kit (Takara Bio, Inc.) was used to synthesize cDNA from isolated total RNA, according to the manufacturer's protocol. Samples were processed in the Applied Biosystems 7500 Real-Time PCR System using TB Green Premix Ex Taq II (cat. no. RR820A; Takara Bio, Inc.). The reaction steps were as follows: i) Pre-denaturation, 95°C for 30 sec; and ii) PCR reaction (40 cycles), 95°C for 5 sec, 60°C for 30 sec. The primer sequences for β-actin and CBXs are shown in Table I. Cycle threshold values were collected to calculate the relative expression of target genes, using the 2^−ΔΔCq method of quantification (17).

SV-HUC-1, 5637 and UMUC3 cells were lysed in NP-40 lysis buffer (cat. no. P0013F; Beyotime Institute of Biotechnology) containing protease inhibitors (Beyotime Institute of Biotechnology). After a 15 min of centrifugation at 15,000 × g at 4°C, the supernatants were collected and protein concentration was assessed using a BCA assay (Thermo Fisher Scientific, Inc.) analyzed in a Multiskan full-wavelength microplate reader (Thermo Fisher Scientific, Inc.). After separation by SDS-PAGE on 10% gels (40 µg protein loaded per lane), proteins were transferred to a PVDF membrane and blocked for 1 h using 5% non-fat dry milk at room temperature. Subsequently, membranes were incubated with primary antibodies overnight at 4°C, and then with the secondary antibody for 1 h at room temperature after three washes in TBS-Tween (containing 0.1% Tween). Primary antibodies were used at the following dilutions: CBX1 (cat. no. 10241-2-AP; Proteintech Group, Inc.) at 1:2,000, CBX2 (cat. no. 15579-1-AP; Proteintech Group, Inc.) at 1:3,000 and CBX7 (cat. no. 26278-1-AP; Proteintech Group, Inc.) at 1:1,000, and GAPDH (cat. no. 2118; Cell Signaling Technology, Inc.) at 1:2,000. The secondary antibody was HRP-linked anti-rabbit IgG (1:1,000; cat. no. 7074; Cell Signaling Technology, Inc.). The protein bands were visualized and semi-quantified using the Gel DOC XR Gel Imaging System and Image lab v5.2 software (Bio-Rad Laboratories, Inc.), respectively.
Human protein atlas (HPA). The HPA (https://www.proteinatlas.org) was used to collect data about protein levels and distribution in BLCA and normal urothelial tissue immunohistochemistry (18). The immunohistochemical images of the expression of CBXs in BLCA and normal tissues (from healthy controls) were extracted from the HPA.

To investigate the prognostic value of CBXs, patient data from the Kaplan-Meier (KM) plotter (http://kmplot.com; dataset name, BLCA; mRNA; 0.05 significance cut-off value) were classified into two groups according to the median expression value of CBX, and the practicality in determining patient prognosis was evaluated. A number of R (version 3.6.1) packages were used for statistical computations and for generating the output graphs (19–22). The ‘survival’ package was used for univariate and multivariate KM analyses (21). Survival curve plots were generated using the ‘survplot’ package. The ‘XML’ and ‘rjson’ R packages were used to load configuration files, the ‘RODBC’ package was used to communicate with the database, and the ‘ggplot2’ package was used to visualize the results (19,20,22). The significance was computed using the Cox-Mantel test to compare two cohorts. The difference between cohorts was numerically characterized by the hazard rate (HR) based on the differential descent rate of the two cohorts. Since the HR is a comparison to the baseline by definition, a relative two-fold drop in one cohort equals a half-fold drop in the other cohorts. Thus, depending on the context, an HR of 2 equals an HR of 0.5. As it is easier to understand an HR above 1 in most cases, an option to invert all HR values below 1 was implemented (23,24). P<0.05 was considered to indicate a statistically significant difference.

GEPIA (http://gepia.cancer-pku.cn/index.html) is a website for gene expression profiling analysis. GEPIA generates a list of genes with similar expression patterns ranked by Pearson correlation coefficient (25). Here, GEPIA was used to select the first 20 genes in BLCA that were similar to the CBX family; a total of 129 genes were identified for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis (26).

The Metascape database (http://metascape.org) was used for enrichment analysis. The express analysis module in Metascape was used for GO and KEGG analysis of CBX and similar genes (26). The relationship between physical proteins captured in BioGrid is the main data source of Metascape (27). Metascape automatically extracts protein-protein interaction networks from these candidates. Then, for each connected network component, the MCODE algorithm is applied to identify the complex of dense connections (28), and Cytoscape is used for visualization (29).

cBioPortal (www.cbioportal.org) is an online open access website resource for analyzing cancer genomics data (30). In the present study, gene mutations of CBXs in BLCA were analyzed using this resource.

TIMER (https://cistrome.shinyapps.io/timer/) is a website for analyzing immune cell infiltration into tumor tissues (31). The present study used the Gene and Survival modules in TIMER to analyze the link between the expression of CBXs and immune cell infiltration, as well as the association of immune cell infiltration with the clinical outcome of patients with BLCA. The ‘Gene’ module was used to explore the correlation between CBXs and the level of immune infiltration in BLCA. Scatterplots were generated, showing the purity-corrected partial Spearman's ρ value and statistical significance. The ‘DiffExp’ module allows the study of differential expression between BLCA tumor and adjacent normal tissues (adjacent tissues from the same patients) for the CBX genes across the BLCA dataset from TCGA. The Cox proportional hazard model (‘survival’ module) was used to explore the clinical relevance of BLCA tumor immune cell subsets and CBX expression levels. The distributions of gene expression levels are displayed using box plots, with the statistical significance of differential expression evaluated using the Wilcoxon test. Up or downregulated genes in tumors compared with in normal tissues (adjacent tissues from the same patients) can be identified for each cancer type.

All collected data were analyzed using SPSS (version 20.0; IBM Corp.) and are expressed as the mean ± SEM. In vitro experiments were repeated ³3 times. The difference between indicated groups was evaluated by Student's t-test, Welch's T-test, Wilcoxon test or ANOVA. Unpaired Student's t-test (parametric) or Wilcoxon signed-rank test (non-parametric) were used to compare the datasets of two groups. One-way ANOVA (parametric) were used to compare the datasets of multiple groups. If three datasets were analyzed, the LSD post hoc multiple comparisons test was performed following ANOVA; otherwise, the Dunnett's post hoc multiple comparisons test was performed. P<0.05 was considered to indicate a statistically significant difference.

The UALCAN, TIMER and ONCOMINE databases were used to evaluate the expression of CBXs in patients with BLCA. Compared with in normal bladder mucosa tissues, BLCA tissues exhibited markedly elevated expression of CBX1 (Figs. 1A and 2), CBX2 (Figs. 1B, 2 and 3), CBX3 (Figs. 1C, 2 and 3), CBX4 (Figs. 1D and 2) and CBX8 (Figs. 1H and 2), and reduced expression of CBX6 (Figs. 1F, 2 and 3) and CBX7 (Figs. 1G, 2 and 3). However, no significant difference was detected for CBX5 (Figs. 1E, 2 and 3) between BLCA and normal bladder mucosa tissues. Moreover, the expression of CBXs in normal SV-HUC-1 bladder cells, and 5637 and UMUC3 bladder tumor cells was evaluated by RT-qPCR. Consistent with the aforementioned results, elevated expression levels of CBX1 (Fig. 4A), CBX2 (Fig. 4B), CBX3 (Fig. 4C), CBX4 (Fig. 4D) and CBX8 (Fig. 4G), and reduced expression levels of CBX6 (Fig. 4E) and CBX7 (Fig. 4F) were detected in BLCA cells compared with in normal SV-HUC-1 cells. These results were also supported by the expression of CBXs at the protein level determined using the HPA (Fig. 5). Additionally, promoter methylation was associated with the progression of bladder tumors. Hypermethylation was detected in the promoters of CBX5 (Fig. 6E), CBX6 (Fig. 6F) and CBX7 (Fig. 6G), whereas hypomethylation was identified in the promoters of CBX1 (Fig. 6A) and CBX2 (Fig. 6B).

The cancer stage was negatively associated with the expression of CBX6 (Fig. 7F) and CBX7 (Fig. 7G), whereas it was positively associated with CBX1 (Fig. 7A), CBX2 (Fig. 7B), CBX3 (Fig. 7C), CBX4 (Fig. 7D) and CBX8 (Fig. 7H). Notably, no association was identified between CBX5 (Fig. 7E) expression and the cancer stages of patients. Hence, it was indicated that CBX1, CBX2, CBX3, CBX4, CBX6, CBX7 and CBX8 expression levels were closely associated the stages of patients with BLCA.

The biological classification of CBXs was investigated via pathway and function enrichment analyses using Metascape. The GO and KEGG analyses predicted the functions of CBX family genes according to the enrichment of CBXs. The enrichment items were divided into four functional groups: Cellular components (Fig. 8A and B), molecular functions (Fig. 8C and D), biological processes (Fig. 8E and F), and KEGG pathways (Fig. 8G and H). CBX and their related genes were enriched for: The ‘mitotic cell cycle’, ‘RNA splicing’, ‘cell division’, ‘PcG protein complex’, ‘chromatin binding’, ‘nuclear matrix’, ‘single-stranded RNA binding’, ‘single-stranded DNA binding’, ‘nuclear chromosome’, ‘spliceosome’, ‘herpes simplex virus 1 infection’, ‘apoptosis’ and ‘amyotrophic lateral sclerosis’.

BLCA patients were ranked according to CBXs expression. Half of the patients with high CBXs expression were included in the high expression group, and the second half of the patients with low CBXs expression were included in the low expression group. Patients with high expression of CBX1 (Fig. 9B), CBX2 (Fig. 9D), CBX5 (Fig. 9J) and CBX6 (Fig. 9L) exhibited poor progression-free survival (PFS), whereas patients with low expression of CBX3 (Fig. 9E) and CBX7 (Fig. 9M) exhibited poor overall survival (OS). Patients with high expression of CBX6 (Fig. 9K) exhibited poor OS. These results suggested that CBX1, CBX2 and CBX7 expression levels were closely related to the prognosis of patients with BLCA. There was no significant difference in OS between patients with high and low expression of CBX1 (Fig. 9A), CBX2 (Fig. 9C), CBX4 (Fig. 9G), CBX5 (Fig. 9I) and CBX8 (Fig. 9O). Additionally, there was no significant difference in PFS between patients with high and low expression of CBX3 (Fig. 9F), CBX4 (Fig. 9H), CBX7 (Fig. 9N) and CBX8 (Fig. 9P).

The protein expression levels of CBX1 were significantly increased in 5637 and UMUC3 cells compared with in SV-HUC-1 cells (Fig. 10B). Similarly, the protein expression levels of CBX2 were significantly increased in UMUC3 cells compared with in SV-HUC-1 cells (Fig. 10C). By contrast, the protein expression levels of CBX7 were significantly decreased in UMUC3 cells compared with in SV-HUC-1 cells (Fig. 10D).

The genetic alterations of CBXs in patients with BLCA were analyzed using the cBioPortal online tool. The results indicated that generally, 48% of patients with BLCA exhibited CBX1-8 genetic alterations. The OncoPrints include missense mutations, truncating mutations, structural variants, amplification, deep deletion, and abnormal expressions. The genetic alteration rates for CBX1-8 were 10, 8, 10, 9, 10, 9, 10 and 8%, respectively (Fig. 11).

Positive correlations were observed between CBX1 and CBX3 expression and the infiltration of dendritic cells, neutrophils, macrophages and CD8⁺ T cells (Fig. 12A and C); CBX2 was positively correlated with macrophage infiltration (Fig. 12B); CBX4 was positively correlated with neutrophil, macrophage and B cell infiltration (Fig. 12D); CBX5 was positively correlated with the infiltration of dendritic cells, neutrophils, macrophages, CD4⁺ and CD8⁺ T cells (Fig. 12E); CBX6 was positively correlated with macrophage, CD4⁺ T cell, CD8⁺ T cell and B cell infiltration (Fig. 12F); CBX7 was positively correlated with the infiltration of macrophages, CD4⁺ T cells and B cells (Fig. 12G); and CBX8 was positively correlated with macrophage infiltration (Fig. 12H). Meanwhile, negative correlations were observed between CBX7 expression and the infiltration of dendritic cells and CD8⁺ T cells (Fig. 12G); and CBX8 with dendritic cell and neutrophil infiltration (Fig. 12H). The present results also found that macrophages, CBX6 and CBX7 were associated with the clinical outcome of patients with BLCA (Table II). However, all correlation values were <0.3 or >-0.3, which indicates that they were weak.