SMC4, CCNB1 and CKS1B as potential targets and new critical biomarkers for the prognosis of human bladder cancer

As a common malignant cancer of the urinary system, the precise molecular mechanisms of bladder cancer remain to be illuminated. The purpose of this study was to identify core genes with prognostic value as potential oncogenes for the diagnosis, prognosis or novel therapeutic targets of bladder cancer. Methods The gene expression proles GSE3167 and GSE7476 were available from the Gene Expression Omnibus (GEO) database. Next, PPI network was built to lter the hub gene through the STRING database and Cytoscape software and GEPIA and Kaplan-Meier plotter were implemented. Frequency and type of hub genes and sub groups analysis were performed in cBioportal and ULCAN database. Finally,We used RT-qPCR to conrm our results. and the normal samples in our results of RT-qPCR.


Introduction
Bladder cancer (BLCA) is one of the most prevalent cancers among all human malignant cancers all over the world (1). It is gured out that there were approximately 549,000 new bladder cancer patients and 200,000 patients with bladder cancer-related deaths in 2018, making bladder cancer become the tenth most frequent cause of cancer death in both genders worldwide (1). Despite surgery, intravesical treatment, and various adjuvant treatments for BLCA (2,3), it has been reported that it is likely to develop into muscle-invasive diseases and has high recurrence (4,5), due to these, the ve-year survival rate of this cancer type is still unsatis ed. Therefore, precisely targeted therapy is particularly important. Although there have been a lot of reports on the exact targets and mechanisms of the occurrence, development and recurrence of bladder cancer for diagnosis, prognosis and treatment. However, chemotherapeutic resistance and low ve-year survival rate still remains(6-8). The mechanism of tumorigenesis is relatively complex. Genomic and its transcriptional abnormalities and genomic methylation are the causes of tumorigenesis (9)(10)(11). Therefore, it is urgent to explore more reliable biomarkers for diagnosis, prognosis or precise therapy and better understanding the potential mechanism.
Due to the development of gene chip detection technology, more and more studies use these techniques to analyze and mine differential genes. (12). Gene expression omnibus (GEO), a comprehensive database of gene expression, is an open data set of gene expression pro les, which stores numberous kinds of tumor gene expression data (13). Therefore, on the basis of these data, we can explore numberous of available things for new research. Furthermore, many bioinformatics studies on bladder cancer have found a large number of diagnostic and prognostic markers of bladder cancer (14)(15)(16), which indicated that bioinformatical methods could help us to identify critical biomarkers.
In the present study, we downloaded two microarray datasets from the GEO database, namely GSE3167 and GSE7476, which included 50 tumor and 12 normal samples. Secondly, the common DEGs of the above two datasets is obtained by using GEO2R online tool in the GEO databases and Venn diagram tool.
The function of DEGs was analyzed by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis in DAVID database, and the important pathways related to DEGs were studied. The protein-protein interaction network was constructed by using the STRING database, and then the most important modules containing core genes were screened from the PPI network by using the MCODE plug-in in Cytoscape software. Moreover, each core gene was furtherly veri ed the expression patterns through GEPIA online tool. Finally, Kaplan-Meier Plotter was utilized to assess the prognosis of the core genes. Frequency and type of hub genes and Subgroups analysis were performed in cBioportal and ULCAN database, and then used the RT-qPCR to con rm the expression of these genes.

Sample collection
Nine bladder Cancer samples and paired adjacent normal tissues were obtained from the Third A liated Hospital of Sun Yat-Sen University. The information of the patients was list in Supplemental_Table_S1.
All fresh tissues were immediately stored at − 80°C after radical resection. Our studies were approved by the ethics committee of the Third A liated Hospital of Sun Yat-Sen University.

Data source
In our study, gene expression pro le datas of GSE3167 and GSE7476 were downloaded from the NCBI-GEO database (http://www.ncbi.nlm.nih.gov/geo) (17). Microarray data of GSE3167, including 41 bladder cancer samples and 9 normal samples, was based on GPL96 ([HG-U133A] Affymetrix Human Genome U133A Array). Based on the GPL 570 platform ([HG-U133_Plus_2] Affymetrix Human Genome U133 Plus 2.0 Array), the GSE7476 dataset contained 9 bladder cancer specimens and 3 normal specimens in this pro le. The dataset information is shown in Table 1, and the clinical information of the samples in each dataset is shown in Supplemental_Table_S2.

Data processing
DEGs between cancer and normal bladder samples were screened by using GEO2R, an online tool of GEO.
Adjust P value < 0.05 and |log FC| >1 were chosen as a threshold criteria. Next, the raw data were processed by venn diagram software, and the common DEGs in two data sets is selected. The DEGs with log FC > 1 and log FC <-1 were considered as an up-regulated gene and down-regulated genes respectively.

GO and KEGG pathway analysis
The Database for Annotation, Visualization, and Integrated Discovery (DAVID: https://david.ncifcrf.gov/) tools was used to analyse Gene ontology(GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment(18). The cut-off standard of functional and related pathway enrichment analysis is P-value < 0.05. Besides, during the GO enrichment analysis among all the common DEGs, we also showed the top 7 gene counts.
Protein-protein interaction (PPI) network and module analysis. The Search Tool for the Retrieval of Interacting Genes (STRING) online tool ( http://string-db.org) was applied to exhibit the PPI information of the common DEGs (19). The data of PPI information in TSV format were downloaded from STRING to view in Cytoscape software as a cutoff criterion that a maximum number of interactors = 0 and con dence score ≥ 0.4. The most signi cant module of PPI network was screened by the MCODE plug-in in Cytoscape(degree cut-off ≥ 2, node score cut-off ≥ 0.2, K-core ≥ 2, and max depth = 100).
Core gene validation by GEPIA GEPIA (Gene Expression Pro ling Interactive Analysis) dataset (http://gepia. cancer-pku.cn/), an online tool based on The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) data (20), was used to verify the expression of core genes identi ed in the module.( P < 0.05 and |log FC| >1).

Survival analysis
In this study, the prognostic value of the core genes was evaluated on the Kaplan-Meier plotter online tool (http://www.kmplot.com). P < 0.05 indicates that there is a statistically signi cant difference in survival between the high expression group and the low expression group, and the gure also shows the 95% con dence interval hazard ratio (HR). cBioportal analysis CBioportal (http://cbioportal.org) is an open access web-based resource for exploring and visualizing multidimensional cancer genome data sets (21). The bladder urothelial carcinoma (TCGA, Firehose Legacy) dataset, mutations, putative copy number alterations from GISTIC, RNASeq V2 RSEM and protein expression Z-scores was selected(RPPA) were selected for further analyses.

UALCAN dabase analysis
UALCAN is an online network database based on TCGA database and contained RNA-seq and clinical data of 31 cancer types (22). We can use the transcriptional and clinical information in the database to do some correlation analysis of clinical features. In this study, we used this database to analyze the difference of expression between tumor group and its subgroups and normal group.

Statistical analysis
The qPCR data was analyzed by Student's t-test in GraphPad 6.0software. The p-value < 0.05 was considered to have statistical differences.

Identi cation of DEGs in bladder cancers
In the present study, we downloaded two gene expression pro les (GSE3167 and GSE7476) from the GEO database. GSE3167 contained 41 tumor cases and 13 normal cases, and GSE7476 included 9 tumor cases and 3 normal cases respectively (Table 1). Based on GEO2R online tools, P < 0.05 and |logFC|>1, 1520 and 2098 DEGs from GSE3167 and GSE7476 respectively were screened (Fig. 1A-B). Then, venn diagram online tool was performed to select the commonly DEGs in all two datasets (Fig. 1C-D). Finally, a total of 251 genes were identi ed, including 173 genes were signi cantly upregulated and 78 genes were signi cantly downregulated (Supplemental_Table_S3).

GO and KEGG Pathway Analysis in bladder cancers
DAVID software was utilized to analyze the gene ontology and KEGG pathway analysis among all 251 common changed DEGs. The results of GO analysis were divided into three groups: biological process (BP), molecular function (MF) and cellular component (CC). In the BP group, up-regulated DEGs were mainly found in response to drug, positive regulation of transcription, DNA-templated, cell division, cell proliferation, and down-regulated DEGs in muscle contraction, positive regulation of GTPase activity, extracellular matrix organization. Additionally, in the CC group, up-regulated DEGs were mainly associated with cytoplasm, nucleus extracellular exosome, down-regulated DEGs mainly with cytoplasm, extracellular exosome, extracellular space. As for the MF group, the results show that the up-regulated DEGs were mainly enriched in protein binding, ATP binding, poly(A) RNA binding, and down-regulated DEGs in actin-binding, heparin binding, oxidoreductase activity ( Fig. 2A-B,Supplemental_Table_S4).
KEGG pathway analysis was also through DAVID online. The results show that the up-regulated differentially expressed genes were signi cantly gathered in Pathways in cancer, Cell cycle, Biosynthesis of antibiotics, p53 signaling pathway, Oocyte meiosis, while down-regulated DEGs in Vascular smooth muscle contraction, Proteoglycans in cancer, Regulation of actin cytoskeleton, Dilated cardiomyopathy, cGMP-PKG signaling pathway and Melanoma (Fig. 2C-D,Supplemental_Table_S5-6).

PPI network construction and signi cant module selection
To further understand the relationship among identi ed DEGs, STRING online tool and Cytoscape software were used. We ltered 210 genes of the 251 commonly altered DEGs into the DEG PPI network (as presented in Fig. 3A), which possessed 210 genes and 1075 edges Among the 210 genes, there were 144 upregulated genes and 66 downregulated genes.
Then Cytotype MCODE (Molecular Complex Detection) plug-in was used to detect signi cant modules in the PPI network. The results showed that 19 nodes were identi ed from the PPI network which were all up-regulated genes (Fig. 3B).
Core genes expression between cancer and normal bladder tissues GEPIA, the website-based GTEx and TCGA database, was utilized to verify the expression level of the 19 core genes between cancerous and normal people. We found that the expression trends were consistent with the two GEO datasets, and 18 of 19 genes were statistically signi cantly upregulated in bladder cancer tissue compared with normal bladder tissue through analysing RNA-Seq pro les of 28 normal and 404 cancer samples from the GTEx and TCGA database (Fig. 4).
Prognostic value analysis of core genes Survival datas of 18 core genes were analyzed on the Kaplan Meier plotter prognostic analysis platform (http://kmplot.com/analysis). According to the results in prognostic analysis platform, we found that 11 of 18 genes were associated with the prognosis. However, the high expression group for SMC4, TYMS, CCNB1, CKS1B, NUSAP1 and KPNA2 had signi cantly unfavorable OS than those in the low expression group while the other ve high expression groups contribute to favorable outcomes, which is con icted with their high expression level in the tumor (Fig. 5).

Frequency and type of hub genes alterations in BLCA
Next, we use the cBioportal database to verify the mutation types and frequencies of six genes in the above steps. As shown in Fig. 6 and Table 2, the mutation frequencies of SMC4, CCNB1, CKS1B, NUSAP1, KPNA2 are 18.45% 7.28% 5.58% 4.61% 9.47% and 14.32% respectively. There is no doubt that gene ampli cation is the largest type of mutation.

Subgroups analysis of patients with bladder carcinoma
We further used the UALCAN database based on TCGA database to evaluate the transcriptional level of SMC4, and showed that the expression of SMC4 mRNA in bladder cancer tissues was signi cantly higher than that in normal tissues (Fig. 7). Further subgroup analysis of multiple clinicopathological features of 408 bladder cancer samples in TCGA, in the subgroup analysis of the two clinical characteristics of disease staging and lymph node metastasis, the transcriptional levels of six hub genes in bladder cancer patients were signi cantly higher than those in healthy people (Fig. 8).

Quantitative real-time PCR validation
After comparing the expression level of six genes between the tumor and normal tissue, we found that the expression levels of SMC4, CCNB1, and CKS1B were upregulated in tumor tissues in 9 pairs of patients( Fig. 9G-I ). Different from the results in the database, our results showed that there were no signi cant change of TYMS, NUSAP1 and KPNA2 in BLCA tissues compared to the normal samples ( Fig. 9J-L ).

Discussion
Bladder cancer is one of the most common cancers worldwide(1). In recent years, great progress, including surgery, intravesical treatment, and various adjuvant immunotherapy has made to treat the patients(2, 3), unfortunately, these methods have limitations, including tumor recurrence, drug resistance, resulting in poor 5-year survival (4,5). Therefore, it is quite important to exploit new therapeutic strategies and biomarkers for diagnosis, prognosis or precise therapy.
In this study, a series of online database and tools were performed to screen useful diagnostic and prognostic biomarkers. Two expression pro le datasets (GSE3167 and GSE7476) were downloaded from the GEO. There were 50 cancer tissues and 12 normal tissues in our present study. Totally, 251 common DEGs (173 upregulated and 78 downregulated) were identi ed. Then, GO and KEGG pathway enrichment analysis of the 251 common DEGs were performed. In GO enrichment analysis, the DEGs genes were mainly enriched in positive regulation of transcription, cell division, cell proliferation, muscle contraction, extracellular exosome, ATP binding and actin binding. These results suggest that these DEGs are involved in the transcriptional regulation and cell proliferation of bladder cancer cells. In KEGG pathway analysis, results revealed that the DEGs were mainly gathered in Pathways in cancer, cell cycle, biosynthesis of antibiotics, p53 signaling pathway, vascular smooth muscle contraction, proteoglycans in cancer, regulation of actin cytoskeleton. As we all know, the shortening of cell division cycle is accompanied by cell proliferation, which is a remarkable feature of tumor cells. at the same time, the increase of material metabolism is necessary for tumor cell proliferation. The p53 signaling pathway has been shown to play an important role in BLCA, Madka(23) and Wang (24) reported that targeting p53 signaling may inhibit bladder cancer cells growth and metastasis in vivo. Most of the Chemotherapeutic drugs in inducing apoptosis of bladder cancer cells through inhibiting p53 pathway (25,26). Cytoskeleton is involved in the deformation of cancer cells, which may be an important basis for cell migration, invasion and metastasis (27). Many studies have been discovered that dysfunction or inhibition of actin cytoskeleton can inhibit migration and invasion in many cancers(28, 29) including bladder cancer (30,31). Therefore, the study of these pathways will help to clarify the potential mechanism of proliferation, invasion and metastasis of bladder cancer, and help to predict tumor progression.
The protein-protein interaction network complex was established to understand the interrelationship of the DEGs via the STRING online database, and there are 210 genes and 1075 edges in the protein-protein interaction network. Then, 19 up-regulated core genes were screened out by Cytotype MCODE plug-in. We then performed a validation of these 19 hub genes using the GEPIA websites. Compared with normal tissues, the result in the GEPIA box plots showed that only 18 hub genes expression levels were statistically signi cantly overexpressed in bladder cancer samples. Furthermore, we found that 6 of 18 genes were signi cantly associated with the prognosis(P < 0.05). In the 6 quali ed genes, high expression group for SMC4, TYMS, CCNB1, CKS1B, NUSAP1, and KPNA2 had signi cantly unfavorable OS than those in the low expression group, suggesting that the six genes may play a crucial role in bladder cancer for diagnosis, prognosis or precise therapy. In the next mutation type analysis, we found that gene ampli cation is the main type of mutation. Further UALCAN analysis found that in the TCGA database, the expression of tumor group, clinical stage and lymph node metastasis were higher than that of the normal group. However, the expression of TYMS, NUSAP1 and KPNA2 gene remained unchanged between the BLCA patients and the normal cases in our qPCR validation. The possible reason is that our sample size is so small, a larger samples was needed to con rm this conclusion.
As a member of the SMC family, structural maintenance of chromosomes 4 (SMC4), is a chromosomal ATPase which plays an important role in maintaining the chromosome structure during chromosome assembly and segregation. Previous studies have demonstrated that SMC4 was overexpression in multiple cancers and promoted proliferation, migration/invasion and tumorigenicity, including primary hepatocellular carcinoma, colorectal cancer, glioma, breast cancer and lung adenocarcinoma (32)(33)(34)(35)(36). In addition, inhibition of SMC4 could suppress cell proliferation (37). Further vitro studies have been reported that overexpression of SMC4 promotes aggressiveness of cancer cells via activating JAK2/Stat3 and TGFβ/Smad pathway (32,34). Besides, higher mRNA expression of SMC4 was also signi cantly associated with unfavorable OS of some cancers patients, indicating SMC4 took part in the tumorigenesis (33,35). However, to date, the expression pattern or prognostic value of SMC4 in BLCA remains unknown. Thus, we are the rst to report signi cant overexpression of SMC4 in BLCA tissues associated with poor survival in BLCA patients.
Cyclin B1(CCNB1), a monitoring protein initiates the process from the G2 phase to mitosis. Many studies reported that CCNB1 is overexpressed in many tumors and promotes tumor cell proliferation, including hepatocellular carcinomas(38), colorectal cancer (39) and bladder cancer (40). Furthermore, it was also pointed out that inhibition of CCNB1 could reduce cell proliferation (41). Similarly, over-expression of CCNB1 was possessed clinical signi cance in the diagnosis and prognosis of various cancers, such as breast cancer (42), pancreatic cancer (43). Our study also indicated the expression of CCNB1 was upregulated in BLCA tissues and reported their poor prognostic value for the patients.
Cyclin-dependent kinase regulatory subunit 1B (CKS1B), as a member of the Cks/Suc1 family, regulate their function of cyclin-dependent protein kinases through binding the catalytic subunit. Frequent upregulation of CKS1B had been found in multiple cancer tissues and cell lines and was associated with the poor prognosis of these patients (44)(45)(46). Recently, targeting of CKS1B by a variety of microRNAs, including miR-204 and miR-1258 has been found to inhibit the proliferation, invasion and migration of cancer cells (47,48). Moreover, CKS1B has been found to play a critical role in chemoresistance via counteraction of Hsp90 and MEK1/2 or JAK2/Stat3 pathways (49,50). Increased CKS1B gene expression signi cantly correlated with tumor stage and grade, indicating that overexpression of CKS1B contributes to the progression of human bladder carcinoma (51). In the present study, higher CKS1B expression was signi cantly related to poor survival of BLCA patients, suggesting that CKS1B may be exploited as a novel and promising therapeutic target for BLCA treatment.

Conclusion
In conclusion, through an integrated bioinformatics analysis of two gene pro les, we identi ed 251 common DEGs (173 upregulated and 78 downregulated), which contain core genes in bladder cancer pathogenesis. Three of the 19 hub genes including SMC4, CCNB1 and CKS1B were ltered out through our analysis and may be potential biomarkers for diagnosis, prognosis or precise therapy in bladder cancer. Tables   Tables 1&2 are provided  Availability of data and materials All analyzed data related to this paper are included in this paper.
Ethics approval and consent to participate Not applicable.

Consent for publication
All the authors have consented for the publication.

Competing interests
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.    Boxplots showing signi cantly expressed 18 genes in bladder cancer patients compared to normal people( * P < 0.05). Red color means cancer tissues and grey means normal tissues. Kaplan-Meier overall survival curve for 18 core genes in bladder cancer and 6 of 18 genes had a signi cantly associated with the poor prognosis(P< 0.05).     The subgroup analysis of cancer stages and lymph node metastasis in bladder carcinoma( * P < 0.05, **p<0.01, ***p<0.001).