Tissue Samples

Fresh samples of cancerous bladder tissue were collected from 19 patients during cystoscopy, at King Faisal Hospital and Research Centre Urology Clinic, before the initiation of any kind of treatment. As control, 5 corresponding normal appearing tissue samples, adjacent to tumour were also obtained for the study. Patients with the history of cystitis and tumour other than bladder region were excluded. Fragments from each tissue sample were embedded in paraffin and stained with H&E stain. The clinicopathological parameters were confirmed by the pathologist. Tumours were staged according to TNM classification system. Detail demographic and clinic-pathologic parameters for all the samples are listed in Table 1. The collected samples were stored at -80^oC until use. All samples were collected after taking informed consent. Ethical approval was obtained from Office of Research (ORA) of King Faisal Specialist Hospital and Research Centre (KFSHRC).

DNA Extraction

The DNA was extracted as per protocol of the Qiagen tissue kit. The quality and quantity of the extracted DNA was confirmed by measuring 260/280 ratios on spectrophotometer (ND-1000; Nano Drop Technologies).

Selection of Genes

In our study, significant difference in DNA hypermethylation levels was observed at APC, BRCA1, CDH1, CDH13, CDKN2A, DAPK1, ESR1, FHIT, MGMT, RASSF1, SOCS1, TIMP3, TP73, VHL, WIF1 between controls and cancerous samples.

It was also observed that CADM1 and DKK3 are differentially methylated in NMIBC versus MIBC samples. In contrast, no significant difference in DNA methylation levels at any gene promoter was found to be associated with tumour grade.

Evidence suggests that these altered DNA methylation patterns are dynamic and respond to environmental influences [17]. To our knowledge, this is the first study which has analysed the methylation patterns in BC, using a panel of 48 candidate genes loci in 19 bladder cancer patients from the Saudi population.

Many studies have identified RASSF1 and WIF1 as candidate tumour suppressor gene which epigenetically gets inactivated during various human carcinogenesis [18]. Their hypermethylation has also been reported in BC which is significantly higher than healthy controls [19]. Recently, the RASSF1 gene has been highlighted as most commonly hypermethylated genes in tumours [20]. Likewise, our study reveals the significant difference in hypermethylation patterns in both these genes between cancerous and control samples; however the hypermethylation levels were much higher in controls which was an unexpected finding. As the controls were taken from the adjacent normal tissue of matched cancerous samples. This suggests that genomic changes and DNA methylation changes in bladder tissue  may occur before the morphological changes [21]. Aberrant methylation patterns in the control samples suggest to be a characteristic of a premalignant lesion as bladder cancer is a multifocal tumour with high recurrence index [22]. A genome-wide DNA methylation study on upper urinary tract urothelial cancer also noted increased DNA methylation in non-cancerous cells at numerous CpG sites without histological changes, suggesting that field effect of tumour agents can induce transformation in neighbouring cells [23]. Thus, this verifies the hypothesis that DNA methylation changes in cancers may pre-exists in normal looking cells and it can be beneficial for the early detection of BC even in the Saudi population. 

TP73 gene was found to be significantly hypermethylated tumour suppressor in this study. This gene belongs to the TP53 family of transcription factor genes and it leads to apoptosis in response to DNA damage [24]. TP73 has been particularly well studied in breast tumours [25]. Overexpression of TP73 has also been reported in bladder and prostate cancers [26]. Interestingly, we showed higher levels of hypermethylation at this gene locus among BC samples of Saudi origin compared to controls. 

TIMP3 gene, encodes the inhibitor of the matrix metalloproteinase, a peptidase involved in degradation of the extracellular matrix [27]. Significant difference in the hypermethylation status of this gene locus found in this study can be considered a significant finding in relation to Saudi Arabia, since previous reports suggested links between hypermethylation of TIMP3 gene with schistosome-infected bladder cancer [28].

Although, our results have confirmed previous studies showing varied hypermethylation frequencies among different genes in BC. Interestingly, one novel tumour suppressor genes ESR1 is also notified showing significant difference in DNA hypermethylation status in BC tissue samples as compared to controls. Hypermethylation of this gene has not been reported earlier in the literature in relation to BC and can be considered specific for the Saudi population. ESR1 encodes the estrogen receptor 1, this protein influences the growth and differentiation function of different tissues in response to estrogen. This gene shows high frequency of genetic variation in many types of human malignant tumours especially in breast tissues [29] but no link with BC was reported earlier.

Several studies have identified methylated genes and panel of loci that are associated with the bladder cancer invasion, such as IPF1, GALR1, TAL1, PENK, TJP2, HOXB2, RASSF1A and RARβ genes [30]. Interestingly, current study shows for the first time that differential hypermethylation of CADM1 and DKK is associated with MIBC.

The CADM1 gene (also known as TSLC1 or IGSF4) codes for cell-to-cell adhesion in Ca+ independent manner. It has been generally investigated as a tumour-suppressor gene in prostate, oesophageal, nasopharyngeal, non-small cell lung and cervical cancers [31,32]. CADM1 is down regulated frequently in many cancers via promoter hypermethylation [33]. However, relevance of its hypermethylation in BC prognosis is not well known. A significant increase in DNA methylation levels of CADM1 among MIBC observed in our study supports its possible down regulation in invasive bladder tumours.

DKK3 encodes a protein that is a member of the dickkopf family. The secreted protein through its interactions with the Wnt signalling pathway is involved in embryonic development. Variety of cancer cell lines have shown decreased expression of this genes and it may function as tumour suppressor [34]. It has been shown previously that hypermethylation of DKK3 in cervical cancer cell lines can lead to activation of Wnt/β-catenin pathway and increased tumorigenesis [35]. Therefore, it would be interesting to decipher its role in invasive bladder cancer. A study with larger sample size can determine its value in prognosticating BC in the Saudi population.

There have also been studies identifying the role of DNA methyltransferase 3B (DNMT3B) in the tumourigenesis of BC through microRNA (miR)-34a promoter methylation, through its downregulation [36]. Matrix Metalloproteinase 11 (MMP11) is also a differentially expressed gene in BCs due to its promoter hypomethylation. MMPs are extracellular enzymes that break down the extracellular matrix [37]. Methylation status of genes such as SOX-1, IRAK3 and LI-MET were shown to be more useful in determining tumour progression than cystoscopy. Furthermore, hypermethylated TIG1, GSTP1 and APC were associated with worse survival rates [14]. Hypomethylation, and consequently downregulation of the promoter region of the AJAP1 gene (which codes for adherens junction-associated protein 1) by the ten-eleven translocation 1 (TET1), indicates poor prognosis due to subsequent activation of the β-catenin signalling pathway leading to BC pathogenesis [38].

Recently, it has been shown that cell free DNA in the urine can be used as a sensitive and specific biomarker for detection of bladder cancer [39,40]. Studies from Rose et al and Wu et al have demonstrated several DNA methylation based urinary biomarkers including ITIH5, ECRG4, HOXA9, PCDh17, POU4F2 and ONECUT2  can be used for diagnosis of bladder cancer [39,40].

In conclusion, in this study we have identified distinctive genes that show evidence of differential methylation between cancerous and normal tissue, non-invasive and invasive cancer tissue specific for the Saudi population. This is  a pilot study where we were able to screen DNA methylation of status of genes which can be used as potential biomarkers for BC diagnosis and prognosis  However, larger studies on BC-associated changes in DNA methylation are required for developing sensitive and specific biomarkers for diagnosis and prediction of the disease. Recent studies have shown that cell free tumour derived DNA in urine and blood samples can be easily detected using specific DNA methylation patterns.  Since the DNA methylation patterns are tumour specific and population specific so identification of DNA methylation patterns specific to Saudi population of bladder cancer is an important first step to develop distinct DNA methylation-based biomarkers specific for the Saudi population.

This is a pilot study to screen out potential DNA methylation markers which can used in future for detection and prognosis of bladder cancer on a large representing sample size of Saudi population. As we were screening 48 tumor suppressors we needed large quantity of good quality DNA (4 microgram) from our biopsy samples to conduct our experiments, which was not available from small size of the normal bladder biopsies. This is the reason we had only 5 control samples taken from the adjacent normal tissues of the bladder cancer patients.    However, we were able to compare the DNA methylation status for 48 genes in three different group which was quite extensive and conducted for the first time in this region. We will utilise data from this study to select significantly differentially methylated genes with larger sample size to develop sensitive and specific tumour markers for BC in future.