Detection of bladder cancer using novel DNA methylation biomarkers in urine sediments

Abstract

Background: Bladder cancer (BCa) remains a lethal malignancy that can be cured if detected early. DNA hypermethylation is a common epigenetic abnormality in cancer that may serve as a marker of disease activity.

Methods: We selected 10 novel candidate genes from the most frequently hypermethylated genes detected by DNA microarray and bisulfite pyrosequencing of bladder cancers and applied them to detect bladder cancer in urine sediments. We analyzed DNA methylation in the candidate genes by quantitative methylation-specific real-time PCR (qMSP) to detect bladder cancer in urine sediments from 128 bladder cancer patients and 110 age-matched control subjects.

Results: Based on a multigene predictive model, we discovered 6 methylation markers (MYO3A, CA10, SOX11, NKX6-2, PENK, and DBC1) as most promising for detecting bladder cancer. A panel of 4 genes (MYO3A, CA10, NKX6-2, and DBC1 or SOX11) had 81% sensitivity and 97% specificity, whereas a panel of 5 genes (MYO3A, CA10, NKX6-2, DBC1, and SOX11 or PENK) had 85% sensitivity and 95% specificity for detection of bladder cancer (area under curve = 0.939). By analyzing the data by cancer invasiveness, detection rate was 47 of 58 (81%) in non-muscle invasive tumors (pTa, Tis, and pT1) and 62 of 70 (90%) in muscle invasive tumors (T2, T3, and T4).

Conclusions: This biomarker panel analyzed by qMSP may help the early detection of bladder tumors in urine sediments with high accuracy.

Impact: The panel of biomarker deserves validation in a large well-controlled prospectively collected sample set.

Figure 1

Scatter plot of bisulfite pyrosequencing results of candidate genes in normal bladders (NB), normal leukocytes (NL), 6 bladder tumor cell lines (BTC) and 26 primary bladder tumors (BCa). We used a mixture of normal bladder DNA from 3 persons (2 males and 1 female) as a control (NB). We also analyzed leukocytes because urine sediment DNA contains a high proportion of leukocyte-derived DNA.

Figure 2

Graph of qMSP results of each gene in urine sediments of bladder cancer patients (n=128) and controls (n=110). The relative level of methylated DNA is depicted as 40-dCt[Ct of specific gene – Ct of mC-LESS (internal control)]. A higher 40-dCt represents more methylation of the target biomarker. All 8 genes show significantly more methylation in tumor cases than controls (P<0.0001).

Figure 3

Receiver operating characteristics (ROC) for bladder cancer detection of the combined dataset (TU = 128 and NU = 110). A. ROC curves of the biomarkers sets (2-5 markers) that showed the highest AUC. Detailed information of the best combined markers was summarized in Table 3. B. Detection of bladder cancer in urine sediments by stages. A case was determined positive for methylations of ≥ 3 markers were hypermethylated. C. AUC curves for the 5 marker set in muscle invasive and non-muscle invasive cancers. D. AUC curves for the 4 marker set in muscle invasive and non-muscle invasive cancers.