DNA Methylation profiles as predictors of recurrence in non muscle invasive bladder cancer: an MS-MLPA approach

Tissue samples from 74 patients (65 males, 9 females) submitted to transurethral resection of primary bladder cancer at the Department of Urology of Morgagni-Pierantoni Hospital in Forlì between 1997 and 2006 were used for the study. All samples were retrieved from the archives of the Pathology Unit of the same hospital. Median age of patients was 73 years (range 39–92): 31 were <70 years and 43 ≥70 years. On the basis of 2004 World Health Organization criteria, final diagnosis was low grade non muscle invasive bladder cancer (NMIBC) in 55 patients and high grade NMIBC in 19 patients. At a median follow up of 5 years 38 patients were still disease-free and 36 had experienced one or more episodes of local recurrence. In this retrospective study, the two subgroups (disease free or relapsed) of patients were equally distributed for sex, age, grade and stage (Table 1).

All patients gave written informed consent for biological samples to be used for research purposes. The study protocol was reviewed and approved by the ‘Area Vasta’ Istituto Scientifico Romagnolo per lo Studio e la Cura dei Tumori (IRST) Ethics Committee.

Five 5-μm-thick sections were obtained from each paraffin-embedded block. Macrodissection was performed on hematoxylin-eosin stained sections and only cancer tissue was used for DNA isolation. Genomic DNA was purified using QIAmp DNA FFPE Tissue (Qiagen, Milan), according to the manufacturer’s instructions.

DNA was also isolated from a human bladder cancer cell line (HT1376) using Qiamp DNA minikit (Qiagen, Milan, Italy), according to the manufacturer’s instructions.

MS-MLPA was performed using at least 50 ng of genomic DNA dissolved in 1XTE buffer (Promega, Madison, WI, USA). DNA isolated from HT 1376 cell line was used as internal control for MS MLPA analysis (Figure 1). The methylation status of 24 tumor suppressor gene promoters was analyzed using the ME001C1 kit (MRC-Holland, Amsterdam, The Netherlands) (Table 2). Two different probes that recognize two different sites of the promoter region were used for genes RASSF1 and MLH. We excluded CDKN2B gene from the analysis because its probe is sensitive to improper Hha1 digestion in FFPE samples. In brief, DNA was denatured (10 min at 98°C) and cooled at 25°C, after which the probe mix was added to the samples and hybridization was performed by incubation at 60°C for 16–18 h. The reaction was divided equally in two vials, one for ligation and the other for ligation-digestion reaction for each tumor. We added a mix composed of Ligase-65 buffer, Ligase-65 enzyme and water to the first vial and a mix of Ligase-65 Buffer, Ligase 65 enzyme, Hha1 enzyme (Promega, UK) and water to the second. The samples were then incubated at 49°C for 30 min. At the end of the ligation and ligation-digestion reactions, samples were amplified by adding a mix of PCR buffer, dNTPs and Taq polymerase. The PCR reaction was performed under the following conditions: 37 cycles at 95°C for 30 sec, 60°C for 30 sec and 72°C for 60 sec. The final incubation was performed at 73°C for 20 min.

Amplification products were analyzed by ABI-3130 genetic Analyzer (Applied Biosystem, UK). Universally methylated and unmethylated genomic DNA was used as positive or negative control, respectively.

Electropherograms obtained were analyzed using Gene Mapper software (Applied Biosystem, UK) and the peak areas of each probe were exported to a home-made excel spreadsheet. In accordance with the manufacturer’s instructions, we carried out “intrasample data normalization” by dividing the signal of each probe by the signal of every reference probe in the sample, thus creating as many ratios per probe as there were reference probes. We then calculated the median value of all probe ratios per probe, obtaining the normalization constant (NC). Finally, the methylation status of each probe was calculated by dividing the NC of a probe in the digested sample by the NC of the same probe in the undigested sample, and by multiplying this ratio by 100 to have a percentage value, as follows:

MS-MLPA technique reproducibility was assessed by performing three independent methylation profile analyses on a bladder cell line (HT1376). The methylation level for each gene was found to be the same in each experiment.

We considered the promoters showing a ratio ≥0.20 as methylated, while those with a ratio <0.20 were regarded as unmethylated. The cut-off was chosen on the basis of experiments performed on the bladder cancer cell line (HT1376) and on data from the literature [21, 22]. We have also performed the analysis on some samples from healthy tissues, to confirm that the background noise was inferior to 0.20 cut-off, such excluding false positive results due to experimental procedure.

The student’s T test was used to assess the methylation index (MI), which was considered as a continuous variable. Logistic regression analysis was performed using the Epicalc of R to evaluate the performance of a panel of gene promoters (HIC1, RASSF1 and GSTP1) in discriminating between recurrent and non recurrent patients. We created logistic regression models with methylation levels of the three gene promoters (HIC1, RASSF1 and GSTP1). Probabilities were calculated as follows: P = exp ((Σ(b_ix_i) + c)/(1 + Σ(b_ix_i) + c), where p is the probability of each case, i = 1 to n; b is the regression coefficient of a given gene, x is the log2-transformed methylation level and c is a constant generated by the model. The ROCR package was used to obtain the ROC curves of the models and area under the curve (AUC) values. Recurrence-free survival was analyzed with the Log-rank test using SAS 9.3 software. All the molecular analyses were performed in a blind manner.