A total of 140 patients with bladder cancer treated at Akita University Medical Center were enrolled in this study. All the patients were histologically diagnosed with urothelial carcinoma of the bladder with specimens obtained from transurethral biopsy or surgical resection. Clinical and histopathological information was obtained from patients’ medical charts, imaging studies and pathological reports. A total of 95 healthy native Japanese men and women over the age of 60 years from Akita Prefecture, who underwent routine community health checkups, were recruited as controls. They were checked by routine urinalysis to rule out urinary tract diseases. The tumor grading system conformed to the World Health Organization grading system (22) and the tumor staging system was based on the American Joint Committee on Cancer TNM classification system (23). This study was approved by the Institutional Review Board of Akita University School of Medicine, and all the subjects provided written informed consent and were asked to provide clinical information, blood and tissue samples for genetic analysis.

Germ line DNA was extracted from peripheral blood lymphocyte (PBL) samples collected from individuals using a QIAamp Blood Kit (Qiagen, Hilden, Germany). Tumor DNA was extracted from bladder tumor tissues obtained at transurethral resection (TURBT). The length of the DATE in the HGF promoter region was determined by a polymerase chain reaction (PCR). The PCR primers (forward, 5′-GGGACAGGTATTGTGGGGCCAAAATAAG-3′; and reverse, 5′-GGGTGTGGTATTGTGGGGCCAAAATAAG-3′) generated a 247-bp product when the length of DATE was 30 deoxyadenosines. PCR reactions were performed in a 25 μl volume containing 20 ng of genomic DNA, 1X PCR buffer (PE Applied Biosystems, Branchburg, NJ, USA), 0.2 mM of each dNTP, 2.5 mM MgCl₂, 50 pM of each primer, and one unit of Ampli-Taq Gold DNA polymerase (PE Applied Biosystems). The PCR amplification conditions were as follows: 10 min at 94°C, followed by 35 cycles of 30 sec at 94°C, 30 sec at 62°C and 60 sec at 72°C, with a final extension of 10 min at 72°C in a thermal cycler (GeneAmp PCR System 9700, Perkin-Elmer, Branchburg, NJ). The samples were purified using the Illustra AutoSeq G-50 Kit (GE Healthcare Life Sciences, Little Chalfont, UK). All purified fragments were subjected to DNA sequencing with forward and reverse primers using the BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Foster City, CA, USA), and analyzed with a Genetic Analyzer (ABI PRISM 310 Genetic Analyzer, Applied Biosystems). Because we could clearly define only the length of the shorter DATE owing to the direct sequencing method, the data of the shorter DATE length were used in further statistical analyses (Fig. 1A). The short DATE was defined as 27 deoxyadenosine repeats or fewer, while the long DATE was defined as more than 27 repeats, as described in the Results.

Total RNA was extracted from approximately 30 mg of tumor tissues obtained by TURBT using TRIzol RNA Isolation Reagent (Life Technologies Inc., Rockville, MD, USA) and was reverse-transcribed to cDNA using a SuperScript VILO cDNA Synthesis Kit (Life Technologies Inc.).

Real-time RT-PCR amplication mixtures (20 μl) contained 2 μl template cDNA, 2X SYBR-Green Master mix buffer (10 μl) (Takara Bio Inc., Otsu, Japan) and 100 μM forward and reverse primers (0.8 μl). Reactions were run on a Thermal Cycler Dice Real-Time System (Takara Bio Inc.) with cycling conditions of 30 sec at 95°C, 40 cycles at 95°C for 5 sec, and 60°C for 60 sec. The primers used were as follows: 5′-ATGATGATGCTCATGGACCCT-3′ (forward) and 5′-CTGGCAAGCTTCATTAAAACC-3′ (reverse) for HGF and 5′-TGATGACATCAAGAAGGTGGTGAAG-3′ (forward) and 5′-TCCTTGGAGGCCATGTGGGCCAT-3′ (reverse) for GAPDH. PCR reactions were simultaneously performed for HGF and the reference GAPDH. The HGF expression levels normalized by the GAPDH were used for further analyses.

Ninety-one bladder cancer specimens obtained at radical cystectomy were subjected to immunohistochemical analysis. The specimens were fixed in 10% buffered formalin and embedded in paraffin. Serial sections (5-μm thick) were deparaffinized in xylene, rehydrated in a graded series of decreasing ethanol concentration, and then rinsed in Tris-buffered saline. For HGF immnostaining, HGF was detected with a goat polyclonal anti-human HGF antibody (1:50; R&D Systems, Minneapolis, MN, USA) as primary antibody. Antigen retrieval treatment was performed at 121°C for 10 min in 10 mM sodium citric acid (pH 6.0), and endogenous peroxidases were blocked using 0.3% hydrogen peroxidase/methanol for 30 min. After washing in phosphate-buffered saline (PBS) for 5 min and 10% bovine serum albumin/PBS for 30 min, the sections were exposed to a primary antibody diluted 1:50 overnight at 4°C. After washing in PBS, a secondary antibody conjugated with anti-goat IgG was applied, followed by incubation at room temperature for 30 min. Immunoreactions were visualized using 3,3′-diaminobenzidine as chromogen. Positive control was represented by a normal hepatic cell showing strong cytoplasmic expression for HGF.

To assess HGF immunoreactivity, we used the following scoring systems. Cytoplasmic HGF staining intensity was scored on a semi-quantitative scale as: 1, weak; 2, moderate; 3, strong; and 4, very strong. The percent of cytoplasmic HGF-positive cells was divided into four groups: 1 (<25%), 2 (25–50%), 3 (50–75%) and 4 (>75%). Total immunoreactivity was finally calculated by multiplying the two scores and was classified into two groups: low expression (≤9) and high expression (>9) (Fig. 1B).

The data was analyzed using SPSS software (version 19.0J, SPSS Inc., Chicago, IL, USA). Two group comparisons were performed with the Mann-Whitney U test for continuous variables and χ² test for categorical variables. A probability of <0.05 was required for statistical significance. Kaplan-Meier survival curves with the log-rank test were used to compare overall survival and disease-specific survival between patients with high and low HGF expression.

We evaluated the length of the DATEs in PBLs of 140 bladder cancer patients and 95 healthy controls. Distributions of the DATE lengths in bladder cancer patients and controls were as follows: in bladder cancer patients, the frequency of 30, 29, 28, 27, 26, 25, 24 and 22 doxyadenosine repeats in the DATE were 3 (2.1%), 15 (10.7%), 52 (37.1%), 48 (34.3%), 13 (9.3%), 5 (3.6%), 2 (1.4%) and 2 (1.4%), respectively. In the controls, the frequency of 30, 29, 28, 27, 26 and 25 doxyadenosine repeats in the DATE were 4 (4.2%), 32 (33.7%), 36 (37.9%), 18 (18.9%), 4 (4.2%) and 1 (1.1%), respectively. Because the median length of DATE was 28 in both groups, we classified the length of DATE into two categories as short (≤27) and long (>27). When we applied our classification, the frequency of individuals with a short DATE was significantly higher in bladder cancer patients than controls (49.3 vs. 24.2%, P<0.001) (Table I).

We assessed the difference in the frequency of short DATEs between bladder tumor tissue and PBLs. In addition, we evaluated the association between the clinicopathological factors of bladder cancer and the presence of a short DATE in bladder tumor tissue. The patients comprised 55 males and 15 females with a mean age of 73.2 years. There were 51 non-muscle invasive tumors (NMIT) (pTa-1) and 19 muscle invasive tumors (MIT) (pT2), and 42 low grade tumors (grade 1 or 2) and 28 high grade tumors (grade 3). The frequency of patients with a short DATE in their PBLs was 41.4% (29/70), while that in their bladder tumor tissue was 52.9% (37/70) (Table II). The frequency of patients with a short DATE in their bladder tumor tissue was significantly higher than that in their PBLs (p=0.008) (Table II). In 25 (35.7%) patients, the length of the DATE in their bladder tumor tissues was shorter than that in their PBLs. In 12 (17.1%) patients, the length of the DATE in their bladder tumor tissues was longer than that in their PBLs. The frequency of the mutation to long DATE was significantly higher than that to short DATE (p=0.047 by binominal test). The results suggested that the DATE was a frequent target of somatic mutation (37/70, 52.9%) in bladder cancers and the mutation from long DATE to short DATE might be dominant and play a role in the progression of bladder cancer. There was no significant difference between NMIT and MIT in the frequency of patients with a short DATE either in PBLs or bladder tumor tissue (Table II). No significant association was observed between the presence of a short DATE in PBLs and tumor grade, while tumor grade was significantly higher in bladder tumor tissue with a short DATE than those without a short DATE (p=0.015, Table II).

We compared HGF mRNA levels between bladder tumor tissue with and without a short DATE. MIT showed significantly higher HGF mRNA expression levels (p=0.019) than NMIT (Fig. 2A). High grade tumors showed significantly higher HGF mRNA expression levels than low grade tumors (p=0.020) (Fig. 2B). HGF mRNA expression was significantly higher in bladder tumor tissue with a short DATE than that without a short DATE (p<0.001) (Fig. 2C).

We evaluated the relationship between HGF immunohistochemical expression in bladder cancer and clinicopathological factors. Overall and disease-free survival rates were compared according to the HGF expression scores in 91 patients who underwent radical cystectomy. The patients comprised 70 males and 21 females with a mean age of 69.8 years. The mean follow-up duration was 19 months (range 0–98). Low and high HGF expression scores were observed in 61 (67.0%) and 30 (33.0%) patients, respectively (Table III). The high HGF expression group had a significantly worse overall survival than the low HGF expression group (p=0.012). Furthermore, although not statistically significant, the high HGF expression group had a tendency towards worse disease-free survival than the low HGF expression group (p=0.056) (Fig. 3).