N-Acetylglucosaminyltransferase V and β1-6 Branching N-Linked Oligosaccharides Are Associated with Good Prognosis of Patients with Bladder Cancer

Bladder cancer is the second most common urologic malignancy. An estimated 57,400 new patients were diagnosed with bladder cancer and ∼12,500 patients died from this disease in the United States (1). Bladder cancers tend to occur in two principal forms: low-grade superficial tumors and high-grade invasive cancer. Most superficial tumors are papillary and often multifocal, occasionally progress to invasive disease, and have a good prognosis. Most invasive tumors are nodular, metastasize during the early phase, and have a poor prognosis. Almost 25% of patients with newly diagnosed bladder carcinoma have muscle invasive disease (2), and of these, ∼50% already have occult distant metastasis at the time of diagnosis (3). At present, however, conventional histopathologic evaluation of bladder cancer (tumor grade and stage) cannot predict accurately the behavior of bladder cancers. With a better understanding of the cell cycle, cell-to-cell and cell-to-extracellular matrix interactions, and improved diagnostic techniques, progress is being made to identify and characterize other potential prognostic markers for transitional cell carcinoma of the bladder. This information may then be used to dictate more aggressive treatment for tumors that are likely to progress and less aggressive treatment for those that are unlikely to progress (4).

Glycosylation status in patients with bladder carcinoma has not been studied extensively, except the roles of NeuAcα3Galβ4GlcCer and sialyl Lewis X (5, 6). N-acetylglucosaminyltransferase V (GnT-V), a key enzyme in the formation of branching of asparagine (N)-linked oligosaccharides, is the most strongly linked to tumor metastasis among the many glycosyltransferases (7–9). We and other groups have purified and cloned the cDNA of GnT-V from rat kidney (10) and a human lung cancer cell (11). Among the several patterns of branching, β1-6 branching of N-acetylglucosamine to α-d-6-mannoside enhances metastasis in an experimental cancer model of mice. In colorectal cancer (12) and brain tumor (13), the expression of GnT-V correlated significantly with distant metastasis. In contrast, hepatocellular carcinoma cases with low or no expression of GnT-V were more likely to show recurrence than cases with high expression (14). Moreover, GnT-V expression was inversely associated with prognosis and histology in non–small cell lung cancers (15). Therefore, the clinical implication of GnT-V expression may differ in each kind of cancer. However, there have been no reports describing clinical or pathologic significance of GnT-V in bladder cancer.

Here, we report that the GnT-V expression defined by its antibody correlates inversely to the aggressive potential of bladder cancer. We also show molecular backgrounds of the altered biological behavior of bladder cancer in terms of mRNA of GnT-V and N-linked oligosaccharide structure. To our knowledge, this is the first report showing close relationship between GnT-V expression and clinicopathologic features of bladder cancer.

Bladder cancer specimens. For immunohistochemical detection of GnT-V, 167 paraffin-embedded surgically removed bladder cancer specimens, including 3 precancerous lesions (dysplasia) and 15 normal bladder tissues, were collected. The specimens were fixed with 10% buffered formalin for 12 hours. The paraffin-embedded samples were cut at 3 μm and subjected to H&E staining and immunohistochemistry.

Bladder cancer patients. To evaluate clinical implication of GnT-V expression, specific cause-related survival of the patients was compared according to the results of immunohistochemistry. The cohort consisted of 57 patients with muscle invasive bladder cancer treated with radical cystectomy at the Department of Urology, Akita University (Akita, Japan). The study group included 43 males and 14 females with the mean age of 66.7 years (range, 25-85). The mean observation period was 33 months (range, 2-125). Histopathologic examination revealed 30 pT₂ tumors and 27 pT₃ tumors, with 4 tumors classified as grade 2 and 53 tumors as grade 3. Written consents were taken from the patients whose tumor specimens were subjected to this study. Tumor staging was based on the American Joint Committee on Cancer staging system (16). The tumor grade was classified according to the WHO system.

Immunohistochemistry. Paraffin-embedded tumor specimens were immunohistochemically examined by avidin-biotin peroxidase method using monoclonal antibody against GnT-V. The paraffin-embedded samples were cut at 3 μm and subjected to H&E staining and immunohistochemistry. Immunohistochemical staining of the biopsy specimens was done as described previously (12, 13) using monoclonal antibody against human GnT-V (12). Briefly, deparaffinized specimens were subjected to immunohistochemical staining of anti-human GnT-V monoclonal antibody as the primary antibody. Anti-mouse immunoglobulin antibody conjugated with horseradish peroxidase (Nichirei, Tokyo, Japan) was used as the secondary antibody, and peroxidase activity was visualized with aminoethylcarbazol solution (Nichirei). A control experiment was done omitting the primary antibody from the staining procedure, and no specific staining was found. Based on the staining status of Golgi apparatus, specimens possessing ≥10% positive cancer cells were judged as GnT-V positive. The results of immunostaining were examined by two independent observers without prior knowledge of the clinical status of the patients. The results were clear-cut, and both observers reached complete agreement with the results of each other.

Statistical analysis. χ² test was used to assess the association of the GnT-V status with clinical and pathologic variables. Cause-specific survival was estimated by Kaplan-Meier curves. Differences between groups were evaluated using the log-rank test. We used the Cox proportional hazards regression analysis to test the association of GnT-V expression with other clinical and pathologic variables for prediction of cause-specific survival. We used the statistical program SPSS 12.0 (SPSS, Chicago, IL).

Reverse transcription-PCR. Randomly selected fresh bladder cancer tissues and two samples of normal bladder epithelium were subjected to reverse transcription-PCR. Bladder cancer samples consisted of 15 samples of superficial cancer and 7 samples of invasive cancer. RNA isolation and reverse transcription-PCR were done as described previously (17). PCR with GnT-V primers yields a 291-bp fragment. The sequences of the primers were 5′ GnT-V primer CTCAGCGCCGTACAGGTCAA and 3′ GnT-V primer CTTGATGAAGTCCCGGCAGG.

PCR with these primers was done in a 50 μL volume with 2.5 μL of the prepared cDNA. A master preparation for a set of simultaneous PCR reaction consisted of 10 pmol of each primer, 5 μL of 10× PCR buffer (Takara, Tokyo, Japan), 4 μL of a mixture containing 2.5 nmol/L of each deoxynucleotide triphosphate, 5 μL DMSO, and 0.5 μL Taq DNA polymerase (Takara) per sample. PCR products were electrophoresed in 1.6% agarose gels and stained with ethidium bromide.

Structural analysis of N-linked oligosaccharides. Four samples of normal bladder epithelium and six bladder cancer samples were subjected to structural analysis of N-linked oligosaccharides. For tumor samples, two superficial bladder cancer samples (pT_a) and four muscle invasive cancer samples (pT_2/3) were used. Standard oligosaccharides were prepared from human transferrin and serum as described previously (18–20). α-Mannosidase, β-N-acetylhexosaminidase, β-galactosidase from jack bean, and pyridylaminated isomaltooligosaccharides were purchased from Seikagaku Co. (Tokyo, Japan).

Samples were cut with razor and boiled with 0.5 mL of 10 mmol/L sodium bicarbonate. Then, lipids were removed with 0.5 mL chloroform and methanol (2:1, v/v). Lyophilized samples were suspended in 100 μL of 0.1 mol/L Tris-HCl (pH 8.0). They were treated with trypsin and chymotrypsin, N-glycosidase, and Pronase sequentially and purified with Bio-Gel P-4 (20). Purified oligosaccharides were pyridylaminated (21, 22), and excess reagents were removed by gel filtration on Sephadex G-15 column (1 × 38 cm, 10 mmol/L NH₄HCO₃).

Pyridylaminated oligosaccharides were analyzed and isolated on HRC-ODS column 6 × 150 mm (Shimadzu, Kyoto, Japan) and TSKgel Amide-80 column 4.6 × 250 mm (Tosoh, Tokyo Japan) by NANOSPACE SI-II (Shiseido, Tokyo Japan) high-performance liquid chromatography systems. Elution on the octadecyl silane (ODS) column was done with 10 mmol/L sodium phosphate (pH 3.8) and 1-butanol at 55°C. 1-Butanol concentration was increased from 0.1% to 0.25% (v/v) linearly in 60 minutes. Solvents on the amide column were 3% acetic acid triethylamine (pH 7.3) and acetonitrile. The ratio of the buffer and acetonitrile was gradually changed from 35:65 to 44:66 (v/v) in 30 minutes at 40°C. Flow rates were 1.0 mL/min and oligosaccharides were detected with fluorescence (excitation, 320 nm; emission, 400 nm). Furthermore, isolated oligosaccharides were treated with sequential exoglycosidase digestions and high-performance liquid chromatography analysis was done with standard oligosaccharides at each step of enzyme treatments, respectively (20).

Immunohistochemical detection of GnT-V and its clinicopathologic significance. Normal bladder epithelia (n = 15) were negative for GnT-V. GnT-V-positive staining was clearly detected in dysplasia and low-grade superficial cancer cells showing granular pattern close to the nucleus, which was consistent with intracellular localization of GnT-V at Golgi apparatus (Fig. 1A and B). On the other hand, the majority of high-grade and invasive cancers were negative for GnT-V (Fig. 1C). Table 1 shows the relationship between GnT-V expression and pathologic features. GnT-V expression inversely correlated with pathologic grade and invasiveness (pT).

GnT-V expression correlated to cause-specific survival of patients with bladder cancer as an independent prognosis predictor. The bladder cancer patients with positive GnT-V staining survived significantly longer than those with negative staining (P = 0.0006, log-rank test). Figure 1D shows Kaplan-Meier curves according to GnT-V staining. Furthermore, univariate and multivariate analyses revealed that GnT-V expression was an independent predictor of prognosis of the patient (Table 2).

m-RNA expression of GnT-V is up-regulated in superficial bladder cancer. GnT-V mRNA was positive in all 15 samples of superficial bladder cancer, whereas it was negative in 7 samples of invasive cancer and 2 normal samples, which was consistent with the results with immunohistochemistry.

Structural analysis of N-linked oligosaccharides of bladder cancer revealed that GnT-V products were rich within superficial bladder cancer. Pyridylaminated oligosaccharide mixtures and samples were separated and analyzed in their eluted position on ODS column (Fig. 2). Next, separated peaks of samples were each analyzed and separated on the amide column, respectively. The structures of isolated oligosaccharides were suggested by their elution positions on the ODS and amide column, and their identities were coincided with known oligosaccharide database (20). Their molar ratios were calculated from peak area of high-performance liquid chromatography analysis. In this way, oligosaccharide structures and ratios were determined as shown in Table 3. Average percentage of oligosaccharide of normal bladder epithelium was calculated. The relative amounts for each oligosaccharide structure were plotted against those found in normal bladder (Fig. 3). In comparison to invasive cancer, the superficial bladder cancer samples were rich in structure codes 310.8 and 410.16, for which biosynthesis by GnT-IV and GnT-V, respectively, are responsible. In structural code 410.16, there is a significant difference between superficial cancer and invasive cancer (P = 0.0210, Student's t test) and between superficial cancer and normal bladder epithelium (P = 0.0026). The remarkably increased amounts of GnT-V product are noted.

It is widely accepted that oligosaccharides expressed on glycoproteins play important roles in carcinogenesis, invasion, and metastasis. Among the glycosyltransferases involved in biosynthesis of oligosaccharides, GnT-V is one of the most intensively characterized enzymes. Originally, GnT-V and β1-6 branching N-linked oligosaccharides have been associated with malignant potential of cancer (9). However, subsequent immunohistochemical studies revealed that GnT-V expression does not always have positive correlation with cancer malignancy (14, 15). In the present study, GnT-V-positive expression incidence was high in low-grade/noninvasive (superficial) bladder cancer and precancerous lesion. These findings support a hypothesis that GnT-V may play an important role in early carcinogenesis of bladder cancer as well as hepatocellular carcinoma (14) and non–small cell lung cancer (15). In non–small cell lung cancer, GnT-V expression correlated inversely to patient survival, which was also the case with bladder cancer and hepatocellular carcinoma. It seems that the clinicopathologic implications of GnT-V vary in each type of cancer.

It has been postulated that implications of GnT-V in tumor malignancy depend on whether its original tissue consists β1-6 branching of N-linked oligosaccharides (15). For instance, when cancers originate from the colon, mammary gland, and esophagus epithelium, which showed no expression of β1-6 branching oligosaccharides, GnT-V expression is associated with malignant potential (12, 23, 24). In these tumors, glycoproteins, such as integrins (25, 26), lysosomal-associated membrane protein 2 (25, 27), and matriptase (28) have been shown to be target glycoproteins of GnT-V. On the other hand, the majority of tissues, including lung, express β1-6 branching oligosaccharides. For the patients with cancer that originates from such tissues, GnT-V is an indicator of good prognosis (15). In the present study, structural analysis has proven that there are small amounts of β1-6 branching N-linked oligosaccharides in normal bladder epithelium. This finding supports the hypothesis that GnT-V and β1-6 branching oligosaccharides may contribute to carcinogenesis of bladder cancer by the same mechanism as hepatocellular carcinoma and non–small cell lung carcinoma.

GnT-V expression detected by immunohistochemistry is not always associated with expression of β1-6 branching N-linked oligosaccharides. In the present study, GnT-V was negative in normal bladder epithelium by immunohistochemistry. Reverse transcription-PCR also revealed negative expression of transcripts of GnT-V in normal bladder epithelium. There are considerable discrepancies between these results and the present structural analysis. This may be due to, at least in part, high sensitivity of structural analysis. It is also possible that another β1-6-N-acetylglucosaminyltransferase, GnT-IX (29), may contribute to the formation of β1-6 branching oligosaccharides in normal bladder epithelium. GnT-IX catalyzes the transfer of N-acetylglucosamine to the 6-OH position of the mannose residue of N-acetylglucosamine β1-2-mannose on both α1-3- and α1-6-linked mannose arms in the core structure of N-glycan (29).

Decreased (low) expression of GnT-V may contribute to altered biological properties of bladder cancer as well as non–small cell lung cancer and hepatocellular carcinoma by decreased synthesis of β1-6 branching oligosaccharides of certain target glycoproteins, resulting in a shorter survival of patients having tumors with low GnT-V expression compared with those having tumors with high GnT-V expression. Therefore, the target glycoproteins of GnT-V in the bladder epithelium should be elucidated.

In the present study, expression of GnT-V mRNA had a consistency with the results with immunohistochemistry. Structural analyses revealed up-regulated production of β1-6 branching N-linked oligosaccharides in low-grade/superficial bladder cancer. Furthermore, clinical evaluation revealed the significant role of GnT-V as a survival predictor of the patients with bladder cancer treated with radical cystectomy.

In conclusion, GnT-V and β1-6 branching N-linked oligosaccharides expressions closely relate to low malignant potential in bladder cancer, which can be applied to risk stratification and selection of additional therapy with radical cystectomy for the patients with invasive bladder cancer.