Young Syrian golden hamsters, 130 to 150 g in body weight, were purchased from Japan SLC (Japan). They were housed five per polycarbonate cage in a clean rack to which pre-filtered air was supplied and maintained at room temperature (22~26℃) under a 12 h light/dark illumination cycle in an approved animal facility at Kangwon National University in Korea. Animals were given a normal diet (Samyang, Korea) and drinking tap water ad libitum throughout the entire experimental period, except during the (4 weeks) administration of the carcinogen in drinking water. All of the hamsters were included in the experiment after one-week acclimatization.

The hamster ChC model was modified according to a previous study. [12] Twenty-nine Syrian golden hamsters were randomly divided into a control and a ChC model group. On the first day of the experiment, the hamsters in the experimental group were infected with 15 metacercariae of liver fluke, Clonorchis sinensis. One day after the parasite infection, they received 15 ppm of dimethylnitrosamine (DMN; Kasei, Japan) in the drinking water for four weeks with a normal diet. Thereafter, the hamsters were given drinking tap water with a normal diet throughout the rest of the study.

To obtain the precancerous liver tissues with dysplastic bile ducts, we sacrificed five control and seven ChC model hamsters at an interim stage of cholangiocarcinogenesis (eight weeks after initiation of the experiment). Thereafter, the remaining five control and 10 ChC model hamsters were maintained for 19 weeks longer (total 27 weeks) to develop cholangiocarcinoma. Necropsy of the hamsters was then performed under ether anesthesia. We grossly examined the liver of each sacrificed hamster and confirmed the liver-fluke infection by identifying adult parasites in the biliary tracts. The experimental design is depicted in Fig. 1.

Hamster liver tissues obtained at eight weeks and 27 weeks after the initiation of the experiment were fixed in 10% buffered neutral formalin for 48 h. After routine tissue processing, the tissues were embedded in a low-melting-point paraffin. Then 3 µm tissue sections were prepared for hematoxylin and eosin (H&E) staining for histological examination and for immunohistochemistry for the TrxR, respectively.

For immunostaining of TrxR, the avidin biotin complex (ABC) method was used. After deparaffinization and hydration, the liver tissue sections were immersed for 30 min in 100% methanol containing 0.3% hydrogen peroxide (Showa, Japan) to block endogenous peroxidase activity. After washing in distilled water and phosphate buffered saline (PBS, pH 7.2), the sections were microwaved in preheated DAKO antigen retrieval solution (pH 6.0) for 15 min at high power, followed by 30 min detergent treatment at room temperature with 0.05% tween 20 in PBS (pH 7.2). After washing in PBS, the tissue sections were incubated in normal blocking serum provided in the Vectastatin Elite avidin-biotin peroxidase complex immunostaining kits. Avidin D and biotin blocking reagents supplied in the Vector Avidin/Biotin blocking kit were applied to the tissue sections according to the manufacturer's instruction to minimize background staining due to endogenous biotin or biotin-binding proteins, lectins, or nonspecific binding substances present within the tissue sections. The sections were then incubated overnight at 4℃ with a primary antibody for TrxR (1 : 500; Upstate Biologicals, USA). For the negative controls, PBS was applied to the sections instead of a primary antibody. The tissue sections were then incubated for 40 min at room temperature with a biotinylated secondary antibody, followed by a 30 min incubation with Vectastatin Elite ABC reagent (Vector, USA) at room temperature. The specific bindings of antibodies within the tissue sections were visualized with 3,3-diaminobenzidine tetrahydrochloride (Dako, Denmark) solution diluted in PBS, and the sections were then counterstained with Mayer hematoxylin. A selected tissue slide, demonstrating a typical positive reaction, was used as a positive control for every batch of TrxR immunostaining.

Two hamsters died at three weeks after initiation of the experiment during DMN treatment. The livers of the hamsters sacrificed during the interim stage of cholangiocarcinoma (eight weeks after initiation of the experiment) were grossly enlarged and yellowish brown in color with multiple small nodules (1~5 mm) on the surface. The common bile ducts and hepatic bile ducts were severely dilated with the adult liver flukes. Histologically, intrahepatic hyperplastic and dysplastic bile ducts of various sizes and shapes were commonly evident in the periportal areas with prominent fibrosis. The hyperplastic bile ducts had a normal structure lined by simple cuboidal bile duct epithelial cells (Fig. 2B), while dysplastic bile ducts were determined based on their atypical duct structures characterized by irregular lumina, multi-layers of epithelial cells, sometimes showing papillary projection, and transformed individual cellular morphology exemplified by taller and larger epithelial cells with larger and heterochromatic nuclei (Fig. 2C). Mitotic figures were quite often evident in the dysplastic bile ducts. Desmoplastic connective tissue with inflammatory cell infiltration surrounded the proliferating hyperplastic and dysplastic bile ducts. Large intrahepatic and hepatic ducts, severely dilated by the presence of adult parasites, showed epithelial cell proliferation with surrounding desmoplastic reaction of connective tissue in which eosinophils, lymphocytes and plasma cells had massively infiltrated. These histopathological findings were consistently noted in all hamster livers sacrificed at the interim stage of cholangiocarcinogenesis. Only one case out of the seven hamsters sacrificed at the interim stage of cholangiocarcinogenesis developed cholangiocarcinoma.

All the hamsters survived up to 27 weeks after initiation of the experiment had developed cholangiocarcinoma in the livers. Histologically, the cholangiocarcinomas that developed were tubular or tubulopapillary cystic types composed of various shapes of tubules or cyst-like structures lined by single to multi-layered tall columnar to cuboidal neoplastic cells and intermittent mucus-producing goblet cells (Fig. 3A-C). The tubules and cysts in general contained relatively large amounts of mucus in their lumina. The tumor cells were aggressively invading adjacent liver parenchyma forming irregular tubular structures or solid sheets of undifferentiated neoplastic cells accompanied by surrounding desmoplastic reaction of connective tissue (Fig. 3A-C). The neoplastic cells had large round nuclei, resulting in a high nucleus/cytoplasm ratio. The supporting connective tissue consisted of bundles of abundant collagen fibers containing interspersed inflammatory cells (Fig. 3A-C).

In the liver tissues at the interim stage of the ChC model, diffuse cytoplasmic overexpression of TrxR was noted in the dysplastic bile duct epithelial cells (Fig. 2c), but the TrxR observed was in general negative or weakly positive in the hyperplastic bile ducts (Fig. 2b). However, at times the immunoreactivity of TrxR in the hyperplastic bile duct epithelial cells was heterogenous and could be negative to strong positive in the same bile duct. In some dysplastic bile ducts, a stronger positive staining was noted at the base of the bile duct cells. The hyperplastic large hepatic bile ducts, containing adult parasites, showed the strongest immunoreactivity for TrxR. The normal bile ducts from the control hamsters were almost negative or very weakly positive for immunoreactivity of TrxR (Fig. 2a).

For all of the tubular types of ChCs that developed, TrxR was diffusely expressed in the cytoplasm of the neoplastic cells (Fig. 3a-c). Compared with the expression level in the dysplastic bile duct cells, TrxR immunoreactivity seemed to be less pronounced in the biliary cancer cells; however, it was still much higher than that observed in normal and type 1 hyperplastic bile ducts. The undifferentiated neoplastic cells forming solid sheets and invading the liver parenchyma totally lost TrxR immunoreactivity in their cytoplasm (Fig. 3c). Type 1 hyperplastic bile ducts showed negative or weakly positive for TrxR, but they were at times strongly positive. Sinusoidal cells showed a strong immunoreactivity for TrxR, while hepatocytes were negative using our immunostaining protocol.

Immunoreactivity of the liver cell components for TrxR is summarized in Table 1.