DEN was purchased from Sigma (St. Louis, MO, USA) and freshly diluted in saline to a final concentration before use.

A total of 36 male Syrian hamsters (Slike Experimental Animal Company, Shanghai, China) at 3 weeks of age and weighing 75–100 g, were housed in accordance with the institutional guidelines for animal experimentation and maintained under standard laboratory conditions with a room temperature of 23±2°C, relative humidity of 60±5% and a 12 h/12 h light/dark cycle. All animals were fed with a standard diet and water ad libitum.

The hamsters were divided into two groups. Hamsters (n=12) injected with normal saline and who drank distilled water were used as control. Model hamsters (n=24) received DEN. DEN was injected subcutaneously twice per week for 15 weeks, followed by withdrawal and normal feeding at 16–24 weeks. From 25 to 40 weeks, DEN was added to the drinking water with an initial concentration of 0.05 g/l at 26 weeks and a final concentration of 0.1 g/l until 40 weeks. Normal drinking resumed from 41 to 50 weeks (Fig. 1A). Two animals from each group were sacrificed at 15, 24 and 45 weeks for histopathological examination. All hamsters were sacrificed by the end of 50 weeks (Fig. 1A).

Animals were weighed weekly during the treatment period and monthly thereafter. At the end of the 50th week, all surviving animals were sacrificed with 1% sodium pentobarbital. Liver, lung, spleen, kidney and heart were examined grossly and microscopically for DEN-induced tumorigenicity. The organs were fixed in 10% phosphate-buffered formalin and processed for histological examination with H&E staining. H&E-stained sections from multiple paraffin blocks (16 per animal) for each period were examined by light microscopy. All nodular lesions were diagnosed histopathologically and counted using serial sections.

Tissues were fixed in 10% phosphate-buffered formalin and processed for α-fetoprotein (AFP) (Envison kit; Maixin-Bio Company, Fuzhou, China). AFP immunostaining was performed according to the steps described in the operational manuals. Dilution of the primary antibody was 1:50. Human AFP-positive HCC tissue served as the positive control, while the negative control was diluted with normal mouse serum in PBS solution as a replacement for the primary antibody control.

The isolated tissue was immediately minced (1 mm³) by scalpel under sterile conditions and incubated with 1 mg/ml collagenase IV (C5138; Sigma) for 30 min at 37°C under gentle agitation (Sanyo, Japan) as previously described (12). Next, the cells were passed through a 200-mesh screen and washed twice before spinning at 1,500 rpm for 5 min at 4°C. Cells were cultured in a culture flask (Corning) in DMEM/F12 medium (Hyclone, Beijing, China) with the addition of 10% fetal bovine serum (Gibco), 100 IU/ml penicillin (Amresco) and 50 μg/ml streptomycin (Amresco). The cell culture was monitored with phase contrast microscopy.

Tumor tissue with high proliferating activity was selected under sterile conditions, rinsed with iced normal saline three times and cut into small pieces of 1×1 mm for xenografts. Under direct vision through surgery, a small piece of tumor tissue prepared within 2 h was inoculated in the left lobe of the liver of male Syrian hamster, and sutured and disinfected conventionally. Caesarean section was performed for observation at 7 and 10 days after transplantation (Nikon ECLIPSE TS100).

All of the data were processed by statistical software SPSS 16.0. Results were expressed as the means ± SE. Statistical analysis was tested using one-way analysis of variance. Differences were considered to denote statistical significance at two-tailed p<0.05.

Fig. 1B shows the survival rate of hamsters. A total of 6 animals treated with DEN died between 45 and 50 weeks, whereas unexpected death was not observed in the control hamsters.

No tumors were observed in the control animals. The gross pathological findings of DEN-treated livers included ascites, hepatomegaly, small hepatic cysts and multiple hepatic nodules. Gross examination at 45 and 50 weeks of the DEN treatment demonstrated macronodular HCC in 33.3% (4/12) of the DEN-treated Syrian hamsters (Fig. 2). The livers with HCC were enlarged with multiple small nodules of 0.2–0.5 cm in diameter (Fig. 2A–D), or with a single large nodule of 3 cm in diameter (Fig. 2E). Light microscopy and H&E staining showed nodular HCC (Fig. 3A). There were frequent bizarre neoplastic cells and abnormal mitotic figures (Fig. 3B). In addition, the HCC nodules were accompanied by chronic inflammatory infiltrates (Fig. 3A) and separated by cords of atrophic hepatocytes (Fig. 3A).

We also observed the invasiveness of macronodular HCC (Fig. 4). Invasion of blood vessels (Fig. 4A) and lymph vessels (Fig. 4B) by HCC was evident in hamsters treated with DEN for 45 weeks and onwards. The blood vessels showed congestion, whereas the lymph vessels were markedly dilated with multiple cystic changes. Immunohistochemical staining revealed that approximately one-third of the HCC cells were positive for AFP (Fig. 5), and the immunoreactivity was localized in the cytoplasm and membrane.

The use of primary cultures from tumor cells is a promising approach for the development of new methods for the diagnostics and therapy of cancer. Fig. 6 illustrates the cultured HCC cells on day 1 (Fig. 6A), day 3 (Fig. 6B) and day 7 (Fig. 6C and D). Clone expansion and confluence was time-dependent in this primary cell culture.

Changes in the DEN-treated livers in addition to HCC were chronic inflammatory cell infiltrates, fatty metamorphosis, ballooning degeneration, focal eosinophilic necrosis, hyperplastic nodules and multilocular cysts (Figs. 7–10). Cirrhosis and dilatation of the common bile duct were not observed in any animals.

From 15 weeks onward, DEN-treated animals showed chronic inflammatory infiltrates, dilated lymph vessels and proliferating bile ducts in the portal area (Fig. 7A). In addition, fatty change and clustering of proliferative bile ducts were found adjacent to the central vein (Fig. 7B). Hepatocytes in the DEN-treated hamsters also underwent degenerative changes, such as nuclear vacuolation, ballooning degeneration (Fig. 7C) and eosinophilic necrosis (Fig. 7D).

Nodular hyperplasia and hyperplastic nodules were encountered in the hamsters at 16–24 weeks of the DEN treatment (Fig. 8). The size of the nodular lesions varied from a pinpoint to 1.5 cm in diameter. The nodular lesions included patches of clear cell hyperplasia coexisting with dilated blood vessels (Fig. 8A), focal clear cell hyperplasia accompanying proliferating bile ducts and chronic inflammatory cells adjacent to the central vein (Fig. 8B), and well-defined nodular hyperplasia arising from the background of chronic inflammation and hepatocyte degeneration (Fig. 8C). In addition, clear cell nodules and focal clustering of microcysts often existed (Fig. 8D).

At 25–50 weeks of the DEN treatment, there were focal bile duct hyperplasia (Fig. 9) and fatty changes (Fig. 10) adjacent to HCC lesions. Focal bile duct dysplasia, with marked dilation and cystic formation (Fig. 9A), fatty changes (Fig. 9B), congestion (Fig. 9C) and clear cell hyperplasia (Fig. 9D), often co-existed. Fatty changes were observed in both early (Fig. 10A) and advanced (Fig. 10B) stage of HCC, suggesting an association between fatty metamorphosis and the hepatic carcinogenesis liver.

Gross and microscopic examination revealed lung tumors in 2/4 hamsters at 25–50 weeks of the DEN treatment. The lung tumor mass was most frequently single rather than multiple, and varied from 0.3 to 1.2 cm in diameter. Light microscopy and H&E staining demonstrated well-differentiated neoplastic cells (Fig. 11A) admixed with chronic inflammatory cell infiltrates (Fig. 11B). The morphology of the neoplastic cells in the lung was different compared to that of HCC. Consistently, immunohistochemistry illustrated that only a few macrophages, but not the neoplastic cells, were positive for AFP (Fig. 12), indicating that the lung tumors may not be HCC metastasis. No tumors were detected in the kidney, pancreas, spleen or other anatomical sites.