**3. F344-congenic analbuminemic rats (F344-alb rats)**

Although NARs are derived from SD rats, the SD rats consist of out-bred strains. Therefore, SD rats may be genetically heterogeneous. We first tried to transplant SD rat hepatocytes into an NAR liver without immunosuppressants. However, this experiment was unsuccessful, most likely because the SD rats were immunogenetically heterogeneous. Cocultured spleen cells from NARs and SD rats displayed a higher rate of cell proliferation compared to cells in NAR/NAR or SD/SD cultures (Yokota & Ogawa, 1978). These results indicate that the immunological rejection may occur after the transplantation of the SD hepatocytes into the NAR liver.

In addition to the variety of abnormalities in serum proteins, lipids and hormones, NARs differ from SD rats in their susceptibility to tumorigenesis induced by chemical carcinogens in various organs. Notably, NARs are highly susceptible to urinary bladder tumors that are induced by *N*-butyl-*N*-(4-hydroxybutyl)nitrosamine (BBN) (Kakizoe et al., 1982) and less susceptible to hepatocarcinogenesis induced by diethylnitrosamine (DEN) and 2 acetylaminofluorene (2-AAF) (Asamoto et al., 1989). It is not clear whether the difference in the susceptibility to chemical carcinogens is the result of the analbuminemia in NARs or of differences in the genetic backgrounds of NARs and SD rats. To reduce the genetic variability, Takahashi et al. (Takahashi et al., 1988) established F344-alb rats with the genetic background of F344 rats. F344-alb rats are highly susceptible to BBN-induced urinary tumors compared to F344 rats (Takahashi et al., 1988) and are equally sensitive to DEN-2- AAF-induced hepatocarcinogenesis (Ohta et al., 1994). Because the only genetic difference between F344-alb rats and F344 rats is the aforementioned 7-base-pair deletion in the albumin gene in the F344-alb rats, the pairing of F344 and F344-alb rats can be used for cell transplantation without using immunosuppressants.

Analbuminemic Rat Model for Hepatocyte Transplantation 127

assays are unsuitable because the serum of F344-alb rats contains proteins with molecular weights similar to that of albumin, and the levels of other proteins may increase to compensate for the lack of albumin (Ohta et al., 1993b). Therefore, conventional gel electrophoretic assays may falsely detect protein levels that are much higher than those of

Cultured hepatocytes differ from hepatocytes *in vivo* with respect to many properties. They often lose specific functions such as the production of albumin and the activity of tyrosine aminotransferase and cytochrome P-450, and gain bile duct epithelium-specific functions such as cytokeratin 19 expression (Block et al., 1996). This difference may arise because gene expression in hepatocytes is strongly influenced by the culture environment, which may activate specific transcription factors (e.g. AP1 and NFB). In addition, the signals mediated by the extracellular matrix influence gene expression in hepatocytes *in vitro* (Serandour et al., 2005; Fasset et al., 2006; Kim et al., 2003). We studied whether the altered phenotype of cultured hepatocytes reverts to that *in vivo* when the cells are reimplanted into the body (Nishikawa et al., 1994). To investigate this problem, two markers were used, one of which is newly expressed and the other is suppressed in cultured hepatocytes, by the F344/ F344-

F344 hepatocytes were cultured on hydrophobic plastic dishes to form spheroidal aggregates. Within 3 days of culture, the hepatocytes formed spheroidal aggregates of approximately 50 to 100 m in diameter, most of which were detached from the bottom and floated freely in the medium. After 5 days of culture, the spheroidal hepatocyte aggregates were harvested and implanted into livers and spleens of F344-alb. The hepatocytes in the liver tissue of F344 rats are positive for p450 (CYP2C6), which is extensively expressed in normal rat hepatocytes (Figure 2C), as well as albumin (Figure 2A), but these cells were completely negative for the placental form of glutathione S-transferase (GST-P), which is one of the glutathione S-transferases that plays an active role in the detoxification of xenobiotics and noxious products generated after tissue damage (Sato, 1999) (Figure 4B). Five days after the start of culture, although albumin expression was maintained in the hepatocytes in spheroidal aggregates (Figure 2D), GST-P was strongly positive in the

After intrahepatic transplantation, the transplanted hepatocytes could be identified by albumin staining within the recipient livers (Figure 2G). On day 5-10 after transplantation, most of the transplanted F344 hepatocytes were located at the portal veins. These cells were firmly attached to their walls and covered by endothelial cells, while some were occasionally observed integrated into the interlobular connective tissue. Most transplanted hepatocyes became completely negative for GST-P staining (Figure 2H), while P-450 was detected in all of the transplanted hepatocytes at an expression level equivalent to that of the surrounding host hepatocytes (Figure 2I). After intrasplenic transplantation, most hepatocytes migrated into the red pulps on days 5-10. These hepatocytes were stained positive for albumin (Figure 2J), still weakly positive for GST-P (Figure 2K) and strongly positive for P-450 which generally gave a more intense signal than in the hepatocytes of F344 livers (Figure 2L).

These results indicated that the phenotype of cultured hepatocytes returned to that of hepatocytes *in vivo* after implantation into intrahepatic and intrasplenic environments.

nucleus (Figure 2E), but P-450 was completely negative (Figure 2F).

**5. Changeability of phenotype of hepatocytes** *in vivo* **and** *in vitro* 

the actual albumin levels.

alb transplantation model.
