Abstract
Butylated hydroxyanisole (BHA) and 1,2-bis(2-pyridyl)ethylene (2PY-e) are phase II drug metabolizing enzyme inducers which cause hepatomegaly without hepatocyte hypertrophy and induce glutathione S-transferase Yp (GST Yp, pi-class GST), which is known as a tumor marker. To evaluate the relationship between GST Yp induction and hepatocyte proliferation, male F344/DuCrj rats were treated with BHA, 2PY-e, or phenobarbital (PB) for three or seven days. All three chemicals caused increases in liver weight after three and seven days. Immunohistochemical examinations revealed that BHA and 2PY-e induced GST Yp in the hepatocytes of the periportal and centrilobular areas at three and seven days, respectively, whereas PB did not. Significant increases in the BrdU labeling indices were found in the livers of rats in each of the three-day treatment groups, but the labeling index of rat livers treated with BHA was decreased to the control level at seven days, although the high labeling indices of 2PY-e and PB persisted at seven days. Double immunostaining confirmed that BrdU-positive nuclei corresponded to GST Yp-positive hepatocytes in both BHA and 2PY-e treated rats. These results suggest that the GST Yp induction caused by BHA or 2PY-e has some kind of relationship with hepatocyte proliferation.
Introduction
The liver is the major organ responsible for the metabolism and detoxification of drugs, and drug-metabolizing enzymes in the liver play central roles in the biotransformation, metabolism, and detoxification of xenobiotics. Phase I enzymes, that is, cytochrome P450s (P450s), exist in the hepatocytic microsomes and have the potential to catalyze numerous xenobiotic biotransformations. However, the metabolites produced by phase I enzymes, such as acetoaminophen, carbon tetrachloride, and benzene, are occasionally toxic (Lindros 1997). Phase II enzymes such as glutathione S-transferases (GSTs) catalyze the detoxification of xenobiotics including carcinogenic chemicals. Glutathione S-transferases are the most important phase II detoxification pathways and play an important role in the cellular defense mechanisms against xenobiotics (Gajewska and Szczypka 1992). Talalay et al. have defined the chemicals that induce both phase I and phase II enzymes as bifunctional inducers and the chemicals that selectively induce phase II enzymes as monofunctional inducers (Talalay et al. 1988). The monofunctional inducers are generally considered to be beneficial because they stimulate the detoxification of potentially carcinogenic chemicals (Hu et al. 2006; Kong et al. 2001; Lee et al. 2007).
Butylated hydroxyanisole (BHA) is a synthetic phenolic antioxidant that is widely used as a food preservative. Butylated hydroxyanisole causes hepatomegaly without hepatocellular hypertrophy and induces phase II enzymes, including GSTs and UDP-glucuronosyl transferase, without inducing phase I enzymes and is classified as a monofunctional inducer (Buetler et al. 1995; Franklin 1991; Makino et al. 1998). Our previous study suggested that 1,2-bis(2-pyridyl)ethylene (2PY-e) also causes hepatomegaly without hepatocellular hypertrophy and induces only phase II metabolizing enzymes without phase I enzyme induction, which indicated that 2PY-e has similar profiles to BHA. Both compounds are defined as monofunctional inducers that characteristically induce GST Yp (pi-class GST), which is not induced by bifunctional inducers such as phenobarbital (PB) (Makino et al. 1998).
Glutathione S-transferases exist in the cytoplasm, the membrane of the mitochondria, the microsomes, and the nucleus of the hepatocyte (Aliya et al. 2003; Stasiecki et al. 1980). Seven distinct classes of cytosolic GSTs have been identified: alpha, mu, pi, sigma, theta, delta, and zeta (Board et al. 1997). Among the isoenzymes, an increased expression of GST-P (pi-class GST) is correlated with the resistance to chemotherapeutic agents in several human tumor cells (Zhao et al. 2005) and not expressed in normal livers, except for the bile duct epithelium in rats. GST-P is expressed only in hepatocytic hyperplastic/preneoplastic nodules and hepatocellular carcinoma in rat and human tumor cell lines and neoplastic tissue specimens (Gajewska and Szczypka 1992; Satoh et al. 2005). It is well known as a tumor marker and is often used in chemical-induced hepatocarcinogenesis in rats as a reliable tumor marker (Ito et al. 2003; Osada et al. 2006; Song et al. 1999).
However, BHA has protective effects against various carcinogens in the liver because of phase II enzyme induction (Buetler et al. 1995), although it has carcinogenicity in the forestomach and urinary bladder, which suggested that the GST Yp induction of BHA treatment does not indicate tumor development. In a previous study, we confirmed that BHA or 2PY-e treatment caused hepatomegaly without cellular hypertrophy and GST Yp induction, which indicated cell proliferation rather than cellular hypertrophy and suggested some correlation between GST Yp induction and hepatocyte proliferation (Makino et al. 1998). Therefore, we investigated the relationship between GST Yp induction and hepatocyte proliferation in rats treated with BHA or 2PY-e, which induce GST Yp, or with PB, which does not induce GST Yp, and suggest that the GST Yp induction has some kind of relationship with hepatocyte proliferation. This model is considered to be useful for investigating the mechanisms of proliferation of GST Yp-expressing cells.
Materials and Methods
Animal Treatment and Tissue Preparation
Male F344/DuCrj rats were purchased from Charles River Japan, Inc., and chemical treatment was started at eleven weeks of age, because male rats can avoid the effect of the estrous cycle on cell proliferation, and cell proliferation stabilizes at this age. The animals were housed individually in mesh cages and maintained in a room controlled at a temperature of 24°C to 28°C and relative humidity 40% to 70%, and under a twelve-hour light-dark cycle. Pellet food (Certified Rodent Diet 5002, PMI Nutrition International, Inc., USA) and tap water were available ad libitum.
After seven days of acclimation, the rats were treated with the chemicals as follows, in accordance with the doses used previously. In two groups treated either three days or seven days, five rats each were treated as follows: control, BHA, PB, and 2PY-e. The BHA rats were given a diet containing 0.7% BHA (Sigma, St. Louis, MO, USA), and the PB rats were given a diet containing 0.08% PB (Tokyo Chemical Industry, Tokyo, Japan) until just before autopsy. The 2PY-e (Tokyo Chemical Industry) given to the 2PY-e rats was solubilized in 1% HCl solution and administered orally at a dose level of 75 mg/kg. The control animals received an equivalent volume of 1% HCl daily for three and seven days.
Three days prior to necropsy, an osmotic pump (Model 2001, 1.0 μg/hr, ALZA, Palo Alto, CA, USA) containing 5-bromo-2-deoxyuridine (BrdU, Sigma, 50 mg/mL dissolved in 40% DMSO) was subcutaneously implanted into each animal. At Day 3 or 7 of chemical treatment, the animals were euthanized with ether, and the livers were removed and weighed. Liver samples were fixed in 10% neutral buffered formalin and embedded in paraffin. Histopathological specimens were prepared and stained with hematoxylin and eosin (HE). The experimental protocol was approved by the Ethics Review Committee for Animal Experimentation of Sankyo Co., Ltd.
Immunohistochemistry
Rabbit polyclonal antibody against rat GST Yp (Biotrin international, Dublin, Ireland, 1:50 dilution) and mouse monoclonal antibody against BrdU (Immunotech, Fullerton, CA, USA, 1:50 dilution) were used as primary antibodies. Deparaffinized liver sections were first incubated for ten minutes with 3% H2O2 in methanol and with normal bovine serum to reduce nonspecific staining. The primary antibodies were applied to the sections for thirty minutes. Secondary antibodies (Dako, Carpinteria, CA, USA) were applied for ten minutes. The primary and secondary antibodies were diluted with phosphate buffered saline (PBS). Horseradish peroxidase (HRP)-labeled streptavidin was then applied for twenty minutes. The above incubations were performed at room temperature. Finally, the sections were visualized by a 3,3’-diaminobenzidine tetrahydrochloride (Dojindo, Kamimashiki, Japan) reaction.
Double immunohistochemistry for GST Yp and BrdU was performed using the same primary antibodies against rat GST Yp and BrdU. Deparaffinized sections were sequentially incubated with 3% H2O2 solution in methanol for ten minutes and 2N HCl for ninety minutes. After the incubation with boric acid-borate buffer to neutralize the acid and 0.05% protease for five minutes, nonspecific binding was blocked using normal bovine serum. The sections were incubated with the mouse monoclonal anti-BrdU antibody and then with rabbit polyclonal anti-rat GST Yp rabbit antibody for thirty minutes. Immunoreactions were visualized with alkaline phosphatase red and ABC peroxidase brown systems for GST Yp and BrdU, respectively.
Labeling Indices of Hepatocytes
Five randomly selected x400 microscopic fields in the centrilobular or periportal areas were analyzed per section. The number of BrdU-positive and -negative hepatocytes was counted in each section (more than 1000 hepatocytes). The labeling indices, expressed as the percentage of positive hepatocytes of the total hepatocytes, were calculated. Similarly, the labeling indices of GST Yp-positive or -negative hepatocytes were calculated on the sections.
Statistical Analysis
Calculations were done with calculation software (Microsoft Excel 2000). Statistical comparisons between the treated groups and the control group were conducted by a Student t test, and p < 0.05 was taken to be statistically significant.
Results
Clinical Signs, Body Weight, and Relative Liver Weight
Treatment with BHA, 2PY-e, or PB for three and seven days did not produce any abnormal clinical signs or significant changes in body weight (Table 1). The liver weights of all the treated groups at three days were increased by about 20% compared with the control group. The liver weights at seven days were increased by 20% to 35% compared with the control. Although the liver weight gains induced by BHA, 2PY-e, and PB were comparable at three days, the gains by 2PY-e continued to increase up to seven days.
Histopathology and Immunohistochemistry
Phenobarbital-treated livers showed centrilobular hypertrophy of hepatocytes. However, BHA- or 2PY-e–treated livers did not show any hypertrophic changes (Figure 1), in spite of the increased liver weights. With all three chemicals, mitotic figures were often observed, except in the livers of rats treated with BHA for seven days. GST Yp was expressed not in hepatocytes, but in the bile duct epithelium of the control animals (Figure 2A). Although BHA induced GST Yp expression in the periportal hepatocytes, 2PY-e induced the expression in the centrilobular hepatocytes (Figures 2B and 2C). The distribution of the GST Yp-positive area was mosaic/heterogeneous, far from the zonal pattern generally observed in the case of CYP induction (Lindros 1997). The intensity of the GST Yp expression was similar between the three-day and seven-day treatments in both the BHA and 2PY-e groups (data not shown). Treatment with PB did not induce GST Yp expression (Figure 2D).
Labeling Indices of Hepatocytes
The results of the BrdU labeling are shown in Figures 3 and 4. Statistically significant increases in the labeling indices were found in both the centrilobular and periportal hepatocytes of all the groups of the three-day treatment. In the BHA- or PB-treated groups, the index values of the periportal hepatocytes were higher than those of the centrilobular hepatocytes (periportal/centrilobular = 2.79 and 1.59, respectively). In the 2PY-e–treated group, in contrast, the index value of the centrilobular hepatocytes was higher than that of the periportal hepatocytes (periportal/centrilobular = 0.60).
The index values of the seven-day treatment of BHA or 2PY-e were decreased compared to those of the three-day treatment, and the values of the BHA-treated group, especially, were drastically decreased to a level similar to the control values. Similar values were recorded between the three-day and seven-day PB treatments. In the 2PY-e–treated group, the index value of the centrilobular hepatocytes was higher than that of the periportal hepatocytes (periportal/centrilobular = 0.69). In the PB-treated groups, the index value of the periportal hepatocytes was similar to that of the centrilobular hepatocytes (periportal/centrilobular = 0.94).
Double Immunostaining for GST Yp and BrdU
Figures 5A and 5B show the results of the double immunostaining of rat livers treated with BHA for three days. In the centrilobular area, few hepatocytes were positive for GST Yp, and most of the BrdU positive nuclei corresponded to the GST Yp-positive hepatocytes (Figure 5A). In the periportal area, the bile duct epithelium and many hepatocytes were positive for GST Yp, and most of the BrdU-positive nuclei corresponded to the GST Yp-positive hepatocytes. Figure 6 summarizes the labeling indices of BrdU counted in either GST Yp-positive or GST Yp-negative hepatocytes of rats treated with BHA for three days. In the centrilobular area, the labeling index of the GST Yp-positive hepatocytes was about ten times higher than that of the GST Yp-negative hepatocytes (GST Yp-positive: 35%, GST Yp-negative: 3%). In the periportal area, the labeling index of the GST Yp-positive hepatocytes was also higher than that of the GST Yp-negative hepatocytes (GST Yp-positive: 40%, GST Yp-negative: 10%). Livers treated with 2PY-e showed that BrdU-positive nuclei were mostly observed in the GST Yp-positive hepatocytes mainly located in the centrilobular area, as in the case of BHA (Figure 5C).
Discussion
All three chemicals caused comparable levels of increased liver weights (20% to 30% increase compared to the control). Phenobarbital caused histopathologically centrilobular hepatocytic hypertrophy, which is considered to be an adaptive change to maintain homeostasis. Adaptive change may be defined as the effect of chemicals on a regulatory pathway leading to modulation of cellular function and structure (Williams and Iatropoulos 2002) and liver enlargement, induction of P450 enzymes, proliferation of organelles, cell hypertrophy, and proliferation occurring dose dependently (Juberg et al. 2006). On the contrary, BHA and 2PY-e did not show apparent morphological changes, but GST Yp induction was observed in the BHA- and 2PY-e–treated groups, but not in the PB-treated group. Similar results were reported in our previous study (Makino et al. 1998). These findings imply that BHA and 2PY-e induce hepatomegaly by cell proliferation, but not by hypertrophy.
Glutathione S-transferase Yp is a kind of GST-P that is expressed in the fetal hepatocytes, but not in adult hepatocytes except for neoplastic or preneoplastic hepatocytes. In the adult liver, GST-P is highly expressed in the hepatic neoplastic foci induced by chemical carcinogens, and it is a well-known tumor marker (Gajewska and Szczypka 1992; Satoh et al. 2005). The oval cells that appear at the regenerating phase after an injury caused by allyl alcohol were also positive for GST-P (Yavorkovsky et al. 1995). Therefore, GST Yp-positive hepatocytes are considered to be immature or to be associated with cell proliferation, but the involvement of this enzyme in cell proliferation is still unclear. The majority of research regarding GST-P expression has focused on carcinogenesis, but only a few reports on GST-P induction under non-neoplastic conditions have been published.
In the present study, BHA and 2PY-e induced GST Yp in different areas of the liver lobule, that is, BHA in the periportal hepatocytes and 2PY-e in the centrilobular hepatocytes. The high BrdU labeling indices were also different between BHA and 2PY-e, that is, BHA in the periportal area and 2PY-e in the centrilobular area. The double immunostaining for GST Yp and BrdU clearly showed higher proliferation activity of GST Yp-positive hepatocytes than GST Yp-negative ones in both BHA-and 2PY-e–treated livers. As shown in Figure 6, the labeling indices of the GST Yp-positive (or -negative) hepatocytes were comparable between the periportal area and in the centrilobular area in BHA. But the number of GST Yp-positive hepatocytes in the periportal area was higher than that in the centrilobular areas, because GST Yp was preferentially induced in the periportal hepatocytes in BHA. These results suggested that the distribution of GST Yp-positive hepatocytes might contribute to the areas that showed high proliferation activity in BHA or 2PY-e. The reason why the GST Yp induction area was different between BHA and 2PY-e is not clear at this time.
There are very few reports describing the distribution of GST Yp induced by other chemical treatment. Aceto et al. have reported the presence of two different-sized GST subunits that cross-react with an antibody that recognizes pi-class GST in human testis (Aceto et al. 1989). It is possible that different isoenzymes of GST Yp composed of distinct subunits that could cross-react with the GST Yp antibody used in this study might be induced by BHA and 2PY-e. However, further study is necessary to clarify this supposition. The labeling indices of PB-treated livers where GST Yp was not induced were comparable between the centrilobular and periportal hepatocytes.
The increased proliferation activities of BHA and 2PY-e were transient, but that of PB was stable. The labeling indices of BHA increased up to three days, but they decreased thereafter to the control level by seven days. The labeling indices of 2PY-e decreased, but not to the control level, which is consistent with an increase in liver weight. The labeling indices of 2PY-e were higher than those of BHA at Day 3 of treatment and it seemed likely that the dose level of 2PY-e was higher than BHA, which is the reason why the labeling indices of 2PY-e did not decrease to the control level at Day 7.
Chemicals mediated by ligand-activated receptors such as AhR, CAR, and PPARα generally cause hepatomegaly and cell proliferation during the initial treatment period (Juberg et al. 2006; Kinoshita et al. 2002; Marsman et al. 1988; Rose et al. 1997; Wada et al. 1992; Yeldandi et al. 1989), which is considered to be an adaptive response (Juberg et al. 2006), but not a carcinogenic effect. Transient proliferation by BHA or 2PY-e treatment implies that these compounds do not have carcinogenic effects. In addition, since GST Yp induction was continued up to seven days, it was considered that GST Yp induction has some kind of relationship with cell proliferation, but that GST Yp itself may not have potential to accelerate cell proliferation.
Batist et al. reported that the GST Yp gene in normal rat liver epithelial cells was expressed more in proliferating cells than in quiescent ones in vitro (Batist et al. 1991), and several hepatocarcinogenicity models using rats showed high proliferation activities in GST-P–positive foci (Short et al. 1997; Song et al. 1999). On the contrary, GST-P null mice had higher basal levels of circulating lymphocytes compared with wild-type mice, and a selective GST-P inhibitor treatment caused a myeloproliferative response in wild-type mice (Ruscoe et al. 2001). Tatematsu et al. reported that the labeling index of the GST-P–positive foci was low and that of the GST–P-negative hepatocytes surrounding the positive foci was high in rats given a single dose of diethylnitrosamine and BHA and then subjected to a partial hepatectomy (Tatematsu et al. 1988). These studies showed conflicting results and did not clearly suggest the effect of GST-P expression on cell proliferation. There are relatively few reports regarding the relationship between GST-P and cell proliferation because most of the studies have been evaluating tumor cells in vivo or in vitro. The models used in this study may be useful to investigate mechanistic considerations of the effects of GST-P on cell proliferation.
The labeling indices of the GST Yp-negative hepatocytes were higher than the values of the control in this study, although those of the GST Yp-negative hepatocytes were lower than those of the GST Yp-positive hepatocytes. These results suggested that GST Yp-negative hepatocytes also increased proliferation activities, although the activities of GST Yp-negative hepatocytes were lower than those of GST Yp-positive hepatocytes. Under normal conditions, since cell growth is controlled by both positive and negative regulators, higher proliferation activities might be derived from the effects of GST Yp on negative regulators of cell growth. TGF-β1 is reported to inhibit the outgrowth of normal epithelial cells in several tissues (Jong et al. 2002; Ko et al. 1998; Sandhu et al. 1997). It has also been reported that the growth of GST-P–positive hepatocytes was not inhibited by TGF- β1, whereas that of GST-P–negative hepatocytes was inhibited in vitro (Stenius 1993). p21 is a member of the cyclin dependent kinase (CDK) inhibitor family, which regulates retinoblastoma (RB) protein phosphorylation thorough CDK inhibition and is induced by TGF-β1 (Hannon and Beach 1994; Malliri et al. 1996; Massague and Wotton 2000). After partial hepatectomy, 2-acetaminofluorene (AAF) blocks the proliferation of hepatocytes through the induction of p21 (Tatematsu et al. 1988; Trautwein et al. 1999). However, in this animal model, oval cells that possess the characteristics of both hepatocytes and bile duct epithelium do not lose their proliferation activity (Petersen et al. 1998). Since oval cells express GST-P, these results may suggest that GST-P–expressing cells have some inhibitory effects on the negative regulation of cell proliferation through p21. Although the investigation of the expression of TGF-β1 or p21 was not conducted in our study, GST Yp-expressed hepatocytes induced by BHA or 2PY-e may have similar inhibitory effects against negative regulators of cell proliferation, which may cause high proliferation activities on GST Yp-expressed hepatocytes.
In summary, the present findings strongly suggest that GST Yp induction by monofunctional inducers, such as BHA or 2PY-e, is closely related to high hepatocyte proliferation. Increased hepatocyte proliferation decreases as a result of repeated treatment, whereas GST Yp induction continues throughout treatment. It seems likely that GST Yp itself does not have proliferation effects. Further investigation is required to explain the precise mechanisms of GST Yp expression and cell proliferation, and our model is considered to be useful for investigation of the mechanisms of the proliferation of GST Yp-expressing cells.