The Influence of Estrogen on Hepatobiliary Osteopontin (SPP1) Expression in a Female Rodent Model of Alcoholic Steatohepatitis

Abstract

Our recent studies suggest that higher neutrophil infiltration in females correlates with increased hepatobiliary expression of osteopontin (OPN) in alcoholic steatohepatitis (ASH). The objective of this study was to understand the role of alcohol in altering estrogen levels in females by examining the effect of ethanol (EtOH) on the estrous cycle and then investigate the potential relationship between estradiol (E2) and hepatobiliary OPN expression in a female rat ASH model. Ovariectomized (OVX) and E2-implanted OVX rats in the ASH group were evaluated for OPN mRNA and protein expression. Low doses of E2 resulted in significant down-regulation of OPN protein and mRNA as compared to the OVX group. However, with increasing doses of E2, there was up-regulation of both OPN mRNA and protein. Osteopontin was localized primarily to the biliary epithelium. Liver injury assessed by serum ALT and histopathology revealed a pattern similar to OPN expression. In all groups, hepatic neutrophilic infiltration correlated positively with OPN expression. Based on these data, we conclude that in our ASH model, low doses of E2 appear to be hepatoprotective, whereas the protective effect appears to diminish with increasing doses of E2, although additional cause and effect studies are needed for confirmation.

Introduction

Osteopontin (OPN, also known as secreted phosphoprotein 1 or SPP1) is an acidic member of the small integrin-binding ligand N-linked glycoprotein (SIBLING) family of proteins involved in cell-to-cell and cell-to-matrix communication (Fisher et al. 2001). This protein, a monomer of 264–301 amino acids with molecular mass ranging from 44 to 80 kDa, is known to play a key role in a variety of inflammatory diseases like glomerular nephritis (Denhardt et al. 2001; Giachelli and Steitz 2000; O’Regan and Berman 2000), inflammation during CC14-induced hepatotoxicity (Kawashima et al. 1999), puromycin-induced toxicity (Denhardt et al. 2001), and nonalcoholic steatohepatitis (Sahai et al. 2004). Recent studies from our laboratory have shown that OPN also plays a significant role in alcoholic steatohepatitis (ASH) (Apte et al. 2005; Banerjee et al. 2006a; Banerjee et al. 2006b; Ramaiah and Rittling 2007a; Ramaiah and Rittling 2007b). In a rodent model of ASH, increased hepatic neutrophil infiltration and liver injury have been reported to be mediated by the higher hepatobiliary expression of OPN (Apte et al. 2005; Banerjee et al. 2006b; Ramaiah and Rittling 2007a; Ramaiah and Rittling 2007b). In addition, females with ASH are also reported to have significantly higher hepatobiliary expression of OPN as compared to males. Higher OPN expression was found to correlate positively with higher neutrophilic infiltration and liver injury in females (Banerjee et al. 2006b).

Although it is recognized that OPN influences hepatic inflammation, the precise mechanism by which OPN is up-regulated in the liver during inflammation is not well understood. Osteopontin is known to be regulated by a variety of hormones (estrogen, progesterone, vitamin D3), cytokines, and growth factors (Craig and Denhardt 1991; Noda et al. 1988; Prince and Butler 1997). Several inflammatory mediators and growth factors like interlukin-1 (IL-1), tumor necrosis factor α (TNF-α), platelet-derived growth factor (PDGF), and transforming growth factor-β (TGF-β) are also known to stimulate OPN gene transcription (Denhardt and Noda 1998). Osteopon-tin expression in rat mammary cells has also been reported to be enhanced by estrogen (El-Tanani et al. 2001), and this function is postulated to be mediated by ERα. Although an estrogen–response element is not present in the OPN promoter, there are seven steroid factor-response element (SFRE)-like sequences in this region (El-Tanani et al. 2001) that have been shown to potentially bind ERα and transactivate the OPN gene.

Clearly, understanding the influence of estrogen on the previously reported hepatobiliary OPN expression during ASH is noteworthy, since estrogen has been shown to play an important role in higher susceptibility of females to alcoholic liver disease. Studies by Iimuro et al. (1997) and Nanji et al. (2001) have shown that female rats fed EtOH have higher levels of endotoxin in plasma than their male counterparts. Because estrogen receptors exist in the intestinal epithelium, estrogen has been reported to affect the permeability of the gut, leading to increased absorption of gut microflora or endotoxin in the blood in females (Kono et al. 2000; Yin et al. 2000), which initiates a cytokine cascade, leading to macrophage activation via membrane CD14 (Schumann et al. 1990) and greater liver injury. Because higher OPN expression in female rodent models of ALD contribute to more severe liver injury, it was reasonable to test the hypothesis that higher estrogen in females also contributes to higher hepatic OPN expression in females. In this study, we report that (1) alcohol alters estrogen levels in females, as indicated by the estrous cycle; and (2) estradiol has a protective effect on liver injury at lower doses, and the protection appear to decrease with increasing doses of estrogen, and the extent of OPN expression caused by estradiol is also dose dependent.

Materials and Methods

Assessment of Estrous Cycle

To determine the effect of ethanol on the estrous cycle, daily cytological vaginal smears of animals were performed before and after feeding of EtOH. Briefly, vaginal samples were taken in the morning (between 9:00 and 11:00 AM) using a small cotton swab moistened with saline. The vaginal swabs were then smeared onto a glass slide and stained with Diff-Quik stain. The different stages of the estrous cycle were determined by microscopic visualization of vaginal cell types, as follows: (a) proestrus, determined by the presence of clusters of round, nucleated epithelial cells; (b) estrus, characterized by keratinized cornified cells; (c) metestrus, defined as the presence of noncornified epithelial cells with a few leukocytes; and (d) diestrus, with a predominance of leukocytes and few round epithelial cells (Goldman et al. 2007; Long and Evans 1922).

Rodent ASH Model

The rat ASH model was based on previous studies from our laboratory involving administration of endotoxin (LPS) after six weeks of oral EtOH in the Lieber De-Carli diet (Apte et al. 2005; Banerjee, Apte et al. 2006). Female Sprague-Dawley (SD) rats, eight weeks old and approximately 220–250 g in weight were purchased from Harlan Sprague-Dawley (Houston, TX, USA), and housed individually in cages (area: 144.5 square inches) in a temperature- and humidity-controlled (temperature: 65°F–80°F; humidity: 45%–50%) animal facility with a twelve-hour light-dark cycle. Rats were entered into the study after a one-week acclimatization period. Age- and weight-matched female SD rats were divided into control, EtOH, control+LPS, and EtOH+LPS (n= 4 each) groups. The experimental rats were fed with EtOH-containing (EtOH = 35.5% of total calories, Apte et al. 2005) Lieber-DeCarli liquid diet for a period of six weeks. The control rats were pair-fed with isocaloric maltose-dextrin diet. After six weeks of feeding, the rats were injected with a single dose of LPS or saline (E. coli 0111:B4: 500,000 EU/mg, 10 mg/kg body weight, intraperitoneally (i.p.) in saline, Sigma Diagnostics, St. Louis, MO, USA), and sacrificed twelve hours later by CO₂ asphyxiation.

In the case of the ovariectomized (OVX) rats, they were allowed to recover for three weeks after surgery, and then they were divided into the following experimental groups (n = 4 in each group): EtOH+LPS (Ovx), EtOH+LPS+0.18 mg E2 (Ovx), EtOH+LPS+0.36 mg E2 (Ovx), EtOH+LPS+0.72 mg E2 (Ovx), EtOH+LPS+1.7 mg E2 (Ovx), control+LPS (Ovx), control+LPS+0.18 mg E2 (Ovx), control+LPS+0.36 mg E2 (Ovx), control+LPS+0.72 mg E2 (Ovx), and control+LPS+1.7 mg E2 (Ovx). The animals in the experimental and the control groups were fed EtOH-containing Lieber-DeCarli diet and isocaloric maltose dextrin diet for a period of six weeks, as mentioned previously. After six weeks of feeding, the rats were injected with a single dose of LPS (E. coli 0111:B4, 10 mg/kg, i.p. in saline, Sigma Diagnostics, St. Louis, MO, USA). The animals in the neutralizing OPN (nOPN) antibody-treated group were injected i.p. with the two doses of nOPN (diluted in PBS, 200 μg/kg body weight, R&D Systems, Minneapolis, MN, USA) antibody, six hours apart, before the LPS injection (last dose of nOPN antibody was injected at the same time as LPS). The rats were then sacrificed by CO₂ asphyxiation twelve hours after the LPS injection. All rats were weighed on the day the study was initiated and weekly thereafter. Animals were provided humane care in compliance with the institutional guidelines (ULACC; University Laboratory Animal Care Committee) of Texas A&M University.

Ovariectomy and Estradiol Supplementation

This surgical procedure was followed to assess the affects of estradiol on ethanol-mediated liver injury in the ASH model. Female Sprague-Dawley rats were anesthetized with ketamine (87 mg/kg body weight) and xylazine (13 mg/kg body weight), and bilateral OVX was performed using a dorsal midline incision ventral to the spine and caudal to the last rib (Jezerski and Sohrabji 2000). In some of the animals (n = 8 in each group), a sixty-day, time-release 17-β estradiol pellet (0.18, 0.36, 0.72, and 1.7 mg; Innovative Research, Clearwater, FL, USA) was inserted subcutaneously prior to closing the surgical incision. The estradiol doses employed in this study were based on the therapeutic and supratherapeutic concentration achieved in blood (Nordell et al. 2003). Based on this study, the 0.18- and 0.36-mg doses were considered to be therapeutic low doses, whereas 0.72 and 1.7 mg were considered higher doses.

Sample Collection and Processing

After CO₂ asphyxiation, blood from the different experimental groups was collected from the dorsal aorta and placed in heparinized tubes. Liver transaminase activities were estimated from a fraction of heparinized plasma (about 0.5 mL), and the remaining plasma was snap-frozen in liquid N₂ and stored at −80°C. Livers were harvested, weighed, and divided into two parts. Slices of the left and median lobes were fixed in 10 % neutral buffered formalin for histopathologic evaluation, whereas remaining liver tissue was snap-frozen in liquid N₂ and stored at −70°C for OPN mRNA and protein expression studies.

Evaluation of Liver Injury

Liver injury was assessed by plasma transaminase activities (alanine aminotransferase; ALT) and histopathology of hematoxylin and eosin (H&E)-stained liver sections, as described previously (Apte et al. 2005; Banerjee et al. 2006a).

Histopathology

Histological evaluation of inflammation mediated by ethanol was evaluated on paraffin-embedded liver sections (one section each from the left and median lobe was examined). Hematoxylin and eosin staining was employed to identify the neutrophils, based on the segmented morphology of the nucleus followed by quantification with naphthol AS-D chloroacetate esterase staining (Sigma Diagnostics, St. Louis, MO, USA), as described previously (Banerjee, Apte, et al. 2006). Briefly, 4-μm-thick, formalin-fixed, paraffin-embedded liver sections were deparaffinized and incubated in the substrate solution (40 mL warm distilled water, 1 mL sodium nitrate, 1 mL fast blue violet LB, 5 mL of trizma 6.3 buffer concentrate, 1 mL naphthol AS-D) for thirty minutes in a 37°C water bath. The slides were kept in the dark during the incubation period, after which they were washed in distilled water and counterstained with Gill’s hematoxylin for forty-five seconds, followed by rinsing in tap water four times. The slides were then dipped three times in 70% alcohol, 100% alcohol, and xylene. The dehydrated sections were then mounted in cytoseal and examined. The red-colored cytoplasmic staining was specific for neutrophils. A blood-smeared slide was used as a positive control for the experiment. To quantify the degree of neutrophilic inflammation (inflammation score), the neutrophilic foci per five high power fields (40×, fields were randomly selected) was counted. The neutrophilic foci (defined as an aggregate of four neutrophils) were quantitated per five 40× fields.

Analysis of OPN mRNA Localization and Expression by In Situ Hybridization

Osteopontin mRNA expression in liver sections was localized by in situ hybridization as previously described (Johnson et al. 1999). Briefly, deparaffinized, rehydrated, and deproteinated liver cross-sections (5 μm) were hybridized with [³⁵S]-radiolabeled antisense or sense OPN cRNA probes. The sense probe was used as a negative control to define nonspecific hybridization. Following washes and RNAaseA digestion, the slides were dipped in Kodak NTB-2 liquid photographic emulsion (Kodak, Rochester, NY, USA), stored at 4°C for five days, developed in Kodak D-19 developer, counterstained with Harris modified hematoxylin (Fisher Scientific, Fairlawn, NJ, USA), dehydrated, and protected with cover slips. Digital photomicrographs of representative bright and dark field images were evaluated with a Zeiss Axioplan2 microscope (Carl Zeiss, Thornwood, NY, USA) fitted with an Axiocam HR digital camera. Individual images were recorded sequentially with AxioVision 4.3 software and in the Zeiss Vision Image (ZVI) file format, and they were subsequently converted to tagged image file (TIF) format. All the figures were assembled in Adobe Photoshop 7.0.1 (Adobe Systems, Inc., San Jose, CA, USA).

Osteopontin Protein Expression

Western Blot Analysis:

Liver cell lysates from control, intact, and OVX ASH groups were prepared in lysis buffer (1% Triton-X-100, 50 mM NaCl, 10 mM Tris, 1 mM EDTA, 1 mM EGTA, 2 mM Na vanadate, 0.2 mM PMSF, 1 mM HEPES, 1 μg/mL leupeptin, and 1 μg/mL aprotinin) and protein concentration was estimated using a Bio-Rad protein assay kit (BioRad, Hercules, CA, USA) according to the manufacturer’s protocol. Briefly, 100 μg of cell lysate was resolved by electrophoresis on a 12% sodium dodecyl sulfate (SDS) polyacrylamide gel (100 v, 1.5 hours) in a running gel buffer containing 25 mM Tris, pH 8.3, 162 mM glycine, and 0.1% SDS. The samples were transferred to nylon membrane for three hours at 500 mA. The membranes were incubated overnight in a mixture of T-TBS with 0.1% tween and 2% milk and OPN antibody (rabbit polyclonal to OPN, 1:1000 dilution, Abcam, Inc., Cambridge, MA, USA). Subsequently, the membrane was incubated in anti-goat secondary antibody for one hour at room temperature. The OPN antibody recognizes both the native (uncleaved) form of OPN (∼66 KD), and the cleaved form of OPN (32 KD; Rittling and Feng 1998). Visualization was carried out with the enhanced chemiluminescence kit (Pierce, Rockford, IL). Glyceraldehyde-3-phosphate dehydrogenase was used as an internal control to ensure equal loading of proteins per well.

Statistics

Group comparisons were performed using one-way analysis of variance (ANOVA) test and by the Tukey post hoc test. Statistical analyses were made using Graph Pad Prism 5.01 software (Graph Pad Software, Inc., La Jolla, CA, version 5.01.). Statistical significance was set at p < .05.

Results

Role of Ethanol in Influencing Estrous Cycle

Consistent with Long and Evans (1922), the estrous cycles in rats fed control diet were found to be about 4.5 days long and regular in about 88% of the animals studied. Only animals cycling for about four days were fed EtOH in Lieber DeCarli diet for six weeks. After the initiation of EtOH feeding, the animals were found to have prolonged estrous cycles (about 5.7 days) due to increased length of the diestrous phase (about three days). The diestrous phase was marked by the predominance of leukocytes and nucleated epithelial cells in the vaginal smear, indicating that EtOH was interfering with the estrous cycle in these rats, leading to alteration of the estrogen level.

Effect of Ovariectomy on Ethanol-mediated Liver Pathology

Both the intact and the ovariectomized females had little or no increase in plasma transaminase (ALT) activity following treatment with EtOH alone. A significant increase in plasma transaminase activity was noted in both the intact and the OVX group following EtOH+LPS treatment as compared to the respective control, EtOH alone, and LPS alone groups (Figure 1A).

When the liver injury was compared between OVX and intact groups, the OVX females in the EtOH+LPS-treated group had approximately 1.5-fold higher plasma transaminase activity as compared to the intact females in the same group. The findings of the H&E-stained liver sections were consistent with the plasma transaminase data, wherein increased multifocal necrosis and neutrophilic infiltration in the OVX EtOH+LPS-treated animals (approximately two-fold) was noted as compared to the intact EtOH+LPS-treated animals (Figures 1B and 2). The animals in the control group that were implanted with estrogen pellets experienced no liver damage (data not shown).

Dose-dependent Effect of Estradiol on Liver Injury and Histology

To assess the effect of estradiol on hepatic injury, plasma ALT transaminase activity and liver histopathology were also evaluated in OVX+E2-implanted rats fed EtOH+LPS. Compared to the OVX group, all animals in the estrogen-implanted group had a significant decrease in plasma transaminase activity. However, a biphasic response to ASH was observed in the OVX+E2-implanted rats. Low doses of E2 (0.18 mg, 0.36 mg) resulted in a significant decrease in plasma transaminase activity as compared to the OVX group. However, the highest E2 dose employed (1.7 mg) resulted in decreased protection as compared to the 0.36-mg and 0.72-mg doses of E2, suggesting a mild elevation of plasma transaminase activity (Figure 3A). Rats treated with the 0.36-mg dose of E2 seemed to have the least liver injury in these groups (Figures 3A, 3B, and 4). The plasma transaminase activity was further confirmed with H&E-stained liver sections, where minimal or no neutrophilic infiltration and lack of multifocal coagulative necrosis was observed in animals treated with 0.36 mg estrogen (Figures 3B and 4). In animals treated with 0.72 mg and 1.7 mg E2, an increase in neutrophilic infiltration and focal areas of necrosis was observed.

Localization and Expression of OPN mRNA by In Situ Hybridization

Because we have shown previously that OPN is the likely mediator of hepatic neutrophil infiltration and liver injury in the intact rat ASH model (Banerjee et al. 2006a), we assessed the effect of estradiol on OPN expression and hepatic neutrophil infiltration. In situ hybridization was carried out to determine the hepatic source of OPN expression in OVX and E2-supplemented, EtOH+LPS-treated rats. Intact, EtOH+LPS-treated animals had a higher OPN mRNA signal (Figure 5A) as compared to the low-dose, estradiol-supplemented (0.18 mg, 0.36 mg), ovariectomized rats (Figures 6A and 6B). However, as compared to the low doses of E2 (0.18 mg, 0.36 mg), a significantly higher OPN mRNA signal was observed in the animals treated with increasing doses (0.72 mg and 1.7 mg) of estrogen (Figures 6C and 6D). The lowest OPN mRNA signal was observed in the animals treated with the 0.36-mg dose of E2. The negative control did not show any nonspecific hybridization (Figure 5B). The localization of OPN was mostly within the biliary epithelium.

Dose-response Effects of Estradiol on Hepatobiliary OPN Expression

Western blotting illustrated a minimal decrease in the level of OPN protein in the OVX animals supplemented with E2 in the ASH model (i.e., EtOH+LPS+E2) as compared to the EtOH+LPS alone group, with the exception of the highest dose of E2 (Figure 7). The animals treated with the highest E2 dose (1.7 mg) experienced >2.5-fold higher expression of OPN protein as compared to the EtOH+LPS-alone group (Figures 7A and 7B). Induction of hepatic OPN protein was also supported by immunohistochemistry (data not shown), where animals treated with the 0.18-mg and 0.36-mg doses of E2 had lower expression of OPN, with maximum expression observed at the highest dose of E2 employed in this study. As previously reported, biliary epithelial cells are the predominant OPN-producing cells in the liver in the rat model of alcoholic liver disease employed in this study (data not shown).

Effect of Neutralizing OPN (nOPN) Antibody on Liver Injury and Neutrophil Infiltration

An nOPN experiment was carried out to assess the role of E2 and OPN in liver injury. Rats treated with nOPN (ovariectomized+nOPN+EtOH+LPS+1.7 mg E2) had a significant decrease in liver injury as determined by plasma ALT, when compared to the untreated group (ovariectomized+EtOH+LPS+1.7 mg E2; Figure 8A). In addition, the nOPN antibody-treated group also had a significant decrease (approximately 85%) in neutrophilic infiltration in the hepatic parenchyma (Figure 8B), indicating that OPN was involved in the severe liver injury and neutrophilic infiltration experienced by the ovariectomized rats implanted with 1.7 mg E2 in the EtOH+LPS-treated group.

Discussion

We have previously reported that female rats in the ASH model experience significantly more severe liver injury than the males. In this model, severe liver injury is accompanied by increased neutrophil infiltration, which appears to be OPN mediated (Banerjee et al. 2006a). However, the precise mechanism by which OPN is increased in the rat ASH model, especially in females, is unknown. Osteopontin expression is known to be regulated by E2 (Lessey 2002; White et al. 2005; White et al. 2006). In addition, estrogen has also been shown to increase susceptibility of females to alcoholic liver disease. However, nothing is known about the role of estrogen in the regulation of OPN expression in alcoholic liver disease, which is the basis for this investigation. In this study, we found that estradiol has a potential protective effect on ethanol-mediated liver injury, although the observed protection seemed to diminish with increasing doses of estrogen.

Alteration of the estrous cycle following alcohol consumption has been previously reported by Rettori et al. (1987) and Emanuele et al. (2001). Rats fed an alcohol diet consistently exhibit a prolonged diestrous phase. The alteration in the estrous cycle in the former study was attributed to alcohol-induced depression in serum luteinizing hormone (LH) and consequent diminished release of LHRH from the hypothalamus. Our results are consistent with these studies, where rats experienced a prolonged diestrous phase following alcohol consumption, suggesting that alcohol is altering the estrogen level in our model. Also, previous studies from this laboratory have reported that females in ALD have higher expression of OPN (Banerjee et al. 2006a), suggesting that expression of OPN is possibly regulated by estrogen.

Although ethanol consumption alters circulating levels of sex hormones, estrogen has been implicated in severe liver injury in females in alcoholic liver disease (Yin et al. 2000). These results were obtained using a continuous intragastric feeding model, in which OVX females with alcoholic liver disease had significantly lower liver injury and inflammation as compared to intact females, and this condition could be reversed with estrogen replacement. This study is somewhat contradictory to results of the present study, where OVX females experienced significantly greater liver injury and inflammation as compared to intact females. Factors such as the type of alcoholic liver disease model used, species of rats employed, and dose of estradiol may have contributed to the observed contradiction. In fact, in this study we showed that higher E2 doses increased hepatic injury and neutrophil infiltration compared to the lower doses. In our model, the level of estrogen in the blood of animals treated with the highest dose of E2 were similar to the levels achieved in the studies by Yin et al. (2000; 224 pg/mL).

In concordance with our studies supporting the protective effect of estradiol, other researchers have shown that estradiol opposes the onset of inflammation in animal models. In an experimental model of autoimmune encephalomyelitis and an animal model of multiple sclerosis, low-dose estrogen therapy was shown to prevent the clinical signs and histopathological lesions of the disease (Bebo et al. 2001; Jansson et al. 1994; Rosette and Karin 1995). Also, other inflammatory conditions such as wound healing (Ashcroft et al. 2003), atherosclerosis (Hogdin and Maeda 2002), ischemia (Dubal et al. 2001), uveitis (Miyamoto et al. 1999), and leukodystrophy (Matsuda et al. 2001) were shown to be significantly influenced by estradiol, where estradiol decreased disease susceptibility and severity of damage.

The precise mechanism by which estradiol is protective during inflammation is worthy of discussion. Studies by Ghisletti et al. (2005), have shown that estrogen blocks inflammation by its action on the transcription factor p65/relA of the NF-κB family. In macrophages, estrogen was reported to block lipopolysaccharide-induced DNA binding and transcriptional activity of p65 by preventing its nuclear translocation, without altering the Iκ-B kinase (IKK) activity. This has been suggested to be mediated by ERα through a nongenomic mechanism, indicating the role of the estrogen-ERα signaling pathway in mediating early inflammatory response (Ghisletti et al. 2005). Certainly, these mechanisms need to be investigated in future studies to ascertain the mechanistic basis for estrogen protection.

In the literature, OPN expression is known to be regulated by estrogen (Craig and Denhardt 1991). In the peri-implantation pig and mouse uterus, estrogen is known to induce OPN mRNA expression in the endometrial luminal epithelium (White et al. 2005; White et al. 2006). Indeed, estrogen and progesterone appear to regulate OPN expression in the human uterus. Similar induction of OPN mRNA has been observed in our model, with higher doses of estrogen. However, lower doses of estrogen decreased OPN expression. This finding is in concordance with the studies by Turner et al. (1990), where estrogen (diethylstilbestrol) was reported to down-regulate OPN levels in the bone tissue of OVX rats. It can be argued that in addition to E2, other hormones such as progesterone may play a role in OPN induction. In fact, there are studies in the literature that have shown that the hormone progesterone also controls OPN expression (Craig and Denhardt 1991; Johnson et al. 2000; White et al. 2006). The OPN gene has been reported to contain a progesterone-response element in its 5^′-flanking region that is likely induced by progesterone in vivo in mice (Craig and Denhardt 1991). Also, OPN is up-regulated in human cytotrophoblasts by progesterone (Omigbodun et al. 1997), indicating that in addition to estrogen, the expression of OPN in our model could also be influenced by progesterone, and this possibility is worthy of future investigation.

In conclusion, estrogen was found to have a protective effect on ethanol-mediated liver injury, although the protection appeared to diminish with increasing doses of E2 in the rat model of alcoholic liver injury. The effects of E2 in this model appear to be mediated by OPN. Based on the deleterious effect of OPN on hepatic inflammation during liver injury in the rat ALD model employed, it is possible that OPN levels can be altered by E2 and that can potentially exaggerate or diminish liver injury. Although further studies are needed to confirm this observation, potential novel therapeutic strategies, such as E2 to influence OPN induction, can be developed to treat human alcoholic liver disease patients.