|Home | About | Journals | Submit | Contact Us | Français|
Alcohol is a significant risk factor for development of hepatocellular carcinoma (HCC). To date, no rodent model has demonstrated formation of hepatic neoplasia in the setting of chronic alcohol consumption alone.
We investigated whether rats selectively bred for high alcohol preference (P rats), allowed free access to water, or water and 10% (v/v) alcohol for 6, 12 or 18 months, develop hepatic neoplasia.
At necropsy, liver tumor incidence and multiplicity were significantly increased in 18-month alcohol-consuming versus water-consuming P rats. These data were confirmed histologically by glutathione-S-transferase pi-class (GSTp) staining. Phosphorylated mitogen-activated protein kinase/extracellular signal-regulated kinase 1/2 (MAPK/ERK) staining was also increased in the sinusoidal lining cells within livers of alcohol-consuming versus water only P rats. In addition, cytochrome p450IIE1 (CYP2E1) mRNA, protein expression/activity, and intrahepatic oxidative stress were significantly increased in alcohol-consuming P rat livers versus water only. In contrast, acetaldehyde dehydrogenase expression decreased in alcohol-consuming versus water only P rats. No significant difference in alcohol dehydrogenase expression was detected.
These data demonstrate that chronic alcohol consumption is associated with hepatic neoplasia, MAPK/ERK activation, increased CYP2E1 activity, and intrahepatic oxidative stress in P rats. Since these rats are well-characterized as a model of alcoholism, these findings identify a novel rodent model of alcohol or “alcoholism”-induced liver neoplasia.
Hepatocellular carcinoma (HCC) is the most common primary liver tumor diagnosed, accounting for ≥600,000 deaths/year (El-Serag and Rudolph, 2007; McKillop and Schrum, 2009). HCC most commonly arises in the setting of exposure to known risk factors of which viral hepatitis (HBV/HCV), aflatoxin exposure, and/or chronic alcohol consumption are the most common (El-Serag and Rudolph, 2007; Schutte et al., 2009). The nature of these risk factors causes distinct patterns of HCC occurrence, the highest incidence being reported in far-Eastern and sub-Saharan countries in which HBV/HCV infection is endemic (Schutte et al., 2009). In contrast, in the United States and Western Europe, chronic alcohol consumption is considered the major underlying risk factor for HCC (El-Serag, 2004; McKillop and Schrum, 2009).
The liver is the major site of alcohol metabolism following ingestion, and many of the detrimental effects of alcohol are directly linked to hepatic alcohol metabolism (McKillop and Schrum, 2009; Wu and Cederbaum, 2003). In the setting of moderate, infrequent alcohol consumption the majority of alcohol is oxidized by alcohol dehydrogenase (ADH) to acetaldehyde, a known carcinogen, which is rapidly oxidized to acetate by acetaldehyde dehydrogenase (ALDH) (Crabb and Liangpunsakul, 2007; McKillop and Schrum, 2009). In the setting of chronic, heavy alcohol consumption cytochrome p450IIE1 (CYP2E1) is induced leading to sustained acetaldehyde production, reactive oxygen species (ROS) generation and increased hepatic oxidative stress (Lu and Cederbaum, 2008). If left unchecked acetaldehyde-ROS adduct formation with proteins, lipids and DNA persists, leading to accumulated genetic instability and hepatocyte transformation (Lu and Cederbaum, 2008; McKillop and Schrum, 2009; Wu and Cederbaum, 2003).
Identifying biochemical pathways involved in alcohol metabolism has been central to our understanding of the progressive nature of alcohol-dependent liver disease (ALD). However, the precise role of alcohol in hepatic carcinogenesis remains to be fully elucidated. In large part, many of the challenges associated with defining the specific role of alcohol in HCC initiation and/or progression are due to the multicellular, intra-hepatic, and extra-hepatic/systemic effects alcohol exerts following consumption (McKillop and Schrum, 2005; McKillop and Schrum, 2009). In addition, the role of alcohol as a hepatic carcinogen has been compromised by the lack of reliable rodent model(s) of alcohol-only induced hepatic neoplasia (Bell et al., 2006; Ponnappa and Rubin, 2000; Siegmund et al., 2003).
Given free choice, most rodents do not voluntarily consume significant amounts of alcohol (Ponnappa and Rubin, 2000); as a result, previous studies have employed models in which rodents are “forced” to consume alcohol (Siegmund et al., 2003). These include the Lieber-DeCarli diet, direct gastric delivery of alcohol by gavage or the Tsukamoto-French intragastric feeding model, weaning and maintenance of animals onto alcohol in drinking water, or continuous alcohol vapor inhalation (Ponnappa and Rubin, 2000; Siegmund et al., 2003). Each of these models has merit in studying ALD; however, none reports the ability to form HCC in the absence of other factors such as dietary modifications and/or chemical insult. In addition, they do not accurately mimic the human condition of alcoholism characterized by alcohol preference, voluntary drinking, and nutrient deficiencies. Forced exposure to alcohol also introduces a stress that is not encountered by human alcoholics who seek and voluntarily consume alcohol.
In the 1970’s, a selective breeding program for high ethanol preference resulted in the derivation of an alcohol-preferring (P) rat line that voluntarily drinks large quantities of alcohol (Li et al., 1993). Another line known as inbred P rats or “iP” rats was also derived that voluntarily consumes moderate quantities of alcohol (Carr et al., 1998). The original P rat line was developed through mass selection from a foundation stock of “outbred” Wistar rats (Li et al., 1993). Animals selected for breeding in the alcohol-preferring P line consumed >5 grams alcohol/kg body weight (BW)/day, and demonstrate a >2:1 preference ratio for alcohol over water. When given free choice between a 10% (v/v) alcohol solution and water, P rats voluntarily consume 6–8g alcohol/kg BW/day (Chester et al., 2004) and attain blood alcohol concentrations (BACs) of 50–150 mg% (Murphy et al., 1986). After adjusting for the 4-fold greater alcohol metabolism rate in rats compared to humans, alcohol intake in P rats is comparable to that required to qualify as an alcoholic using DSM-IV-TR criteria (2000). That is, intake of 6 grams alcohol/kg BW/day is equivalent to an intake of 1.5 grams alcohol/kg BW/day in humans (~8–9 standardized alcoholic drinks/70kg person/day). In comparison, iP rats consume approximately half as much alcohol (3–4 g/kg BW/day) as the P rats.
Previous studies report that P rats meet all five criteria of an animal model of alcoholism (McMillen, 1997). They maintain alcohol consumption in the presence of other palatable solutions (Lankford et al., 1991), find alcohol reinforcing as evidenced by the fact that they will orally self-administer alcohol (Murphy et al., 1989), consume alcohol for its central nervous system reinforcing effects (Murphy et al., 1988), and develop physical dependence with chronic free-choice drinking (Kampov-Polevoy et al., 2000). In light of the extensive characterization of the P rat line, this model represents the best rodent model of alcoholism currently available. However, the effect of alcohol on hepatocarcinogenesis in these rat lines has not been studied.
In this study, we employed P and iP rats given free access to water and standard rodent chow in the presence or absence of water containing 10% (v/v) alcohol. We confirm significantly increased alcohol consumption in P rats compared to iP rats. Furthermore, after 18 months of alcohol consumption, the incidence and multiplicity of hepatic tumors increased in alcohol-consuming P rats compared to alcohol-consuming iP rats and P rats with access to water only. Hepatic tissue from alcohol-consuming P rats was further characterized by increased MAPK/ERK phosphorylation, CYP2E1 expression/activity and intrahepatic oxidative stress.
An Ethanol L3K assay kit was purchased Genzyme Diagnostics (Framingham, MA). Horseradish peroxidase polymer conjugate was purchased from Zymed (San Francisco, CA). Primary antibody against GSTp was purchased from Biogenex. Primary antibody against phospho-ERK was purchased from Cell Signaling (Danvers, MA). CD68 specific antibody (clone ED1) was purchased from AbD Serotec (Raleigh, NC). Anti-Ki67 antibody, liquid DAB+ substrate chromogen, and the avidin-biotin and protein blocks were purchased from Dako (Carpinteria, CA). Hematoxylin QS counterstain was purchased from Vector Laboratories (Burlingame, CA). Radioimmuno precipitation assay (RIPA) buffer, HALT protease and phosphatase inhibitors were purchased from Thermo Fisher Scientific (Rockford, IL). Antibodies specific against CYP2E1 were purchased from Abcam (Cambridge, MA), and against ADH, ALDH1A1, and β-actin from Santa Cruz Biotechnology (Santa Cruz, CA). All secondary antibodies were purchased from Santa Cruz Biotechnology. TRIZOL reagent was purchased from Invitrogen (Carlsbad, CA). The IMPROM II™ transcription system (used for first strand cDNA synthesis) and GoTaq green master mix (for PCR amplification) were purchased from Promega (Madison, WI). A thiobarbituric acid reactive substances (TBARS) kit was purchased from Cayman Chemical Co. (Ann Arbor, MI). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO).
Male rats (250–300g starting weight) from the alcohol-preferring P and iP lines (Indiana Alcohol Research Center Animal Production Core, Indianapolis, IN) were used for these studies. All experiments were approved by the Institutional Animal Care and Use Committee and conform to the Federal Care and Use of Laboratory Animals guidelines.
Animals were housed individually with standard bedding and rat chow in a temperature/humidity-controlled room with 12-hour light/dark cycles. At 6 months of age, alcohol-naïve rats were randomly assigned to control (water only) or alcohol access groups. Rats in the alcohol access group were given continuous 24-hour free choice between two calibrated glass bottles: one containing water, the other containing 10% (v/v) alcohol. The alcohol containing solution was made weekly using 96% (v/v) ethanol diluted with water to give a final alcohol content of 10% (v/v). Rat weights and alcohol intake were recorded weekly.
After 6, 12 or 18 months (P rats) or 18 months (iP rats) of water alone or water with alcohol access, animals were euthanized by carbon monoxide asphyxiation. Blood was obtained by cardiac puncture and plasma prepared and stored (−80°C). The liver was excised, weighed and grossly examined for tumors. Representative sections from three different lobes were fixed in 10% (v/v) formalin and paraffin embedded for histology. Histopathological evaluation was performed by a certified pathologist, blinded to treatment groups. Additional liver tissue was frozen in liquid nitrogen and stored (−80°C) for biochemical analyses.
Blood alcohol concentrations (BACs) were measured using the Ethanol L3K assay according to the manufacturer’s protocol. Samples were assayed in triplicate, and relative ethanol concentration calculated from known standards and expressed in mmol/L. Measurement of plasma alanine aminotransferase (ALT), aspartate aminotransferase (AST) and albumin levels were performed by Clarian Pathology Laboratories (Indianapolis, IN).
Anesthesia was induced and maintained by isoflurane inhalation. Ultrasound images of the liver were obtained and measured using a Vevo 770 (VisualSonics, Toronto, Canada) micro-ultrasound probe (Fortune-Fry Micro-Imaging Ultrasound Core, Indiana University). Animals were allowed to recover and returned to their respective experimental groups.
Liver sections, 4 µm in thickness, were deparaffinized and hydrated. Antigen retrieval was done in 140 mM citrate buffer (pH 6.0) in a decloaking chamber (BioCare, Walnut Creek, CA) at 115°C for 3 minutes. The slides were transferred to boiling deionized water and cooled for 20 minutes at room temperature (RT). After blocking of endogenous peroxidase, endogenous biotin and non-specific proteins, the slides were incubated with primary antibody (GSTp 1:100, 2 hours RT; Ki-67 1:50, 3 hours RT; P-ERK 1:500, 30 minutes RT; CD68 1:100, 30 minutes RT), washed in Optimax, followed by incubation with horseradish peroxidase polymer conjugate (10 minutes, RT). Staining was visualized using a DAB+ substrate chromogen solution and hematoxylin QS counterstain. Preneoplastic lesions were defined as clusters of hepatocytes (> 10 cells) with clearly defined lesion borders that stained positively for GSTp. GSTp-positive foci were quantified using image analysis and sterologic software (Bioquant, Nashville, TN). P-ERK positive cells were counted in three fields (400X) with the highest density of staining. Staining was expressed as a percentage of the total number of cells per field.
Representative liver tissues from animals within each group were homogenized in radioimmune precipitation (RIPA) buffer containing HALT protease and phosphatase inhibitors. Total liver lysates were centrifuged (10,000 rpm, 10 minutes, RT) and protein concentrations determined by the method of Bradford and equalized using RIPA buffer. Individual lysates as well as pooled samples (equal aliquots of individual lysates from each experimental group) were stored at −80°C prior to analysis. Protein expression was determined in individual as well as pooled samples by SDS-PAGE and Western blot as previously reported (Hennig et al., 2009). Primary antibodies were diluted 1:5000 (CYP2E1), 1:2000 (ADH) or 1:1000 (ALDH) in 5% (w/v) nonfat dry milk dissolved in 0.05% (v/v) Tween-20/Tris-buffered saline (TBS). Band intensities of the individual and pooled samples were determined by densitometric analysis using NIH-Image software, and equal protein loading confirmed by stripping/re-probing membranes using an anti-β-actin antibody.
Total hepatic RNA was extracted using TRIZOL as per the manufacturer’s instructions. Total RNA was quantified using a Nanodrop 2000c (Thermo Scientific, Wilmington, DE), adjusted to an equal concentration, and combined for each experimental group. To perform RT-PCR, 1µg of total RNA was reverse transcribed by Improm II™ using random hexamers to generate complementary DNA (cDNA) as per the manufacturer’s protocol. The cDNA was precipitated and re-suspended in nuclease-free water and 100ng cDNA was used in each reaction with the following gene-specific primers: CYP2E1 forward primer (fp) 5’CCTACATGGATGCTGTGGTG 3’ and reverse primer (rp) 5’CTGGAAACTCATGGCTGTCA 3’; β-actin (fp) 5’GAGCTATGAGCTGCCTGACG 3’, and (rp) 5’GGATGTCAACGTCACACTTC 3’. PCR reactions consisted of 95°C for 2 minutes followed by 35 cycles of 95°C, 55°C and 72°C for 45 seconds each with a final elongation at 72°C for 2 minutes. Amplified products were resolved using a 1.5% (w/v) agarose gel and stained with ethidium bromide before conversion to digital images (Brandon-Warner et al., 2009). The level of each PCR product was determined using Quantity One software.
CYP2E1 activity in hepatic tissue was assessed as rate of p-nitrophenol hydroxylation to 4-nitrocatechol as previously reported (Brandon-Warner et al., 2009; Koop, 1986). Briefly, tissue was placed in KH2PO4 (pH 6.8, 4°C), sonicated, and total protein determined and corrected to 25mg/ml. The reaction mixture (100µM p-nitrophenol and 1mM NADPH) was added to each sample, assayed in duplicate, and incubated for 10 minutes at 37°C. The reaction was terminated by adding 300µl of 60% (v/v) 1N perchloric acid. Samples were centrifuged (10,000×g, 5 minutes, 4°C) and the supernatant removed. 100µl of 10M NaOH was added to the supernatant and samples read on a spectrophotometer at 540nm. Finally, activity was calculated against a standard curve of known 4-nitrocatechol concentrations.
As a marker of intrahepatic oxidative stress, thiobarbituric acid reactive substances (TBARS) were measured using a commercially available kit as per the manufacturer’s instructions. Samples were prepared and assayed in duplicate.
Statistical methods including T-test or one-way ANOVA with Dunnett post tests were performed using Prism 3.02 software (Graphpad, San Diego, CA). p<0.05 was considered significant.
Analysis of alcohol consumption confirmed that P rats reproducibly consumed significantly more alcohol than iP rats when presented with identical bottles containing either water alone or 10% (v/v) alcohol (8.9 ± 0.4 vs. 3.3 ± 0.2 mg/kg BW/day, Figure 1A). Plasma blood alcohol concentrations (BAC) were elevated with no significant difference between P rat groups drinking alcohol for 6, 12, and 18 months (Figure 1B). No significant differences in plasma AST, ALT or albumin were detected in water- versus alcohol-drinking P rats at 6, 12 or 18 months (data not shown). Sequential body weight measurement of P rats drinking water or alcohol demonstrated that both groups gained weight at an equal rate throughout the experimental period (Figure 1C), and no significant differences in liver:body weight ratios were detected at necropsy (Figure 1D).
Gross visualization of the liver at necropsy failed to detect neoplasms in P rats given free access to water or 10% (v/v) alcohol for 6 months (Table 1A). At 12 months, one hepatic tumor (1/12 [8.3%]) was detected in P rats given free access to 10% (v/v) alcohol; no gross lesions were observed in P rats with access to water only. However, after 18 months, a highly significant increase in gross liver lesions was observed in P rats given access to alcohol compared to water only (15/18 [83.3%] vs. 1/20 [5.0%] respectively, Table 1A and Figure 2A). Tumor multiplicity was also significantly higher in P rats chronically drinking alcohol for 18 months compared to water only (1.94 ± 1.40 vs. 0.05 ± 0.20 gross lesions/animal respectively, Table 1B). The incidence and multiplicity of macroscopically visible liver nodules were further confirmed by performing high-resolution ultrasound in a subset of P rats that drank alcohol for 18 months (Figure 2B and Table 1C). By ultrasound, average liver lesion size was determined to be 3.6 ± 1.5 mm in P rats given access to alcohol (Table 1C). In contrast, examination of livers from iP rats did not demonstrate any detectable gross neoplastic lesions in animals with free access to water only or alcohol for 18 months (0/5 per group [0%]).
Histological analysis was performed using representative sections from alcohol- or water-consuming animals after 6, 12 and 18 months of drinking. Hematoxylin and eosin (H&E)-stained liver sections of P rats given free access to alcohol for 18 months revealed well-demarcated expansile nodules compressing the adjacent non-tumorous hepatic parenchyma. The tumor nodules appeared well-differentiated (Figures 3A and B). Staining for reticulin fibers, an established criterion for diagnosis of well-differentiated HCC (Yang et al., 2004), demonstrated an absence of reticulin fibers in the neoplastic nodules. Conversely, reticulin fibers were present in normal numbers in adjacent liver (Figure 3C).
To further analyze focal masses detected in hepatic tissue, we performed immunohistochemistry using an antibody against glutathione-S-transferase pi-class (GSTp; marker of preneoplasia/neoplasia) to localize GSTp-positive (GSTp+) lesions in P rats allowed free access to water or alcohol (Figure 4A). Because the rats were enrolled in this study at 6 months of age, the existence of a low level of background preneoplastic lesions associated with age is expected (Schulte-Hermann et al., 1983). Significantly increased numbers of microscopic GSTp+ foci were detected in P rats after 18 months of drinking alcohol compared to the water only group (Figure 4B). In addition, net cross-sectional diameter of GSTp+ foci was significantly greater in P rats allowed access to alcohol for 12 and 18 months compared to water only counterparts (Figure 4C). Similar results were observed in a parallel series in which P rats, beginning at a younger age of 6 weeks, were given free access to water or alcohol for 12 and 18 months (Supplementary Figure 1).
The number of Ki-67 positive nuclei within GSTp+ lesions in P rats drinking alcohol for 6 and 12 months significantly increased compared to water only counterparts, confirming increased alcohol-dependent hepatocyte proliferation in P rats consuming alcohol (Figure 4D). Increased staining of phosphorylated mitogen-activated protein kinase/extracellular signal-regulated kinase 1/2 (P-MAPK/ERK) was also observed in sinusoidal lining cells (Kupffer or endothelial cells) within livers of alcohol-consuming rats relative to water only P rats; endothelial cells lining the central vein also expressed P-ERK (Figures 5A & B). To determine whether the P-ERK-positive cells were Kupffer cells, consecutive serial slides were stained with either CD68 or P-ERK (Figure 5B). CD68-positive Kupffer cells were identified within the sinusoids; however, cells that positively stained for P-ERK did not correspondingly stain positive for CD68. These data suggest that P-ERK-positive cells are predominantly sinusoidal endothelial cells. In addition, Western blot analysis was performed to determine P-ERK expression in liver tissue Lysates. Using this approach, there was no significant difference in phosphorylated or total ERK levels between the water- and alcohol-drinking groups (data not shown).
Although no gross neoplastic lesions were present in the livers of iP rats, microscopic GSTp+ hepatic focal precancerous lesions were quantified (Figure 6). These data demonstrated a trend towards more precancerous lesions in alcohol-consuming iP rats compared to water only counterparts at 18 months (Figure 6A). The number of large (>15,000µm2) GSTp+ foci and the area occupied by GSTp+ foci were significantly greater in the alcohol groups (Figures 6B and C).
Western blot analysis was initially performed with pooled hepatic lysates using specific antibodies against ADH, ALDH and CYP2E1 (Figure 7A). These data demonstrated no detectable differences in ADH expression in liver tissue resected from P rats with or without access to alcohol, naïve P rats (6 months of age), or normal Wistar rats (strain from which the P rat line was originally derived). Similarly, there were no differences in ALDH expression detected between P rats given access to water, naïve P rats, and Wistar rats. However, ALDH expression was decreased in P rats after 18 months of free access to alcohol (78% decrease compared to naïve P-rats at 6 months of age). These results were also confirmed by quantifying ADH/ALDH expression following Western blot analysis of individual hepatic lysates. No significant difference in ADH expression was observed; a trend toward lower ALDH levels in the 12- and 18-month alcohol groups was apparent (Figure 7B). Analysis of CYP2E1 expression in the pooled fractions demonstrated a ~2.5-fold increase in CYP2E1 protein in P rats allowed free access to alcohol for 6, 12 and 18 months. The data generated with pooled fractions was confirmed by analyzing samples individually and showed a significant increase in CYP2E1 expression in P rats consuming alcohol (Figure 7B). Similarly, hepatic CYP2E1 mRNA expression increased in P rats allowed free access to alcohol compared to the water only group after 6, 12 and 18 months (Figure 7C).
CYP2E1 activity was measured as the ability to hydroxylate p-nitrophenol to 4-nitrocatechol. These data demonstrate no significant differences in CYP2E1 activity in P rats allowed access to water only compared to naïve P-rats or control Wistar rats (Figure 7D). In contrast, significantly increased CYP2E1 activity was measured in liver tissue from P rats allowed access to alcohol for 6 months, rising to significantly higher activity at 12 and 18 months (Figure 7D). To assess potential relevance of CYP2E1 activity to hepatic REDOX status, hepatic thiobarbituric acid reactive substances (TBARS) were measured. No significant difference in TBARS was observed in P rats allowed access to water only (6, 12 or 18 months) compared to naïve P-rats or control Wistar rats (Figure 7E). Conversely, significant increases in TBARS were measured in liver tissue from P rats allowed access to alcohol compared to age-matched P rats allowed access to water only, a maximum increase in TBARS being measured in the 18-month alcohol access group.
HCC is one of the most common causes of mortality from solid organ malignancy. Developing an animal model of alcohol-induced liver neoplasia would be an invaluable tool for studying the role of alcohol consumption in HCC and possibly other alcohol-related cancers. Since most rodents do not voluntarily consume alcohol (Lieber et al., 1989; Ponnappa and Rubin, 2000; Siegmund et al., 2003), the effects of alcohol have been assessed in rodents receiving forced, chronic alcohol exposure via alcohol-containing liquid diets as the sole source of fluid, or continuous housing in alcohol vapor inhalation chambers (Becker and Veatch, 2002; Siegmund et al., 2003). Force feeding alcohol in a liquid diet, such as the Lieber-DeCarli diet, has become standard protocol for assuring high alcohol intake in rodents. With this diet, the only source of both nutrients and water is a liquid diet that can be supplemented with alcohol (Lieber et al., 1989). Although the animals have an aversion to alcohol, they must consume this diet or starve (Lieber et al., 1989; Ponnappa and Rubin, 2000). The Lieber and DeCarli diet also provides isocaloric maltodextrin in control-fed animals, minimizing nutrient deficiencies that are characteristic of human alcoholism. When given over a period of 4 to 6 weeks, this diet does not cause liver necrosis, inflammation, or fibrosis, but can result in fatty liver (DeCarli and Lieber, 1967; Ponnappa and Rubin, 2000). The important limitations of forced alcohol exposure models are that they do not mimic the human condition of alcoholism, characterized by a preference for alcohol and voluntary consumption of alcohol which is accompanied by nutrient deficiencies. These significant differences make translation of the findings from forced alcohol exposure in rodents to human alcoholics difficult.
Another variable that should be considered when relating animal models of alcohol consumption to human drinkers is that the effect of alcohol on hepatic tumorigenesis may be influenced, in part, by pattern of alcohol intake. Blood alcohol concentration (BAC) resulting from voluntary alcohol drinking is characterized by moderate/high BAC associated with each alcohol drinking bout (“intermittent”). This pattern of drinking may allow periods of liver recovery and/or regeneration after alcohol-induced hepatocyte necrosis. This differs from the BAC produced by forced alcohol vapor inhalation (continuous elevation of BAC). In addition to route of administration, whether alcohol is given simultaneously or separately from food may also affect hepatic tumorigenesis. Nonetheless, the results of prior studies with forced alcohol exposure are important and serve as the basis for the present study.
The P rat model employed in our study meets all five criteria of an animal model of alcoholism. One of the strengths of the model is that it mimics human behavior because the rat voluntarily drinks large amounts of alcohol, comparable to amounts consumed by alcoholic patients. We showed that gross neoplastic lesions developed almost exclusively in the livers of P rats that had been drinking alcohol for 18 months. Tumor incidence, multiplicity, and size (determined either grossly or by ultrasound) were increased in P rats drinking alcohol for 18 months compared to the water drinkers. To our knowledge, this is the first time an animal model of chronic alcohol intake/alcoholism has demonstrated formation of liver neoplasia. Histopathologically, we have classified the neoplastic lesions as well-differentiated HCC, but they also bear similarity to adenoma or dysplastic foci.
While these data demonstrate hepatic carcinogenesis in a rat model of voluntary alcohol consumption, no evidence of alcohol-induced hepatic cirrhosis occurred in P rats drinking alcohol. In the clinical setting, HCC associated with alcohol abuse most commonly arises in the setting of cirrhosis, and the risk for HCC in decompensated alcohol-induced cirrhosis approaches 1% per year (Morgan et al., 2004). Importantly, HCC may also occur in persons with alcohol-induced liver disease who do not have cirrhosis (Donato et al., 2002; Morgan et al., 2004). Of interest, the iP rats, which consume approximately half as much alcohol as P rats, did not develop gross neoplastic liver lesions. This result supports studies which suggest that chronic alcohol use is a risk factor for HCC in humans, whereas moderate use (1–3 drinks/day) is not (Yu and Yuan, 2004). As such, it will be of considerable interest to expand these studies to address the incidence and degree of HCC that arises in this model when underlying hepatic cirrhosis is induced.
Because the rats in this study were 6 months of age when they began consuming alcohol, the existence of a low level of background preneoplastic lesions associated with age is expected (Schulte-Hermann et al., 1983). Livers from water-drinking rats had a background level of GSTp+ lesions that increased in number and size in alcohol-drinking iP and P rats. This suggests that alcohol may promote the growth of preneoplastic focal lesions from pre-existing initiated cells that arise in the liver as the rat ages. This is supported by our results from a related study in which P rats were given free access to water or alcohol beginning earlier at 6 weeks of age (Supplementary Figure 1). The number of GSTp+ lesions was significantly higher in P rats enrolled in the study at 6 months of age (Figure 4B) compared to 6 weeks of age after drinking water for 18 months (corresponding to 24 vs. 19.5 months of age respectively). Similarly, GSTp+ lesion size was significantly greater in P rats enrolled in the study at 6 months (Figure 4C) versus 6 weeks of age after drinking water for 12 or 18 months. Thus, in the absence of alcohol, number and size of background preneoplastic hepatic lesions increases with age. In P rats enrolled at 6 weeks of age, tumor incidence was 50% (data not shown) after drinking alcohol for 18 months compared to 83% with enrollment at 6 months of age. Taken together, our results suggest that hepatic tumorigenesis is influenced by age at which alcohol consumption is initiated.
In the liver, alcohol is metabolized to acetaldehyde by ADH, with acetaldehyde subsequently being converted to acetate by ALDH (Crabb and Liangpunsakul, 2007; McKillop and Schrum, 2009). Both reactions result in the production of NADH, leading to reactive oxygen species (ROS) production and oxidative stress (Crabb and Liangpunsakul, 2007). ROS can cause peroxidation damage to lipids, proteins and DNA (Wu and Cederbaum, 2003). Alcohol metabolism and alcohol-induced oxidative stress also generate reactive aldehydes (malondialdehyde [MDA] and 4-hydroxy-2-nonenal [HNE]) in addition to acetaldehyde, all of which can form potentially harmful adducts (Niemela, 2001). While ADH/ALDH are involved in the oxidation of most alcohol consumed, cytochrome P450IIE1 (CYP2E1) becomes more important for metabolism during chronic alcohol consumption (Lu and Cederbaum, 2008). CYP2E1 activity is linked to alcohol-associated toxicity through the production of ROS and further oxidative stress. In H4IIE HCC cells, it has been reported that ethanol, at doses similar to blood ethanol content observed following moderate ethanol consumption (10–25 mmol/L), significantly increases CYP2E1 expression and activity without affecting ADH or ALDH levels in vitro (Brandon-Warner et al., 2009).
In the present study, we evaluated whether alcohol preference was determined by differences in the basal levels of alcohol metabolizing enzymes. The protein levels of CYP2E1, ADH and ALDH were similar in the Wistar background control rats, alcohol-naïve P rats at 6 months of age, and P rats drinking water for 6, 12 and 18 months (corresponding to 12, 18 and 24 months of age respectively). This suggests that differences in basal liver enzyme levels do not contribute to alcohol preference or predispose the rats to alcohol drinking, nor does expression of these enzymes change with age in the P rat model. In contrast, long-term alcohol drinking induced CYP2E1 protein levels, decreased ALDH levels, but did not alter ADH levels in the P rats. The alcohol-induced changes in CYP2E1 protein levels were confirmed by increased CYP2E1 mRNA expression as well as enzymatic activity. The combined effect of higher CYP2E1 and lower ALDH level/activity would be predicted to result in the accumulation of acetaldehyde, ROS and oxidative stress. This was supported by the detection of elevated MDA levels, indicative of ROS formation and subsequent lipid peroxidation, in the livers of alcohol-drinking P rats. The timing and extent of these intracellular changes coincided with the detection of preneoplastic and neoplastic changes in the liver after 12 months and especially after 18 months of alcohol drinking. Although our results provide evidence that these events may be associated, causal association remains to be proven experimentally in this model.
Finally, we report that staining of phosphorylated MAPK/ERK increased in sinusoidal lining cells within the livers of rats consuming alcohol for 18 months. We further determined that P-ERK-positive cells within the sinusoids appear to be predominantly endothelial, not Kupffer, cells based upon staining serial slides with P-ERK and CD68. Studies to explore the significance of this finding are ongoing in our laboratory.
In summary, we report that chronic alcohol consumption is associated with hepatic neoplasia, MAPK/ERK activation, increased CYP2E1 activity, and intrahepatic oxidative stress in P rats. This novel animal model of alcohol-induced liver neoplasia will likely be an invaluable tool for further elucidating the mechanisms which support this specific etiology and may further our understanding of the general pathogenic mechanisms of alcohol-induced cancers.
Supplementary Figure 1. Analysis of GSTp+ hepatic lesions in P rats given free access to water or alcohol beginning at 6 weeks of age. A) The number of GSTp+ lesions/cm2 was determined in alcohol- and water-consuming P rats after 12 and 18 months (n=3–8 animal/group). B) Cross-sectional diameter (µm) of GSTp+ lesions was measured in P rats after 12 and 18 months (n=21–121 lesions per group). Results are expressed as mean ± SEM. *p< 0.05.
Financial Support: NIH AA16360, Clarian Values Grant, and Indiana Genomics Initiative (INGEN, supported in part by Lilly Endowment Inc.) of Indiana University (CMS); NIH AA10709 and AA007611 (JCF); NIH AA016858 (IHM)
Disclosures: None of the authors have any potential conflicts of interest.