Laboratory of Liver Diseases, National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD
Address reprint requests to: Bin Gao, M.D., Ph.D., Laboratory of Liver Diseases, National Institute on Alcohol Abuse and Alcoholism/National Institutes of Health, 5625 Fishers Lane, Bethesda, MD 20892. E-mail:firstname.lastname@example.org; fax: 301-480-0257.
Potential conflict of interest: Nothing to report.
See Editorial on Page 1526
Chronic plus binge ethanol feeding acts synergistically to induce liver injury in mice, but the mechanisms underlying this phenomenon remain unclear. Here, we show that chronic plus binge ethanol feeding synergistically up-regulated the hepatic expression of interleukin-1β and tumor necrosis factor alpha and induced neutrophil accumulation in the liver, compared with chronic or binge feeding alone. In vivo depletion of neutrophils through administration of an anti-Ly6G antibody markedly reduced chronic-binge ethanol feeding-induced liver injury. Real-time polymerase chain reaction analyses revealed that hepatic E-selectin expression was up-regulated 10-fold, whereas expression of other neutrophil infiltration-related adhesion molecules (e.g., P-selectin, intercellular adhesion molecule 1, and vascular cell adhesion molecule 1) was slightly up- or down-regulated in this chronic-binge model. The genetic deletion of E-selectin prevented chronic-binge ethanol-induced hepatic neutrophil infiltration as well as elevation of serum transaminases without affecting ethanol-induced steatosis. In addition, E-selectin-deficient mice showed reduced hepatic expression of several proinflammatory cytokines, chemokines, and adhesion molecules, compared to wild-type mice, after chronic-binge ethanol feeding. Finally, the expression of E-selectin was highly up-regulated in human alcoholic fatty livers, but not in alcoholic cirrhosis. Conclusions: Chronic-binge ethanol feeding up-regulates expression of proinflammatory cytokines, followed by the induction of E-selectin. Elevated E-selectin plays an important role in hepatic neutrophil infiltration and injury induced by chronic-binge feeding in mice and may also contribute to the pathogenesis of early stages of human alcoholic liver disease. (Hepatology 2013;58:1814–1823)
More than 95% of heavy drinkers develop fatty liver, but only 20%-40% develop steatohepatitis and progress to cirrhosis and hepatocellular carcinoma.[1, 2] The underlying mechanisms that render some individuals more susceptible to severe forms of alcoholic liver disease (ALD) are not clear, and many risk factors may be involved.[1-3] These potential risk factors include sex, obesity, dietary factors, smoking, and non-sex-linked genetic factors.[1-3] Additionally, several epidemiologic studies have suggested that the drinking pattern also significantly influences the pathogenesis of ALD in humans.[4-7] For example, drinking outside mealtime, the consumption of multiple types of drinks, and a mixed drinking pattern were significantly associated with an increased risk of ALD.[6, 7] However, the mechanisms through which drinking pattern affects ALD pathogenesis remain largely unknown. Previously, we have demonstrated that chronic ethanol feeding for 10 days plus a single binge dose of ethanol delivered by gavage synergistically elevated serum levels of alanine aminotransferase (ALT), induced steatosis, and up-regulated expression of proinflammatory cytokines, compared with chronic or gavage ethanol feeding alone. This pattern of chronic-binge feeding reproduces closely the drinking behaviors in many heavy drinkers, which may be associated with an increased risk of ALD.[6, 7, 9, 10] In the current study, we further characterized liver inflammation in this mouse model of chronic-binge feeding. Our results showed that chronic-binge feeding synergistically induced hepatic neutrophil infiltration, but not macrophage infiltration, compared with chronic or binge feeding alone.
Neutrophil infiltration is a hallmark of alcoholic hepatitis[11-13] and has been shown to correlate with severity of alcoholic hepatitis. Feeding rats with a diet including ethanol also increased hepatic neutrophil infiltration, and depletion of neutrophils reduced liver injury in ethanol-fed rats. This suggests that neutrophil infiltration likely contributes to hepatocellular damage, possibly by killing hepatocytes through production of reactive oxygen species (ROS) and proteases. However, the mechanisms through which neutrophils are recruited to the liver during alcoholic liver injury remain incompletely understood. Neutrophil recruitment is controlled by a multistep adhesion cascade that involves multiple adhesion molecules and their ligands, which are expressed on endothelial cells (ECs) and neutrophils, respectively.[17-19] These include the endothelium-expressed molecules, E-selectin (SELE) and P-selectin (SELP), vascular cell adhesion molecule 1 (VCAM-1), and intercellular adhesion molecule 1 (ICAM-1) and the neutrophil-expressed proteins, L-selectin (CD62L) and E- and P-selectin ligands, including E-selectin ligand 1 (ESL-1), P-selectin glycoprotein ligand 1 (PSGL-1), and CD44.[17, 18] To examine the mechanisms underlying the synergistic effects of chronic-binge feeding on neutrophil infiltration, we evaluated the expression of an array of adhesion molecules and their ligands. Our results revealed that among these factors, E-selectin had the highest induction (up to 10-fold) in the liver after chronic-binge ethanol feeding.
E-selectin, also called CD62 antigen-like family member E (CD62E), endothelial-leukocyte adhesion molecule 1, or leukocyte-EC adhesion molecule 2, is an adhesion molecule that is specifically expressed on activated ECs. Through interacting with an array of ligands, including ESL-1, CD44, and PSGL-1, E-selectin plays a critical role in the transition from slow rolling to arrest as well as promotion of the efficient transendothelial migration and activation of neutrophils.[17, 20, 21] In the present study, we demonstrated that the genetic disruption of E-selectin abolished chronic-binge ethanol-induced neutrophil infiltration, liver injury, and inflammation, suggesting that E-selectin plays a critical role in inducing neutrophil recruitment, liver inflammation, and injury after chronic-binge feeding.
Materials and Methods
Mice and Ethanol Feeding Protocols
E-selectin-deficient (SELE−/−) mice and wild-type (WT) controls (C57BL/6J) were purchased from The Jackson Laboratory (Bar Harbor, ME). All animal experiments were approved by the National Institute on Alcohol Abuse and Alcoholism Animal Care and Use Committee. Eight- to twelve-week-old female mice were subjected to one of three different feeding protocols. The first feeding protocol involved chronic feeding, in which mice were initially fed the control Lieber-DeCarli diet (Bio-Serv, Frenchtown, NJ) ad libitum for 5 days to acclimatize them to a liquid diet. Then, mice were allowed free access to the ethanol Lieber-DeCarli diet (Bio-Serv) containing 5% (vol/vol) ethanol for 10 days, and control-fed groups were pair-fed with an isocaloric control diet. The second feeding protocol involved binge feeding, in which mice were gavaged with a single dose of ethanol (5 g/kg body weight [b.w.]) or isocaloric dextrin-maltose in the early morning and sacrificed 9 hours later. The third feeding protocol (the NIAAA model) involved chronic-binge feeding,[8, 22] in which mice were fed the control or ethanol Lieber-DeCarli diet for 10 days, as described for the chronic feeding group. On day 11, ethanol- and pair-fed mice were gavaged in the early morning with a single dose of ethanol (5 g/kg b.w.) or isocaloric dextrin-maltose, respectively, and sacrificed 9 hours later. WT and SELE−/− mice exhibited comparable daily alcohol intakes. Mean body weights were similar in all groups at the end of the experiments.
Results are expressed as the means ± standard error of the mean of 5-21 mice per group. Group comparisons were performed using an unpaired t test or one-way analysis of variance, followed by Tukey's multiple comparison test. Correlations were analyzed using Spearman's rank correlation test. P < 0.05 was considered statistically significant.
Additional methods are described in the Supporting Materials.
Chronic-Binge Feeding Synergistically Induces Liver Injury and Inflammation in Mice
We previously demonstrated that chronic-binge feeding caused significant liver injury that peaked at 9 hours after gavage in C57BL/6 mice, and that serum ALT and aspartate aminotransferase (AST) levels were higher in female than male mice. Consequently, we used female mice in this study to further explore the mechanisms underlying the synergistic effect of chronic-binge feeding on liver injury. Alone, ethanol feeding for 10 days (chronic) or a single gavage of ethanol (binge) induced only a mild elevation of serum ALT and AST in female C57BL/6J mice (Fig. 1A). In contrast, as expected, chronic-binge ethanol feeding resulted in more severe liver injury, as indicated by much higher serum ALT and AST levels (Fig. 1A). Furthermore, chronic-binge feeding, but not 10-day chronic feeding or single ethanol gavage alone, up-regulated hepatic expression of the proinflammatory cytokines, tumor necrosis factor alpha (TNF-α), interleukin (IL)−1β, and IL-6, and the chemokines, monocyte chemotactic protein 1 (MCP-1), macrophage inflammatory protein (MIP)−1α, MIP-1β, and MIP-2 (Fig. 1B,C). Interestingly, hepatic expression of IL-10 was up-regulated after chronic feeding alone, but not after binge or chronic-binge feeding.
Chronic alcohol consumption is known to induce cytochrome P450 (CYP)2E1, deplete glutathione, and increase production of reactive oxygen species and OS in the liver. We hypothesized that these changes may contribute to the increased susceptibility of 10-day alcohol-fed mice to binge-induced liver injury. Therefore, we investigated the effect of 10-day alcohol feeding on these parameters. Feeding mice with an ethanol diet for 10 days markedly induced CYP2E1 protein expression, but did not affect hepatic messenger RNA (mRNA) expression of nicotinamide adenine dinucleotide phosphate oxidase subunits (e.g., Cyba, Cybb, Ncf1, and Ncf2) or hepatic levels of malondialdehyde and glutathione (Supporting Fig. 1A-D). These results indicate that chronic ethanol feeding for 10 days is sufficient to up-regulate liver CYP2E1 protein expression, but not OS. Such up-regulated CYP2E1 likely contributes to the increased susceptibility of 10-day ethanol-fed mice to binge-induced liver injury.
Chronic-Binge Feeding Induces Hepatic Neutrophil Infiltration in Mice
The above-mentioned data show that expression of MCP-1, MIP-1α/β, and MIP-2, which are potent chemoattractants for monocytes and/or neutrophils, was up-regulated after chronic-binge feeding. Therefore, we analyzed hepatic expression of markers for these inflammatory cells in ethanol-fed mice. Hepatic gene expression levels of the macrophage marker, F4/80, and the monocyte/macrophage marker, CD68, were increased after chronic ethanol feeding (Fig. 2A). After gavage of a single dose of ethanol, gene expression of F4/80 was decreased, whereas CD68 was not altered. This is consistent with previous findings that short-term ethanol feeding increased apoptosis of F4/80-positive Kupffer cells (KCs). Expression of F4/80 and CD68 genes was unchanged and slightly increased, respectively, after chronic-binge feeding. In contrast, chronic-binge feeding, but not chronic or single ethanol gavage alone, strongly up-regulated hepatic expression of the neutrophil marker, Ly6G (Fig. 2B). Immunohistochemical (IHC) staining for myeloperoxidase (MPO) further confirmed that a large number of MPO+ neutrophils had infiltrated livers of chronic-binge-fed mice, compared with chronic or binge ethanol-fed mice (Fig. 2C and Supporting Fig. 2). Interestingly, most neutrophils were distributed throughout zones 1-3 and appeared to infiltrate into liver parenchyma (Supporting Fig. 2A,B). A small number of inflammatory foci were also observed in livers of chronic-binge–fed mice (Supporting Fig. 2C,D). Fluorescence-activated cell sorting (FACS) analyses also confirmed higher percentages of neutrophils in livers of chronic-binge–fed mice than in pair-fed mice (Fig. 2D). Furthermore, neutrophils isolated from livers of chronic-binge–fed mice exhibited enhanced cell-surface expression of CD11b, but reduced surface expression of CD62L, compared with those isolated from their pair-fed controls (Fig. 2E). Because neutrophil activation is associated with up-regulation of CD11b[24, 25] and down-regulation of CD62L expression, the results in Fig. 2E therefore suggest that liver-infiltrating neutrophils are activated in response to chronic-binge feeding.
We next investigated whether the observed increase in hepatic neutrophil infiltration contributed to chronic-binge feeding-induced liver injury by depleting neutrophils. Figure 3A shows that injection of anti-Ly6G antibody (Ab), which has been shown to specifically deplete neutrophils, reduced hepatic expression of Ly6G mRNA by 90% without affecting hepatic expression of F4/80 mRNA in the chronic-binge–fed mice. Depletion of neutrophils was further confirmed by IHC analyses showing that the number of MPO+ cells was reduced by 90% after anti-Ly6G Ab treatment (Fig. 3B). Finally, anti-Ly6G Ab treatment markedly reduced serum ALT and AST levels, but did not affect hepatic triglyceride (TG) levels (Fig. 3C,D). These data suggest that liver-infiltrating neutrophils participated in chronic-binge feeding-induced liver injury, but not steatosis.
Chronic-Binge Feeding Up-Regulates Hepatic E-Selectin Expression in Mice
To explore the mechanisms underlying the chronic-binge–mediated induction of neutrophil infiltration into the liver, hepatic expression of several adhesion molecules was examined. Chronic-binge feeding resulted in the highest fold induction of E-selectin (10-fold), whereas expression of SELP and ICAM-1 was only up-regulated by approximately 2-fold, and the expression of VCAM-1 was actually down-regulated after chronic-binge feeding (Fig. 4A). Interestingly, hepatic E-selectin was not up-regulated in the chronic or gavage-alone groups (Fig. 4B). In addition, expression levels of several E-selectin ligands (e.g., ESL-1 and CD44) and the enzymes involved in E-selectin ligand biosynthesis (e.g., FucT-IV) were significantly, albeit less than 2-fold, higher in the chronic-binge feeding groups than in the pair-fed groups (Supporting Fig. 3). Finally, hepatic expression levels of E-selectin and Ly6G were positively correlated in chronic-binge–fed mice (Fig. 4C).
E-Selectin Deficiency Reduces the Hepatic Neutrophil Infiltration Induced by Chronic-Binge Feeding
To test whether E-selectin was involved in the hepatic neutrophil recruitment induced by chronic-binge feeding, WT and SELE−/− mice were subjected to chronic-binge feeding. E-selectin deficiency did not alter hepatic expression of F4/80 or CD68 in pair- or ethanol-fed mice (Fig. 5A). In contrast, hepatic expression of the neutrophil marker, Ly6G, was strongly induced in ethanol-fed WT, but not SELE−/−, mice (Fig. 5B). IHC staining confirmed that a lower number of MPO+ neutrophils infiltrated livers of chronic-binge–fed SELE−/− mice than those of WT mice (Fig. 5C and Supporting Fig. 4). FACS analyses (Fig. 5D) showed that the percentage of neutrophils was significantly lower in livers of chronic-binge–fed SELE−/− mice, compared to WT mice. Finally, the cell-surface expression of CD11b was lower, whereas that of CD62L was higher, in liver neutrophils from chronic-binge–fed SELE−/− mice, compared to those of WT mice (Fig. 5E), suggesting that chronic-binge–induced neutrophil activation is attenuated in SELE−/− mice.
E-Selectin Deficiency Protects Mice From Chronic-Binge Feeding-Induced Liver Injury and Inflammation
We then investigated whether the reduced hepatic neutrophil infiltration observed in SELE−/− mice was associated with decreased liver injury. Indeed, chronic-binge feeding-associated elevation of serum ALT and AST levels was prevented in chronic-binge–fed SELE−/− mice, compared to WT mice (Fig. 6A). However, the level of hepatomegaly was comparable between chronic-binge–fed WT and SELE−/− mice (Fig. 6B). Further analyses revealed that micro- and macrovesicular steatosis, as well as hepatic TG content, were similar in both groups of mice (Fig. 6C and Supporting Fig. 5). Next, we evaluated the effect of E-selectin deficiency on chronic-binge feeding-induced hepatic expression of inflammatory mediators. Although levels of TNF-α, MIP-1α, and MIP-1β mRNAs (Fig. 6D,E) were similarly up-regulated in livers of ethanol-fed WT and SELE−/− mice, induction of IL-1β, IL-6, MCP-1, and MIP-2 (Fig. 6D,E) was significantly blunted in livers of SELE−/− mice. Furthermore, chronic-binge feeding-induced hepatic expression of the adhesion molecule, ICAM-1, was prevented in SELE−/− mice (Fig. 6F). Expression of VCAM-1 was similarly decreased in both groups after chronic-binge feeding (Fig. 6F). Collectively, these results indicate that SELE−/− mice were resistant to chronic-binge feeding-induced hepatocellular damage and inflammation, but not steatosis, compared to WT mice.
Hepatic Expression of E-selectin Is Up-Regulated in Patients With Alcoholic Fatty Liver, but Not With Alcoholic Cirrhosis
Expression of E-selectin and other adhesion molecules in normal human liver, alcoholic fatty liver (AFL), and alcoholic cirrhosis samples was examined. Expression of E-selectin was markedly up-regulated in AFL samples, but not in cirrhosis samples, compared to normal livers (Fig. 7). In contrast, hepatic expression of SELP and VCAM-1 was increased in alcoholic cirrhosis, but not in AFL. Expression of ICAM-1 was comparable in these three groups. In addition, as expected, transforming growth factor beta (TGF-β) expression was markedly elevated in cirrhosis samples, but not in AFL samples, whereas expression of IL-1β was up-regulated in both groups (Fig. 7B). Finally, E-selectin expression levels were positively correlated with levels of the neutrophil marker, MPO, in these liver samples. These data suggest that E-selectin expression is up-regulated in the early stages, but not the late stages, of human ALD.
In the current study, we have demonstrated that chronic-binge feeding synergistically up-regulates hepatic expression of E-selectin, which induces hepatic neutrophil accumulation and subsequently promotes liver injury and inflammation. We have integrated all of these findings into a model (Fig. 8) depicting the important role of E-selectin in neutrophil recruitment and liver injury upon chronic-binge feeding.
It has been known for many years that alcoholic hepatitis is associated with the accumulation of neutrophils in the liver[11-13]; however, in mice, chronic feeding with a Lieber-DeCarli diet or acute administration of ethanol only induces mild hepatic neutrophil infiltration. In this study, we demonstrated that chronic-binge feeding synergistically induced neutrophil accumulation in the liver, compared with chronic or binge ethanol feeding alone. Liver histology and MPO staining analyses revealed that most neutrophils were distributed throughout zones 1-3, and only a small number of inflammatory foci was observed in livers after chronic-binge feeding (Supporting Fig. 2A-D). Interestingly, human alcoholic steatohepatitis is often associated with many inflammatory foci in the liver, which are formed as a result of hepatocyte necrosis and likely contribute to removal of necrotic cells. The reason for the small number of inflammatory foci in the chronic-binge model was partly because this model represents the early stages of liver injury and is associated with mild hepatocyte injury. Moreover, we investigated the mechanisms underlying induction of hepatic neutrophil infiltration after chronic-binge feeding by examining expression of several adhesion molecules that are related to neutrophil recruitment. Among these molecules, we found that E-selectin had the highest level of induction in the liver after chronic-binge feeding.
Although E-selectin has been shown to play an important role in leukocyte recruitment,[17, 20] its roles in hepatic neutrophil infiltration and injury are unclear. For example, several studies have reported that blockade of E-selectin with an E-selectin-neutralizing Ab inhibited leukocyte recruitment and liver injury induced by D-galactosamine (D-Gal)/lipopolysaccharide (LPS) or concanavalin A (Con A), but other studies have suggested that E-selectin does not contribute to liver injury in these models[29, 30] or in an LPS-induced liver injury model. Our findings here suggest that E-selectin is a critical factor for chronic-binge feeding-induced hepatic neutrophil recruitment and injury. First, expression of E-selectin was positively correlated with infiltration of neutrophils in the liver; second, SELE−/− mice had reduced neutrophil infiltration, as demonstrated by MPO staining, FACS analyses, and real-time PCR analyses; and, finally, SELE−/− mice had reduced liver injury. The next question is why E-selectin may not contribute to hepatic neutrophil recruitment and injury in the LPS- or Con A–induced hepatitis models,[29, 30] but does play a significant role in response to chronic-binge feeding. The most likely answer is that the Con A and D-Gal/LPS models are associated with strong liver inflammation, during which many adhesion molecules other than E-selectin are highly up-regulated in the liver and subsequently contribute to neutrophil recruitment. In contrast, in the chronic-binge feeding model, liver inflammation is mild, and E-selectin is dramatically up-regulated (10-fold), whereas expression levels of many other adhesion molecules are up-regulated by less than 2-fold or even down-regulated. Thus, E-selectin plays a critical role in neutrophil recruitment in the context of chronic-binge feeding.
Currently, the mechanism by which E-selectin is highly up-regulated after chronic-binge feeding is not fully understood. E-selectin is not constitutively expressed by ECs, but its expression is highly up-regulated by inflammatory mediators, such as TNF-α, IL-1, and LPS. Hepatic expression of TNF-α and IL-1 was strongly induced after chronic-binge feeding (Fig. 1), and their expression correlated positively with the hepatic expression of E-selectin (Supporting Fig. 6). This suggests that chronic-binge feeding synergistically up-regulates the hepatic expression of TNF-α and IL-1, which may contribute to the induction of hepatic E-selectin.
Neutrophils have been shown to induce hepatocellular damage by generating ROS and cytotoxic mediators,[11, 33] but these cells also contribute to liver repair by removing necrotic debris[33-35] and promoting liver regeneration. Our results from the current study suggest that in the chronic-binge feeding model, neutrophils exacerbate liver injury. First, chronic-binge feeding not only induced neutrophil accumulation, but also stimulated neutrophil activation (Fig. 2); second, SELE−/− mice had reduced neutrophil infiltration and reduced liver injury after chronic-binge feeding, compared to WT mice (Figs. 5,6); third, mice treated with anti-Ly6G Ab were specifically depleted for neutrophils, but not macrophages, and were subsequently protected against liver injury induced by chronic-binge feeding (Fig. 3). Finally, chronic ethanol feeding has been shown to induce hepatocyte apoptosis. In the chronic-binge feeding model, we observed that the number of terminal deoxynucleotidyl transferase dUTP nick end labeling–positive hepatocyte nuclei was only slightly, but significantly, elevated, compared to pair-fed mice, and was comparable between ethanol-fed WT and SELE−/− mice (Supporting Fig. 7), although ethanol-fed SELE−/− mice had significantly lower serum ALT levels than ethanol-fed WT mice (Fig. 6). This suggests that hepatocyte necrosis, but not apoptosis, was most likely the major cause of hepatocyte cell death in this acute-on-chronic alcoholic liver injury model, which is consistent with the fact of neutrophil-mediated hepatocellular oncotic necrosis.
Although our findings suggest an important role of E-selectin in neutrophil infiltration and liver injury in the chronic-binge feeding model, the roles of E-selectin in human ALD remain unclear. By using IHC analyses of 28 ALD patients with severe alcoholic hepatitis or cirrhosis, Adams et al. demonstrated that E-selectin staining was generally weak in these patients, except in 2 patients with strong staining.[37, 38] However, in these studies, expression levels of E-selectin were not quantified. In the current study, by using real-time quantitative PCR, we demonstrated that expression of E-selectin was significantly up-regulated in human AFLs, but not in alcoholic cirrhotic livers, compared to normal livers, although expression of IL-1β was elevated in both groups. The reason why E-selectin is not up-regulated in cirrhotic livers, despite increased inflammation, may be the result of elevated TGF-β (Fig. 7), a cytokine that is known to effectively suppress expression of E-selectin in ECs. Taken together, E-selectin may play a role in promoting neutrophil infiltration and liver injury in the early stages of ALD without fibrosis.
The authors thank Dr. Philippe Gual (INSERM, U1065, Equipe 8, Complications Hépatiques de l'Obésité, Nice, France) for his great suggestions and helpful discussion during this study.