Potential conflict of interest: Nothing to report.
Gut-derived, endotoxin-mediated hepatocellular damage has been postulated to play a crucial role in the pathogenesis of alcohol-induced liver injury in rodents. Endotoxins induce production of tumor necrosis factor α (TNF-α) by Kupffer cells via Toll-like receptor (TLR) 4 and contribute to liver injury. This study addressed the contribution of other TLRs and ligands to alcoholic fatty liver. C57Bl6/J mice were fed a modified Lieber-DeCarli diet. Serum aminotransferase measurements, histological analysis, and quantification of liver TNF-α and TLR1-9 messenger RNA (mRNA) were performed. The effect of TLR ligands on liver injury was assessed in vivo. Neomycin and metronidazole or diphenyleneiodonium sulfate (DPI) were administered to evaluate the role of gut bacteria and NADPH oxidase activity, respectively, in hepatic TLR expression. Enteral ethanol (EtOH) exposure induced steatosis and increased liver weight, aminotransferase levels, and expression of TLR1, 2, 4, 6, 7, 8, and 9 liver mRNA. Injection of lipoteichoic acid, peptidoglycan (PGN), lipopolysaccharide (LPS), loxoribine, and oligonucleotide containing CpG (ISS-ODN) increased TNF-α mRNA expression more in the livers of EtOH-fed mice than in control mice. PGN, LPS, flagellin, and ISS-ODN induced liver inflammatory infiltrate in EtOH-fed mice but not control mice. Addition of antibiotics reduced the severity of alcoholic fatty liver without affecting TLR expression, whereas daily DPI injections reduced the EtOH-mediated upregulation of TLR2, 4, 6, and 9 mRNA. In conclusion, EtOH-fed mice exhibited an oxidative stress dependent on upregulation of multiple TLRs in the liver and are sensitive to liver inflammation induced by multiple bacterial products recognized by TLRs. (HEPATOLOGY 2006;43:989–1000.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Nearly 70% of alcoholic subjects develop steatosis, 20%-30% develop steatohepatitis, and 10% develop cirrhosis, the latter of which is a major worldwide cause of morbidity and mortality.1 Other than abstinence, the only acceptable treatment for end-stage liver disease is liver transplantation. Understanding the pathogenesis of alcohol-induced liver injury may lead to the development of new therapeutic options.
Several authors have suggested that gut bacteria play a crucial role in the development of alcoholic liver disease (ALD). Oral administration of nonabsorbable antibiotics or probiotics to rats has been reported to reduce ethanol-induced liver injury.2, 3 Alcohol ingestion disrupts gastrointestinal barrier function and subsequently induces the diffusion of luminal bacterial products into the portal blood.4, 5 In addition, a correlation between intestinal permeability, endotoxemia, and liver injury was demonstrated in animal models and patients suffering from ALD.6, 7
Immune host–bacteria interactions have also begun to be better characterized. Recently, it was shown that innate immune cells recognize conserved pathogen-associated molecular patterns through pattern recognition receptors, among which the family of Toll-like receptors (TLRs) occupies an important place.8 Ten mammalian TLRs (TLR1-10) have been identified. Individual TLRs have been shown to recognize specific patterns of microbial components. The specific ligands of these TLRs have begun to be better characterized: thus, TLR4 recognizes lipopolysaccharides (LPSs) from gram-negative bacteria.9 TLR2 is essential for the recognition of microbial lipopeptides, and TLR1 and 6 combined with TLR2 distinguish the subtle differences between triacyl- and diacyl-lipopeptides.10 TLR3 is involved in the recognition of viral double-stranded RNA, and TLR5 recognizes bacterial flagellin.11, 12 TLR7 and TLR8 bind viral single-stranded RNA.13 Lastly, TLR9 recognizes bacterial and viral DNA-containing unmethylated CpG motifs.14 Most activated TLRs recruit an adapter molecule, myeloid differentiation factor 88, via their cytoplasmic Toll/interleukin 1 receptor domain; this leads to the recruitment and activation of the complex involving interleukin 1 receptor–associated kinase 1 and 4 and tumor necrosis factor receptor–associated factor 6, which activates both nuclear factor κB and mitogen-activated protein kinase pathways and finally leads to the production of proinflammatory mediators such as tumor necrosis factor α (TNF-α) and interleukin 1β.15
Uesugi et al.16 showed that LPS-hyporesponsive mice of the C3H/HeJ strain that carry a mutation in the Toll/interleukin 1 receptor domain of the TLR4 gene develop less severe early alcohol-induced liver injury than wild-type mice. In their animal model, the expression of functional TLR4 is essential for alcohol-induced expression of TNF-α in the liver, which is responsible for hepatocyte injury via TNF-α receptor I (p55).17 Kupffer cells, the resident liver macrophages, play a crucial role in the hepatic immune response to gut-derived endotoxins and alcohol-induced liver injury.18, 19 The activity of macrophagic NADPH oxidase and the subsequent generation of oxidants are essential for the development of this liver injury and for the increased liver expression of the LPS coreceptor, CD-14.20, 21 In addition, chronic ethanol (EtOH) exposure, in vitro or in vivo, enhanced LPS-induced TNF-α expression in macrophages by increasing the transcription and stabilization of the RNA transcript.22, 23 However, little is known about the potential role of TLRs (other than TLR4) and their respective ligands in alcoholic fatty liver.
To investigate the role of TLRs in the pathogenesis of ALD, we studied the effect of 10 days of EtOH feeding on their expression in mouse liver, and also their functional role by challenging mice fed EtOH and control diets with their specific ligands. To determine the mechanism of EtOH-induced liver TLR messenger RNA (mRNA) expression, we treated the mice with oral antibiotics or subcutaneous injections of diphenyleneiodonium sulfate (DPI), an NADPH oxidase inhibitor.
Eight-week-old female C57Bl6/J mice (Charles River, Brussels, Belgium) were fed a liquid diet adapted from the classical Lieber-DeCarli EtOH diet (Table 1).24, 25 Animal care was provided in accordance with National Institutes of Health guidelines. Initially, all mice were given the control (CTR) liquid diet ad libitum for 1 week. The EtOH-fed mice were allowed access to the EtOH diet ad libitum for 10 days, and the CTR mice were pair-fed with the CTR diet.
Table 1. Composition of the Basic Experimental Liquid Diet
NOTE. This is a modified version of the Lieber-DeCarli diet.24 For the EtOH diet, 62.7 mL of 100% v/v ethanol (28% EtOH-derived calories) was added. For the control diet, an additional 92 g of dextri-maltose was added to maintain an isocaloric balance. The composition of the control high-fat liquid diet was 56% carbohydrate, 28% fat, and 16% protein.
Manufacturer: MP Biomedicals, Echewege, Germany.
AIN-76 mineral mix
AIN-76A vitamin mix
Alphacel non-nutritive bulk
EtOH- and CTR-fed mice were injected intraperitoneally to assess the liver immune response to TLR ligands26 after 10 days of experimental diet with 1 of the following: 60 or 120 μg lipoteichoic acid (LTA) (TLR2/TLR1 complex ligand)27 from Bacillus subtilis; 60 or 120 μg peptidoglycan (PGN) (TLR2/TLR6 complex ligand)28 from Bacillus subtilis; 60 or 120 μg polyinosine-polycytidylic acid (polyIC) (TLR3 ligand); 30 or 60 μg LPS (TLR4 ligand) from Escherichia coli serotype 055:B5 (Sigma-Aldrich, Bornem, Belgium); 60 μg flagellin (TLR5 ligand) from Bacillus subtilis; 60 or 120 μg 7-Allyl-8-oxoguanosine (loxoribine) (TLR7 ligand)29 (InvivoGen, San Diego, CA); or 30 or 60 μg immunostimulatory synthetic oligodeoxynucleotide analogs containing unmethylated CpG (ISS-ODN: 5′-TGACTGTGAACGTTCGAGATGA-3′) (TLR9 ligand)30 (Eurogentec, Seraing, Belgium). Mice were killed 8 hours after injection. Control mice for the experiments with LTA, PGN, polyIC, LPS, flagellin, and loxoribine received at the same time the same volume of vehicle (saline solution) intraperitoneally. In the CpG experiment, CTR mice were injected with 60 μg CTR oligonucleotides (5′-TGACTGTGAAGGTTAGAGATGA-3′) (Eurogentec).
Neomycin (350 mg/L) and metronidazole (600 mg/L) were added to the liquid diet from 3 days before starting the EtOH diet until the end of the protocol.31
We administered 1 mg/kg−1/d of the NADPH inhibitor, DPI (MP Biomedicals), or vehicle (5% glucose, 200 μL/d) subcutaneously for 10 days. This dose was previously reported to be well tolerated and effective in reducing alcohol-induced liver injury in the Tsukamoto-French enteral alcohol feeding model.32
Serum Ethanol Determination.
Serum EtOH levels were determined using REA Ethanol (Abbott Laboratories, North Chicago, IL).
Assessment of Hepatotoxicity.
Blood was sampled and mice were killed via cervical dislocation. The liver was excised and weighed, and part of it was fixed in formaldehyde. The remainder was frozen at −80°C. Paraffin sections were stained with hematoxylin-eosin and examined by the same investigator (N. N.) in a blinded manner. Liver pathology was scored as described by Nanji et al.33 with some modifications. Steatosis (the percentage of liver cells containing fat) was scored as follows: <25% = 1, <50% = 2, <75% = 3, >75% = 4. Steatosis and inflammatory foci (satellitosis plus nonspecific inflammatory foci) were counted on 20 fields (magnification ×100). Steatosis and inflammation scores were calculated as the mean per field.
Serum levels of alanine aminotransferase (ALT) were measured with a commercially available kit (Boehringer Mannheim, Mannheim, Germany) based on methods recommended by the International Federation of Clinical Chemistry.
TLR Messenger RNA Quantification via Real-Time Reverse-Transcriptase Polymerase Chain Reaction.
Frozen liver samples were homogenized in lysis solution with a MagNalyser (Roche Diagnostics, Brussels, Belgium). Total RNA was then extracted with the High Pure RNA Tissue Kit (Roche Diagnostics) according to the manufacturer's protocol, which included DNase treatment. The mRNA quantification was performed using a 2-step real-time reverse-transcriptase polymerase chain reaction (LightCycler, Roche Diagnostics). The primers and probes were designed using Primer3 software (Whitehead Institute for Biomedical Research, Cambridge, MA) (Table 2). We used 18s rRNA as the housekeeping gene (Applied Biosystems, Foster City, CA). Copy numbers were calculated as previously described.34, 35
Table 2. Primer and Probe Sequences
NOTE. TLR real-time reverse-transcriptase polymerase chain reaction primers and 5′-(6-Fam) TaqMan probe (Tamra phosphate)-3′ sequences are shown.
Blood samples were collected from the portal vein using heparinized syringes and were then centrifuged at 1,200 rpm for 10 minutes. Plasma samples were diluted and heated for 30 minutes at 70°C. Plasma endotoxin was measured with a Limulus Amebocyte Lysate test kit (Kinetic-QCL; BioWhittaker, Walkersville, MD).
At the end of protocol, mice feces were weighed and homogenized in 1 mL phosphate-buffered saline. Serial dilutions of fecal suspensions were placed onto CHROMagar (CHROMagar, Paris, France). The culture plates were incubated for 24 hours at 37°C in air containing 5% CO2, and colonies of bacteria were counted.
Measurement of Liver Reduced Glutathione and Lipid Peroxidation.
Glutathione levels and lipid peroxidation product (malonedialdehyde and 4-hydroxyalkenals) were determined using the colorimetric Bioxytech GSH-400 assay and LPO-586 assay (OxisResearch, Portland, OR) respectively according to the manufacturer's instructions. Protein concentration was measured using a dye-binding assay (BioRad, Richmond, CA). The results were expressed in nanomoles per milligram of protein.
Data are expressed as medians. Significance was assessed using the Mann-Whitney U test. For multiple comparisons, the groups were first compared using the nonparametric Kruskal-Wallis test. The Mann-Whitney U test was then used for post hoc analysis. A P value of less than .05 was considered significant. Analyses were performed using SPSS 11.0 software (SPSS, Chicago, IL).
EtOH Diet Induces Severe Steatosis in 10 Days.
As previously published for this protocol, both EtOH- and CTR-fed mice initially lost weight, but after 7 days of the experimental diet, their body weights became stable.36 This evolution was similar for the 2 groups of mice. Before the start of the EtOH diet and at the end of the protocol, the median body weights of the EtOH- and CTR-fed groups were 19.7 (range 18.4–20.2) and 19.5 (range 18.4–22.1) g and 17.3 (range 15.7–19.9) and 16.8 (range 16.3–19.4) g, respectively. After 10 days, the median serum EtOH levels were 139.1 (range 18.9–330) mg/dL. The EtOH diet induced an increase in liver weight (normalized by total body weight) compared with the CTR diet (Table 3). Hepatomegaly was associated with the development of macrovacuolar and microvacuolar steatosis (nearly 75% of hepatocytes containing fat; median steatosis score of 4 [range 3.55–4]) and no significant inflammatory infiltration or necrosis disclosed by histology (Fig. 1). Absence of neutrophilic accumulation was confirmed by the absence of detection of myeloperoxidase mRNA expression in the liver (data not shown). Significant increases in serum ALT levels (Table 3) and liver TNF-α mRNA expression were found in EtOH-fed mice compared with CTR-fed mice (21,627 [8,576–151,605] vs. 3,659 [203–6,597] copies of TNF-α normalized by rRNA 18s; P < .001).
Table 3. Features of Alcoholic Fatty Liver
CTR-Fed Mice (n = 13)
EtOH-Fed Mice (n = 13)
NOTE. Data are expressed as the median (range) from 3 independent experiments with homogeneous results. Steatosis (percentage of liver cells containing fat) was scored as follows: <25% = 1, <50% = 2, <75% = 3, and >75% = 4. Steatosis and inflammatory foci (satellitosis plus nonspecific inflammatory foci) were counted on 20 fields (magnification × 100). Steatosis and inflammation scores were calculated as means per field. Significance was determined via Mann-Whitney U test (P <.01).
EtOH Diet Upregulates Liver Expression of TLR1, 2, 4, 6, 7, 8, and 9.
Hepatic expression of TLRs mRNA was measured after 10 days of EtOH exposure. Significant increases in TLR1, TLR2, TLR4, TLR6, TLR7, TLR8, and TLR9 mRNA levels were observed in the livers of EtOH-fed mice compared with CTR-fed mice. (Table 4) On the other hand, TLR3 and TLR5 mRNA expression was not statistically different in the 2 groups.
NOTE. Data are expressed as the median (range) from 3 independent experiments with homogeneous results. Values are expressed as the number of TLR mRNA copies normalized by the number of rRNA 18s copies. Significance was determined via Mann-Whitney U test (P < .01).
Effect of Challenge With TLR Ligands on Liver TNF-α mRNA Expression.
To assess the functional relevance of the upregulation of several TLR mRNAs, TLR ligands were administered intraperitoneally to mice after 10 days of liquid diet. Liver TNF-α mRNA levels were measured 8 hours after the challenge. All ligands (LTA, PGN, LPS, loxoribine, and ISS-ODN) whose respective TLR mRNAs were up-regulated by EtOH administration induced a larger increase in TNF-α mRNA expression in the livers of EtOH-fed mice than CTR-fed mice (Fig. 2). On the other hand, administration of polyIC (TLR3 ligand) and flagellin (TLR5 ligand), whose TLR expression was not modified by EtOH dietary intake, induced similar increases in TNF-α mRNA expression in both EtOH- and CTR-fed mice.
Effect of Challenge With TLR Ligands on Liver Injury.
After challenging mice with each TLR ligand, serum ALT levels were always higher in EtOH-fed mice than in CTR-fed mice. PolyIC, LPS, and flagellin challenges induced a dramatic increase in serum ALT levels in EtOH-fed mice compared with EtOH-fed mice given vehicle, suggesting that these ligands exert a hepatotoxic effect after sensitization with EtOH (Fig. 3). In contrast, the challenges with 30 μg LPS and 60 μg flagellin did not modify serum ALT levels in CTR-fed mice. However, a small increase in ALT levels was observed in CTR-fed mice given 60 μg LPS and polyIC compared with CTR-fed mice given vehicle. The LTA, PGN, loxoribine, and ISS-ODN challenges were not associated with a significant change in serum ALT levels at either of the doses tested in CTR- and EtOH-fed mice.
Fat accumulation in the liver was not affected by TLR ligand administration, as shown by semiquantitative assessment of steatosis (Table 5). On the other hand, administration of PGN, polyIC, LPS, flagellin, or ISS-ODN (60 μg) raised the number of nonspecific inflammatory and satellitosis foci in the livers of EtOH-fed mice compared with EtOH-fed mice given vehicle or CTR oligonucleotides (P < .05) (Fig. 4). In CTR-fed mice, only the injection of polyIC increased the number of inflammatory foci compared with CTR-fed mice given vehicle. Inflammatory infiltration induced by the administrations of PGN, LPS, flagellin, or ISS-ODN (60 μg) was significantly higher in EtOH-fed mice than in CTR-fed mice (P < .05). On the other hand, there was no significant difference between the number of inflammatory foci induced by polyIC injection in CTR- and EtOH-fed mice.
Table 5. Effects of TLR Ligands on Alcoholic Fatty Liver
NOTE. Semiquantitative assessment of fat accumulation and inflammatory infiltration in the liver of CTR- and EtOH-fed mice (4–6 mice per group) 8 hours after intraperitoneal injections of TLR ligands. Data are expressed as the median (range). Statistical comparisons were made using the Kruskal-Wallis test. Post hoc analysis was conducted using the Mann-Whitney U test.
Upregulation of Liver TLR mRNA in EtOH-Fed Mice Is Not Directly Dependent on Gut Flora.
In our model, the 10-day EtOH regimen did not increase plasma endotoxin levels in the portal vein that were below detection levels (5 EU/mL) These data cannot exclude a small and/or transient increase of gut permeability in our model. To assess the role of gut bacteria—which could be the source of undetectable amounts of endotoxin or other bacterial products—in EtOH-induced overexpression of liver TLRs, mice were treated with 2 broad-spectrum antibiotics: neomycin and metronidazole. The addition of these antibiotics to the EtOH liquid diet reduced slightly the amount of EtOH consumed per day (16.34 [7.2–22.25] for EtOH-fed mice vs. 10.48 [5.11–19.87] g/kg−1/day−1 for antibiotics [AB] + EtOH-fed mice). To maintain a similar quantity of EtOH intake in the different groups, the amounts of diet of EtOH-, CTR-, and AB + CTR–fed mice were matched with those of AB + EtOH–fed mice. Fecal culture of CTR- and EtOH-fed mice showed bacterial growth of enterobacter–citrobacter (5.86 [5.45–6.52] log CFU/mg feces) and Klebsiella (4.25 [3.36–6.28] log CFU/mg feces) after 48 hours on CHROMagar orientation. Cultures from EtOH-fed mice treated with neomycin and metronidazole, however, showed no growth of Klebsiella and a clear decrease of enterobacter–citrobacter (2.62 [2.32–2.8] log CFU/mg feces), underlining a clear reduction of gram-negative bacteria. Addition of antibiotics reduced the severity of EtOH-induced fatty liver—as shown by the significant decreases in liver weight, ALT levels, and steatosis scores—which is in agreement with previously published data2 (Table 6). Although antibiotic administration reduced the severity of EtOH-induced liver injury, a significant decrease in EtOH-induced expression of liver TLR mRNA was not observed (Table 7). Similarly, TLR expression in CTR-fed mice was not modified by AB administration (data not shown). The expression of TLR1, 2, 4, 6, 7, 8, and 9 mRNA in the livers of AB + EtOH–fed mice remained significantly higher than in the livers of CTR-fed mice and AB + CTR–fed mice. Therefore, these data suggest that liver TLR mRNA expression does not seem to be directly dependent on gut bacteria, despite a clear effect of antibiotics on liver injury.
Table 6. Effects of Antibiotics on Alcoholic Fatty Liver
CTR (n = 7)
CTR + AB (n = 7)
EtOH (n = 6)
EtOH + AB (n = 6)
NOTE. Data are expressed as the median (range). Significance was determined via Kruskal-Wallis test. Post hoc analysis was conducted using the Mann-Whitney U test. Abbreviations: CTR, control; AB, antibiotics (neomycin + metronidazole); EtOH, ethanol; ALT, alanine aminotransferase.
Increased TLR2, 4, 6, and 9 mRNA Expression in EtOH-Fed Mice Is Dependent on NADPH Oxidase–Induced Oxidative Stress.
As previously demonstrated in other models of ALD, hepatic oxidative stress was observed in our model, as illustrated by an increase of lipid peroxidation product levels (e.g., malonedialdehyde, 4-hydroxyalkenals) (1.25 [0.72–2.79] vs. 1 [0.63–1.56] nmol/mg protein; P < .05) and a decrease of antioxidant levels (e.g., glutathione) (3.98 [2.91–5.21] vs. 5.36 [3.9–6.46] nmol/mg protein; P < .01) in the livers of EtOH-fed mice. On the other hand, this oxidative stress was not associated with an induction of CYP2E1 mRNA in the liver (3.76 [1.22–8.19] 106 vs. 5.49 [2.46–11] 106 number of copies mRNA CYP2E1 normalized by rRNA 18s for EtOH- and CTR-fed mice, respectively; P value not significant). Because reactive oxygen species originating from NADPH oxidase activity are thought to play a key role in alcohol-induced liver injury, we investigated their possible role in the upregulation of alcohol-induced TLRs.20 For this purpose, DPI, a known NADPH oxidase inhibitor, was administered subcutaneously to mice on a daily basis. As previously reported,32 DPI administration clearly reduced the severity of alcohol-induced fatty liver, as shown by the significant decreases in liver weight, serum ALT levels, and steatosis score induced by the EtOH diet (Table 8). DPI treatment also significantly reduced the TLR2, 4, 6, and 9 mRNA upregulation caused by EtOH exposure (Fig. 5). On the other hand, it did not significantly change alcohol-induced upregulation of TLR1, TLR7, and TLR8 mRNA (P = 0.573, .055, and .083, respectively). Liver TLR3 and TLR5 mRNA expression, the baseline levels of which were not increased by alcohol ingestion, were also not affected by DPI.
Table 8. Effects of NADPH Oxidase Inhibition by DPI Administration in Alcoholic Fatty Liver
CTR + Vehicle (n = 9)
CTR + DPI (n = 8)
EtOH + Vehicle (n = 10)
EtOH + DPI (n = 8)
NOTE. Data are expressed as the median (range). Significance was determined via Kruskal-Wallis test. Post hoc analysis was conducted using the Mann-Whitney U test.
The findings of the present study suggest that an alcohol diet induces steatosis and upregulation of several TLRs in the liver that are not directly dependent on gut bacterial flora, but are dependent on NADPH oxidase activity (at least for some TLRs). This upregulation is associated with a stronger response to TLR ligands in terms of TNF-α mRNA synthesis and induction of hepatitis.
In rodents, the TLR pathway takes a prime place in the pathogenesis of early alcohol-induced liver injury in rodents. Indeed, CD14-deficient and nonfunctional tlr4 mice are protected against the murine-adapted Tsukamoto-French model of ALD.16–38 We demonstrate that chronic EtOH feeding induces severe steatosis without liver inflammation and clear upregulation of TLR1, TLR2, TLR4, TLR6, TLR7, TLR8, and TLR9 in the liver. The regulation of TLR expression has not been well characterized and is largely different between mice and humans.39 An et al.40 demonstrated that LPS induces upregulation of TLR2, TLR4, and TLR9 gene expression on mouse dendritic cells. In this study, we tested the hypothesis that gut-derived LPS or other bacterial products increase hepatic TLR gene expression. Clearly, the administration of broad-spectrum antibiotics reduced the severity of alcohol-induced fatty liver but did not modify alcohol-induced TLR gene expression. These results strongly suggest that the observed upregulation of liver TLRs is not directly dependent on gut bacteria. The reactive oxygen species produced during alcohol-induced liver injury modulate the activity of signal transduction pathways involved in the innate immune response, especially the mitogen-activated protein kinase and activator protein 1 pathways,41, 42 and are responsible for the alcohol-induced increase in the liver expression of CD-14, another pattern recognition receptor.21 We therefore postulated that reactive oxygen species might have an impact on TLR mRNA expression. To test this possibility, we administered DPI (an inhibitor of NADPH oxidase, CYP2E1 not being induced) the major early source of oxidants due to EtOH, to alcohol-exposed mice. DPI dramatically reduced the severity of alcohol-induced fatty liver and blunted the increase in the expression of liver TLR2, 4, 6, and 9 mRNA expression. Surprisingly, DPI was unable to blunt the upregulation of TLR1, 7, and 8, suggesting that the regulation of some but not all TLRs is dependent on reactive oxygen species production. There are significant differences between the promotor sequences and gene regulatory elements of these TLR genes, and little is known about their transcriptional regulation. Further studies are necessary for clearer characterization of the mechanisms regulating these receptors during chronic alcohol exposure. Moreover, because antibiotics do not modulate liver TLR expression at the opposite of DPI while both are reducing liver injury, stimuli other than bacterial products are probably necessary for the NADPH oxidase activation during chronic alcohol exposure in this model.
The recent analysis of transgenic mouse lines expressing the complete murine TLR4 gene supports the concept that TLR transcriptional regulation is a key process in the immunological response to TLR ligands.43 Consequently, to assess the functional relevance of TLR mRNA upregulation in our model, TLR ligands were administered intraperitoneally to mice after 10 days on the experimental diets. With regard to the TLRs that were overexpressed after EtOH exposure, administration of their respective specific ligands resulted in the overexpression of TNF-α mRNA in the livers of EtOH-fed mice compared with the livers of CTR-fed mice given the same doses of TLR ligands. The ligands concerned were LTA (TLR2/TLR1), PGN (TLR2/TLR6), LPS (TLR4), loxoribine (TLR7), and ISS-ODN containing CpG motif (TLR9). The administration of polyIC (TLR3) and flagellin (TLR5) induced similar TNF-α mRNA levels in the livers of EtOH-fed mice and CTR-fed mice, because these TLRs were not upregulated by EtOH exposure. Therefore, the results of these experiments suggest that in this model, the upregulation of specific TLRs increases TNF-α expression in response to each TLR's specific ligand. This is to be considered with the inhibition induced by acute EtOH exposure, on the TLRs pathways,44 underlining the fact that the duration of exposure is critical for the modulation of innate immunity by EtOH.
In mice sensitized by EtOH feeding, different patterns of liver injury were demonstrated through challenge with different TLR ligands (Table 9). As previously reported in the rat Lieber-DeCarli model,45 we showed that the livers of EtOH-fed mice were more sensitive to LPS challenge than the livers of CTR-fed mice. Thus, the LPS-induced liver inflammatory infiltrate was significantly larger, and the increase in ALT serum levels was significantly higher in EtOH-fed mice than in CTR-fed mice and correlated with the higher TNF-α expression. The latter finding is in agreement with the findings of others who have demonstrated that the Kupffer cells of EtOH-fed rodents generate more TNF-α in response to LPS compared with CTR animals.46 Similarly, the livers of EtOH-fed mice were more sensitive to flagellin challenge than the livers of CTR-fed mice in the absence of gene upregulation of the flagellin receptor (TLR5). Indeed, flagellin administration induced a large increase in ALT levels and in the inflammatory infiltrate in the livers of EtOH-fed mice, but had no effect in CTR-fed mice given the same dose of flagellin. Paradoxically, we did not observe any significant difference between the flagellin-induced production of TNF-α in the 2 groups. A possible explanation of this discrepancy might be that the effector mechanisms through which LPS and flagellin trigger liver injury are very different. LPS has a greater effect on the release of TNF-α, but flagellin is a more potent inducer of oxidative stress damage.47 Flagellin might therefore act synergistically with EtOH on oxidative stress and liver injury. We observed with interest that the PGN and CpG challenges induced the recruitment of leukocytes in the livers of EtOH-fed mice but not CTR-fed mice, but did not alter systemic ALT levels. In EtOH-fed mice, the accumulation of inflammatory foci was parallel to the larger production of TNF-α. The absence of correlation between liver inflammatory processes and circulating ALT levels is also observed in patients suffering from alcoholic hepatitis.48 On the other hand, whereas the LTA and loxoribine challenges induced higher TNF-α production in EtOH-fed mice than in CTR-fed mice, they did not modify the histological features observed in EtOH- and CTR-fed mice given vehicle. This underlines the fact that TLR-mediated liver leukocyte recruitment is not only TNF-α–dependent. We noted with interest that only polyIC induced a rise in ALT levels and liver inflammatory infiltration in both groups. As recently reported, polyIC administration itself causes natural killer–dependent liver injury and leukocyte recruitment (in contrast to other TLR ligands), and only the TLR3 pathway seems to be independent of myeloid differentiation factor 88.49 The different effects of various ligands on alcoholic fatty liver can be explained by the fact that ligand recognition by distinct TLRs elicits alternative immune effector mechanisms.
Table 9. Patterns of Response to TLR Ligands: Summary of the Effects of TLR Ligands on Liver TNF-α mRNA Expression, Inflammatory Infiltration, and Serum ALT Levels
Upregulation of Liver TLR in EtOH-Fed Versus CTR-Fed Mice
Increased TNF-α Response to Ligand in EtOH-Fed Versus CTR-Fed Mice
Increased Serum ALT
NOTE. “Increased Inflammation” and “Increased Serum ALT” columns indicate increase of inflammation score or ALT in mice given TLR ligand versus mice given vehicle.
The present study assessed the role of different TLRs through their liver gene expression and in vivo response to different specific ligands. Quantification of cell surface expression at the protein level and dissection of the specific signaling cascade of each TLR in different liver cell populations (i.e., Kupffer cells and hepatocytes) remains difficult but will certainly help, in the future, to further unravel the observations of our study. Moreover, in this study, the doses and the route of TLR ligand administration might not be clinically relevant in light of intestinal translocation. The detection and quantification of TLR ligands in the blood of ALD patients may strengthen our hypothesis.
In conclusion, in this model of alcoholic fatty liver, severe steatosis is the visible part of a “first hit” on the liver, combined with an upregulation of TLRs (TLR1, 2, 4, 6, 7, 8, and 9). This upregulation is associated with a state of TNF-α hyperresponse to their specific ligands. The administration of microbial products (PGN, polyIC, LPS, flagellin, and unmethylated CpG) to EtOH-exposed animals may induce the “second hit” on the liver with inflammatory infiltrate and satellitosis formation. The innate immune response to pathogenic molecular patterns through pattern recognition receptors seems to take a primary role in the pathogenesis of alcoholic fatty liver in mice, and may lead to the development of new potential targets for the treatment of ALD. Bacterial and viral compounds other than LPS and TLR4 might be involved in the worsening of alcohol-induced liver injury and may trigger alcoholic hepatitis in humans.
The authors thank Vincent Vercruysse for helpful technical support.