Potential conflict of interest: Nothing to report.
The innate immune system has been implicated in the pathogenesis of alcoholic liver disease. Although innate immunity is usually considered an early response to injury, previous work implicating innate immunity in ethanol-induced liver injury focuses primarily on long-term ethanol exposure. We investigated the early period of ethanol exposure to determine whether there were temporal associations between activation of innate immune responses and known correlates of liver injury. Female C57BL/6 mice were allowed free access to an ethanol-containing Lieber-DeCarli diet or were pair-fed a control diet. Within 4 days of ethanol exposure, we observed a striking spike in expression of hepatic proinflammatory cytokines—including tumor necrosis factor α (TNF-α), interleukin-6, and interferon-γ—prior to hepatic triglyceride accumulation or increased plasma alanine aminotransferase activities, as well as before the induction of cytochrome P450 2E1 or oxidative stress. This early spike in inflammatory cytokines coincided with deposition of C3b-iC3b/C3c (C3b) in the liver. This deposition, resulting from the cleavage of the third component of the complement system (C3), is evidence for activation of complement in response to ethanol. C3−/− mice were protected from the early, ethanol-induced increase in hepatic TNF-α expression. Ethanol increased C3b deposition in mice deficient in C3a receptor or C5a receptor, as well as in wild-type mice depleted of hepatic macrophages; however, there was no increase in hepatic TNF-α in the absence of C3a receptor, C5a receptor, or hepatic macrophages. In contrast, the absence of Toll-like receptor 4 (TLR-4) had no effect on the early, ethanol-induced increase in either C3b or TNF-α. Conclusion: We have identified a complement- and macrophage-dependent, but TLR-4 independent, phase in the pathogenesis of ethanol-induced liver injury. (HEPATOLOGY 2009.)
The development of alcoholic liver disease is a complex process involving both parenchymal and nonparenchymal cells in the liver, as well as the recruitment of other cell types to the liver in response to damage and inflammation.1 Accumulating evidence indicates that different components of the innate immune response are activated during chronic ethanol exposure. For example, Kupffer cells, the resident macrophages in the liver, play an essential role in chronic ethanol-induced liver injury.2 Endotoxin/ lipopolysaccharide, a component of gram-negative bacterial cell walls, is an important activator of Kupffer cells during long-term ethanol exposure, resulting in increased production of proinflammatory cytokines and chemokines,3, 4 as well as reactive oxygen species.5 Mice lacking components of the Toll-like receptor 4 (TLR-4) complex (TLR4−/− or CD14−/−), as well as tumor necrosis factor receptor I (TNFRI−/−), are protected from chronic ethanol-induced steatosis, as well as inflammation and early stages of fibrosis.3, 4
Complement is also implicated in chronic ethanol-induced liver injury.6-8 Chronic ethanol feeding in mice for 6 to 8 weeks increases the activation of the third component of the complement pathway (C3), increasing the concentration of C3a in the plasma,6 as well as deposition of C3 in the liver.9 Mice lacking C3 (C3−/−) do not develop steatosis in response to chronic ethanol feeding.6, 8 In contrast, while mice lacking the fifth component of the complement system (C5−/−) still exhibit increased hepatic triglycerides in response to long-term ethanol exposure, they are completely protected from ethanol-induced increases in alanine aminotransferase (ALT) and inflammatory cytokines.6 Mice lacking CD55/DAF, a complement regulatory protein, exhibit exacerbated steatosis, ALT, and inflammatory cytokines in response to chronic ethanol feeding, suggesting that CD55/DAF acts as a brake against chronic ethanol-induced liver injury.6
Studies of the role of the innate immune response in ethanol-induced liver injury have usually focused on relatively long periods (4 to 8 weeks) of ethanol exposure. However, activation of innate immune responses can be an early response to a variety of insults. In particular, activation of complement is a rapid response to infection, tissue injury, or stress; rapid activation of complement serves to orchestrate different aspects of innate immunity, as well as to provide links to the adaptive immune response.10 Therefore, we hypothesized that activation of innate immune responses, including increased expression of proinflammatory cytokines and activation of complement, would likely be an early response to ethanol exposure. Here we report that ethanol exposure results in an early, transient increase in both complement activation products and tumor necrosis factor α (TNF-α) expression, prior to the accumulation of triglycerides in the liver or increases in ALT in the plasma. Importantly, Kupffer cells, as well as both C3a receptor or C5a receptor, were required for the early, transient increase in TNF-α expression during ethanol feeding. However, in contrast to the contributions of TLR-4 to the long-term, chronic effects of ethanol in liver,3 the early ethanol-induced activation of complement and TNF-α expression was independent of TLR-4. These data illustrate the specific, dynamic interactions between multiple components of the innate immune response at early times in the hepatic response to ethanol.
ALT, alanine aminotransferase activity; C3, third component of the complement system; C3aR, C3a receptor; C3b, C3b-iC3b/C3c; C5, fifth component of the complement system; C5aR, C5a receptor; CYP2E1, cytochrome P450 2E1; 4-HNE, 4-hydroxynonenal; IFN-γ, interferon-γ; IL-6, interleukin-6; SEM, standard error of the mean; TLR-4, Toll-like receptor 4; TNF-α, tumor necrosis factor α.
Materials and Methods
Female mice (8 to 10 weeks old) of each genotype were used in the studies. C57BL/6J, TNFI−/− (B6.129-Tnfrsf1atm1Mak/J), and C3−/− (B6.129S4-C3129S4-C3tm1Crr/J) mice were purchased from the Jackson Laboratory (Bar Harbor, Maine). C3aR−/− and C5aR−/− mice were on a mixed wild-type background (WTaR) as described.11, 12TLR4−/− were on a C57BL/6 background.13 (See Supplementary Material for further details on Materials.)
All procedures using animals were approved by the Case Western Reserve University or Cleveland Clinic Institutional Animal Care and Use Committees. Eight- to 10-week-old female mice were housed in shoe box cages (two animals per cage) with microisolator lids. (Details of animal husbandry are provided in Supplemental Material). The ethanol-fed group was allowed free access to an ethanol-containing diet with increasing concentrations of ethanol: 1% (vol/vol) (1%, day 2) and 2% each for 2 days (2%, day 2), then 4% ethanol for 7 days (4%, day 3; 4%, day 7), and finally 5% ethanol for up to 4 weeks (5%, week 1, week 2, week 3, or week 4).6, 14 The 5% ethanol diet provides 26.9% of calories as ethanol. Control mice were pair-fed diets that iso-calorically substituted maltose dextrins for ethanol over the entire feeding period. Plasma ethanol concentrations were measured in samples collected 3 hours into the feeding/dark cycle. In one experiment, mice were intravenously injected with 0.2 mL of a 1-mg/mL suspension of clodronate encapsulated in liposomes as described15 prior to the start of ethanol feeding to selectively deplete Kupffer cells.16
Plasma Measurements and Liver Triglycerides.
Plasma samples were assayed for ALT and ethanol using commercially available enzymatic assay kits (Diagnostic Chemicals, Ltd, Oxford, CT) following the manufacturer's instructions. Total liver triglycerides were measured using the Triglyceride Reagent Kit from Pointe Scientific Inc. (Lincoln Park, MI).
Immunohistochemical analyses of C3b-iC3b/C3c (C3b), TNF-α, 4-hydroxynonenal (4-HNE), and F4/80 were performed in liver sections using standard techniques (see Supplemental Material).
Preparation of Liver Lysates.
Frozen liver tissue (0.5-1.0 g) was homogenized in 10 mL/g tissue in lysis buffer exactly as described.6 Protein concentrations were measured using the BCA assay. Lysates were then used to measure cytochrome P450 2E1 (CYP2E1) expression via western blotting and cytokine concentrations via enzyme-linked immunosorbent assay as described.6
Values are reported as the mean ± standard error of the mean (SEM). Because of the logistics of animal care, several feeding trials were performed and combined for the final data analysis. Data were analyzed by general linear models procedure (SAS, Carey, IN), blocking for trial effects when data from more than one feeding trial were used in a data set. Data were log-transformed if needed to obtain a normal distribution. Follow-up comparisons were made via least square means testing.
To begin to address the interactions between innate immune responses and ethanol-induced liver injury, we allowed female C57BL/6 mice ad libitum access to the Lieber-DeCarli ethanol diet with increasing concentrations of ethanol (Fig. 1). The ethanol-fed mice continued to grow at the same rate as their pair-fed controls (Fig. 1A). In the ethanol-fed mice, hepatic triglycerides (Fig. 1B) and plasma ALT (Fig. 1C), a marker of hepatocyte injury, were increased as early as 7 days (4%, day 3). Liver triglycerides continued to accumulate at a steady rate over the entire ethanol feeding period (Fig. 1B). Although plasma ALT was highest during the first week of exposure to 5% ethanol (Fig. 1C), plasma ALT remained elevated over the ethanol feeding period.
It is well known that expression of several inflammatory cytokines is increased in ethanol-induced liver injury; TNF-α, in particular, plays an essential role in the development of ethanol-induced liver injury.4, 17 Analysis of our ethanol-fed mice showed that TNF-α protein in the liver was transiently increased after only 4 days of ethanol feeding (2%, day 2) and dropped below baseline after 7 days (4%, day 3) (Fig. 1D). Interferon-γ (IFN-γ) and interleukin-6 (IL-6) expression also exhibited a similar rapid and transient induction early in the response to ethanol (Fig. 1E-F). After this initial response, hepatic TNF-α and IFN-γ increased again after 5%, week 4, of ethanol feeding (Fig. 1D-E), typical of previous reports of increases in expression of hepatic TNF-α after chronic ethanol feeding.6 Immunohistochemical staining revealed that TNF-α protein was expressed throughout the liver (Fig. 2A); the staining intensity paralled TNF-α concentrations determined by enzyme-linked immunosorbent assay (Fig. 1D). Colocalization studies with F4/80, a macrophage-specific marker, and TNF-α demonstrate that F4/80 positive cells had higher TNF-α immunostaining after ethanol feeding (2%, day 2) (Fig. 2B).
In order to evaluate the pathophysiological significance of this early peak in hepatic TNF-α, we assessed markers of ethanol-induced liver injury in TNFRI−/− mice after ethanol feeding for 11 days (4%, day 7), just after the transient elevation in hepatic TNF-α. In wild-type mice, hepatic triglycerides (Fig. 2C), plasma ALT (Fig. 2D), and accumulation of 4-HNE adducts (Fig. 2E), a measure of lipid peroxidation, were increased at day 11 (4%, day 7) of ethanol exposure. In contrast, whereas hepatic triglycerides still accumulated in TNFRI−/− mice (Fig. 2C), ALT and 4-HNE adducts were not increased by ethanol feeding in the absence of TNF receptor I (Fig. 2D-E).
We next addressed the molecular mechanism for the early ethanol-induced increase in hepatic TNF-α. When mice consumed 1% or 2% ethanol, plasma ethanol concentrations, measured 3 hours into the dark/feeding period, were undetectable (Fig. 3A). In contrast, increased ethanol concentrations in the plasma were observed when the ethanol concentrations in the diet were 4% or higher (Fig. 3A). CYP2E1 expression was induced after long-term ethanol feeding (5%, week 4), but not yet increased at day 4 (2%, day 2) (Fig. 3B). Similarly, whereas 4-HNE adducts, a measure of lipid peroxidation, were increased at 5%, week 4, no adducts were detected at day 4 (2%, day 2) (Fig. 3C). Neither plasma adiponectin concentrations nor the concentration of endotoxin in platelet-rich plasma were affected by ethanol feeding at day 4 (2%, day 2) (data not shown).
Recent studies by ourselves6 and others8 have causally implicated complement activation in chronic ethanol-induced liver injury. Because complement activation generates C3a and C5a, anaphylatoxins that increase inflammatory cytokine production, we next investigated the time course of complement activation during ethanol feeding. We have previously shown that 6 weeks of ethanol feeding increases serum C3a concentration.6 Here we have taken the alternative approach and used antibodies generated to detect neo-epitopes revealed on the C3b portion of C3 after cleavage.18 Immunohistochemistry showed a rapid and transient increase in C3b deposition at day 4 (2%, day 2) (Fig. 4A), paralleling the kinetics of TNF-α expression (Fig. 2). This was followed by a later accumulation of C3b at 5%, week 4, of ethanol feeding (Fig. 4A). At day 4, in contrast to the localization of TNF-α, C3b deposition was concentrated in sinusoidal areas of the liver.
If complement activation is required for ethanol-induced increases in TNF-α at these early time points, then TNF-α protein should not be detected in C3−/− mice during ethanol feeding. As expected in the absence of C3 protein, C3b deposition was not detectable in the livers of C3−/− mice at day 4 (2%, day 2) of ethanol feeding (Fig. 4B). Moreover, ethanol-induced increases in TNF-α protein were suppressed in C3−/− compared with wild-type mice at day 4 (2%, day 2) (Fig. 4B), arguing that complement activation is required for the early ethanol-induced increase in TNF-α expression.
Complement activation generates C3a and C5a, which upon interaction with their receptors, C3aR or C5aR, induces expression of inflammatory cytokines. Therefore, we next made use of mice lacking either C3aR (C3aR−/−) or C5aR (C5aR−/−) to determine if one or both of these receptors is required for the early increase in hepatic TNF-α expression that accompanies ethanol feeding. Whereas wild-type mice exhibited increased hepatic TNF-α expression after 4 days (2%, day 2) of ethanol feeding (Fig. 5), no increase occurred in either C3aR−/− or C5aR−/− mice (Fig. 5). C3b accumulation was detected in livers from C3aR−/− and C5aR−/− mice after ethanol feeding at day 4 (data not shown), indicating that the abrogation of ethanol-induced TNF-α expression was due to the absence of C3aR or C5aR in these knockout strains, rather than the absence of complement activation.
Previous work has shown that ethanol-induced increases in lipopolysaccharide and its interaction with TLR-4 on Kupffer cells are important contributors to chronic ethanol-induced hepatic TNF-α production and subsequent liver injury.2, 4, 17 Moreover, TLR-4 and C3a receptor/C5a receptor signaling synergistically increases the expression of cytokines.19, 20 Therefore, we next investigated whether interactions among complement, Kupffer cells, and TLR-4 were required for the early ethanol-induced increase in TNF-α expression. We treated mice with clodronate to deplete Kupffer cells prior to ethanol exposure.16 Localization of F4/80 demonstrated a profound decrease in Kupffer cells 4 days after clodronate treatment in both ethanol- and pair-fed mice (Fig. 6A). While Kupffer cell depletion did not affect ethanol-induced C3b deposition (Fig. 6B), the absence of Kupffer cells abolished the ethanol induction of hepatic TNF-α expression (Fig. 6C). In contrast, the absence of TLR-4 had no effect on the early ethanol-induced increase in C3b deposition or TNF-α expression (Fig. 7).
The pathophysiological development of ethanol-induced liver injury follows a characteristic pattern of steatosis, hepatocyte injury, and inflammation, which in some cases progresses to fibrosis and cirrhosis. Whereas this process can take years in humans, in animal models the early stages of steatosis and hepatocyte injury can be reproduced over a period of weeks. Activation of hepatic macrophages by lipopolysaccharide, increased expression of inflammatory cytokines (particularly TNF-α), activation of complement, and the generation of reactive oxygen species are all important contributors to chronic ethanol-induced liver injury.3-5 However, the dynamics of these different contributors are not understood. Here we have identified an early increase in hepatic TNF-α expression that precedes the ethanol-induced accumulation of hepatic triglycerides and increased plasma ALT, as well as the induction of CYP2E1 and the accumulation of 4-HNE adducts, which result from oxidative stress. This early, transient increase in TNF-α was not observed in mice lacking C3, C3aR, or C5aR, indicating that ethanol-mediated activation of complement and interaction of C3a and C5a with their receptors contributed to the early induction of TNF-α expression (Fig. 8). Activation of complement by ethanol did not involve hepatic macrophages; however, macrophages were required for the early, ethanol-induced increase in hepatic TNF-α. Interestingly, TLR-4 was not required for either the early, ethanol-induced increase in C3b deposition or TNF-α accumulation. Taken together, our data suggest that ethanol exposure rapidly activates multiple components of innate immunity, including complement, Kupffer cells, and inflammatory cytokine production. The synergistic interaction among these pathways is a critical mechanism contributing to ethanol-induced liver injury.
Recent studies have identified the complement pathway as an important regulator of chronic ethanol-induced liver injury, with both C3−/− and C5−/− mice protected from specific consequences of long-term ethanol feeding.6, 8CD55/DAF−/− mice, lacking an important complement regulatory protein, exhibit exacerbated injury after long-term ethanol feeding.6 Interestingly, the absence of C3 or C5 differentially protects from chronic ethanol-induced liver damage. After 6 weeks of ethanol feeding, C3−/− mice, though protected from ethanol-induced steatosis, still exhibit increased expression of inflammatory cytokines, including TNF-α, IL-6, and IFN-γ, as well as moderately increased ALT. In contrast, chronic ethanol feeding does not increase liver cytokines or ALT in C5−/− mice, even though they develop steatosis.6 These data suggest that C3 and C5 differentially contribute to steatosis, hepatocyte injury, and increased production of inflammatory cytokines during long periods of ethanol exposure.
We further characterized the role of complement activation in the hepatic response to ethanol, identifying an essential role for both C3aR and C5aR in an early, complement-dependent increase in hepatic TNF-α expression. C3a and C5a are potent anaphylatoxins, capable of increasing inflammatory cytokine responses,21 which contribute to a number of chronic inflammatory diseases, including myocardial ischemia reperfusion, respiratory distress syndrome, arthritis, and antibody-dependent type II autoimmunity.21, 22 In the healthy liver, Kupffer cells and stellate cells express both C3a and C5a receptors. C5a receptor expression can be induced in hepatocytes in response to inflammatory cytokines23 or in regenerating hepatocytes.24 Whereas chronic ethanol feeding (5%, week 4) did not increase total hepatic C3aR or C5aR messenger RNA (data not shown), absence of either C3aR or C5aR prevented the early, ethanol-induced increase in hepatic TNF-α (Fig. 5). Although the present studies demonstrate that both C3a and C5a contribute to the early, ethanol-induced increase in TNF-α, our earlier studies suggest that C5a, rather than C3a, is more important for sustained ethanol-induced increases in inflammatory cytokine expression after chronic ethanol feeding.6 The greater effect of C5a in hepatic injury over longer periods of ethanol exposure could relate to its ability to recruit neutrophils to sites of infection/injury.25, 26 Alternatively, although C3a has numerous proinflammatory properties, it also has anti-inflammatory activity under some conditions. For example, C3a receptor knockout mice show increased sensitivity in a model of endotoxemia, suggesting important anti-inflammatory functions for the C3a receptor.27
Whereas both C3aR and C5aR were required for the early, ethanol-induced increase in hepatic TNF-α, TLR-4 was not required for this response. This is in contrast to the critical role of TLR-4 in mediating increased TNF-α expression and liver injury in response to chronic ethanol exposure.3, 4 The difference in the role of TLR-4 in early versus chronic ethanol-induced TNF-α expression may be due to the relatively low doses of ethanol and/or the short period of ethanol exposure used in this study. Thus, whereas our data demonstrate that the early increase in hepatic cytokine expression is primarily dependent on complement activation, additional factors, including activation of TLR-4 and elevated production of reactive oxygen species,3, 4 have been shown to contribute to sustained inflammatory cytokine production during the later phases of ethanol-induced liver injury.
Activation of systemic complement can be initiated via the classical, lectin, and alternative pathways, although additional extrinsic pathways for activation of C5 by proteases can play a role in some pathophysiological conditions.28, 29 Local production and activation of C5 by Kupffer cells also contributes to type II autoimmune injury in mice.21 More studies will be needed to determine if localized activation of C5 within the complex cellular environment of the liver contributes to ethanol-induced liver injury.
Previous reports demonstrated that TNFRI−/− mice are protected from chronic ethanol-induced liver injury, using intragastric infusion of ethanol over 4 weeks.3, 4 We exposed TNFRI−/− to ethanol for a shorter time (4%, day 7), when increased markers of injury were initially observed in wild-type mice, in order to more directly assess the role of the early increase in hepatic TNF-α in mediating subsequent ethanol-induced liver injury. The TNFRI−/− mice were completely protected from ethanol-induced increases in ALT and 4-HNE adduct accumulation at day 11, but still exhibited hepatic steatosis. These data support the hypothesis that the early increase in TNF-α is required for increased reactive oxygen species production and hepatocyte injury. This dichotomy between the role of TNF-α in ethanol-induced increases in ALT and steatosis is consistent with a growing body of evidence suggesting that distinct mechanisms contribute to ethanol-induced increases in ALT compared with triglyceride accumulation. For example, C5−/− mice are protected from increased cytokine and ALT, but not triglyceride accumulation, during chronic ethanol exposure.6 Alternatively, at these early times of ethanol exposure, hepatic steatosis may be more dependent on shifts in NADH/NAD ratios, resulting from ethanol metabolism, leading to decrease β-oxidation, rather than mechanisms for triglyceride accumulation that are dependent on complement and inflammatory cytokines that involve the up-regulation of genes regulating fatty acid synthesis and down-regulation of genes involved in fatty acid oxidation, as well as an impairment of secretion of very low-density lipoprotein.30
Our studies demonstrate that complement is necessary for early, ethanol-induced TNF-α expression and that Kupffer cells are also involved in mediating the early effects of ethanol on inflammatory cytokine expression in the liver. The contribution of Kupffer cells is likely due to C3aR and C5aR signal transduction during the early period of ethanol exposure and not related to activation of TLR-4. The present study thus illustrates the complex integration of multiple elements of the innate immune system during the response of the liver to ethanol. Remarkably, the absence of a single component, such as a deficiency in C3aR or C5aR, prevents the early, ethanol-induced increase in TNF-α expression. These data thus add to our understanding of the highly orchestrated response of the liver to insult, such as the injury that occurs in response to ethanol.
The authors are grateful to Dr. Shizuo Akira (Osaka University, Osaka, Japan) for the generous gift of TLR-4–deficient mice.