Complement C3 contributes to ethanol‐induced liver steatosis in mice

Abstract

Background. It is becoming increasingly clear that liver steatosis, a typical early consequence of alcohol exposure, sensitizes the liver to more severe inflammatory and fibrotic changes. On the other hand, activation of the key complement component C3, a central player in causing inflammation and tissue damage, is also known to be involved in the regulation of lipid metabolism. This prompted us to study the development of alcoholic liver steatosis in mice lacking C3 (C3−/−).

Results. Both C3−/− and normal C3+/+ mice were fed a steatosis‐promoting high‐fat diet with or without ethanol for 6 weeks. The diet without ethanol caused moderate liver steatosis in C3−/− but not in C3+/+ mice. As expected, ethanol‐containing diet caused marked macrovesicular steatosis and increased the liver triglyceride content in C3+/+ mice. In contrast, ethanol diet tended to reduce steatosis and had no further effect on liver triglycerides in C3−/− mice. Furthermore, while in normal mice ethanol significantly increased the liver/body weight ratio, liver malondialdehyde level and serum alanine aminotransferase (ALT) activity, these effects were absent or small in C3−/− mice. A separate experiment with mice on chow diet confirmed the aberrant steatotic effect of ethanol in C3−/−mice: 4 hours after acute dosing of ethanol the liver triglyceride level had increased by 138% in C3+/+ mice (P<0.001), but only by 64% in C3−/− mice (n.s.).

Conclusion. In C3−/− mice alcohol‐induced liver steatosis is absent or strongly reduced after chronic or acute alcohol exposure. This suggests that the complement system and its component C3 contribute to the development of alcohol‐induced fatty liver and its consequences.

• C3 deficiency
• chronic alcohol
• complement
• liver steatosis
• mice

Abbreviations
ALT	=	alanine aminotransferase
C3	=	complement component 3
C3adesArg	=	ASP acylation‐stimulating protein
MDA	=	malondialdehyde
LPS	=	lipopolysaccharide
TG	=	triglyceride

Introduction

Alcoholic liver disease advances from fatty liver to inflammation, necrosis, fibrosis and cirrhosis. Steatosis, the first consequence of alcohol exposure, seems to sensitize the liver to various inflammatory stimuli, such as bacterial endotoxin, thus acting as a gateway for more severe damage 1. Alcohol abuse is frequently associated with elevated endotoxin levels, which have been found to correlate with pathology 2. In addition to infectious triggers, oxidative stress, ischemic damage and apoptotic or injured cells can aggravate ethanol‐induced hepatic steatosis and inflammation.

The inflammatory response includes activation of the complement system, which leads to the generation of inflammation‐inducing anaphylatoxins (complement activation products C3a, C4a and C5a). Activation of the classical or alternative pathway leads to cleavage of the complement component C3 to the pro‐inflammatory C3a fragment. C3a can be cleaved to C3a_desArg by carboxypeptidase N or R 3. C3a_desArg has also been referred to as acylation‐stimulating protein, ASP 4. Both ASP and its precursor C3a have been shown to be involved in the regulation of fatty acid uptake to adipose tissue and body lipid homeostasis 5. This has been established in experiments with C3−/− mice, which cannot generate ASP and manifest various metabolic alterations in their lipid metabolism. The C3−/− mice on atherosclerosis prone low‐density lipoprotein receptor/apolipoprotein double deficient background exhibit increased serum triglyceride concentrations 6. The C3−/−mice show also a delayed postprandial triglyceride clearance, an effect that is normalized by administration of ASP 7.

In addition to systemic effects on lipid metabolism, C3a and also the related anaphylatoxin C5a have recently been suggested to participate in the process of liver regeneration via activation of nuclear factors and cytokines 8. Both C3a and C5a mediate, e.g., lipopolysaccharide (LPS)‐induced pathogenic insults by stimulating the production of prostanoids and pro‐inflammatory cytokines by liver Kupffer cells 9,10. It has been shown that liver inflammatory and steatotic effects of alcohol exposure are to a large extent mediated via Kupffer cell activation 11,12.

On the basis of the above it is conceivable that complement factor C3 has multiple roles in the liver, pro‐inflammatory or regenerative, in responding to exposure by toxins like alcohol. C3 activation could promote inflammatory reactions as well as interact with lipid homeostasis. We have previously found that chronic alcohol administration to rats leads to complement activation and deposition of C1, C3 and C8 in the liver 13, and that the pathological effects are different in the livers of complement C6‐deficient animals 14. Here we report our further investigations of the role of complement in alcohol‐induced steatosis and other signs of liver damage by using mice lacking the C3 complement component.

Key messages

• In mice lacking complement component C3 the normal alcohol‐induced liver steatosis is absent or strongly reduced, suggesting a role for complement and C3 in the development of alcohol‐induced fatty liver and its consequences.

Methods

Animal treatment

C3−/− mice were backcrossed onto the C57BL/6 genetic background (Charles River, Uppsala, Sweden) for 13 generations (15). Heterozygous mice were then intercrossed to generate homozygous C3−/− mice. Male C3−/− mice were used for the experiments. Normal male C57BL/6 C3+/+ mice, initially weighing 24–27 g, were obtained from Harlan, the Netherlands. For the study all mice were individually housed in M2 plastic cages under specific pathogen‐free conditions and fed standard rodent food ad libitum for 5 days until initiation of the diet experiment. At the age of 6–7 weeks all mice received a diet based on a modified liquid diet protocol, further modified into a gel by addition of agar 16. Briefly, the modified high‐fat/low‐carbohydrate liquid diet was based on the commercial Lieber‐DeCarli diet (LD 101A; Purina Mills, Richmond, IN, USA), which provided 50% of the calories as previously described 17. In the modified version the fat content is increased from 35% (calories) to 44% by adding extra corn oil and the protein content maintained at 16% by adding casein (technical grade; Sigma, St Louis, MO, USA). Also vitamins and minerals are added to equal the composition of the commercial Lieber‐DeCarli diet, but no carbohydrate, so that its content is reduced from 11% to 5.5%. Twelve mice from both strains received ethanol diet. The content of ethanol was gradually increased from 2% to 5.3% (final), which corresponded to 34.5% of total diet calories. Control mice (n = 8 in both strains) received a diet containing 40% carbohydrate (maltodextrose) to equicalorically replace ethanol. Agar (Lab M Agar No 2, Amersham MC 6, 0.5% (w/w)) was added as described 16. All mice received their diet (20–25 g) in tilted Falcon tubes equipped with a 2×2 cm hole. Control mice were pair‐fed the same amount of control diet. During the 6 weeks of agar diet administration all mice also had access to a water bottle during the experiment. At termination, the mice were anaesthetized with sodium pentobarbital (60 mg/kg i.p.), blood samples collected by heart puncture and plasma separated and stored at −20°C. Pieces of livers were rapidly frozen or were collected in buffered formalin and embedded in paraffin.

In the acute ethanol exposure experiment C57BL/6 and C3−/− mice (n = 6–7) were given one dose of ethanol (5 g/kg; 20% solution) by intragastric intubation. Control animals (n = 6–7) received a corresponding amount of water. Four hours later the animals were sacrificed and liver samples removed as above.

The study was approved by the Committee for Animal Experimentation at the University of Helsinki.

Liver histopathology and biochemical assays

The efficiency of the chronic ethanol exposure was tested by analyzing blood samples (25 µL) taken between 8–9 a.m. from the saphenous vein. Average blood ethanol levels were calculated from 12+12 individual samples taken weekly during weeks 3–6 from cohorts of 3+3 C3−/− and C3+/+ mice. Ethanol was determined by head‐space gas chromatography 18. For histology, formalin‐fixed paraffin embedded liver pieces were cut in 6‐μm sections and stained with hematoxylin/eosin. Steatosis was evaluated by three persons and graded blindly from 0–4 as follows: 1 = <25% of cells containing fat; 2 = 26%–50%; 3 = 51%–75%; 4 = >75%.

For assay of liver triglycerides (as glycerol), 1 mL of liver methanol‐chloroform‐mixed homogenate was washed with sodium chloride; the resultant extract was dried and dissolved in 200 µL of tetraethyl ammoniumhydroxide/95% ethanol (1:28). After incubation at 60°C for 30 min the extracts were subjected to hydrolysis by mixing with 200 µL 50 mM HCl. Glycerol and serum alanine aminotransferase (ALT) activity were measured enzymatically by using commercial kits (Boehringer‐Mannheim, Germany). The liver concentration of malondialdehyde, an indicator of lipid peroxidation, was determined by the thiobarbituric method 19.

Statistical analysis of the data

The data are expressed as means±SD. Student's t test was used to test statistical difference between groups. Pathological scores were compared using the Mann‐Whitney U test. A P‐value<0.05 was considered statistically significant.

Results

In the experiment with chronic ethanol administration, both C3+/+ and C3−/− mice were given either ethanol‐containing diet (12+12 mice) or control diet (8+8 mice). The diets were relatively rich in fat, and in the control diet ethanol was equicalorically substituted with carbohydrates. The mice from all four groups consumed 15–16 g of the agar gel diet per day. Although all groups of mice gained weight during the 6‐week gel diet treatment, the weight gain in both ethanol‐treated groups of mice was moderate (0.8±1.8 g) as compared to the control mice (6.7±1.9 g). The final weights of the C3+/+ and C3−/− groups of mice on control diet did not differ, but in those on ethanol diet the C3−/− mice weighed about 6% less than the C3+/+ animals (P<0.05) (Table I). The daily intake of ethanol, as calculated from the consumption of ethanol diet, was 23–24 mg per g body weight in both ethanol‐treated groups. Analysis of ethanol from blood samples obtained from some mice at weekly intervals confirmed that the alcohol exposure was roughly the same (average 43 mM) in C3−/− and C3+/+ animals.

Table I. Effect of chronic ethanol feeding on body parameters of C3 +/+ and C3 −/− mice. Mice were fed either high‐fat liquid diet or ethanol (EtOH) high‐fat liquid diet for 6 weeks and liver steatosis was assessed from tissue sections as described in Methods.

	C3+/+ (n = 8)	C3+/+ EtOH (n = 12)	C3−/− (n = 8)	C3−/− EtOH (n = 12)
Body weight (g)	32.76±1.60	26.86±1.72	32.29±2.37	25.18±1.28 ^a,b
Liver weight (g)	1.25±0.05	1.36±0.14	1.34±0.14	1.15±0.07 ^a,b
Liver/body weight, %	3.82±0.20	5.08±0.36 ^a	4.14±0.31 ^b	4.55±0.18 ^a,b
Steatosis score (0–4)	1.6±0.5	2.4±0.5 ^a	2.8±0.6 ^b	1.5±0.4 ^a

^a P<0.05 for the effect of ethanol; ^bP<0.05 for the effect of genotype.

The hepatic effects of both control and ethanol diet were clearly different between C3+/+ and C3−/− mice. The high‐fat diet by itself caused marked macrovesicular steatosis in C3−/− mice, in contrast to C3+/+ mice (Figure 1; Table I). The strain‐specific diet effect was also seen as a significantly increased content of triglycerides in the liver (Figure 2). As a probable consequence, the relative liver weight was also higher in C3−/− than in C3+/+ mice (Table I). Interestingly, in animals on control diet, the serum ALT activity was significantly higher (P<0.05) in C3−/− than in C3+/+ mice (Figure 2).

Figure 1. Representative micrographs of liver sections after 6 weeks of feeding either control high‐fat liquid diet to C3+/+ (A) and C3−/− (B) mice or ethanol‐high‐fat liquid diet to C3+/+ (C) and C3−/− (D) mice. Formalin‐fixed sections stained with hematoxylin/eosin were used. Note the marked mixed micro‐ and macrovesicular steatosis after control diet in C3−/− mice (B) as compared to C3+/+ mice (A). While ethanol feeding caused significant steatosis in C3+/+ mice (C) only minor steatosis was seen in C3−/− mice (D).

Figure 2. Effect of chronic ethanol feeding on the concentration of triglycerides(A) and malondialdehyde (B) in the liver and on serum alanine aminotransferase activity (C) in C3+/+ and C3−/− mice. Animals were treated with ethanol (n = 12) or control (n = 8) diet for 6 weeks. (* P<0.05 for the effect of ethanol; # P<0.05 for the effect of genotype.)

In normal C3+/+ mice, the ethanol diet caused expected changes in the liver. There were significant increases in steatosis and triglyceride content, and the liver/body weight ratio had increased by 33%. In addition, both serum ALT activity and the concentration of malondialdehyde, which reflects lipid peroxidation, were doubled (P<0.05). These ethanol effects were markedly different in C3−/− mice. The steatosis score was even lower than in mice on ethanol‐free diet, and there was no effect of ethanol diet on liver triglycerides. The liver/body weight ratio increased only by 10%, and there was no significant effect of ethanol treatment on serum ALT activity or liver malondialdehyde concentration.

To elucidate whether the aberrant effect of ethanol on liver lipids in C3−/− mice was specific to the high‐fat diet composition combined with chronic ethanol exposure, a separate study was undertaken. Both C3+/+ and C3−/− mice kept on chow diet were intragastrically intubated a single intoxicating dose of ethanol. Four hours later the content of triglycerides in the liver had increased by 138% in the normal C3+/+ mice (P<0.001) as compared to mice intubated with water. In contrast, in livers of C3−/− mice the increase was only 64% (P<0.1) (Figure 3). Thus, following both acute and chronic ethanol intake C3+/+ and C3−/− mice respond differently, implying a role for C3 and the complement system in regulating liver steatosis.

Figure 3. Effect of acute ethanol challenge on the content of liver triglycerides(TG) in C3+/+ and C3−/− mice (n = 6–7 in each group). An intoxicating dose of ethanol (5 g/kg) was intubated intragastrically to mice on chow diet 4 hours before termination and removal of liver specimens. (** P<0.01 for the effect of ethanol; # P<0.05 for the effect of genotype.)

Discussion

The present data demonstrate that mice deficient in complement component C3 respond to ethanol differently than normal C3‐sufficient mice. Ethanol exposure to C3−/− mice on high‐fat diet reduces the hepatic steatosis score and does not increase the accumulation of triglycerides, a phenomenon normally seen in rodents as well as in man 1. The aberrant response of the C3−/− mice was also observed after acute ethanol administration. While there is published evidence for abnormal lipid metabolism in C3 deficiency 6, the response to ethanol has not been reported before. Our data suggest that C3 contributes to ethanol‐induced liver fatty infiltration that can progress to inflammatory and fibrotic damage and ultimately to cirrhosis. Thus, resolving the mechanisms of interaction between C3 activities and liver lipid metabolism might be helpful in the future treatment of both alcoholic and non‐alcoholic liver disease.

It is apparent that the complement system has multiple roles both in causing injury and in repair processes of liver tissue. Complement activation can trigger tissue damage by inflammation or contribute to homeostasis via clearance and disposal of damaged cellular material. In our initial study with rats, we observed that chronic ethanol treatment caused deposition of complement components C1, C3 and C8 but reduced expression of the complement regulators Crry and CD59 in the livers 13. These changes could contribute to liver damage. However, in a subsequent study we found that chronic ethanol exposure to complement C6‐deficient rats resulted in more liver damage than in C6‐sufficient animals 14. This suggested to us that an intact terminal pathway of the complement system can also be important in the protection of the liver from external insult. Our present study, applying a similar alcohol exposure model in mice, suggests that an intact complement system, and especially the presence of C3, contributes to liver damage rather than being protective. This contribution to damage seems to be indirectly mediated via steatosis, which is considered to sensitize the liver to inflammatory attack. This concept is supported by our findings that in C3+/+ mice, but not in C3−/− mice, chronic ethanol administration increased serum ALT activity and liver malondialdehyde concentration. Malondialdehyde is an indicator of the degree of lipid peroxidation, which gives rise to reactive oxygen radicals injurious to cells. Unlike with C3 our previous study on rats suggested that the C6 component is protective against liver damage. This could be due to the different animal model used. However, the fact that C3 has a different and more central role in the complement cascade than C6 may explain why C6 and an intact terminal pathway protected against damage. Together, however, the studies suggest a dual role for the complement system where, on one hand, it contributes to inflammation and tissue damage and, on the other hand, participates in the clean‐up process and tissue repair. The absence of the protective liver homeostasis maintaining function in C3−/− mice may have contributed to the fatty infiltration observed in livers of these mice on diet without ethanol. In the acute experiment ethanol significantly increased the hepatic level of triglycerides in C3+/+ but not in C3−/− mice. This supports the notion that the different response of the C3−/− mice is independent of their dietary status and not due to a ceiling effect on triglyceride accumulation in the C3−/− mice kept on high‐fat diet.

Many studies indicate that the alternative pathway (AP) of the complement system influences lipid metabolism. Thus, in patients with an antibody against the AP C3 convertase C3bBb, called C3 nephritic factor (C3Nef), partial lipodystrophy occasionally accompanies membranoproliferative glomerulonephritis type II 20. C3Nef causes hypercatabolism of the AP and drastically lowers C3 levels. Interestingly, in a series of 35 patients with partial lipodystrophy, low C3 levels and renal disease were found in 83% and 22% of the patients, respectively 21. Hypercatabolism of C3 and consequent depletion of C3 thus seems to be related to a redistribution of fat from peripheral tissue to the liver and the lower abdominal area. The reasons for this association remain unknown.

Complement C3 activation leads to the formation of C3a, which is subsequently converted to C3a_desArg, alias ASP, following removal of the C‐terminal arginine. ASP has been suggested to participate in the clearance of postprandial triglycerides and to promote uptake of free fatty acids to peripheral fat tissue. It is therefore possible that the absence of C3 and ASP reduces the risk for peripheral obesity but allows steatosis in the liver, as we have observed in our present mouse model. In the presence of ethanol, however, the situation seems to be somewhat different. In livers of the C3+/+ mice, but not in the C3−/− mice, ethanol increases fat accumulation and this is accompanied by more hepatotoxic and inflammatory changes. This suggests that an intact complement system contributes to these changes. Such changes could be a consequence of enhanced production of inflammatory signals by Kupffer cells in response to ethanol. For instance, Kupffer cells are known to be activated by circulating LPS originating from the gut as a result of ethanol‐induced increased intestinal leakage. Although moderate activation of Kupffer cells can contribute in repairing injury, excessive activation may be detrimental and lead to aggravation of injury. Since the liver is the major organ synthesizing complement components, hepatic dysfunction will lead to decreased synthesis of major complement components. Indeed, in humans an association between liver damage and decreased activity of the complement system is frequently observed 22. As a consequence, this will lead to an increased susceptibility to infections caused e.g. by pneumococci and meningococci.

In conclusion, the complement system and especially its component C3 appear to have an important function in the development of ethanol‐induced acute and chronic fatty liver. It is becoming increasingly clear that liver steatosis sensitizes the liver to inflammatory attack and fibrotic lesions. Consequently, clarifying the mechanisms whereby C3 contributes to steatosis may ultimately help in understanding the pathogenesis of alcoholic liver damage.

Acknowledgements

This study was supported by a grant to SM from the Finnish Foundation for Alcohol Studies. We thank Dr Matti Jauhiainen and Dr Vesa Olkkonen for fruitful discussions and Dr Antti Väkevä for expert help in producing the digital images.