In humans, chronic ethanol consumption leads to a characteristic set of changes to the metabolism of lipids in the liver that is referred to as an "alcoholic fatty liver" (AFL) 1. This condition is characterized by an increase in liver weight 2, the accumulation of triglycerides and changes in expression of genes involved in lipid metabolism 3. In severe cases, these changes eventually lead to inflammation 4 and steatohepatitis 5 and are considered risk factors for cirrhosis and liver cancer in humans. Most of these changes are also observed in rodents 6 allowing us to use chronic alcohol exposure in rodents as a model to understand the human condition.

In recent years, many of the genes responsible for the generation of circadian rhythms have been identified 7. Many of these genes have been found to be expressed in the liver where the transcription of a number of key genes is regulated on a circadian time scale 8. Mice containing a mutation in the transcription factor, CLOCK, are unable to generate circadian rhythms in behavior and hepatic gene expression 910. Interestingly, recent studies suggest that Clock-mutant mice exhibit abnormal triglycerides 11, cholesterol metabolism 12, and become obese 13. Therefore, to understand the potential interactions between ethanol consumption, lipid metabolism and the circadian clock, we investigated the effect of chronic ethanol intake on the lipid metabolism of Clock-mutant mice.

The liver plays a key role in the metabolism of alcohol and is also sensitive to the deleterious effects of chronic alcohol consumption. In humans, chronic alcohol consumption leads to a characteristic set of changes to the metabolism of lipids in the liver that is referred to as an "alcoholic fatty liver". In humans, AFL is characterized by an increase in liver weight, accumulation of triglyceride and changes in the expression of genes involved in lipid metabolism. In severe cases, these metabolic changes result in the enlargement and fibrillization of the liver and are considered risk factors for cirrhosis and liver cancer.

Previous work suggests that there may be links between mutations in the circadian system and liver metabolism. The transcription of many genes in liver, including those involved in lipid metabolism, is regulated on the circadian time scale 822. Furthermore, mice with mutations in core elements of the circadian timing system also exhibit deficits in liver function. In the best studied example, the Clock-mutant mice 923 exhibit abnormal triglyceride 11, abnormal cholesterol metabolism 12, and become obese over time 13. In humans, common genetic variations in CLOCK transcription factor are associated with fatty liver disease 24. To further understand the potential interactions between ethanol consumption, lipid metabolism, and the circadian clock, we investigated the effect of chronic ethanol intake on the lipid metabolism of Clock-mutant mice. We used a previously established protocol in which mice are given drinking water with 15% ethanol over the course of 8 weeks 25.

We found that ethanol exposure increased the weight of the liver in Clock-mutant, but not wild-type mice (Fig. 1). Previous studies have found that alcohol consumption leads to increase in weight of liver of humans 2 and rats 19. This difference in liver weight between the two genotypes could not be obviously explained by consumption. The body weight between the two genotypes was not different with or without ethanol exposure (Table 2). Finally, we saw little evidence for liver damage in the mice exposed to this treatment regime. ALT and AST are important enzymes produced by the liver and serum levels of these enzymes are widely used biomarkers of liver health 20. Alcohol consumption leads to increase in ALT and AST of humans 26 and rats 27, but not in wild-type or Clock-mutant mice in the present study (Table 2). Thus, the 8-week exposure to ethanol caused a modest and selective increase in the weight of the liver of Clock-mutant mice without causing overt liver damage. The mechanisms underlying toxicity of ethanol associated with clock gene expression should be discussed from the view point of pharmacokinetics as well as pharmacodynamics. The circadian rhythm of alcohol dehydrogenase activity was previously reported in the rat liver 28. Therefore, alcohol metabolism might be disturbed in Clock-mutant mice, and resulted in increase of blood ethanol level. However, ethanol intake, and importantly, serum levels of ethanol were not different between the wild-type and Clock-mutant mice (Table 2). The hypothermic, hypnotic, and also lethal effect of ethanol showed circadian variations 29, suggesting that there is circadian rhythm in pharmacodynamics of ethanol. Taken all evidences together, it is suggested that we should consider the ethanol toxic effect in the circadian rhythms of the pharmacokinetics and pharmacodynamics, although mice did not show any evidence for a daily rhythm in the levels of liver triglyceride.

Triglyceride is one of the most important lipids in the liver and triglyceride in liver and serum are widely used measures of lipid metabolism 30. Ethanol treatment also increased triglyceride content of liver in Clock-mutant mice as well as wild-type controls (Fig. 2). The ethanol-induced increase in triglycerides was larger in the mutants than in the wild-type mice. Previous studies in humans 31 and rats 32 have also reported that ethanol increases levels of triglyceride. There were the circadian rhythm in Aco and Pparα, but not in Acc1 and Mtp. Analysis of oil red O staining provides a qualitative measure of the lipid content in the liver as higher levels of lipids leads to stronger staining 11. We found again that the ethanol-treated group exhibited stronger staining than untreated and that this increase in staining was more robust in the Clock-mutant mice (Fig. 3). Therefore, both the triglyceride measurements and the histological analysis suggest that ethanol increases the lipid content in the liver and that this effect is more robust in the Clock-mutant mice.

CLOCK is a transcription factor and Clock-mutant mice exhibit a number of changes in the transcription of gene in the liver 33. The ethanol-induced increase in lipid content of the liver of Clock-mutant mice could be explained by changes in triglyceride synthesis, degradation, or transport from the liver to blood. Therefore, we used RT-PCR methods to measure the expression of some of the genes involved in triglyceride synthesis, degradation, or transport (Fig. 4). Triglyceride is synthesized by fatty acids and glycerol released by the glycolytic pathway 34. In triglyceride synthesis, ACC1 1 is the rate limiting enzyme of Acyl-CoA synthesis from fatty acid in mouse. We found that the Clock-mutant mice had higher baseline levels of Acc1 and that ethanol-exposure produced a strong increase in the expression of this gene. Another important player in this pathway is Aco, which is involved in triglyceride degradation by peroxisome β-oxidation 35. In the present experiment, gene expression of Aco in the liver of Clock-mutant mice was significantly reduced by ethanol intake. Thus, a reduction of β-oxidation by the Clock mutation could also contribute to the high liver triglyceride content. Finally, Mtp is a gene involved in the triglyceride transport from the liver to blood 36. Ethanol-exposure decreases expression of Mtp in the liver of Clock-mutant mice. A reduction of triglyceride export from the liver through reduced Mtp expression may also explain the increase of triglyceride content in the liver of Clock-mutant mice. This gene expression data suggests that exposure of Clock-mutant mice to ethanol produces a coordinated set of changes in the expression of genes involved in triglyceride synthesis, degradation and transport.

We also found that the expression of Pparα, which is clock-controlled gene, was reduced by ethanol intake in the Clock-mutant, but not in wild-type mice. It has been reported that Pparα knock-out mice receiving ethanol for eight weeks developed severe fatty liver 37. Therefore, the ethanol-induced decrease in Pparα expression could be responsible for the abnormal lipid accumulations in the liver of Clock-mutant mice. Interestingly, both Aco and Mtp genes have PPRE sequences in their promoter regions and receive a transcriptional regulation by Pparα 3839. Thus, ethanol-induced reduction in Pparα could explain the reduction in the transcription of Aco, and Mtp. The findings that the expression of these genes were reduced in untreated Clock-mutant mice in comparison with wild-type mice suggests that CLOCK is important for maintaining higher expression levels of these metabolic genes. In other words, these metabolic genes are controlled by Clock genes through PPARα. The Clock knock-out mouse has no central circadian phenotype 40, but Clock gene seems to be indispensible for the functioning of peripheral tissue clocks 41. CLOCK protein has also been shown to be a histon acetyl-transferase and this gene-regulatory function42, which is not necessarily only acting on other clock genes like Bmal1, may have to do with the liver phenotype shown here.

Thus, the ethanol-induced enlargement of the liver and the increase of liver triglyceride in Clock-mutant mice could be an early marker for alcohol-induced liver failure. By itself, the Clock-mutation did produce some modest changes in the levels of some of the genes involved in lipid metabolism. These transcriptional changes did not appear to result in any altered liver function at least with the parameters that we could measure. On the other hand, when we challenged the mutant mice with chronic exposure to ethanol, we found the mutant mice exhibited an increase in liver weight and lipid content. These mutant mice also exhibited ethanol-induced changes in expression of genes involved in triglyceride synthesis and degradation that could explain the effects that we observed. These changes in liver lipid content are similar to those reported in humans with AFL and, thus we propose, that Clock-mutant mice could serve as a good model for studying AFL formation.

We found that ethanol treatment produced a number of changes in the liver of Clock-mutant mice without impacting the wild-type controls. First, we found that 8 weeks of exposure to ethanol in the drinking water increased the weight of the liver in Clock-mutant mice. Ethanol treatment also increased triglyceride content of liver in Clock-mutant and wild-type mice. This increase was larger in the mutant mice. Finally, ethanol treatment altered the expression of a number of genes related to lipid metabolism in the Clock-mutant mice.

The authors declare that they have no competing interests.

TK suggested the primary hypotheses, performed the oil red O staining and real time RT-PCR, and wrote much of the manuscript. TT performed measuring body weight, liver weight, liver TG. SS contributed to the hypotheses and design of the study, and wrote parts of the manuscript. All authors read and approved the final manuscript.