Key Laboratory of Human Disease Comparative Medicine, Ministry of Health, Institute of Laboratory Animal Science, Chinese Academy of Medical Sciences and Comparative Medical Center, Peking Union Medical College, Beijing, China
Potential conflict of interest: Nothing to report.
This work was supported by grants from the National Natural Science Foundation of China (31030038 to P.L., 30800555 to L.X., and 30925017 to P.L.) and the National Basic Research program (2011CB910800) and National High Technology Research and Development program (2010AA023002) from the Ministry of Science and Technology of China.
High levels of dietary saturated fat have been closely associated with the development of hepatic steatosis, but the factors that mediate this process remain elusive. Here, we observed that the level of cell death-inducing DNA fragmentation factor-alpha-like effector a (Cidea) expression was highly correlated with the severity of hepatic steatosis in humans. Overexpression of Cidea in mouse liver resulted in increased hepatic lipid accumulation and the formation of large lipid droplets (LDs). In contrast, mice with a Cidea deficiency had decreased lipid accumulation and alleviated hepatic steatosis when they received a high-fat-diet feeding or in ob/ob mice. Furthermore, the knockdown of Cidea in livers of ob/ob mice resulted in significantly reduced hepatic lipid accumulation and smaller LDs. Importantly, we observed that Cidea expression in hepatocytes was specifically induced by saturated fatty acids (FAs), and such induction was reduced when sterol response element-binding protein (SREBP)1c was knocked down. In contrast, the overexpression of SREBP1c restored the saturated FA-induced expression of Cidea. In addition, we observed that the stability of Cidea protein in hepatocytes increased significantly in response to treatment with FAs. Conclusion: Cidea plays critical roles in promoting hepatic lipid accumulation and in the development of hepatic steatosis by acting as a sensor that responds to diets that contain FAs. (Hepatology 2012;56:95–107)
Abundant epidemiological and experimental evidence has linked an excessive nutrient supply to the development of metabolic diseases, including obesity, type 2 diabetes, cardiovascular disease, and hepatic steatosis.1 Hepatic steatosis is characterized by the accumulation of excess amounts of hepatic neutral lipids, resulting from abnormal hepatic lipid metabolism.2 Mice with deficiencies in leptin or its receptor (ob/ob or db/db mice, respectively) or high-fat-diet (HFD) feeding develop hepatic steatosis because of increased food intake and higher plasma lipid levels.3-5
The composition of dietary lipids, including the balance between free fatty acids (FFAs) and triacylglycerols (TAGs), the ratio of saturated versus unsaturated fatty acids (FAs), and the molecular structures of unsaturated FAs, can control the degree of hepatic steatosis. Previous reports show that animals fed with saturated FAs develop severe hepatic steatosis.5 In contrast, treatment of animals or human patients with polyunsaturated fatty acids (PUFAs), such as omega-3 PUFAs and/or docosahexanoic acids (DHAs), alleviates hepatic steatosis and improves insulin sensitivity.6, 7 The molecular mechanisms by which FAs exert differential roles in hepatic steatosis are complex and controversial. FAs can modulate the gene expression involved in lipid and lipoprotein metabolism in the liver.8 PUFAs, such as arachidonic acids, eicosapentaenoic acids (EPAs), and/or DHAs, can specifically inhibit the expression of sterol response element-binding protein (SREBP)1c,8-10 whereas saturated FAs are reported to enhance SREPB1c expression,11 possibly by an increased endoplasmic reticulum (ER) stress response.12, 13 However, the identities of the molecules that sense the effects of dietary saturated FAs to initiate the induction of hepatic steatosis remain unclear.
The cell death-inducing DNA fragmentation factor-alpha-like effector (CIDE) proteins (e.g., Cidea, Cideb, and Cidec [or fat-specific protein of 27KD (Fsp27), the homolog of Cidec in the mouse])14 are lipid-droplet (LD)-associated proteins that have emerged as important regulators of lipid storage and the formation of large LDs in adipocytes and hepatocytes.15-19 Mice with a deficiency in Cidea, Cideb, or Fsp27 exhibit a higher energy expenditure and enhanced insulin sensitivity, as well as being resistant to HFD-induced obesity and diabetes.15, 16, 19, 20 Transcriptional regulation of the CIDE proteins is complex and appears to be tissue specific. The promoter regions of Cidea and Fsp27 contain the response elements characteristic of peroxisome proliferator-activated receptor (PPAR)α/γ and can be activated by a PPAR agonist in mouse liver21, 22 or in human adipocytes.23 Cidea has also been shown to be up-regulated in the presence of insulin; this up-regulation may be mediated by SREBP1c.24, 25 In addition, Cidea expression is regulated by PPARγ transcriptional coactivator 1 alpha through its interaction with other transcription factors or cofactors.26, 27
Interestingly, expression levels of both Cidea and Fsp27 are markedly elevated in livers of HFD-fed and ob/ob mice.19, 22, 28 Their expression is also increased in livers of mice with a hepatocyte-specific phosphatase and tensin homolog deficiency, which develop hepatic steatosis.29 In contrast, expression levels of Cidea and Fsp27 are decreased in several genetically modified animals that are resistant to hepatic steatosis.30, 31 Therefore, expression of both Cidea and Fsp27 in the liver is correlated with the development of hepatic steatosis in mice. Fsp27 has been shown to be a direct mediator of PPARγ-dependent hepatic steatosis.22 However, the role of Cidea in hepatic steatosis is controversial.22, 24 Multiple lines of evidence reveal that Cidea promotes large LD formation and TAG accumulation in various nonhepatic cell types.15, 17, 32 However, the physiological role of Cidea in the control of lipid storage and the development of hepatic steatosis, as well as the molecular mechanism of the up-regulation of Cidea during the development of hepatic steatosis, remain unclear.
Here, we observed that Cidea expression was markedly increased in human hepatic steatosis. Using various genetically modified animal models, we demonstrate that Cidea is a crucial player in the development of hepatic steatosis. In addition, we observed that Cidea expression was specifically increased in hepatocytes in response to saturated FA intake; this up-regulation was likely mediated by SREBP1c. We also observed that the stability of the Cidea protein in hepatocytes was significantly increased in response to FA treatment. Overall, we have elucidated a novel pathway for FA-induced hepatic steatosis that is mediated by Cidea.
ACC1, acetyl-coenzyme A carboxylase 1; BAT, brown adipose tissue; CE, cholesterol ester; Cide, cell death-inducing DNA fragmentation factor-alpha-like effector; CHX, cycloheximide; DGAT, diacylglycerol O-acyltransferase; DHA, docosahexanoic acid; ELOVL6, elongation of very long chain fatty acids protein 6; EPA, eicosapentaenoic acid; ER, endoplasmic reticulum; FA, fatty acid; FAS, fatty acid synthase; FFA, free fatty acid; Fsp27, fat-specific protein of 27KD; H&E, hematoxylin and eosin; HFD, high-fat diet; IHC, immunohistochemistry; LA, lenoleic acid; LD, lipid droplet; LNA, linolenic acid; mRNA, messenger RNA; ND, normal diet; OA, oleic acid; PA, palmitic acid; PIO, pioglitazone; PPAR, peroxisome proliferator-activated receptor; PUFA, polyunsaturated fatty acid; RT-PCR, reverse-transcription polymerase chain reaction; SA, stearic acid; SEM, standard error of the mean; siRNA, small interfering RNA; SREBP, sterol response element-binding protein; TAG, triacylglycerol; VLDL, very-low-density lipoprotein; WAT, white adipose tissue; WT, wild type; WY, WY-14643.
Materials and Methods
Cidea−/− mice were generated and maintained as previously described,15, 19 and wild-type (WT) C57BL/6 mice were used as controls. The double-knockout mice of ob/ob/Cidea−/− and ob/ob/Cideb−/− were generated by crossing Cidea−/− or Cideb−/− mice with leptin+/− mice. Detailed protocols for animal experiments are described in the Supporting Materials and Methods. Mouse experiments were performed in the animal facility of the Center of Biomedical Analysis at Tsinghua University (Beijing, China).
Human Liver Specimens.
Human liver specimens were collected from 15 patients from Xijing Hospital, The Fourth Military Medical University (Xian, China). Experiments were performed in accord with ethical requirements of The Fourth Military Medical University, and subjects were given written informed consent. Methods for hematoxylin and eosin (H&E) staining, immunohistochemistry (IHC), and reverse-transcription polymerase chain reaction (RT-PCR) analysis are described in the Supporting Materials and Methods.
All statistical analyses were performed using GraphPad Prism V4.0 (GraphPad Software, Inc., La Jolla, CA). Consolidated data are expressed as mean ± standard error of the mean (SEM), and P values were calculated using the nonparametric Student t test. Values of P < 0.05 were considered statistically significant.
Additional methods are described in the Supporting Materials and Methods.
Increased Cidea Expression in Human Hepatic Steatosis and Liver-Specific Expression of Cidea Promotes Hepatic Lipid Accumulation.
To evaluate the potential role of Cidea in the development of hepatic steatosis, we examined the expression levels of all three of the CIDE proteins in the livers of leptin (ob/ob)-deficient and HFD-fed mice. Cideb was abundantly expressed in the livers of normal diet (ND)-fed mice and was maintained at similar levels in the livers of HFD-fed and ob/ob mice (Fig. 1A). In contrast, Cidea and Fsp27 were not detected in livers of ND-fed mice, but were markedly elevated in livers of HFD-fed mice (Fig. 1A) and were further increased in livers of ob/ob mice (Fig. 1A), corresponding to higher TAG storage and more severe hepatic steatosis in ob/ob mice (Supporting Fig. 1A-C). Interestingly, messenger RNAs (mRNAs) for Cidea and Cidec were also detected in human liver specimens that showed steatotic morphology, but not in the healthy nonsteatotic livers (Fig. 1B). In addition, levels of Cidea and Cidec mRNA were correlated with the severity of human hepatic steatosis (Fig. 1B and Supporting Fig. 1D). Furthermore, Cidea protein was detectable on the surface of LDs of the liver secretion showing severe steatosis (Fig. 1C). Therefore, both Cidea and Cidec/Fsp27 are markedly up-regulated in steatotic livers of humans and mice, which strongly correlates with the development of hepatic steatosis.
To examine the role of Cidea in promoting hepatic lipid storage, we ectopically expressed Cidea in the liver cell line, AML12 (Supporting Fig. 1E), and observed a significant increase in the accumulation of larger LDs (Fig. 1D) and cellular TAG levels (Fig. 1E). When Cidea was specifically targeted to the livers of WT mice (Supporting Fig. 1F), levels of hepatic TAGs were significantly increased (Fig. 1E), and LDs were larger relative to those in mice that expressed green fluorescent protein (Fig. 1F). Expression of Cidea in the liver did not affect gross morphology, size of LDs, and amounts of TAGs in either white adipose tissue (WAT) or brown adipose tissue (BAT) (data not shown). Therefore, hepatic expression of Cidea resulted in increased lipid storage and the accumulation of large LDs in the liver.
Cidea Deficiency Alleviates the Hepatic Steatosis Caused by HFD Feeding and Leptin Deficiency.
To further evaluate the physiological role of Cidea in the formation of hepatic steatosis, we treated Cidea-deficient mice with an ND or HFD. Under the ND feeding condition, liver morphology and levels of hepatic TAGs were similar between WT and Cidea−/− mice (Fig. 2A,B). However, when fed with a HFD, livers of Cidea−/− mice contained lower amounts of lipids (approximately 40% lower) and fewer and smaller LDs (Fig. 2A,B). The decreased expression of PPARγ and genes in the FA-synthesis pathway (acetyl-coenzyme A carboxylase 1 [ACC1], fatty acid synthase [FAS], and elongation of very-long-chain fatty acids protein 6 [ELOVL6]) was observed in livers of HFD-fed Cidea−/− mice (Supporting Fig. 2A), whereas expression levels of genes in the FA β-oxidation, oxidative phosphorylation, and lipolysis pathways were similar between WT and Cidea−/− mice (Supporting Fig. 2A). mRNA levels of SREBP1c were slightly decreased (Supporting Fig. 2A). However, protein levels of mature nuclear form of SREBP1c were significantly decreased (Supporting Fig. 2B), correlating well with the decreased expression of FAS and ACC. Consistent with a previous report,15 levels of TAGs and sizes of LDs in BAT and WAT of Cidea−/− mice were lower than those in WT mice fed with a HFD (Supporting Fig. 2C,D). These data indicate that a Cidea deficiency resulted in reduced hepatic lipid accumulation and alleviated HFD-induced hepatic steatosis.
To further confirm the roles of Cidea in controlling hepatic steatosis, we generated Cidea and leptin double-deficient (ob/ob/Cidea−/−) mice. Livers of ob/ob/Cidea−/− mice had smaller and fewer LDs (Fig. 2C) and decreased levels of TAGs and cholesterol esters (CEs) (Fig. 2D,E). Interestingly, expression levels of SREBP1c, PPARα/γ, and de novo synthesis genes of FAs were significantly decreased in livers of the ob/ob/Cidea−/− mice (Fig. 2F). Expression levels of hepatic genes in lipolysis and mitochondrial oxidation pathways were similar between ob/ob and ob/ob/Cidea−/− mice (Supporting Fig. 3A). In addition, levels of TAGs and sizes of LDs were reduced in BAT of ob/ob/Cidea−/− mice relative to those of ob/ob mice (Supporting Fig. 3B,C). Consistent with the similar levels of hepatic Cideb in WT and ob/ob mice (Fig. 1A), sizes of LDs and levels of TAGs and CEs in livers of Cideb and leptin double-deficient (ob/ob/Cideb−/−) mice were similar to those in ob/ob mice (Fig. 2C-E). Overall, these data indicate that a Cidea deficiency results in reduced lipid accumulation and ameliorates the hepatic steatosis induced by an HFD or a leptin deficiency.
Liver-Specific Knockdown of Cidea Reduces Hepatic TAG Accumulation in ob/ob Mice.
To eliminate the possibility that the reduced hepatic lipid accumulation in ob/ob/Cidea−/− mice was the result of indirect contributions of other tissues, such as BAT where Cidea is expressed at high levels, we knocked down Cidea specifically in livers of ob/ob mice. Knockdown of Cidea in livers of ob/ob mice (a 60% reduction in Cidea protein level; Fig. 3A and Supporting Fig. 4A) significantly reduced serum and hepatic levels of TAGs and LD sizes relative to those of the control (Fig. 3B-D). Furthermore, liver-specific knockdown of Cidea increased oxygen consumption and overall energy expenditure (Fig. 3E,F and Supporting Fig. 4B,C). In contrast, liver-specific knockdown of Cidea did not affect food intake, body weight, serum levels of FFAs, hepatic expression of Fsp27 and Cideb, and cellular levels of TAG and LD sizes in WAT and BAT (Fig. 3A-D and Supporting Fig. 4A,D). Knocking down of Cidea in primary ob/ob hepatocytes (Supporting Fig. 5A) also led to the accumulation of smaller LDs and reduced hepatic TAG levels (Supporting Fig. 5B,C). Overall, these data strongly indicate that Cidea plays a crucial role in promoting hepatic lipid accumulation and in the formation of hepatic steatosis in animals fed with an HFD or harboring a leptin deficiency.
Expression of Cidea Is Specifically Induced by Saturated FAs.
Next, we sought to understand the molecular mechanisms governing Cidea high expression in the liver during HFD feeding or in leptin-deficient mice. Consistent with their increased protein levels, hepatic Cidea and Fsp27 mRNA levels were markedly increased in livers of HFD-fed and ob/ob mice (Fig. 4A). Levels of mRNA encoding ADRP and tail-interacting protein of 47KD in livers of HFD-fed or ob/ob mice were also increased, albeit to a much lesser extent than that of Cidea and Fsp27 (Fig. 4A).
We then monitored the expression profiles of Cidea and Fsp27 during the course of HFD treatment. Induction of hepatic Cidea mRNA levels was observed 2 days after HFD treatment and continued to increase with further HFD feeding (Fig. 4B). However, induction of Fsp27 expression was only observed in livers of animals treated with an HFD for 1 month (Fig. 4B). Hepatic Cideb mRNA levels were similar before and after HFD treatment (Fig. 4B). Concomitant with increased Cidea mRNA levels, levels of serum FFAs were increased after 2 days of HFD feeding (Fig. 4C). Hepatic TAG levels were increased 2 weeks after HFD feeding (Fig. 4D). These data indicated that the expression of the CIDE family proteins was differentially regulated by an HFD and that Cidea gene expression was the most sensitive to dietary fat treatment.
We further evaluated the expression of the CIDE family proteins in response to various types of FAs in isolated ob/ob hepatocytes. When treated with saturated FAs, mRNA levels of Cidea were induced 2.5- and 2.0-fold by palmitates (PAs) and stearates (SAs), respectively (Fig. 5A). PAs and SAs also enhanced Cidea expression in AML12 cells (Fig. 5B). However, levels of Cidea mRNA were not induced either in ob/ob hepatocytes or AML12 cells, by unsaturated FAs, including oleic (OA), linoleic (LA), linolenic (LNA) acids, or EPA (Fig. 5A,B). WY-14643 (WY) and pioglitazone (PIO), agonists for PPARα and PPARγ, respectively, also did not induce Cidea expression. Interestingly, mRNA levels of Fsp27 in ob/ob hepatocytes and AML12 cells were not affected by FA treatment, but were significantly enhanced by PPAR agonists (Fig. 5A,B). Cideb expression was not affected by treatment with FAs or PPAR agonists (Supporting Fig. 6A,B). In addition, the expression level of Cidea (but not Fsp27 and Cideb) was up-regulated in the primary hepatocytes that were isolated from mice treated with an HFD for 2 days and incubated with saturated FAs (Fig. 5C and Supporting Fig. 6C). Consistent with increased gene expression, Cidea protein levels were higher in ob/ob hepatocytes treated with saturated FAs relative to the control cells (Fig. 5D and Supporting Fig. 6D). Fsp27 protein levels were also increased in cells treated with PPAR agonists (Fig. 5D and Supporting Fig. 6E). Interestingly, despite no effects on inducing Cidea mRNA level, OAs, LAs, and LNAs were able to increase Cidea and Fsp27 protein levels (Fig. 5D and Supporting Fig. 6D,E), suggesting a post-transcriptional regulation of Cidea and Fsp27 by these FAs. Overall, these data indicated that gene expression of the CIDE family members was differentially regulated by dietary FAs and PPAR agonists, and that Cidea expression was specifically induced by saturated FAs.
SREBP1c Mediates Saturated FA-Induced Cidea Expression in Hepatocytes.
To identify the transcription factor(s) responsible for saturated FA-induced Cidea expression in hepatocytes, we checked expression levels of several key transcriptional regulators in livers of HFD-fed mice. We observed that levels of hepatic SREBP1c mRNA and its downstream target genes (i.e., FAS and ACC1) were increased in animals fed with HFDs for 2 days and continued to increase with HFD treatment (Fig. 6A and Supporting Fig. 7A), which correlated well with the increased Cidea expression. In addition, protein levels of the mature nuclear form of SREBP1c were significantly increased in livers of HFD-fed mice (Fig. 6A and Supporting Fig. 7C). mRNA levels and its nuclear form of SREBP1c were also increased in ob/ob hepatocytes treated with PAs and SAs (Fig. 6B and Supporting Fig. 7D). Hepatic expression of other transcriptional regulators, including SREBP2, PPARα, and liver X receptor alpha, were not affected by HFD treatment (Supporting Fig. 7B).
The strong correlation between expression levels of SREBP1c and Cidea suggests that SREBP1c may serve as a transcriptional activator for saturated FA-induced Cidea expression. To test this hypothesis, we first overexpressed SREBP1c in AML12 cells and observed that Cidea (but not Fsp27 and Cideb) expression was significantly increased (Supporting Fig. 8A). The addition of PAs further enhanced this expression (Supporting Fig. 8A). Next, we knocked down SREBP1c in ob/ob hepatocytes (an 80% reduction in mRNA levels; Supporting Fig. 8B) and observed that mRNA levels of FAS, one of its downstream targets, were also reduced (Supporting Fig. 8B). Interestingly, only slightly decreased levels of basal Cidea mRNAs were observed in SREBP1c knock-down hepatocytes in the absence of PAs (Fig. 6C-E). mRNA and protein levels of Fsp27 and Cideb were not affected by the knockdown of SREBP1c (Fig. 6D and Supporting Fig. 8C-E). Importantly, mRNA and protein levels of Cidea were significantly reduced in the SREBP1c-knockdown ob/ob hepatocytes treated with PAs (Fig. 6C-E). Consistently, the hepatic TAG level was reduced in SREBP1c knock-down hepatocytes treated with PAs (Fig. 6F). In contrast, the knockdown of SREBP1c did not affect Cidea mRNA and protein levels (Fig. 6C-E) and hepatic TAG accumulation in the presence of OAs (Fig. 6F). These data indicate that SREBP1c is an important mediator of saturated FA-induced Cidea expression and hepatic lipid accumulation.
Cidea Proteins Are Stabilized by FA Treatment and Lipid Synthesis in Hepatocytes.
During the course of our analysis, we noted that the increase in the Cidea protein levels was higher than the corresponding increase in its mRNA levels in the presence of both saturated and unsaturated FAs (Fig. 5D), which suggested that Cidea protein stability may be increased in the presence of FAs. To test this possibility, we first treated primary ob/ob hepatocytes with OAs or PAs and then incubated them with cycloheximide (CHX), which inhibits protein synthesis. OA treatment significantly prolonged the half-life of Cidea, which was increased from 40 to 80 minutes (Fig. 7A,B). Consistent with our previous study using adipocytes,33 the half-life of Fsp27 in ob/ob hepatocytes was also increased in the presence of OAs (Fig. 7A,B). Half-lives of Cidea and Fsp27 in ob/ob hepatocytes were also increased in the presence of PAs (Supporting Fig. 9A,B). In contrast, Cideb was a relatively stable protein; its stability was not affected by FA treatment (Fig. 7A).
Because FAs are usually converted into TAGs and stored in LDs, we checked whether FA-induced Cidea stability in hepatocytes was dependent on lipid synthesis by knocking down diacylglycerol O-acyltransferase (DGAT)1 and DGAT2, which are enzymes that catalyze the final step of TAG synthesis. Levels of DGAT1/2 in ob/ob hepatocytes were decreased by small interfering RNAs (siRNAs) specific for DGAT1/2 (Fig. 7C), and levels of Cidea and Fsp27 proteins and their half-life also decreased significantly (Fig. 7D-F and Supporting Fig. 9C). Similar results were observed in AML12 cells that overexpressed HA-Cidea (Supporting Fig. 9D-F). These data indicated that Cidea and Fsp27 were stabilized by treatment with FFAs and by lipid synthesis, which provided a positive feedback mechanism that promoted lipid storage and liver steatosis in hepatocytes.
CIDE family proteins are important regulators of various aspects of lipid metabolism, including control of lipid storage and LD size in adipocytes (by Cidea and Fsp27)15, 19 and control of very-low-density lipoprotein (VLDL) lipidation in the liver (by Cideb).18 Here, we explored the physiological function of Cidea in hepatic steatosis and identified a molecular pathway for its up-regulation by dietary FAs. We demonstrated that Cidea expression was highly correlated with the development of hepatic steatosis in humans, and that hepatic overexpression of Cidea results in a significantly increased hepatic lipid accumulation and large LDs. In contrast, Cidea-deficient mice were resistant to hepatic steatosis caused by HFD feeding or a leptin-deficiency. Furthermore, liver-specific knocking down of Cidea in ob/ob mice resulted in less lipid accumulation and alleviated hepatic steatosis. These data clearly demonstrate that Cidea plays pivotal roles in promoting lipid accumulation and hepatic steatosis in humans and mice. These data are also consistent with the role of Cidea in promoting lipid accumulation and LD growth in adipocytes15, 17, 32 and in isolated primary hepatocytes.24 The negative effect of overexpressing Cidea in primary hepatocytes observed by Matsusue et al.22 may be a result of the short duration of Cidea expression, lower levels or activity of Cidea, or lack of OA treatment in their experiments. Fsp27 also mediates the development of hepatic steatosis, because the knockdown of Fsp27 in livers of ob/ob mice reduced hepatic lipid storage.22 Therefore, both Cidea and Fsp27 likely contribute to the development of hepatic steatosis in humans and mice.
Unlike Cidea and Fsp27, Cideb is constitutively expressed in the liver, and its expression is not affected by HFD feeding or FA incubation or the development of hepatic steatosis in mice or humans. We have previously shown that under ND conditions, a Cideb deficiency results in decreased VLDL secretion.18 Interestingly, we observed here that hepatocytes deficient for both Cideb and leptin (ob/ob/Cideb−/−) had similar LD sizes and levels of TAG accumulation relative to ob/ob mice. This may be the result of the compensating effects of the significantly increased expression of Cidea and Fsp27 in livers of ob/ob mice. Therefore, Cideb appears to play an important role in controlling lipid homeostasis by regulating hepatic lipid storage and VLDL secretion under normal physiological conditions, when Cidea and Fsp27 are not expressed. Under pathological conditions, such as long-term HFD feeding or a leptin deficiency, Cidea and Fsp27 are highly expressed in the liver and are responsible for dramatically increased hepatic lipid storage and the development of severe hepatic steatosis. Therefore, the CIDE proteins appear to have differential functions in the promotion of hepatic lipid homeostasis.
Although up-regulation of both Cidea and Fsp27 was observed in livers of HFD-fed and ob/ob mice and in humans, our results indicate that the factors that control their transcriptional programs are different. A time-course analysis revealed that hepatic expression of Cidea, but not Fsp27, was correlated with the increase in serum FFA level after HFD feeding. In addition, the expression of Cidea, but not Fsp27, in primary hepatocytes and AML12 cells was up-regulated by saturated FAs, but not by mono- and polyunsaturated FAs or PPARα/γ agonists. Therefore, Cidea appears to be a specific mediator of dietary saturated FA-induced hepatic steatosis.
We also demonstrated that saturated FA-induced Cidea expression is likely mediated by SREBP1c, based on the following evidences. First, Cidea and SREBP1c expression is highly correlated in livers of HFD-treated mice and in isolated primary hepatocytes treated with FAs. Second, overexpressing SREBP1c induced Cidea expression, and this induction was further boosted by saturated FAs. Most important, knocking down of SREBP1c led to a marked abrogation of saturated FA-induced Cidea expression. In contrast, expression levels of Fsp27 and Cideb in hepatocytes were not affected by the overexpression or knockdown of SREBP1c. The mechanism of the up-regulation of SREBP1c by saturated FAs is not clear. Saturated FAs have been shown to induce ER stress,12 which may result in enhanced SREBP1c cleavage and increased nuclear activity.13 Consistent with this, the increased levels of the mature nuclear form of SREBP1c were observed in HFD- or saturated FA-treated hepatocytes. Interestingly, we only observed a slightly reduced basal expression of Cidea in SREBP1c knock-down hepatocytes in the absence of FA treatment. It is possible that SREBP1c may play a minor role in mediating basal Cidea expression in hepatocytes. Hepatic Cidea expression is also reported to be induced by PPARα/γ agonists.21 However, this induction was not easily recapitulated in isolated primary hepatocytes.22 It is possible that the induction of hepatic Cidea expression by a PPARα/γ agonist is dependent on the presence of both PPARα/γ and other specific cofactors, such as mediator 1, which bridges PPARγ and RNA polymerase II.34
Another interesting observation that explains the high levels of Cidea and Fsp27 in livers of HFD-fed and ob/ob mice is the drastically increased stability of these proteins in the presence of FFAs. This phenomenon is likely the result of an increased incorporation of FAs into TAG and to the formation of large LDs, because the knockdown of DGAT1/2, the enzymes responsible for TAG synthesis, abrogates the FFA-induced stabilization of Cidea and Fsp27. Levels of intermediate lipids in the TAG-synthesis pathway, including DAG, may also affect Cidea and Fsp27 stability. Increased Fsp27 stability in the presence FFA has also been observed in white adipocytes and led to an increase in lipid storage capacity.33 Therefore, enhanced Cidea stability in hepatocytes may provide a positive feedback to promote hepatic lipid storage and the development of hepatic steatosis.
Overall, our current data demonstrate that the gene expression and protein stability of CIDE family proteins are differentially regulated in the liver in response to various stimuli (Fig. 8). Cideb is the only CIDE family protein expressed at high levels under normal physiological conditions in liver, and its expression and stability are not affected by FFA or PPARγ agonists. Therefore, Cideb is likely to be a crucial regulator of hepatic lipid homeostasis under physiological conditions, possibly through the control of VLDL lipidation. Hepatic Cidea protein levels are markedly increased under HFD feeding and in ob/ob mice because of saturated FA-induced gene expression and elevated protein stability. Thus, Cidea is an important sensor of dietary FAs and a mediator of saturated FA-induced hepatic steatosis. In combination with the previous findings, our current data also indicate that Fsp27 may play dual roles in mediating PPARγ and FA-induced liver steatosis. Elucidating the mechanism of CIDE-regulated hepatic lipid storage may provide novel therapeutic opportunities for intervening in human hepatic steatosis and the associated pathologies.
The authors thank members of Peng Li's laboratory for helpful discussion.