ATF4 Protein Deficiency Protects against High Fructose-induced Hypertriglyceridemia in Mice*

Background: Hypertriglyceridemia is the most common lipid disorder with incompletely understood mechanisms. Results: ATF4 deficiency attenuates lipogenesis in the liver and protects against high fructose-induced hypertriglyceridemia in mice. Conclusion: ATF4 plays a pivotal role in regulating hepatic lipid metabolism. Significance: ATF4 is a contributing factor for the pathogenesis of hypertriglyceridemia. Hypertriglyceridemia is the most common lipid disorder in obesity and type 2 diabetes. It results from increased production and/or decreased clearance of triglyceride-rich lipoproteins. To better understand the pathophysiology of hypertriglyceridemia, we studied hepatic regulation of triglyceride metabolism by the activating transcription factor 4 (ATF4), a member of the basic leucine zipper-containing protein subfamily. We determined the effect of ATF4 on hepatic lipid metabolism in Atf4−/− mice fed regular chow or provided with free access to fructose drinking water. ATF4 depletion preferentially attenuated hepatic lipogenesis without affecting hepatic triglyceride production and fatty acid oxidation. This effect prevented excessive fat accumulation in the liver of Atf4−/− mice, when compared with wild-type littermates. To gain insight into the underlying mechanism, we showed that ATF4 depletion resulted in a significant reduction in hepatic expression of peroxisome proliferator-activated receptor-γ, a nuclear receptor that acts to promote lipogenesis in the liver. This effect was accompanied by a significant reduction in hepatic expression of sterol regulatory element-binding protein 1c (SREBP-1c), acetyl-CoA carboxylase, and fatty-acid synthase, three key functions in the lipogenic pathway in Atf4−/− mice. Of particular significance, we found that Atf4−/− mice, as opposed to wild-type littermates, were protected against the development of steatosis and hypertriglyceridemia in response to high fructose feeding. These data demonstrate that ATF4 plays a critical role in regulating hepatic lipid metabolism in response to nutritional cues.

Hypertriglyceridemia is the most common lipid disorder in obesity and type 2 diabetes. It results from increased production and/or decreased clearance of triglyceride-rich lipoproteins. To better understand the pathophysiology of hypertriglyceridemia, we studied hepatic regulation of triglyceride metabolism by the activating transcription factor 4 (ATF4), a member of the basic leucine zipper-containing protein subfamily. We determined the effect of ATF4 on hepatic lipid metabolism in Atf4 ؊/؊ mice fed regular chow or provided with free access to fructose drinking water. ATF4 depletion preferentially attenuated hepatic lipogenesis without affecting hepatic triglyceride production and fatty acid oxidation. This effect prevented excessive fat accumulation in the liver of Atf4 ؊/؊ mice, when compared with wild-type littermates. To gain insight into the underlying mechanism, we showed that ATF4 depletion resulted in a significant reduction in hepatic expression of peroxisome proliferator-activated receptor-␥, a nuclear receptor that acts to promote lipogenesis in the liver. This effect was accompanied by a significant reduction in hepatic expression of sterol regulatory elementbinding protein 1c (SREBP-1c), acetyl-CoA carboxylase, and fatty-acid synthase, three key functions in the lipogenic pathway in Atf4 ؊/؊ mice. Of particular significance, we found that Atf4 ؊/؊ mice, as opposed to wild-type littermates, were protected against the development of steatosis and hypertriglyceridemia in response to high fructose feeding. These data demonstrate that ATF4 plays a critical role in regulating hepatic lipid metabolism in response to nutritional cues.
Hypertriglyceridemia is the most common lipid disorder in obesity and type 2 diabetes. Hypertriglyceridemia results from increased production and/or decreased clearance of triglyceride (TG) 4 -rich lipoproteins (1)(2)(3). Hypertriglyceridemia along with its metabolic sequelae of plasma accumulation of TG-rich lipoprotein remnants is atherogenic, accounting for increased risk of coronary artery disease in obesity and type 2 diabetes (4 -10). Nevertheless, the underlying pathophysiology of hypertriglyceridemia is poorly understood. Critical for triglyceride metabolism is the liver, in which TG-rich very low density lipoproteins (VLDL-TG) are assembled for subsequent secretion into the blood. In response to fatty acid influx, the liver increases VLDL-TG assembly and secretion. This effect, which is regulated by insulin and fatty acid substrate availability, has been viewed as a mechanism for staving off excessive fat accumulation in the liver of normal individuals (1,11,12). In insulin-resistant subjects, this mechanism becomes impaired, contributing to the pathogenesis of hepatic steatosis and hypertriglyceridemia, two pathological conditions that are commonly intertwined in obesity and type 2 diabetes.
To better understand the pathophysiology of hypertriglyceridemia, we studied hepatic regulation of lipid metabolism by activating transcription factor 4 (ATF4). ATF4, also known as cAMP-response element-binding protein 2, belongs to a family of basic leucine zipper-containing proteins (13). Expressed abundantly in the bone, ATF4 plays a critical role in bone homeostasis (14 -19). ATF4 promotes type I collagen production and promotes osteoblast proliferation and differentiation and mediates parathyroid hormone stimulation of bone formation (14,16,17,20,21). ATF4 also regulates osteoclast differentiation and bone resorption via direct and indirect (via the up-regulation of receptor activator of NF-B ligand in osteoblasts and bone marrow stromal cells) mechanisms (18,19).
There is emerging evidence that ATF4 contributes to the regulation of energy homeostasis. ATF4 is shown to modulate the bioactivity of osteocalcin, a bone hormone that is known to stimulate insulin secretion from pancreatic ␤-cells in response to glucose (22). As a result, Atf4 Ϫ/Ϫ mice are associated with augmented glucose-stimulated insulin release, accounting for enhanced glucose tolerance (22). ATF4 deficiency also stimulates lipolysis in adipose tissue, contributing to increased energy expenditure in Atf4 Ϫ/Ϫ mice (23). Atf4 Ϫ/Ϫ mice are lean, due to reduced fat mass (22,23). This effect protects against the development of glucose intolerance and insulin resistance in Atf4 Ϫ/Ϫ mice in response to high fat feeding (24,25). ATF4 is also expressed in the brain (14). There is nascent evidence that hypothalamic ATF4 plays an important role in modulating insulin action in the liver via the vagus nerve system (26). ATF4 is up-regulated in response to amino acid or glucose deprivation in HepG2 cells, arguing for the notion that ATF4 integrates nutritional cues to carbohydrate metabolism, and ATF4 acts as a nutrient sensor (13,27,28). Nonetheless, it remains unknown as to how ATF4 impacts lipid metabolism and whether ATF4 contributes to the pathogenesis of hypertriglyceridemia.
In this study, we characterized the role of ATF4 in hepatic lipid metabolism in Atf4 Ϫ/Ϫ mice and wild-type littermates. We focused our studies on the liver, because the liver plays a pivotal role in lipid homeostasis. We determined the effect of ATF4 depletion on three distinct pathways that govern lipid metabolism in the liver, namely lipogenesis, VLDL-TG production, and fatty acid oxidation in Atf4 Ϫ/Ϫ mice fed regular chow or provided with free access to fructose drinking water. Excessive fructose consumption is associated with hypertriglyceridemia in animal models as well as human subjects (29 -34). We hypothesized that ATF4 deficiency would improve TG metabolism and protect against fructose-elicited hypertriglyceridemia in mice.

EXPERIMENTAL PROCEDURES
Animal Studies-Atf4 ϩ/Ϫ heterozygous mice were bred for generating homozygous Atf4 null (Atf4 Ϫ/Ϫ ) mice and wild-type (WT) littermates (Atf4 ϩ/ϩ ) in Black Swiss background, as described (18). Mice were fed regular rodent chow and water ad libitum in sterile cages with a 12-h light/dark cycle. For fructose feeding, mice were fed regular chow with free access to drinking water containing 30% (w/v) fructose. For blood chemistry, mice were fasted for 16 h and tail vein blood samples were collected into capillary tubes pre-coated with potassium/EDTA (Sarstedt, Nümbrecht, Germany). The study was performed in male Atf4 Ϫ/Ϫ and sex/age-matched wild-type littermates. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Pittsburgh.
VLDL-TG Production Assay-Mice were fasted for 16 h, followed by intravenous injection of tyloxapol (0.5 g/kg, Sigma) to inhibit systemic TG clearance. Aliquots of tail vein blood were taken at different times for determining plasma TG levels. VLDL-TG production rates were defined as the amount of hepatic TG produced per unit time, as described (35).
Fat Tolerance Test-Mice were fasted for 16 h, followed by an oral bolus of olive oil (10 l/g). Aliquots of blood (25 l) were taken from the tail vein at different times for determining plasma TG, as described (36).
Glucose Tolerance Test-Mice were fasted for 5 h, followed by intraperitoneal injection of glucose (2 g/kg). Blood glucose levels were measured prior to and at different times after glucose injection.
Insulin Tolerance Test-Mice were injected intraperitoneally with recombinant regular human insulin (Lilly) at the dose of 0.75 units/kg body weight, followed by the determination of blood glucose levels at different times.
Hepatic mRNA Superarray-Aliquots of total RNA (5 g) isolated from liver tissues were subjected to Superarray analysis, using the ABI 7900 HT Fast Real Time PCR system (Applied Biosystems, Foster City, CA). This assay allows simultaneous determination of 96 mRNA species in a given sample. We used the mouse insulin signaling and carbohydrate metabolism pathway (PAMM-030, BABioscience, Valencia, CA) according to the manufacturer's instructions for quantifying mRNA levels of genes in insulin action and glucose and lipid metabolism in the liver.
FPLC Fractionation of Lipoproteins-Aliquots (400 l) of plasma pooled from Atf4 Ϫ/Ϫ mice and WT littermates were applied to two head-to-tail linked Tricorn high performance Superose S-6 10/300GL columns using an FPLC system (GE Healthcare), followed by elution with PBS at a constant flow rate of 0.25 ml/min. Fractions (500 l) were eluted for determining TG and cholesterol levels, as described (35,36).
Hepatic Lipid Content-Aliquots of liver tissue (20 mg) were homogenized in 400-l HPLC-grade acetone. After incubation with agitation at room temperature overnight, aliquots (50 l) of acetone/extract lipid suspension were used for the determination of triglyceride concentrations using the Infinity triglyceride reagent (Thermo Electron). Hepatic lipid content was defined as milligrams of triglyceride/g of total liver proteins, as described (35,37).
Cell Culture-HepG2 cells were purchased from American Type Culture Collection (ATCC, Manassas, VA). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Lonza, Walkersville, MD) and were transduced with Adv-ATF4 or control Adv-Empty vector as described (38). All adenoviral vectors were produced in HEK293 cells, as described (36).
Mouse Primary Hepatocytes-Mouse primary hepatocytes were isolated from C57BL/6J mice using the protocol, as described (38). The liver was infused in situ through the inferior vena cava first with 10 ml of HBSS supplemented with EGTA (1 mM), then with 10 ml of EGTA-free HBSS, and finally with 10 ml of HBSS supplemented with collagenase, type V (1.95 mg/ml, Sigma). The liver was harvested, minced, and incubated at 37°C for 10 min. The cells were dispersed in HBSS, filtered through a 70-m nylon mesh, and washed three times with HBSS. Hepatocytes were plated in collagen-coated 12-well plates at 2 ϫ 10 5 cells/well and cultured in hepatocyte maintenance medium (Lonza) supplemented with dexamethasone, insulin, and GA-1000 according to the manufacturer's instructions (Lonza), as described (38). For adenovirus transduction, primary hepatocytes were incubated with Adv-ATF4 or control Adv-Empty vector at a predefined dose (200 pfu/cell). After 24 h of incubation, primary hepatocytes and conditioned medium were collected for analysis.
Statistics-Statistical analyses of data were performed by analysis of variance using StatView software (Abacus Concepts). Analysis of variance post hoc tests were performed to study the significance between different conditions. Data were expressed as the mean Ϯ S.E. p values Ͻ0.05 were considered statistically significant.
A significant reduction in plasma TG levels could arise from increased systemic TG clearance and/or decreased VLDL-TG secretion from the liver. To distinguish these possibilities, we treated mice with tyloxapol, an antagonist of lipoprotein lipase, to block systemic TG hydrolysis, followed by the determination of hepatic VLDL-TG output. As expected, plasma TG levels increased with time following tyloxapol administration in Atf4 Ϫ/Ϫ and WT control mice ( Fig. 2A). However, the rate of VLDL-TG production, defined as plasma TG increment per unit time, remained unchanged in Atf4 Ϫ/Ϫ versus WT groups (Fig. 2B).
To account for the underlying physiology of improved TG metabolism in Atf4 Ϫ/Ϫ mice, we performed a fat tolerance test. Effect of ATF4 depletion on TG production and glucose metabolism. Atf4 Ϫ/Ϫ mice and WT littermates (n ϭ 7 per group) at 12 months of age were fasted for 16 h, followed by intravenous injection of tyloxapol for inhibiting systemic TG clearance. Aliquots of tail vein blood (25 l) were taken at different times for determining plasma TG levels (A). Data were used for calculating relative hepatic VLDL-TG production rates, defined as the amount of hepatic TG produced per unit time (B). In addition, Atf4 Ϫ/Ϫ and WT mice were fasted for 16 h, followed by an oral bolus of olive oil. Aliquots of blood from tail vein (25 l) were taken at different times for determining postprandial TG clearance (C). Mice were sacrificed at 15 months of age for determining epididymal fat mass (D). Body weight was compared between Atf4 Ϫ/Ϫ and WT littermates at 15 months old (E). Relative fat masses were calculated by normalizing total fat mass to body weight (BW) (F). Food intake during a 24-h period was determined under ad libitum conditions at 12 months old (G). Blood glucose levels under nonfasting condition (H) and after 16 h of fasting (I) were determined at 12 months of age. For glucose tolerance test, mice were fasted for 5 h, followed by intraperitoneal injection of glucose (2 g/kg). Blood glucose levels before and after glucose injection were determined (J). For insulin tolerance test, mice were injected intraperitoneally with insulin (0.75 units/kg), followed by determining blood glucose levels (K). *, p Ͻ 0.05, and **, p Ͻ 0.001 versus control. NS, not significant.
Mice were orally gavaged with a bolus of olive oil, followed by the determination of plasma TG levels. In contrast to WT littermates, Atf4 Ϫ/Ϫ mice exhibited significantly improved plasma TG profiles during the fat tolerance test (Fig. 2C). Furthermore, Atf4 Ϫ/Ϫ mice had significantly reduced epididymal fat mass (Fig. 2D). This effect persisted even after normalizing to body weight (Fig. 2, E and F). Atf4 Ϫ/Ϫ mice, as opposed to WT littermates, had relatively higher food intake after normalizing to body weight (Fig. 2G).
As control, we determined the impact of ATF4 on glucose metabolism. ATF4 depletion resulted in a significant reduction in blood glucose levels under both fed (Fig. 2H) and fasting (Fig.  2I) conditions, accompanied by enhanced glucose tolerance (Fig. 2J) and whole-body insulin sensitivity (Fig. 2K) in Atf4 Ϫ/Ϫ mice versus WT controls.
ATF4 Depletion Attenuates Lipogenesis in Liver-To assess the effect of ATF4 depletion on hepatic lipid metabolism, we determined the expression levels of key genes in hepatic lipogenesis. We detected a significant reduction in hepatic mRNA levels of sterol regulatory element-binding protein 1c (Srebp-1c) (Fig. 3A), the master regulator of lipogenesis in Atf4 Ϫ/Ϫ mice versus WT littermates. This effect contributed to the diminution in hepatic expression of Srebp-1c targets, namely Acc (Fig. 3B) and Fas (Fig. 3C) in the liver of Atf4 Ϫ/Ϫ mice.
We then determined the effect of ATF4 depletion on hepatic expression of MTP and apolipoprotein B (apoB), two key functions in hepatic VLDL-TG production. ApoB is the core component of VLDL particles. MTP catalyzes the rate-limiting step in transferring lipid to nascent apoB polypeptides for VLDL-TG assembly (12, 39 -41). We did not detect significant differences in hepatic mRNA levels of Mttp (Fig. 3D) and ApoB (Fig. 3E), consistent with the lack of alterations in hepatic VLDL-TG production rates in Atf4 Ϫ/Ϫ mice versus WT littermates (Fig. 2, A and B).
Furthermore, we determined the impact of ATF4 depletion on hepatic expression of key functions in fatty acid oxidation, including peroxisome Ppar-␣, and its two downstream targets Cpt1 and Acox1. We detected about 30% reduction in hepatic Ppar-␣ (Fig. 3F) and Acox1 levels (Fig. 3G) in Atf4 Ϫ/Ϫ mice, although this reduction did not reach a significant level, when compared with WT control. In contrast, hepatic Cpt1 mRNA levels remained unchanged in Atf4 Ϫ/Ϫ versus control mice (Fig.  3H). These data indicate that ATF4 depletion impacts hepatic lipid metabolism preferentially by attenuating lipogenesis in the liver.

Gene symbol
To provide further mechanistic underpinning for altered hepatic lipid metabolism in Atf4 Ϫ/Ϫ mice, we subjected liver tissues to mRNA Superarray analysis for determining hepatic expression of genes in insulin signaling and carbohydrate metabolism. This assay revealed a 2-fold reduction in hepatic Ppar-␥ mRNA abundance in the liver of Atf4 Ϫ/Ϫ mice (Table  2), consistent with the reduction of hepatic fat content in Atf4 Ϫ/Ϫ mice versus WT littermates (Fig. 3I).
Atf4 Ϫ/Ϫ Mice Are Protected against Fructose-induced Hypertriglyceridemia-Excessive fructose consumption is known to induce lipid disorder, culminating in the development of hyperlipidemia and hepatic steatosis in both humans and rodents (29 -34). Based on the above findings, we hypothesized that ATF4 deficiency would protect mice from developing hypertriglyceridemia in response to fructose feeding. To test this hypothesis, we provided Atf4 Ϫ/Ϫ mice and WT littermates with free access to drinking water containing 30% fructose. Fructose drinking for 8 weeks resulted in hypertriglyceridemia, accompanied by elevated plasma cholesterol levels in WT mice (Table 1). In contrast, Atf4 Ϫ/Ϫ mice maintained significantly lower plasma levels of TG and cholesterol after 8 weeks of fructose drinking (Table 1). To corroborate these findings, we subjected plasma of Atf4 Ϫ/Ϫ and WT mice to FPLC-mediated gel filtration chromatography for the fractionation of lipoproteins. We showed that Atf4 Ϫ/Ϫ mice exhibited markedly lower VLDL-TG levels ( Fig. 4A) with concomitant reduction in LDL cholesterol and HDL cholesterol levels (Fig. 4B). Atf4 Ϫ/Ϫ mice also displayed significantly lower nonesterified fatty acid levels, consistent with their improved lipid profiles in response to fructose consumption (Table 1).
Atf4 Ϫ/Ϫ Mice Stave Off Fructose-induced Steatosis-To determine the potential beneficial effect of ATF4 depletion on hepatic steatosis, we quantified hepatic lipid content following 8 weeks of fructose drinking. Atf4 Ϫ/Ϫ mice, as opposed to WT littermates, had significantly reduced hepatic TG content (Fig.  4C). Furthermore, Atf4 Ϫ/Ϫ mice had significantly reduced fat mass (Fig. 4D), accompanied by less weight gain (6.6 Ϯ 0.3 g) relative to WT littermates (10.5 Ϯ 0.5 g, p Ͻ 0.01) during the 8-week high fructose feeding. This effect on fat mass was maintained after normalizing to body weight (Fig. 4, E and F). To support these findings, we determined the effect of ATF4 depletion on hepatic expression of genes in lipogenesis versus fatty acid oxidation in response to fructose drinking. Atf4 Ϫ/Ϫ mice were associated with attenuated hepatic lipogenesis, culminating in a significant reduction in hepatic Ppar-␥ expression (Fig. 5A). Hepatic expression of Srebp-1c (Fig. 5B), Acc (Fig.  5C), and Fas (Fig. 5D) mRNA levels was also reduced in Atf4 Ϫ/Ϫ mice, but the degree of reduction did not reach significant levels, when compared with WT littermates. Instead, we detected a significant reduction in hepatic mRNA levels of Ppar-␣ (Fig.  5E), Cpt1 (Fig. 5F), and Acox1 (Fig. 5G), three key functions in fatty acid oxidation, correlating with the reduction of hepatic fat content (Fig. 4C) and plasma nonesterified fatty acid levels ( Table 1) in Atf4 Ϫ/Ϫ mice. Hepatic expression of MTP (Fig. 5H) and apoB (Fig. 5I), two enzymes catalyzing the rate-limiting step of VLDL-TG production in the liver, remained unchanged in Atf4 Ϫ/Ϫ mice versus WT littermates.
To provide further evidence that hepatic ATF4 depletion attenuated lipogenesis and protected against high fructose-induced steatosis, we determined hepatic abundance of key proteins in lipogenesis versus fatty acid oxidation pathways. We showed that hepatic levels of lipogenic proteins Ppar-␥ (Fig.  6A), Fas (Fig. 6B), and Acc (Fig. 6C) were significantly reduced, consistent with reduced fat accumulation in the livers of Atf4 Ϫ/Ϫ mice (Fig. 4C). In contrast, Srebp-1c nuclear protein levels remained unchanged (Fig. 6D). Hepatic Ppar-␣ protein levels were also reduced in Atf4 Ϫ/Ϫ mice (Fig. 6E), although the reduction did not reach a significant level in comparison with WT controls. We did not detect significant differences in hepatic Acox1 (Fig. 6F) and Cpt-1 (Fig. 6G) protein levels in Atf4 Ϫ/Ϫ mice versus WT littermates.
Effect of ATF4 Gain-of-Function on Hepatic Lipid Metabolism in Mouse Primary Hepatocytes-To provide further physiological underpinning for the hypothesis that ATF4 is pivotal for regulating hepatic lipid metabolism, we determined the effect of ATF4 gain-of-function on hepatic lipogenesis. We postulated that ATF4 would stimulate lipogenesis and instigate hepatic steatosis. To test this hypothesis, we prepared primary mouse hepatocytes, which were subsequently transduced with Adv-ATF4 at a moderate dose for elevating hepatic ATF4 expression levels by 2-3-fold, when compared with Adv-empty control vector-treated hepatocytes. We detected a significant elevation of TG in both cells (Fig. 7A) and conditioned medium (Fig. 7B) of primary hepatocytes that were pretreated with Adv-ATF4 vector. This effect correlated with the induction of hepatic lipogenesis, as evidenced by significantly increased Srebp-1c, Acc, and Fas mRNA levels in primary hepatocytes with elevated ATF4 expression (Fig. 7C). In contrast, hepatic mRNA levels of Ppar-␣, Cpt-1, and Acox1 in fatty acid oxidation remained unchanged (Fig. 7C). Likewise, hepatic ATF4 production did not result in significant changes in Ppar-␥ mRNA expression in primary hepatocytes (Fig. 7C).
To corroborate these findings, we subjected ATF4 and control vector-treated primary hepatocytes to immunoblot analysis. We confirmed that adenovirus-mediated ATF4 production resulted in a significant induction in Srebp-1c (Fig. 7D), Fas (Fig. 7E), and Acc (Fig. 7F) protein levels, without alterations in Ppar-␥ protein expression in primary hepatocytes (Fig. 7G). Consistent with their mRNA expression, Ppar-␣ and Acox1, Cpt1 protein levels remained unchanged in ATF4 vectortreated primary hepatocytes (data not show).

DISCUSSION
In this study, we characterized the role of ATF4 in hepatic lipid metabolism in Atf4 Ϫ/Ϫ mice. We showed that Atf4 Ϫ/Ϫ mice, as opposed to wild-type littermates, exhibited significantly reduced fasting plasma TG and VLDL-TG levels. Atf4 Ϫ/Ϫ mice also had significantly improved postprandial TG profiles during fat tolerance. In response to high fructose feeding, Atf4 Ϫ/Ϫ mice were protected against the development of hypertriglyceridemia. To gain insight into the underlying mechanism, we showed that ATF4 depletion resulted in a significant reduction in hepatic expression of lipogenic genes, without significantly affecting the expression of hepatic genes that are involved in fatty acid oxidation or VLDL-TG production in the liver. This effect contributed to the attenuation of hepatic lipogenesis and prevention of excessive fat infiltration in the liver of Atf4 Ϫ/Ϫ mice. We recapitulated this finding in Atf4 Ϫ/Ϫ mice under both regular chow and high fructose feeding conditions. In keeping with our finding, two independent studies demonstrate that ATF4 deficiency protects mice from developing steatosis in response to high fat feeding (24,25). Furthermore, we showed that ATF4 gain-of-function was associated with increased TG synthesis and secretion secondary to augmented lipogenesis in primary hepatocytes of C57BL/6J mice. Together these data characterize ATF4 as a significant factor for regulating hepatic lipid metabolism in response to nutritional cues.
The underlying mechanism by which ATF4 loss-of-function curbs hepatic lipogenesis and improves lipid metabolism is presently unknown. It is plausible that this observed improvement in lipid metabolism is secondary to enhanced glucose catabolism in Atf4 Ϫ/Ϫ mice. Implicit in this assertion is that ATF4 exudes a direct impact on glucose metabolism via osteocalcin, a bone hormone that is secreted specifically from osteoblasts (22). ATF4 acts in cooperation with the forkhead transcription factor FoxO1 to stimulate osteoblast expression of embryonic stem cell phosphatase (Esp), an enzyme that catalyzes osteocalcin carboxylation, resulting in inactivation of osteocalcin (22,45). As a result, ATF4 loss-of-function is asso-FIGURE 7. Hepatic protein expression in primary hepatocytes. Primary hepatocytes of C57BL/6J mice were transduced in culture with Adv-ATF4 (ATF4) and control Adv-Empty (Control) vectors at a predefined dose (200 pfu/cell). Each condition was run in triplicate. After a 24-h incubation, both hepatocytes and conditioned medium were harvested for determining TG concentration in cells (A) and medium (B). In addition, aliquots of ATF4 and control hepatocytes were subjected to real time qRT-PCR assay for determining hepatic mRNA expression. Hepatic mRNA levels, after normalizing to ␤-actin mRNA levels, were compared between control and ATF4 groups (C). Other aliquots of cells were subjected to Western blot analysis for determining hepatic protein levels of Srebp-1c (D), Fas (E), Acc (F), and Ppar-␥ (G), using actin protein as control. *, p Ͻ 0.05 and **, p Ͻ 0.001 versus control. ciated with enhanced insulin secretion from ␤-cells, contributing to a significant improvement in both fasting and postprandial glucose profiles in Atf4 Ϫ/Ϫ mice (22). We recapitulated these findings, demonstrating that ATF4-deficient mice exhibited significantly lower fasting blood glucose levels and enhanced glucose tolerance. Despite having relatively higher food intake, Atf4 Ϫ/Ϫ mice, as opposed to WT littermates, are lean, in accordance with enhanced energy expenditure (22,24). ATF4 depletion seems to enhance insulin sensitivity in the liver (26). The improvement in glucose homeostasis along with enhanced hepatic insulin sensitivity contributes to improved lipid metabolism in Atf4 Ϫ/Ϫ mice (23,25).
As an alternative mechanism, ATF4 exerts a direct effect on hepatic lipid metabolism. In support of this mechanism, we showed that ATF4 depletion curbs hepatic expression of Ppar-␥, a nuclear receptor that functions in complex with retinoid X receptor to promote lipogenesis in the liver (42, 46 -48). Hepatic Ppar-␥ expression is up-regulated in mice with hepatic steatosis (44,49) or dietary obesity (50 -52). This observation is recapitulated in humans, as obese patients with hepatosteatosis exhibited markedly increased PPAR-␥ activity in the liver (53,54). Conversely, mice with liver-specific Ppar-␥ deletion are refractory to fat-induced steatosis, due to the reduction of lipogenesis in the liver (42,47). Thus, ATF4 may act through PPAR-␥ to modulate hepatic lipid metabolism in ATF4-deficient mice. Consistent with this notion, we showed that hepatic PPAR-␥ expression was significantly decreased, coinciding with the attenuation of hepatic lipogenesis and reduction of hepatic fat infiltration in Atf4 Ϫ/Ϫ mice under both regular chow and high fructose feeding conditions.
Nonetheless, ATF4 gain-of-function did not seem to impact Ppar-␥ expression in mouse primary hepatocytes, despite the effect that ATF4 stimulated lipogenesis, contributing to increased TG synthesis and secretion. This suggests that ATF4 alone is not sufficient to promote hepatic Ppar-␥ expression, or alternatively, ATF4-mediated effect on Ppar-␥ expression is through an indirect mechanism. It is noteworthy that ATF4 does not bind to DNA alone. Instead, ATF4 binds together with CCAAT/enhancer-binding protein (C/EBP) to the C/EBP-ATF4-response element (CARE) in target promoters (13). ATF4 binds to one-half of the CARE motif, and the other half of the CARE is bound by C/EBP. Although ATF4 acts in concert with C/EBP to enhance the transcription of target genes (13,55), it is unknown how ATF4 cooperates with C/EBP for binding at the cognate site and for trans-activating promoter activity, nor is it clear whether C/EBP binding facilitates the recruitment of ATF4 at the CARE site or whether the occupancy of the CARE by both ATF4 and C/EBP is necessary for enhancing promoter activity. A recent study shows that ATF4 interacts with the forkhead transcription factor FoxO1 in regulating osteocalcin expression in osteoblasts (45). Further studies are needed to determine the molecular interplay of ATF4 with C/EBP or FoxO1 in regulating hepatic lipid metabolism.
In conclusion, we showed that ATF4 depletion diminished hepatic lipogenesis with little impact on hepatic VLDL-TG production or fatty acid oxidation. This effect prevented excessive fat accumulation in the liver and protected against the development of hypertriglyceridemia in fructose-fed ATF4-deficient mice. Our data characterize ATF4 as a contributing factor for hypertriglyceridemia. Further investigation is warranted to know whether ATF4 is a potential therapeutic target for treating hypertriglyceridemia in type 2 diabetes.