Role of Pregnane X Receptor in Obesity and Glucose Homeostasis in Male Mice*

Background: PXR is a xenobiotic nuclear receptor that defends against toxic agents. Results: In male mice fed a HFD, the mouse PXR gene promoted obesity, whereas mice lacking the PXR or possessing the human transgene were hyperglycemic. Conclusion: The impact of PXR on HFD-induced obesity and hyperglycemia is species-dependent. Significance: The current data provide in vivo significance of PXR in metabolic syndrome. Clinical obesity is a complex metabolic disorder affecting one in three adults. Recent reports suggest that pregnane X receptor (PXR), a xenobiotic nuclear receptor important for defense against toxic agents and for eliminating drugs and other xenobiotics, may be involved in obesity. Noting differences in ligand specificities between human and mouse PXRs, the role of PXR in high fat diet (HFD)-induced obesity was examined using male PXR-humanized (hPXR) transgenic and PXR-knock-out (PXR-KO) mice in comparison to wild-type (WT) mice. After 16 weeks on either a control diet or HFD, WT mice showed greater weight gain, whereas PXR-KO mice gained less weight due to their resistance to HFD-induced decreases in adipose tissue peroxisome proliferator-activated receptor α and induction of hepatic carnitine palmitoyltransferase 1, suggesting increased energy metabolism. Interestingly, control-fed PXR-KO mice exhibited hepatomegaly, hyperinsulinemia, and hyperleptinemia but hypoadiponectinemia and lower adiponectin receptor R2 mRNA levels relative to WT mice. Evaluation of these biologic indicators in hPXR mice fed a control diet or HFD revealed further differences between the mouse and human receptors. Importantly, although HFD-fed hPXR mice were resistant to HFD-induced obesity, both PXR-KO and hPXR mice exhibited impaired induction of glucokinase involved in glucose utilization and displayed elevated fasting glucose levels and severely impaired glucose tolerance. Moreover, the basal hepatic levels of the gluconeogenic enzyme phosphoenolpyruvate carboxykinase 1 were increased in hPXR mice compared with WT mice. Altogether, although the mouse PXR promotes HFD-induced obesity, the hPXR mouse carries a genetic predisposition for type 2 diabetes and thus provides a model for exploring the role of human PXR in the metabolic syndrome.

Obesity is a complex metabolic disorder that affects every segment of the United States population, with one in three adults defined as clinically obese (1). Although a high caloric diet is a major cause of obesity and insulin resistance, the specific genes that determine sensitivity to dietary obesity remain elusive (2,3). Nuclear receptors are a class of intracellular transcription factors activated by various ligands that regulate multiple metabolic pathways (4). Although several nuclear receptors have been identified as important regulators of lipid metabolism, previous research on obesity has mainly focused on the peroxisome proliferator-activated receptors (PPARs) 2 (5,6).
Although initially characterized as a xenobiotic-sensing nuclear receptor, several lines of evidence now suggest that the nuclear hormone receptor pregnane X receptor (PXR; NR1I2) also plays an important role in lipid, glucose, and energy metabolism (7,8). First, PXR is a sensor that regulates the coordinate expression of genes associated with endobiotic and xenobiotic clearance (9,10). Second, relevant PXR ligands can alter plasma lipid levels in patients, where activated PXR alters expression of genes involved in lipid homeostasis (11,12). Third, PXR is involved in cross-talk with various hormone-responsive transcription factors (e.g. forkhead box O1 (FoxO1), forkhead box A2 (FoxA2), cAMP-response element-binding protein (CREB), and PPAR␥ coactivator 1␣ (PGC-1␣)) to decrease energy metabolism via down-regulation of gluconeogenesis (8,13,14). Fourth, PXR is one of several nuclear receptors responsive to bile acids that regulates cholesterol metabolite toxicity (15,16). Finally, a recent report indicates that PXR gene variants are associated with disease severity in nonalcoholic fatty liver disease, which is a contributor to the metabolic syndrome (17). Although PXR is highly expressed in the liver, a major organ for lipogenesis, fatty acid ␤-oxidation, glucose metabolism, and lipid secretion, it is not known whether PXR-mediated gene regulation plays a role in diet-induced obesity (7,18).
Interestingly, two recent reports on the in vivo role of PXR in obesity induction reached different conclusions (19,20). Although the mouse PXR ligand, pregnenolone 16␣-carbonitrile (PCN) inhibited high fat diet (HFD)-induced obesity in AKR/J mice, Pxr-null mice were resistant to diet-induced obesity, suggesting that both PXR activation and loss of the receptor protect against obesity (19,20). Two independent groups have generated mice that lack PXR leading to changes in the expression of PXR target genes (10,21). The Pxr-null mice established by Xie et al. (21) and used in the aforementioned studies was generated by deleting exons 2 and 3 encoding amino acid residues 63-170 in the PXR DNA binding domain.
Understanding the role of PXR in HFD-induced obesity is further complicated by the existence of a second Pxr-null mouse line (10). Staudinger et al. (10) independently developed these Pxr-null mice by deleting the first coding exon of the Pxr gene, including the translation start site and the first zinc finger of the PXR DNA binding domain (amino acids 1-63). These two lines of Pxr-null mice appear normal when maintained under standard laboratory conditions; however, PCN can no longer up-regulate the expression of cytochrome P450 3a11 (Cyp3a11) mRNA in the livers of either line of these Pxr-null animals (10,21). However, the two lines differ in the basal transcription levels of Cyp3a11 (10). Pxr-null animals developed by Xie et al. (21) match wild-type (WT) Cyp3a11 gene transcription, whereas Staudinger et al. (10) Pxr-null mice possess elevated basal Cyp3a11 mRNA levels (10). The intriguing differences between PXR strains and the seeming contradiction that both PXR activation and loss protect against diet-induced obesity together strongly argue for the need of further detailed mechanistic studies on the mechanism(s) by which this nuclear receptor regulates energy homeostasis.
Valid animal models are a prerequisite for developing new drugs to treat obesity and its associated metabolic disorders. Unfortunately, differences between the mouse and human ligand binding domain sequences result in species-specific responses to ligand activation of human and mouse PXR (22). Thus, chemical ligands that activate human PXR typically have little effect on the mouse form of this receptor and vice versa (23,24). To address these differences experimentally, PXR-humanized (hPXR) transgenic mice were developed to provide a more valid in vivo model of human xenobiotic responses (10,21,25). Although the mouse PXR may promote obesity, the function of human PXR in obesity is unknown (19,24).
To explore the function of the mouse and human PXR genes in obesity, weaned male WT, hPXR transgenic, and the Pxr-null [PXR-knock-out (PXR-KO)] mice developed by Staudinger et al. (10) were placed on a HFD for 16 weeks and analyzed for a variety of features related to obesity, diabetes, and lipid and glucose metabolism. Our studies suggest a marked difference between the mouse and human PXR in the metabolic control of obesity and type 2 diabetes and unique phenotypes for each of the three PXR genotypes. Importantly, both PXR-KO and hPXR mice share a general loss of glucose tolerance that suggests their value as complementary models for the study of type 2 diabetes and related symptoms of metabolic syndrome in conjunction with moderate but not severe obesity.

EXPERIMENTAL PROCEDURES
Animals-Breeding-pairs of adult male and female C57BL/6 mice (which served as WT) were purchased from The Jackson Laboratory (Bar Harbor, ME). Only male mice were used in these studies. The mice were housed (3-5 mice/cage) in polycarbonate cages on racks directly vented via the facility exhaust system at 22°C with a 12/12-h light/dark cycle at the Animal Resources Complex at North Carolina Central University. Breeding pairs of both PXR-KO and hPXR mice on a C57BL/6 background were transferred from the colony housed at the National Cancer Institute (National Institutes of Health, Bethesda, Maryland) (10,25). Age-matched male WT, PXR-KO, and hPXR mice were each randomly separated at weaning (ϳ3 weeks old) into two groups (n ϭ 7-13) for either diet: control diet (12% of calories as fat; catalogue #D03032702) or a HFD (45% of calories from fat; catalogue #D03032705) from Research Diets Inc. (New Brunswick, NJ) for 16 weeks. The detailed diet compositions were previously described except that HFD in this study provided 45% of calories from lard and soybean oil instead of 43% (26). Furthermore, the HFD and control chow diets were matched in the type of fat and other nutrients. The food consumption and the change in body weight were monitored weekly. All procedures, in accordance with the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals, were approved by the North Carolina Central University Institutional Animal Care and Use Committee. After 16 weeks on their respective diets, mice were sacrificed without starvation, and sections of liver, brown adipose tissue (BAT, removed from the back), and white adipose tissue (WAT, comprising epididymal, perirenal, and mesenteric fat) were rapidly removed, weighed, snap-frozen in liquid nitrogen, and stored at Ϫ80°C. A part of the fresh liver tissue was immediately homogenized, and the remaining portion was fixed in 10% formalin for hematoxylin and eosin (H&E) staining. Blood samples collected by cardiac puncture were centrifuged at 3000 rpm for 15 min to collect serum for liver enzyme, triglycerides, glucose, leptin, adiponectin, and insulin analysis.
Serum Alanine Aminotransferase (ALT), Glucose, Leptin, Adiponectin, Insulin, and Triglyceride Measurements-Blood was collected by cardiac puncture, and serum was processed and stored at Ϫ80°C. Serum ALT activity and glucose levels were determined using the Cholestech LDX analyzer (Cholestech Corp., Hayward, CA) (27). Serum leptin and insulin levels were determined using enzyme-linked immunosorbent assay (ELISA) kits according to protocols provided by the kit manufacturers (Crystal Chem Inc., Dowers Grove, IL). Serum adiponectin levels were also determined using ELISA kits (Millipore Corp., Billerica, MA). Serum triglyceride levels were quantified using a triglyceride test kit (Wako Pure Chemical Industries, Richmond, VA).
Hepatic Triglyceride Levels-Total liver lipids were extracted from 100 mg of liver homogenate using methanol and chloro-form as previously described (28). Hepatic triglycerides were quantified using a triglyceride test kit (Wako Pure Chemical Industries).
Glucose Tolerance Test (GTT) and Insulin Tolerance Test (ITT)-GTT and ITT were performed on WT, PXR-KO, and hPXR male mice (n ϭ 4 -6/group) fed the control diet or HFD for 15-16 weeks. Glucose and insulin tolerance tests were conducted on fasting animals (overnight for GTT and 4 h for ITT). For the GTT, a single dose of D-glucose (10% solution in water, 10 ml/kg) was injected intraperitoneally. For the ITT, a single dose of human insulin (HumulinN, Eli Lilly and Co., Indianapolis, IN) (0.75 units/kg, 3 ml/kg in saline) was injected intraperitoneally. Glucose levels were measured from blood collected from the tail vein immediately before and at 15, 30, 60, 90, and 120 min for (GTT) or 15, 30, and 60 min (for ITT) after the injections using Contour TS blood glucose strips (Bayer HealthCare LLC, Mishawaka, IN).
Statistical Analysis-Data are presented as the means Ϯ S.E. (n ϭ 7-13). Statistical analysis was performed using one-way analysis of variance followed by post hoc tests. A p value of Ͻ 0.05 was considered statistically significant.

RESULTS
The Mouse PXR Enhances HFD-induced Obesity Compared with the Human PXR-The role of PXR in the widely studied HFD-induced model of obesity was examined. As expected, body weight differences between mice of the same genotype fed a control diet and those fed a HFD were statistically significant for all groups from week 2 onward (Fig. 1). The growth curve was the same for all the three PXR genotypes on a HFD until after week 8, when differences in rate of body weight gain between WT and the other genotypes (hPXR and PXR-KO mice) emerged. The body weight differences between WT mice fed a HFD and both PXR-KO and hPXR mice fed a HFD were statistically significant from week 10 onward to the end of the 16-week study ( Fig. 1). There were no differences in body weight between PXR-KO and hPXR mice fed the HFD at any time in the 16-week study period ( Fig. 1). At the end of the study period, the HFD-fed WT mice weighed significantly more than both the HFD-fed PXR-KO and hPXR mice fed the same diet ( Fig. 1 and Table 1), indicating that the mouse PXR gene contributes to HFD-induced obesity to a greater extent than the human PXR gene ( Fig. 1 and Table 1). Of importance, there was no significant difference in weight gain between the WT and PXR-KO mice fed control chow diet ( Fig. 1 and Table 1). However, the growth curve was significantly lower between control chow-fed hPXR and both WT (from week 4 onward to the end of week 12) and PXR-KO mice (from week 4 to week 10) (Fig. 1). After week 10, the growth curve was the same between control chow-fed hPXR and PXR-KO mice to end of the 16-week study ( Fig. 1). In agreement with weight gain patterns, food intake was the same for the control chow-fed WT, PXR-KO, and hPXR mice (Table 1). However, although both WT and PXR-KO mice ate more on a HFD than on a control diet, the food intake did not vary with diet among hPXR mice (Table 1). HFD increased BAT and WAT mass in all three genotypes of mice compared with mice fed control diet (Table 1). However, we observed significant differences in WAT mass between WT and PXR-KO mice exposed to HFD (Table 1). PXR deficiency led to hepatomegaly in PXR-KO mice on control diet compared with WT mice or hPXR mice fed a control diet (Table 1). When liver weight is expressed as percent of body weight, a significant increase was found in PXR-KO mice fed the control diet compared with both WT and hPXR mice fed the control diet (Table  1). However, HFD did not significantly alter the liver-to-body weight ratio in the three PXR mouse genotypes compared with control-fed littermates ( Table 1).
The PXR Gene Modulates Lipid Storage-Because the PXR gene is highly expressed in liver and intestine, the effects of PXR levels on liver composition and pathology were examined in response to HFD and control diet (32). Histologic scoring via H&E staining of lipid accumulation showed that control chow diet-fed mice of all three genotypes possessed fewer stored lipid droplets than did their HFD-fed counterparts (Fig. 2). However, the nature and extent of lipid packaging varied. Macrovesicular steatosis, observed as a single vacuole displacing the nucleus in each hepatic cell, was most prominent in WT mice ( Fig. 2, a-c). In contrast, microvesicular steatosis, multiple smaller vacuoles in each hepatic cell without significant nuclear displacement, was more prominent in PXR-KO and hPXR mice, especially those receiving a HFD ( Fig. 2, a--c) (33). However, inflammation, necrosis, and fibrosis were absent in all three HFD-fed genotypes. All three genotypes showed similar increases in hepatic triglyceride levels on a HFD, whereas basal triglyceride levels were elevated in control-fed PXR-KO mice (Fig. 2d).
The Mouse PXR Gene Protects against HFD-induced Hyperglycemia-Both fasting and fed glucose levels did not vary with diet among WT mice (Fig. 3, a and b). However, HFD feeding increased fasting blood glucose levels in both PXR-KO and hPXR mice by 64 and 44%, respectively (Fig. 3, a and b). Furthermore, in the fed state, serum glucose levels were elevated by HFD in both PXR-KO and hPXR mice by 19 and 43%, respectively (Fig. 3b). Compared with control-fed animals, serum insulin levels in HFD-fed animals increased by 13.7-, 7.6-, and 10.6-fold in WT, PXR-KO, and hPXR mice, respectively (Fig. 3c). As reported previously, basal insulin levels were elevated (1.8-fold) in PXR-KO mice compared with WT mice (Fig. 3c) (8). It was also observed for the first time that serum insulin levels in control chow-fed hPXR mice were elevated (1.9-fold).
PXR Deficiency Leads to Decreased Basal Serum Adiponectin but Increased Leptin Levels-Serum levels of ALT, a marker of liver injury, were increased after HFD ingestion by 3.1-, 1.9-, and 1.9-fold in WT, PXR-KO, and hPXR mice, respectively (Fig.  4a). Whereas liver injury occurred in all three genotypes of mice fed a HFD, a differential hepatotoxic phenotype was observed. PXR-KO mice are more resistant to HFD-induced liver injury than WT mice (Fig. 4a). Serum triglyceride levels were elevated (42%) only in HFD-fed hPXR mice (Fig. 4b). Our findings that the PXR gene influences circulating insulin levels suggest the possibility that PXR also influences serum adipokine levels. Increasing dietary fat in rodents is known to enhance adiposity and elevate plasma leptin levels (34,35). We noted that HFD promoted hyperleptinemia in all three PXR mouse genotypes; specifically, HFD-feeding increased leptin levels in WT (4.7-fold), PXR-KO (2.4-fold), and hPXR (2.2-fold) FIGURE 1. Body weights of mice during a 16-week feeding of a control diet or a HFD. Male WT, Pxr-null (PXR-KO), and PXR-humanized (hPXR) transgenic mice were fed a control or a HFD for 16 weeks. Body weight was assessed weekly. Data represent the mean Ϯ S.E. (n ϭ 7-13). * P Ͻ 0.05 between WT mice fed control diet and hPXR mice fed the control diet. # P Ͻ 0.05 between mice fed control and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO mice fed a HFD. † P Ͻ 0.05 between WT mice fed a HFD and hPXR mice fed a HFD.
above the levels found in control-fed mice (Fig. 4c). Basal serum leptin levels were higher in control-fed PXR-KO mice (2.7-fold) and hPXR mice (2.3-fold) than WT mice on control diets, recapitulating the serum insulin patterns we observed (Figs. 3c and 4c). These data indicate that even on a control diet, PXR-KO and hPXR mice are modestly insulin-and leptin-resistant as compared with their WT littermates. The elevated basal serum leptin levels in both PXR-KO and hPXR mice fed the control diet suggest that the mouse PXR indirectly suppresses leptin secretion in adipose tissue. In contrast, basal serum adiponectin levels were decreased in control-fed PXR-KO (26%) compared with WT mice fed a control diet (Fig. 4d), and adiponectin levels were further decreased when these mice were placed on a HFD (Fig. 4d). Although basal serum adiponectin levels were not significantly different between WT and hPXR mice fed a control diet, like PXR-KO mice, HFD feeding led to reduced adiponectin levels in hPXR animals (74% compared with hPXR controls) but not in HFD-fed WT mice (Fig. 4d).  , macrovesicular steatosis score (c), and hepatic triglyceride levels (d) from different treatment groups were performed as described under "Experimental Procedures." HFD feeding produced both macrovesicular (indicated by arrows) and microvesicular steatosis (indicated by arrowheads) only in the WT mice. The pattern of steatosis was mainly microvesicular steatosis in hPXR and PXR-KO mice fed a HFD. However, inflammation, necrosis, and fibrosis were absent in all three genotypes fed the HFD. Data represent the means Ϯ S.E. (n ϭ 4 -6). * P Ͻ 0.05 between WT mice fed control diet and PXR-KO mice fed the control diet. # P Ͻ 0.05 between mice fed control diet and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO or hPXR mice fed a HFD.
Glucose Tolerance Is Impaired in Both HFD-fed PXR-KO and hPXR Mice-Impaired glucose tolerance is the earliest identifiable metabolic abnormality in the pathogenesis of type 2 diabetes (36). To establish whether glucose clearance was influenced by altered PXR expression, intraperitoneal glucose tolerance tests were performed among the three control-and HFD-fed PXR genotypes. Serum glucose levels peaked at 30 min after glucose challenge for all three PXR genotypes fed control chow (Fig. 5). WT mice (Fig. 5a) fed a HFD cleared systemic glucose spikes more efficiently and more rapidly than their PXR-KO (Fig. 5c) and hPXR (Fig. 5e) counterparts. Glucose elimination was severely impaired in both HFD-fed PXR-KO and hPXR mice, and the total systemic glucose (represented by the area under the curve (AUC) significantly increased by 125 and 107%, respectively, as the result of excess dietary lipids, indicative of glucose intolerance (Fig. 5g). The AUC in HFD-fed WT mice increased by only 28% compared with WT mice on the control diet (Fig. 5g). Because of the marked glucose intolerance in HFD-fed PXR-KO and hPXR mice (Fig. 5, c and e), we evaluated the impact of the PXR genotypes on insulin capacity to lower blood glucose levels using the intraperitoneal insulin tolerance test (Fig. 5, b, d, and f). Insulin administered to HFD-fed WT mice was only slightly impaired in its ability to decrease blood glucose levels with an AUC increase of 34% (Fig. 5h). HFD-fed hPXR mice exhibited an even greater decrease in insulin sensitivity, with an increase in AUC of 64% compared with littermates fed the control diet (Fig. 5, f and h). Unexpectedly, there was no significant difference in insulin sensitivity between control and HFD-fed PXR-KO mice (Fig. 5, d and h). Consistent with this, the ITT AUC for HFD-fed PXR-KO mice was significantly lower than that for both HFD-fed WT and hPXR mice (Fig. 5h).
The Pepck1 Gene and Protein Are Constitutively Activated in hPXR Mice-In the current study hyperglycemia occurred in the fed and fasted states in both HFD-fed PXR-KO and hPXR mice, suggesting an inability to down-regulate gluconeogenesis. To explore this possibility, hepatic levels of gluconeogenic and glycolytic enzymes were examined using real-time PCR and Western blot analysis (37). In comparison with WT mice, the gene and protein levels of PEPCK1, the key gluconeogenic enzyme in control and HFD-fed PXR-KO animals, were slightly but not significantly elevated (Fig. 6, a and e) (37). A similar but more statistically significant relationship was observed in comparing control-fed WT and hPXR mice; constitutive Pepck1 mRNA and protein levels were elevated 2.4-and 1.6-fold, respectively (Fig. 6, b and f). In contrast, hepatic Pepck1 protein but not mRNA expression was up-regulated by HFD only in WT mice (Fig. 6, a, b, e, and f).
The basal protein levels of G6Pase, also involved in gluconeogenesis, were increased in control chow-fed PXR-KO and hPXR mice by 1.5-and 1.7-fold, respectively; however, it was not statistically significant compared with control chow-fed WT mice (Fig. 6, g and h) (37). HFD significantly increased G6Pase protein levels in both WT and hPXR mice, with a somewhat larger expression levels in the hPXR animals ( Fig. 6g and  h). In contrast, hepatic G6Pase was uninducible in HFD-fed PXR-KO mice (Fig. 6, g and h). The basal glucokinase mRNA, but not the protein levels, were lower in hPXR (53%) mice compared with control chow-fed WT mice (Fig. 6c). Although the HFD diet increased hepatic glucokinase protein levels by 36 -56% in WT mice, glucokinase protein expression was inhibited in HFD-fed PXR-KO and hPXR mice by 47 and 38%, respectively, leading to dramatic differences in glucokinase protein levels among the three PXR mouse strains fed a HFD (Fig. 6, i and j).
Basal Expression of Ppar␥ and Cyp4a14 mRNA Is Reduced in Both PXR-KO and hPXR Mice but Highly Inducible by HFD-PPARs are ligand-activated transcription factors that primarily regulate genes involved in lipid metabolism, adipocyte differentiation, and insulin action (38). The composition of PPAR isoforms may vary in fatty livers as previously reported (38). Given that this study reports hyperinsulinemia and liver injury among all HFD-fed PXR genotypes (Figs. 3c and 4a), we explored , and PXR-humanized (hPXR) transgenic mice were fed control diet (white) or a HFD (black) for 16 weeks. Serum levels of glucose and insulin in the fasted and fed state were determined as described under "Experimental Procedures." Data represent the means Ϯ S.E. (n ϭ 5-9). * P Ͻ 0.05 between WT mice fed control diet and PXR-KO or hPXR mice fed the control diet. # P Ͻ 0.05 between mice fed control diet and HFD. † P Ͻ 0.05 between WT or PXR-KO mice fed a HFD and hPXR mice fed a HFD.

Male PXR in Obesity and Type 2 Diabetes
FEBRUARY 7, 2014 • VOLUME 289 • NUMBER 6 changes in the expression of Ppar␣, Ppar␥, and select target genes due to PXR genotype and diet. Basal Ppar␥ mRNA levels were significantly lower in both PXR-KO (36%) and hPXR (38%) mice compared with WT mice (Fig. 7, a and b). Although HFD ingestion did not have any significant effect on the Ppar␥ mRNA levels in WT mice, it increased Ppar␥ mRNA levels in the livers of PXR-KO (5-fold) and hPXR (3.3-fold) mice (Fig. 7,  a and b).
Although HFD exposure modestly reduced hepatic Cyp2e1 mRNA levels in all genotypes, dramatic differences in hepatic mRNA and protein expression of Cyp414, a Ppar␣ target gene, were observed (39). First, on a control diet, only WT mice expressed robust Cyp4a14 mRNA (Fig. 7, c and d) (mRNA in PXR-KO and hPXR was 8 and 2% of WT Cyp4a14 mRNA, respectively) even though protein expression across all PXR genotypes was relatively low. HFD spiked CYP4A14 protein expression in all genotypes (WT 4.2-fold, PXR-KO 3.0-fold, and hPXR 2.1-fold; Fig. 7, m and n). HFD also induced robust Cyp4a14 mRNA up-regulation in both PXR-KO (9-fold) and hPXR (9-fold) mice (Fig. 7, c, d, m, and n).
The basal hepatic Ppar␣ mRNA levels did not vary among WT, PXR-KO, and hPXR mice fed the control chow and were not altered by HFD (Fig. 7, g and h). Similarly, basal Cpt-1 mRNA levels did not vary among the three PXR genotypes (Fig.  7, i and j). However, although HFD had no effect on Cpt-1 gene expression in WT and hPXR mice, HFD induced a significant increase (79%) in Cpt-1 mRNA levels in PXR-KO mice (Fig. 7i). Furthermore, the constitutive Acox-1 mRNA levels were lower in both PXR-KO (47%) and hPXR (49%) mice compared with WT mice (Fig. 7, k and l); however, Acox-1 mRNA levels in all three PXR genotypes were unaffected by HFD (Fig. 7, k and l). Somewhat surprisingly, real-time quantitative PCR analysis in our study showed neither suppression nor induction of Ppar␣ by PXR deficiency or HFD (Fig. 7, g and h). However, basal Ppar␥ mRNA levels, like Cyp4a14, were reduced in both PXR-KO and hPXR mice that could be up-regulated by HFD. Therefore, it seems that basal Cyp4a14 gene expression and up-regulation by HFD is associated with PPAR␥, not PPAR␣ (Fig. 7, a-h) (39).
Both PXR Deficiency and hPXR Gene Lead to Reduced Hepatic AdipoR2 mRNA Levels and Enhanced AMPK Activation-The current data show that although serum adiponectin levels remained unchanged in HFD-fed WT mice, they were markedly decreased in both PXR-KO and hPXR mice fed a HFD, suggesting the involvement of hypoadiponectinemia in these mice (Fig. 4d). Furthermore, basal serum adiponectin levels were lower in control-fed PXR-KO compared with WT mice fed a control diet (Fig. 4d). Adiponectin mediates its effects by binding to two receptors, AdipoRI and AdipoR2 (40). We examined the contribution of diet and PXR genotype to the gene expression of the two receptors AdipoRI and AdipoR2 and to the activation of AMPK␣, known to be closely involved in the regulation of adiponectin expression and function. Quantitative real-time PCR indicated no significant differences in Adi-poR1 mRNA levels among control-fed PXR genotypes, whereas hPXR mice showed a slight reduction in AdipoR1 mRNA expression in response to HFD (Fig. 8, a and b). Surprisingly, the control-and HFD-fed AdipoR2 mRNA levels were significantly lower in both PXR-KO (34% of WT on control diet) and hPXR mice (41% of WT on control diet) (Fig. 8, c and d).
AMPK␣ is a serine/threonine kinase and functions as a metabolic regulator that phosphorylates a variety of enzyme-associated energy metabolism and promotes insulin sensitivity (41). We hypothesized that the hyperglycemia and severely impaired glucose tolerance observed in both HFD-fed PXR-KO and

. Blood glucose levels during GTTs and ITTs in male WT, Pxr-null (PXR-KO), and hPXR transgenic mice fed control diet or a HFD for 16 weeks.
GTT and ITT were performed after mice (WT, PXR-KO, and hPXR) had been on the control diet or HFD for 15 and 16 weeks, respectively (n ϭ 4 -6/group). Glucose and insulin tolerance tests were conducted on fasting animals (overnight for GTT and 4 h for ITT). For the GTT, a single dose of D-glucose (10% solution in water, 10 ml/kg) was injected intraperitoneally. For the ITT, a single dose of human insulin (HumulinாN, Eli Lilly) (0.75 units/kg, 3 ml/kg in saline) was injected intraperitoneally. Glucose levels were measured from blood collected from the tail vein immediately before and at 15, 30, 60, 90, and 120 min (for GTT; a, c, and e) or 15, 30, and 60 min (for ITT) after the injections using Contour TS blood glucose test strips (Bayer HealthCare). Blood glucose levels for the ITT were normalized to those at 0 min in each group (100%) (b, d, and f). AUC for the GTT (g) and ITT (h) were calculated using Sigma Plot 12.0 (Systat Software Inc, San Jose, CA). # P Ͻ 0.05 between mice fed control diet and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO or hPXR mice fed a HFD. IPGTT, intraperitoneal glucose tolerance test; IPITT, intraperitoneal insulin tolerance test. hPXR mice was the result of altered AMPK activation. Therefore, levels of active (phosphorylated Thr-172) hepatic AMPK␣ were monitored in the three mouse PXR genotypes on both control and HFD diets. We found that AMPK␣ in both PXR-KO and hPXR mice was constitutively activated, unlike in WT mice in which AMPK␣ was activated by HFD alone (Fig. 8, e and f). Given that we observed higher basal levels of serum leptin in both PXR-KO and hPXR mice (Fig. 4d), it is interesting to note that recent reports suggest that some of the metabolic effects of leptin may be mediated through activation of AMPK (42). It is not known whether the increase in phosphorylated AMPK we observed in the livers of both control-fed PXR-KO and hPXR mice is directly related to the high serum leptin levels also noted in these mice.

PXR-KO Mice Are Resistant to HFD-induced Decreases in WAT GLUT4
Protein Expression-Impaired glucose disposal in peripheral tissues is linked to abnormal insulin signaling. GLUT4 is selectively expressed in insulin target tissues such as WAT and is primarily responsible for the marked increase in glucose disposal after insulin administration (43,44). Because this study noted normal reduction in blood glucose levels in response to insulin injection in HFD-fed PXR-KO mice compared with both WT and hPXR mice fed the HFD (Fig. 5, d and  h), we used an immunoblot analysis to test whether GLUT4 expression could explain this anomaly. Specifically, GLUT4 protein expression was measured in WAT from all three PXR genotypes fed HFD and control diets. HFD reduced GLUT4 expression in WT (34%) and hPXR (37%) mice relative to GLUT4 levels in their control chow-fed littermates (Fig. 9, a  and b). In contrast, HFD had no effect on GLUT4 expression in PXR-KO mice (Fig. 9a). Constitutive expression of the WAT GLUT4 protein was slightly, but significantly decreased in control chow-fed PXR-KO (14%) compared with WT mice (Fig.  9a). AMPK␣ (phospho-Thr-172) in WAT did not vary among genotypes or diets except for a slight reduction in active AMPK␣ in HFD-fed hPXR mice (Fig. 9, c and d).
PXR-KO Mice Are Resistant to HFD-induced Decreases in Adipose Tissue-PPAR␣ Levels-PPAR␣ is a ligand-activated transcription factor that binds fatty acids and activates the transcription of genes that regulate lipid metabolism (45). Therefore, we examined the association between PXR genotype, diet, and adipose (WAT and BAT) PPAR␣ protein expression. In control-fed animals, WAT PPAR␣ protein expression did not vary among the three PXR genotypes (Fig. 10, a and b). HFD down-regulated WAT PPAR␣ protein in WT (ϳ47%) and hPXR (51%) but not PXR-KO mice (Fig. 10, a and b). These findings suggest a role for both mouse and human PXR in HFDinduced inhibition of WAT PPAR␣ expression (Fig. 10a). Furthermore, although there was no difference in WAT PPAR␥ and C/EBP␣ protein expression among control-fed animals (Fig. 10, a and b), HFD feeding decreased WAT PPAR␥ and C/EBP␣ protein levels in all three PXR genotypes compared with their respective controls (Fig. 10, a and b).
The basal expression of PPAR␣ protein in BAT was somewhat higher in PXR-KO (71%) compared with WT mice (Fig.  10c), whereas the BAT PPAR␣ protein levels in hPXR mice was slightly lower (76% compared with control WT mice) (Fig. 10d). HFD ingestion significantly decreased BAT PPAR␣ protein levels only in WT mice by ϳ46% (Fig. 10, c and d). After HFD, the BAT PPAR␣ protein levels in both PXR-KO and hPXR mice were significantly higher than in WT mice (Fig. 10, c and d). In rodents and human infants, a major defense mechanism against obesity is BAT, which is involved in energy dissipation, and the protein responsible for BAT thermogenesis is UCP1 (46). Western blot analysis was used to examine protein expression of UCP-1. The basal expression of UCP1 protein in BAT was increased in both PXR-KO (19%) and hPXR (41%) mice compared with WT mice on control chow diet; however, these increases were not statistically significant (Fig. 10, e and f). In both WT and PXR-KO mice HFD induced UCP1 protein by 40 and 78%, respectively (Fig. 10e), whereas the UCP1 protein in hPXR mice was unaffected by the HFD (Fig. 10f).

DISCUSSION
Several observations in this study suggest that the PXR gene plays a role in the induction of human obesity, the suppression of hyperglycemia, and the development of type 2 diabetes. The following findings were found in PXR-KO mice that lack this receptor. 1) Similar to WT mice, they were hyperphagic but gained less weight and WAT when fed a HFD. 2) They develop less HFD-induced liver injury (evidenced by serum ALT levels) even though livers in HFD-fed PXR-KO mice were significantly larger than their WT or hPXR counterparts. 3) They have higher control levels of serum insulin and leptin but lower serum adiponectin and AdipoR2 mRNA levels. 4) They accumulate higher levels of hepatic triglycerides (also previously reported (8). 5) They have increases in HFD-induced hepatic Cpt-1 mRNA levels. 6) They fail to induce glucokinase and exhibit fed and fasting hyperglycemia as well as severely impaired glucose tolerance when given a HFD. Taken together, these observations suggest that the mouse PXR gene directly or indirectly promotes weight gain and uncouples diet-induced obesity from type 2 diabetes.
Previous evidence indicates that PXR is capable of downregulating the expression of gluconeogenic genes (7,13,14). Furthermore, previous reports revealed that treatment with the mouse PXR activator PCN decreased blood glucose levels in fasting WT mice but not in PXR-KO mice, suggesting that PXR  (a and b) and glucokinase (c and d) by real-time PCR using the SensiFast SYBR Hi-ROX mix as described in under "Experimental Procedures." Liver homogenate (40 g/lane) from mice fed control diet or a HFD were electrophoresed on 10% SDS-polyacrylamide gels, transferred to PVDF membranes, and incubated with anti-PEPCK1, anti-G6Pase, and anti-glucokinase antibody as indicated in representative blots (e-j). Quantitative analysis of PEPCK1 protein after normalizing with ␣-tubulin (e and f), G6Pase protein after normalizing with ␣-tubulin (g and h), and glucokinase protein after normalizing with ␣-tubulin (i and j) in WT, PXR-KO, and hPXR mice fed control diet (white) or a HFD (black) for 16 weeks. The intensities of the bands were quantified using the AlphaImager 2200 Documentation and Analysis System Software (Alpha Innotech). Data represent the mean Ϯ S.E. (n ϭ 3-6). * P Ͻ 0.05 between WT mice fed control diet and PXR-KO or hPXR mice fed the control diet. # P Ͻ 0.05 between mice fed control and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO or hPXR mice fed a HFD.

Male PXR in Obesity and Type 2 Diabetes
FEBRUARY 7, 2014 • VOLUME 289 • NUMBER 6 may serve as a therapeutic target for type 2 diabetes and its medical complications (8). In the current study all three HFDfed PXR genotypes displayed glucose intolerance; however, the magnitude and time course varied. Most striking was the severe glucose intolerance and hyperglycemia observed in both HFDfed PXR-KO and hPXR mice, unlike the normoglycemia and moderate glucose intolerance of their WT counterparts. Although the basal levels of both mRNA and protein of the key gluconeogenic enzyme Pepck1 were significantly increased in the livers of chow-fed hPXR mice, HFD ingestion did not further increase Pepck1 expression. However, protein levels of G6Pase, which produces glucose from glucose 6-phosphate, allowing its release from the liver into circulation, were signif-icantly increased in HFD-fed hPXR mice; the contrast in G6Pase expression between control and HFD was especially marked in hPXR mice compared with WT mice (37). Unexpectedly, neither PXR deficiency nor HFD significantly modulated the Pepck1 and G6Pase gene and/or protein in PXR-KO  Hepatic PPAR␥, cytochrome P450 4a14 (Cyp4a14), Cyp2e1, PPAR␣, Cpt-1, Acox-1 mRNA, and immunoblot analysis of hepatic CYP4A14 and CYP2E1 protein levels in male WT, Pxr-null (PXR-KO), and hPXR transgenic mice fed control diet (white) or a HFD (black) for 16 weeks. Total hepatic RNA was isolated, and mRNA levels were determined by real-time PCR using the SensiFast SYBR Hi-ROX mix as described under "Experimental Procedures." Liver homogenate (40 g/lane) from mice fed control diet or a HFD were electrophoresed on 10% SDS-polyacrylamide gels, transferred to PVDF membranes, and incubated with anti-CYP4A14 or anti-CYP2E1 antibody as indicated in representative blots (m and n). Quantitative analysis of CYP4A14 protein levels plotted after normalizing with ␣-tubulin (m and n) in WT, PXR-KO, and hPXR mice. The intensities of the bands were quantified using the AlphaImager 2200 Documentation and Analysis System Software (Alpha Innotech). Data represent the mean Ϯ S.E. (n ϭ 3-6). * P Ͻ 0.05 between WT mice fed control diet and PXR-KO or hPXR mice fed the control diet. # P Ͻ 0.05 between mice fed control and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO or hPXR mice fed a HFD.  (a and b) and AdipoR2 (c and d) by real-time PCR using the SensiFast SYBR Hi-ROX mix as described under "Experimental Procedures." Liver homogenate (40 g/lane) from mice fed control diet or a HFD were electrophoresed on 10% SDS-polyacrylamide gels, transferred to PVDF membranes, and incubated with anti-phosphorylated-unphosphorylated forms of AMPK␣ antibody as indicated in representative blots (e and f). Quantitative analysis of phosphorylated-AMPK␣ (Thr-172) protein levels were plotted after normalizing with AMPK␣ (e and f) in WT, PXR-KO, and hPXR mice. The intensities of the bands were quantified using the AlphaImager 2200 Documentation and Analysis System Software (Alpha Innotech). Data represent the mean Ϯ S.E. (n ϭ 3-6). * P Ͻ 0.05 between WT mice fed control diet and PXR-KO or hPXR mice fed the control diet. # P Ͻ 0.05 between mice fed control and HFD. ‡ P Ͻ 0.05 between WT mice fed a HFD and PXR-KO or hPXR mice fed a HFD. mice. In the liver, glucokinase catalyzes a reaction that counters G6Pase, the phosphorylation of glucose to produce glucose-6phosphate. Glucokinase activity determines the rate of glucose utilization and glycogen synthesis, and glucokinase-deficient mice die with severe hyperglycemia (47,48). In this study HFD increased glucokinase protein but not gene expression in WT mice. However, HFD-induced increases in glucokinase protein levels were inhibited in both PXR-KO and hPXR mice. Thus, it is speculated that the hyperglycemia and impaired glucose tolerance seen in both HFD-fed PXR-KO and hPXR mice was due to inhibition of glucokinase induction leading to reduced glucose catabolism and disposal in the livers of these mice. To explore additional mechanisms that might account for the hyperglycemia and severe glucose intolerance observed in both HFD-fed PXR-KO and hPXR mice, levels of adiponectin and its receptors in liver were examined.
Reduce expression of AdipoR1 and AdipoR2 in obesity correlates with secretion of adiponectin (40). In this study we noted a decrease in the basal circulating levels of adiponectin and in the mRNA levels of the liver-specific AdipoR2 in PXR-KO mice, consistent with the HFD-induced hyperglycemia and severely impaired glucose tolerance found in PXR-KO mice. However, when ITTs were performed, HFD-fed PXR-KO mice showed a normal reduction in blood glucose in response to insulin injection, suggesting that peripheral glucose uptake is not compromised in these mice. We also noted that HFD-fed WT and hPXR, but not PXR-KO mice, expressed reduced levels of WAT GLUT4, the insulin-responsive glucose transporter. Previous reports indicate that visceral fat accumulation is a major factor that determines insulin resistance (49). In this study, HFD-fed PXR-KO mice accumulated less WAT compared with HFD-fed WT mice, which may correlate with improved insulin sensitivity in the PXR-KO mice. HFD also reduced WAT GLUT4 protein levels to the same extent in both WT and hPXR mice, suggesting that peripheral insulin resistance may be similar in these two strains of mice. Although hepatic and peripheral insulin resistance is tightly associated, dissociation between the two has been reported (50,51). Therefore, the severe hepatic glucose intolerance but normal ITT observed in HFD-fed PXR-KO mice is consistent with a slightly improved glucose transport by GLUT4 in WAT coupled with a leaner phenotype based on low visceral fat accumulation exhibited by the PXR-KO mice.
Adiponectin is a signaling protein secreted by adipose tissue and transcribed by adipocyte transcription factors such as PPAR␥ and C/EBP␣ (52). We sought to understand the mech- anisms by which PXR deficiency leads to decreased serum adiponectin levels in PXR-KO mice. Immunoblot analysis of WAT PPAR␥ and C/EBP␣ revealed that the basal levels of these transcription factors did not differ between WT and PXR-KO mice. Given the fact that the PXR gene is not expressed in WAT, the mechanism by which endogenous PXR regulates adipokines such as leptin and adiponectin is important and remains to be determined. Taken together, these data indicate that decreased serum adiponectin levels and AdipoR2 gene expression as well as inhibition of glucokinase induction, but not induction of PEPCK1 or G6Pase, contribute to the hyperglycemia and glucose intolerance observed in HFD-fed PXR-KO mice.
In the liver activation of AMPK results in decreased levels of plasma glucose and enhanced fatty acid ␤-oxidation (53,54). In this study HFD ingestion activated hepatic AMPK only in WT mice. In contrast, activation of AMPK in WAT and liver was reduced in HFD-fed hPXR mice compared with control chowfed hPXR mice. Similar to PXR-KO mice, basal AdipoR2 mRNA levels were lower in hPXR mice compared with WT mice. Furthermore, HFD decreased WAT GLUT4 protein levels, hepatic mRNA levels of both AdipoR1 and AdipoR2, and serum adiponectin levels in hPXR mice. Together, these data suggests that insulin resistance, hyperglycemia, and impaired glucose tolerance in HFD-fed hPXR mice can be largely accounted for by 1) overexpression of PEPCK1, 2) marked induction of G6Pase protein, 3) inhibition of glucokinase induction, 4) decreased GLUT4 protein, and 5) lack of AMPK activation.
The current study demonstrating that mouse PXR promotes obesity is consistent with a recent report by He et al. (19), who employed a different genetic strain of PXR-KO mouse. Together, both studies indicate that ablation of the PXR gene protects against obesity. However, in the animal model used by He et al. (19), loss of PXR inhibited HFD-induced type-2 diabetes and hepatic steatosis, which is contrary to the pro-diabetic effect of PXR ablation observed in the line of PXR-KO mice used in the current study (10). In our study increased hyperglycemia and impaired glucose tolerance were supported by elevated fasting and fed serum glucose levels and GTT. Hepatic lipid accumulation in our HFD-fed PXR-KO mice were also confirmed by H&E staining of liver sections and hepatic triglyceride levels. The reason for the discrepancy in HFD-induced type 2 diabetes and hepatic steatosis between our study and the findings of He et al. (19) is not known. It is possible that it could be due to the location of the deletion in the PXR gene or differences in the genetic backgrounds of the two strains of PXR-KO mice. It should be noted that our data are consistent with previous reports that the PXR gene can repress transcription factors involved in gluconeogenesis, thereby inhibiting type 2 diabetes (7,13,14). Although Pepck1 expression was not significantly increased, factors that promote glucose intolerance and type 2 diabetes including steatosis, inhibition of glucokinase induction, and low AdipoR2 mRNA levels were present in our HFD-fed PXR-KO mouse model.
Our current findings that the mouse PXR promotes obesity is contrary to a recent report in which the mouse PXR ligand, PCN, was found to protect against HFD-induced obesity in AKR/J mice (20). PXR exhibits tissue-specific expression patterns and plays opposing roles in cellular defenses and lipid synthesis (7,23). Reports have shown that PXR ligands can up-regulate both lipogenic (PPAR␥) and antilipogenic (insulininduced gene-1) proteins in liver (55,56). In addition, intestinal activation of PXR has cross-tissue effects and increases hepatic steatosis (57). Interpretation of PXR ligand-induced obesity is further complicated by the fact that PXR ligand treatment may trigger multiple receptors leading to additional transcriptional consequences compared with the Pxr-null animals (10,58,59). Thus, the discrepancy between our current results and those of Ma and Liu (20) may be due to their intraperitoneal administration of PCN, the mouse strain origin of the PXR receptor, and the AKR/J mouse genetic background. Further studies will be needed to conclusively determine whether activation of PXR protect against or aggravate obesity, lipogenesis, and insulin resistance using WT and PXR-KO mice fed a HFD supplemented with the rodent-specific PXR ligand, PCN.
We are the first to characterize the effects of a HFD on mice expressing the human PXR. Prior studies established distinct differences between the mouse and human PXRs in their response to ligand activation even though they display similar tissue expression patterns (23,25). HFD-fed hPXR mice resembled their WT counterparts in that both genotypes have lower WAT PPAR␣ protein levels than HFD-fed PXR-KO mice and fail to induce hepatic Cpt-1 gene expression. However, we noted that many features of this HFD-induced weight gain differed dramatically between WT and hPXR mice.
Unlike mouse PXR, hPXR suppresses HFD-induced obesity. However, significant resistance to HFD-induced weight gain in hPXR mice only occurred after 8 weeks of HFD exposure. One explanation for this pattern might be the involvement of steroid hormones or steroid hormone receptors as the male mice approach puberty. A major difference between WT and hPXR mice could be energy intake; unlike both WT and PXR-KO mice, HFD-fed hPXR mice did not increase their food consumption. Furthermore, the lower growth curve observed in HFD-fed hPXR mice may be associated with the higher basal serum leptin levels in the hPXR mice compared with control chow-fed WT mice. A relationship between basal leptin levels and the response of leptin to dietary fat has been suggested as a predictor of development of subsequent obesity (60,61). Leptin increases outflow from the sympathetic nervous system to brown fat leading to diet-induced thermogenesis, increases metabolic rate, and dissipates excess caloric intake as heat (62,63). Indeed, in the current study basal protein levels of UCP1, expressed exclusively in BAT, were constitutively induced in hPXR mice (46). Thus, the high BAT UCP1 expression will promote thermogenesis and increase energy dissipation, which may contribute to less body weight gain in both control chow and HFD-fed hPXR mice compared with the WT mice.
Adipocytes are the primary site for energy storage and accumulation of triglycerides during ingestion of excess energy. Reports indicate that activation of PXR is also associated with suppression of several PPAR␣ target genes involved in fatty acid ␤-oxidation and elevation of energy expenditure (45). In this study decreases in the protein levels of PPAR␣ in both WAT and BAT were observed in the HFD-fed WT mice; in contrast, in PXR-KO mice, WAT PPAR␣ levels remained unchanged but were up-regulated in BAT in HFD-fed PXR-KO mice. PPAR␣ protein levels in HFD-fed hPXR mice were significantly reduced in WAT but not in BAT. Furthermore, hepatic Cpt-1 mRNA levels were up-regulated in PXR-KO mice fed the HFD but not in either WT or hPXR mice. CPT-1 controls the rate of mitochondrial ␤-oxidation and regulates the deposition or oxidation of fatty acids in the liver (64). Thus, failure to up-regulate expression of PPAR␣ transcriptional programs, which drive the transcription of genes involved in peroxisomal and mitochondrial oxidation of fatty acids, could contribute to lipid accumulation and obesity in HFD-fed WT and hPXR mice. Conversely, induction of liver Cpt-1 mRNA levels and high BAT PPAR␣ protein levels in PXR-KO would increase fatty acid ␤-oxidation and energy expenditure, thereby suppressing body weight gain in PXR-KO mice fed the HFD as seen in the present study.
The morphology pattern of steatosis provides insights to the severity of liver dysfunction (65). Macrovesicular steatosis is most common and occurs in nonalcoholic fatty liver disease (66). In contrast, microvesicular steatosis is generally a more severe form of steatosis that often occurs from defects in fatty acid ␤-oxidation found in a variety of conditions such as toxic-ity of several medications and alcohol ingestion (67,68). The molecular mechanisms responsible for steatohepatitis are poorly understood, but the more common macrovesicular formation found in HFD-fed WT animals should not be as severe as HFD-driven formation of microvesicular steatosis in hPXR and PXR-KO mice. This result is somewhat surprising in view of the more modest HFD-induced increase in serum ALT levels in these two PXR genotypes as compared with WT, thus suggesting milder hepatotoxicity. ACOX-1 is the first and the ratelimiting enzyme involved in peroxisomal ␤-oxidation of very long chain fatty acids (69). Disruption of the Acox-1 gene in mice results in the development of severe microvesicular hepatic steatosis (70). Interestingly, in this study hepatic Acox-1 mRNA levels were markedly reduced in both PXR-KO and hPXR mice fed either control chow or HFD consistent with our observations of microvesicular steatosis in these two mice strains. In addition, the protein levels of CYP4A14, which metabolizes long chain fatty acids to toxic dicarboxylic acids that serve as substrates for peroxisomal ␤-oxidation, were elevated in all the three HFD-fed mouse genotypes (71). Thus, it is possible that the microvesicular steatosis observed in both HFD-fed PXR-KO and hPXR mice is related to the accumulation of un-metabolized dicarboxylic acid due to reduced ACOX-1 peroxisomal ␤-oxidation leading to mitochondrial damage and steatosis. Together, our data indicate that human PXR is involved in diet-induced hepatotoxicity possibly through CYP4A14-mediated microsomal fatty acid metabolism.
In conclusion, PXR deficiency and hPXR expression in a C57BL/6 background has allowed us to generate a diet-induced model for type 2 diabetes. When subjected to a HFD, both PXR-KO and hPXR mice exhibit moderate obesity but possess elevated glucose, insulin, and leptin levels and display glucose intolerance and hypoadiponectinemia; these are the attributes of type 2 diabetes. As such, HFD-fed PXR-KO and hPXR mice should be valuable and clinically relevant addition to the existing in vivo models of type 2 diabetes. Although there are clear differences between the molecular pathological events occurring from mouse and human PXR-driven HFD-induced obesities, the interplay between both WT and hPXRs and PPAR␣ remains intact in both strains. Therefore, this model provides a useful starting point for further exploration of the molecular mechanisms of type 2 diabetes. The current data provide in vivo significance of PXR in metabolic syndrome and indicate that PXR gene could be a therapeutic target for the prevention of metabolic diseases.