The metabolic syndrome is a constellation of findings, including central obesity, insulin resistance, dyslipidemia, hypertension, and hepatic steatosis, which predispose to diabetes, cardiovascular disease, and cancer. As the prevalence of diabetes, obesity, and the metabolic syndrome reach staggering proportions, much attention has been focused on their etiologies and the relationship among them (1,2). Although both genetic and environmental factors clearly play a role, exactly how these factors interact to produce the metabolic syndrome and its various components remains unclear.
In most humans, insulin resistance appears to be polygenic and heterogeneous (3). Thus, there are multiple genes that potentially contribute to the phenotype, and the development of disease in any individual may involve only a specific subset of these genes that varies from population to population. High-risk populations predisposed to the development of obesity and insulin resistance, such as Pima Indians and Mexican Americans, are thought to be enriched for clusters of genes acting together to produce the metabolic syndrome in the context of an appropriate environmental trigger, such as the high-fat western diet.
Dysregulation of hepatic lipid metabolism may play a central role in the pathogenesis of the metabolic syndrome. McGarry (7) has proposed that increased synthesis of lipids by the liver produces insulin resistance in other tissues, such as muscle. Increased storage of lipids in the liver results in fatty changes that are now known to be a feature of the metabolic syndrome (8). These changes form a spectrum of pathology, labeled nonalcoholic fatty liver disease, ranging from simple benign steatosis to nonalcoholic steatohepatitis, which can progress to cirrhosis and liver failure (9). It is thought that nonalcoholic fatty liver disease may now be the most common cause of cryptogenic cirrhosis in this country (10). Patients with the metabolic syndrome also typically have increased triglycerides and decreased HDL (11). Dyslipidemia is closely tied to the cardiovascular morbidities associated with the metabolic syndrome and may be attributed, at least in part, to the aberrant handling of lipids by the liver (12).
To understand how genes interact with dietary fat to produce the changes in lipid metabolism that occur in the metabolic syndrome, we used two strains of mice, representing differences in susceptibility to the development of insulin resistance. C57B1/6 (B6) mice have previously been shown to develop diabetes when subjected to genetically induced insulin resistance due to a double heterozygous deletion of one insulin receptor allele and one insulin receptor substrate-1 allele (13,14). 129Sv (129) mice on the other hand are protected from diabetes when carrying the same insulin receptor/insulin receptor substrate-1 double heterozygous defect. In the present study, B6 and 129 mice were placed on two extremes of diet: a low-fat diet (LFD; 14% calories from fat) and a high-fat diet (HFD; 55% calories from fat). We have compared the effects of genetic and dietary factors not only on glucose, but also on serum and hepatic lipid profiles and hepatic lipogenic gene expression, to better understand how these factors alter lipid metabolism and to identify the key elements controlling the progression of the metabolic syndrome.
RESEARCH DESIGN AND METHODS
Six-week-old male C57B1/B6 and 129S6/SvEvTac (Taconic) mice were placed on a low-fat high-carbohydrate (NIH#31; Taconic) or high-fat low-carbohydrate diet (TD93075; Harlan Teklad). The LFD derives 14% calories from fat, 25% calories from protein, and 61% calories from carbohydrates and was found to contain 1.5% saturated fatty acids, 2.7% monounsaturated fatty acids (MUFAs), and 0.6% polyunsaturated fatty acids by weight. The HFD derives 55% calories from fat, 21% calories protein, and 24% calories from carbohydrates and was found to contain 4.2% saturated fatty acids, 5.0% MUFAs, and 11.2% polyunsaturated fatty acids. The LFD and HFD have 15.4 and 7.1 mg cholesterol per 100 g, respectively. Arachidonic acid was undetectable in both diets. The mice were maintained on a 12-h light-dark cycle; unless otherwise indicated, serum samples were taken and mice were killed between 9:00 and 11:00 A.M., in the nonfasted state at ~6 months of age. Insulin levels were measured in plasma samples of random fed mice using the Crystal Chem ELISA kit and mouse insulin standards. Three independent cohorts were used to perform these experiments.
Serum lipid analysis. Equal volumes of serum from three to four mice fasted for 6 h (beginning in the morning) were pooled. Cholesterol and triglycerides were measured using Sigma kits 352 and 339, adapted for microtiter plates. Additionally, the serum was subjected to fast-performance liquid chromatography (FPLC) as previously described (15), and cholesterol was measured in the eluted fractions. Serum lipid analysis was performed by the Lipid, Lipoprotein and Atherosclerosis Core of the Vanderbilt Mouse Metabolic Phenotyping Centers.
Immunohistochemistry. Livers from the dead animals were frozen in liquid nitrogen, embedded in an optimal temperature cutting compound, and cut into 6-[micro]m sections. Hematoxylin/eosin and Oil-Red-O staining was performed using standard techniques.
Hepatic lipid analysis. Hepatic lipid analysis was performed by the Lipid, Lipoprotein and Atherosclerosis Core of the Vanderbilt Mouse Metabolic Phenotyping Centers. Lipids were extracted, filtered, and recovered in the chloroform phase. Individual lipid classes were separated by thin-layer chromatography using Silica Gel 60 A plates and visualized by rhodamine 6G. Phospholipids, triglycerides, and cholesterol esters were scraped, methylated, and analyzed by gas chromatography (16,17).
Oligonucleotide microarrays. Total RNA (25 [micro]g) Was pooled from two to three animals to make cRNA as described previously (18). cRNA (15 [micro]g) Was hybridized on Affymetrix murine chips U74Av.2, with four chips representing each group. Data were analyzed using MAS v5, with each chip being normalized to an average intensity of 1,500.
Real-time PCR. Total RNA was extracted and purified using the RNeasy kit (Qiagen) and used to direct cDNA synthesis using the RT for PCR kit (Clontech). RT-PCR was performed using SYBR green master mix (ABI), 5% of the cDNA synthesis reaction, and 300 nmol/l of the relevant primers. Sterol regulatory element-binding protein (SREBP)-1c and SREBP-la primers were isoform specific and have been previously described (19). Other primers were as follows: suppressor of cytokine signaling (SOCS)-3, 5'-CCTCGGGGACCATAGG AG-3' and 5'-AACTTGCTGTGGGTGACCAT-3'; SREBP-2, 5'GCGTTCTGGAGAC CATGGA-3' and 5'-ACAAAGTTGCTCTGAAAACAAATCA-3'; peroxisome proliferator-activated receptor-[gamma] coactivator (PGC)-1[alpha], 5'-GTCAACAGCAAAAG CCACAA-3' and 5'-TCTGGGGTCAGAGGAAGAGAg-3'; and PGC-1[beta], 5'-CCCTGT CCGTGAGGAACG-3' and 5'-ATCCATGGCTTCGTACTTGC-3'. The primers were found to amplify linearly. Because common housekeeping genes such as TATA-binding protein and the ribosomal protein 36B4 varied between swains, expression was normalized to the input RNA and calculated as a function of [2.sup.-Ct].
Immunoblotting of SREBP-1. Nuclear protein extracts of mouse liver were prepared as described by Sheng et al. (20). For each condition, equal portions of two to three mouse livers were pooled to produce nuclear extracts; these experiments were done in duplicate or triplicate. Immunoblotting was performed per the Amersham ECL detection system kit protocol, except that the washing solutions were supplemented with 0.1% SDS (wt/vol), 1% (vol/vol) Nonidet P-40, and 0.5% (wt/vol) sodium deoxycholate. Antibodies against mouse SREBP-1 have been previously described (21).
Stearoyl-CoA desaturase enzymatic activity. Conversion of [[1-[sup.14]C]stearoyl-CoA to [1-[sup.14]C]oleate was used to measure stearoyl-CoA desaturase (SCD) enzyme activity from microsomes prepared from individual liver extracts as previously described (22).
Immunoblotting of SOCS-3. Approximately 100 mg frozen liver from 16-week-old male mice, fed an HFD for 6 weeks, was homogenized in 25 mmol/l Tris 7.4, 2 mmol/l [Na.sub.3]V[O.sub.4], 10 mmol/l NaF, 10 mmol/l [Na.sub.4][P.sub.2][O.sub.7], 1 mmol/l EGTA, 1 mmol/l EDTA, 1% NP40, and one protease inhibitor tablet (Complete Protease Inhibitor tablets; Roche) in 50 ml and subjected to ultracentrifugation at 50,000 rpm for 45 min in a TLA100.2 rotor. Protein concentration was determined using a Bradford Assay (Bio-Rad). Protein (75 [micro]g) was subjected to SDS-PAGE, and immunoblotting was performed using a Roche Chemiluminescence Kit with antibodies against SOCS-3 (Santa Cruz).
