Type 2 diabetes in humans is associated with increased de novo lipogenesis (DNL), increased fatty acid (FA) fluxes, decreased FA oxidation, and hepatic steatosis. In this condition, VLDL production is increased and resistant to suppressive effects of insulin. The relationships between hepatic FA metabolism, steatosis, and VLDL production are incompletely understood. We investigated VLDL-triglyceride and -apolipoprotein (apo)-B production in relation to DNL and insulin sensitivity in female ob/ob mice. Hepatic triglyceride (5-fold) and cholesteryl ester (15-fold) contents were increased in ob/ob mice compared with lean controls. Hepatic DNL was increased ∼10-fold in ob/ob mice, whereas hepatic cholesterol synthesis was not affected. Basal rates of hepatic VLDL-triglyceride and -apoB100 production were similar between the groups. Hyperinsulinemic clamping reduced VLDL-triglyceride and -apoB100 production rates by ∼60% and ∼75%, respectively, in lean mice but only by ∼20% and ∼20%, respectively, in ob/ob mice. No differences in hepatic expression of genes encoding apoB and microsomal triglyceride transfer protein were found. Hepatic expression and protein phosphorylation of insulin receptor and insulin receptor substrate isoforms were reduced in ob/ob mice. Thus, strongly induced hepatic DNL is not associated with increased VLDL production in ob/ob mice, possibly related to differential hepatic zonation of apoB synthesis (periportal) and lipid accumulation (perivenous) and/or relatively low rates of cholesterogenesis. Insulin is unable to effectively suppress VLDL-triglyceride production in ob/ob mice, presumably because of impaired insulin signaling.
Type 2 diabetes is associated with increased de novo lipogenesis (DNL), decreased plasma fatty acid (FA) oxidation, and increased FA flux from peripheral tissues to the liver (1). These factors may all contribute to hepatic steatosis and increased hepatic VLDL production, two characteristic hallmarks of type 2 diabetes (2), and are probably related to hepatic insulin resistance (i.e., an insensitivity of hepatic metabolic processes to the effects of insulin). The relative contribution of the various pathways in hepatic lipid metabolism to the development of a fatty liver and disturbances in VLDL production is unknown but may, at least in part, be related to the localization of these processes within the liver. FA synthesis and triglyceride (TG) accumulation occur predominantly in the perivenous areas (zone 3) of the liver, whereas FA oxidation is more associated with the periportal areas (zone 1) (3,4). VLDL secretion has so far not been restricted to a specific hepatic zone.
Leptin-deficient ob/ob mice develop a fatty liver, insulin resistance, and hyperlipidemia (5,6). The contribution of hepatic lipoprotein production to the development of hyperlipidemia in these mice is not clear. Hyperglycemia together with an increased glycolytic activity in ob/ob mice may lead to an increased availability of acetyl-CoA residues for DNL (7), as supported by increased hepatic expression levels and activities of glycolytic enzymes, i.e., glucokinase (8,9), phosphofructokinase (7), and pyruvate kinase (7,8). Furthermore, elevated plasma free fatty acid (FFA) levels (9) and increased hepatic expression of fatty acid translocase (FAT or CD36) and plasma membrane–fatty acid binding protein have been reported in this model (10). Increased endoplasmic reticulum–associated acetyl-CoA synthase activity may increase the FA availability for esterification rather than for oxidation (10), which could contribute to increased TG and cholesteryl ester formation.
Insulin resistance seen in type 2 diabetes is associated with increased VLDL production (2). Acute hyperinsulinemia reduces VLDL production in healthy volunteers (11,12) but not in type 2 diabetic patients (2) and obese individuals (11). Despite the insulin-resistant condition and an increased hepatic TG content in ob/ob mice, a decreased VLDL-TG production rate under basal fasted conditions has been reported in this model (13,14). However, increased VLDL-TG secretion in ob/ob mice associated with enhanced expression and activity of the microsomal TG transfer protein has also been reported (15). The reason for these discrepant observations is unknown. The impact of insulin on VLDL production in the ob/ob mouse model has not been reported previously. Therefore, we quantified hepatic DNL and cholesterol synthesis using mass isotopomer distribution analysis (MIDA) and related the synthesis rates to VLDL-TG and -apoB production rates determined under basal conditions and during hyperinsulinemic clamps in ob/ob mice and lean littermates. Hepatic insulin signaling and expression levels of genes encoding transcription factors and important enzymes involved in FA and cholesterol metabolism, VLDL formation, and insulin signaling were studied to provide a mechanistic basis for our findings.
RESEARCH DESIGN AND METHODS
Animals.
Female ob/ob and lean littermates were purchased from Harlan (Zeist, the Netherlands) and housed in a light- and temperature-controlled facility. Experimental protocols were approved by the local experimental ethical committee for animal experiments.
Analytical kits.
Plasma and hepatic TG, cholesterol, and glucose levels were determined by commercially available kits (Roche, Mannheim, Germany). Plasma and hepatic phospholipid concentrations and plasma FFA concentrations were determined with the Phospholipid-kit and NEFA-C kit, respectively (Wako Chemical, Neuss, Germany). Plasma insulin was determined by a radioimmunoassay (RI-13K; Linco Research, St. Charles, MO).
Experimental procedures.
Female ob/ob and lean mice, weighing between 51–63 and 24–28 g, respectively, were ad libitum fed normal diet (RMH-B 2181; Hope Farms BV, Woerden, the Netherlands) enriched with 2% [1-13C]-acetate (Isotec, Miamisburg, OH). After 11 days, mice were fasted for 4 h and anesthetized with halotane, and livers were excised. A portion of abdominal fat was also collected. Liver and fat tissue were immediately frozen in liquid nitrogen and stored at −80°C. Blood was collected by heart puncture and immediately placed on ice in EDTA-containing tubes and centrifuged 10 min at 5,000 rpm at 4°C.
Hyperinsulinemic clamp.
To study effects of insulin on lipoprotein metabolism, a second group of lean and ob/ob mice received a hyperinsulinemic clamp or a saline infusion under anesthesia after a 9-h fast. Based on euglycemic insulin clamps performed in rats and mice by Hawkins et al. (16) and Rossetti et al. (17), in which insulin concentrations were fixed at ∼25 ng/ml, we used a single infusate to establish hyperinsulinemia and euglycemia. The procedure was tested in pilot experiments. The infusate contained insulin (18 mU · kg−1 · min−1; Novo Nordisk, Bagsvaerd, Denmark), somatostatin (1.5 μg · kg−1 · min−1; UCB, Breda, the Netherlands), and glucose (25 mg · kg−1 · h−1; Merck, Darmstadt, Germany). All solutions were freshly prepared in saline containing 1.5% BSA (Sigma, St. Louis, MO). Blood glucose concentration was determined with a GlucoTouch glucose analyzer (LifeScan, Beerse, Belgium). The total infusion time was 2 h. After 1 h, mice received a Triton WR1339 injection (Sigma) as a 12% wt/wt solution dissolved in saline, in a dose of 5 ml/kg lean body wt (19). Blood samples were taken before (t0) and 30 and 60 min after Triton injection. At the end of the experiment, a large blood sample was obtained by heart puncture for isolation of VLDL particles (see below). VLDL production rates were calculated from the slope of the linear TG versus time curves (Fig. 1D). Because mouse liver secretes TG-rich lipoproteins in the form of intermediate-density lipoprotein (IDL)-like and VLDL particles, we used a solution of 15.3% NaCl and 35.4% KBr (final concentration 0.65% and 1.52%, respectively) in saline with a density <1.019 g/ml to isolate VLDL/IDL. Plasma (0.2 ml) was mixed with 0.8 ml of the NaCl-KBr solution and centrifuged for 100 min at 120,000 rpm (627,000g) and 4°C in an ultracentrifuge (rotor TLA 120.2; Beckman). Tubes were sliced at 1.5 cm and the top-fraction, containing VLDL/IDL, was collected and frozen at −80°C until composition analyses. VLDL/IDL particle size was determined using a Submicron Particle Sizer (Nicomp, Santa Barbara, CA).
Liver lipid analysis.
Histology.
The localization of hepatic TG and apolipoprotein (apo)-B (apoB) mRNA were visualized as indicators of TG deposition and VLDL formation, respectively. Hepatic morphology was visualized by standard hematoxylin-eosin staining, and neutral lipids were visualized by Oil-Red O. The apoB in situ hybridization technique was similar to that previously described for apoE (21). The pGEM-3z vector containing apoB cDNA was a gift from Dr. H.M. Princen (Gaubius Laboratory, TNO Prevention and Health, Leiden, the Netherlands).
MIDA.
MIDA allows quantitation of the biosynthesis of polymers in vivo and is described in detail elsewhere (22). The enrichment of the pool of acetyl-CoA precursor units (p) that have entered newly synthesized cholesterol and palmitate during feeding of a [1-13C]-acetate–enriched diet can be calculated by comparison with a theoretical table generated using binomial expansion and known isotope frequencies of the atomic isotopes. When the enrichment of the acetyl-CoA pool is known, it becomes possible to calculate the fraction (f) of newly synthesized cholesterol and palmitate molecules in plasma or tissues. To determine the absolute amount of newly synthesized hepatic cholesterol and palmitate, we multiplied f by the total amount of hepatic free cholesterol and palmitate, respectively. Oral labeling of acetyl-CoA pools with 13C-acetate was described by Jung et al. (23). By label-feeding for longer periods of time, pools with a slow turnover, such as connective tissue and arterial walls, are also being labeled. After 10 days of labeling, pools with a rapid turnover will have reached a steady state, but pools with a slow turnover might not have. One must realize therefore that this may lead to a certain underestimation in synthesis rates in ob/ob mice, mainly for palmitate, because pool sizes of palmitate are larger in these animals.
Gas chromatography/mass spectrometry analysis.
Plasma cholesterol was extracted and derivatized as described (24). Cholesterol-trimethylsilyl derivatives were separated on an HP 5890 Plus gas chromatograph (Hewlett-Packard, Palo Alto, CA) using a 30 m × 0.25 mm (0.2 μm film thickness) DB5 ms column (J&W Scientific, Falson, CA) inserted into the ion source of a Quadruple mass spectrometer (model SSQ 7000; Finnigan Matt, San Jose, CA). The mass fragment mass-to-charge ratios 368, 369, 370, and 371 were monitored by selected ion recording.
For analysis of the methyl esters of palmitate (25), the same gas chromatography/mass spectrometry mode described above was used, equipped with a 20 m × 0.18 mm AT1701 column (0.4 μm film thickness; Alltech Associates, Deerfield, MI). They were analyzed at a mass-to-charge ratio of 271, 272, and 273 (mass isotopomers M0, M1, and M2) using chemical ionization and selected ion recording.
ApoB quantification.
ApoB100 concentrations were quantified by comparison to an IDL apoB100 standard isolated from healthy human subjects (26). Because human LDL does not contain apoB48, we were unable to accurately quantify apoB48 levels by this procedure; these levels were estimated. Isolated VLDL samples (10 μl) were delipidated with methanol and diethylether and dried under nitrogen. Delipidated lipoproteins were reduced in SDS sample buffer (8 mol/l urea, 10 mmol/l Tris base, 2% SDS, 10% glycerol, and 5% β-mercaptoethanol) and separated by SDS-PAGE using 4–15% gradient gels (Ready gels; Bio-Rad, Hercules, CA). Gels were either subjected to silver-staining (26) or were used for Western blot analysis. Proteins were transferred onto nitrocellulose membranes (Hybond ECL; Amersham Pharmacia Biotech, Buckinghamshire, U.K.). Blots were stained with the primary polyclonal antibody against human apoB, raised in sheep (dilution 1:100,000; Roche) and secondary IgG anti-sheep antibody conjugated with horseradish-peroxidase activity (dilution 1:10,000; Calbiochem, San Diego, CA).
Hepatic gene expression.
Total RNA was isolated from ∼30 mg tissue using Trizol methodology (Life Technologies, Paisley, U.K.) followed by the SV Total RNA Isolation System (Promega, Madison, WI). RNA was converted to single-stranded cDNA by a RT procedure with M-Mulv-RT (Boehringer Mannheim, Mannheim, Germany), and mRNA levels were quantified by real-time PCR using the ABI Prism 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). Procedures were optimized for the particular genes using appropriate forward and reverse primers (Life Technologies) and a template-specific 3′-tetramethylrhodamine, 5′-6-carboxyfluorescein–labeled Double Dye Oligonucleotide probe (Eurogentec, Seraing, Belgium). Calibration curves were run on serial dilutions of a 8× concentrated cDNA solution, resulting in a series containing 8×, 4×, 2×, 1×, 0.5×, 0.125×, 0.062×, and 0.031× the cDNA present in the assay incubation. Both assay and calibration incubations were done simultaneously. The fluorescence data obtained were processed using the software program ABI Sequence Detector v1.6.3 (Applied Biosystems). All quantified expression levels were within the linear part of the calibration curves and calculated using these curves. The primers and probe sets used are listed in Table 1.
Hepatic insulin signaling.
For analysis of insulin receptor (IR)-β and insulin receptor substrate (IRS)-1, IRS-2, and IRS-3 phosphorylation, liver tissue was homogenized in RIPA buffer (30 mmol/l Tris, pH 7.5, 1 mmol/l EDTA, 150 mmol/l NaCl, 0.5% Triton X-100, 0.5% deoxycholate, 1 mmol/l sodium orthovanadate, 10 mmol/l sodium fluoride, and protease inhibitors [Complete; Boehringer Mannheim]) using a Ultraturrax mixer followed by centrifugation (14 krpm; 15 min; 4°C). Protein content of supernatants was determined using the BCA kit (Pierce, Rockford, IL). A total of 25 μg protein was analyzed by immunoblotting for expression of the IRβ subunit (Transduction Laboratories, Lexington, KY), IRS-1 (27), IRS-2 (28), and IRS-3. Anti–IRS-3 antibody was obtained from rabbits immunized with a recombinant His-tagged IRS-3 fusion protein produced from pET16B-IRS-3 (aa198–494 of rat IRS3) as described by Ouwens et al. (27).
Statistical analysis.
All values reported are means ± SD; statistical significance implies P < 0.05. Because of small sample sizes, all statistical analyses involved nonparametrical Mann-Whitney U tests.
RESULTS
Animal characteristics.
Mean body weight was 26 ± 1 vs. 58 ± 5 g in the lean and ob/ob mice, respectively (P < 0.05). Fasting plasma glucose, insulin TG, cholesterol, and FFA concentrations were elevated in ob/ob mice (Table 2). Excess TG and cholesterol in ob/ob plasma was predominantly found in VLDL-sized fractions upon fast-protein liquid chromatography separation (data not shown). Liver weight (2-fold), hepatic TG (5-fold), total cholesterol (∼2.6-fold), free cholesterol (∼1.6-fold), cholesteryl ester (∼15-fold), and glycogen levels (∼1.8-fold) were all increased in ob/ob mice. No differences in hepatic phospholipid and glucose-6-phosphate levels were detected between lean and ob/ob mice (Table 2).
Neutral fat deposition (not shown in Fig. 2) in ob/ob mice was clearly associated with the perivenous (zone 3) area of liver lobules resulting in enlarged fat-laden hepatocytes in these parts of the liver (Fig. 2A vs. B). To check whether localization of fat in ob/ob liver was compatible with that of apoB gene expression, Apob mRNA was visualized by in situ hybridization in lean and ob/ob mouse liver (Fig. 2C and D, respectively). Apob mRNA was present in the entire liver lobe, but a stronger signal was observed in the periportal zone of the liver both in lean and ob/ob mice, suggesting zonal differentiation between VLDL formation and TG deposition in ob/ob mouse liver.
De novo lipogenesis and cholesterol synthesis.
Palmitate and cholesterol synthesis rates in lean and ob/ob mice fed a [1-13C]-acetate–enriched diet are summarized in Table 3. The enrichment of the acetyl-CoA pool and fractional synthesis rate values could not be calculated in adipose tissue of ob/ob mice because of low isotopic enrichments. Enrichments of the hepatic acetyl-CoA pools for DNL and cholesterogenesis were similar between lean and ob/ob mice. Hepatic fractional DNL was increased 1.7-fold in ob/ob mice compared with lean controls. The absolute amount of newly synthesized hepatic palmitate was 10-fold higher in livers of ob/ob mice than in those from controls. DNL values in adipose tissue of lean controls indicate that adipocytes may significantly contribute to total DNL in mice. The absolute amounts of newly synthesized cholesterol were similar in livers of both groups.
Expression levels of key genes involved in DNL (i.e., fatty acid synthase [Fas] and acyl-CoA carboxylase [Acc]), were clearly increased in livers of ob/ob mice. The mRNA levels of the transcription factor sterol regulatory element–binding protein (SREBP)-1c was also increased, but gene expression of liver X receptor (LXR) and carbohydrate responsive element–binding protein (CHREBP), recently implicated in the control of Fas expression, were significantly decreased. Expression level of the sterol regulatory element–binding protein-2 (SREBP-2), involved in control of cellular cholesterol homeostasis, was decreased in ob/ob mice liver, but this did not result in reduced expression of its target gene HMG-CoA reductase (Hmgr). Hepatic expression levels of the transcription factor peroxisome proliferator–activated receptor (PPAR)-α was decreased, but the expression level of PPAR-γ was increased in ob/ob mouse liver. Genes involved in ketogenesis and β-oxidation that are controlled by PPAR-α (i.e., mitochondrial HMG-CoA synthase [Hmgs], carnitine palmitoyl transferase-1a [Cpt1a], and medium-chain acyl dehydrogenase [Mcad]) tended to be decreased in ob/ob mice, suggesting a decreased β-oxidation in ob/ob mouse liver (Fig. 3B).
Hepatic VLDL production under basal conditions and hyperinsulinemic clamp.
During the clamp, plasma glucose levels were fixed at fasting (9 h) plasma concentrations that were reached within 60 min. Average plasma glucose levels were 7 ± 1 and 15 ± 1 mmol/l for lean and ob/ob mice, respectively (Fig. 4A). Plasma insulin increased to stable levels of 34 ± 3 and 34 ± 2 ng/ml in lean and ob/ob mice, respectively (Fig. 4B). Saline-infused mice maintained their fasting insulin level during the clamp (0.7 ± 0.4 and 5 ± 2 ng/ml in lean and ob/ob mice, respectively). Plasma FFA levels decreased in the insulin-infused mice only (Fig. 4C). Although ob/ob mice showed higher basal plasma FFA concentrations, insulin reduced plasma FFA levels to a similar concentration as in lean mice within 60 min.
After 60 min of saline infusion or hyperinsulinemia, Triton WR1339 was injected to determine VLDL-TG and apoB100 production rates. Basal VLDL-TG production rates were similar in lean and ob/ob mice (64 ± 14 and 52 ± 7 μmol · kg−1 · h−1, respectively; Fig. 1A). Acute hyperinsulinemia reduced the VLDL-TG production rate to 27 ± 1 μmol · kg−1 · h−1 (−58%) in lean mice but only to 41 ± 1 μmol · kg−1 · h−1 (−21%) in ob/ob mice (Fig. 1A). ApoB100 production showed a similar pattern as VLDL-TG production rates. Insulin suppression of apoB100 production was much more pronounced in lean mice than in ob/ob mice (Fig. 1B). The apoB100/B48 ratio in nascent VLDL particles, as determined by intensity scanning of Western blots (Fig. 1E), was much higher in ob/ob mice than in lean controls; this ratio decreased upon insulin infusion.
Expression of genes encoding apolipoproteins involved in VLDL assembly and secretion (i.e., apoB and apoE) were similar in liver of lean and ob/ob mice. Despite the increased apoB100/B48 ratio in ob/ob mouse VLDL, expression of Apobec-1, encoding the Apob mRNA-editing protein, was not different between both groups, indicating a posttranscriptional upregulation of editing activity in ob/ob mice. Expression of the genes encoding the microsomal TG transfer protein and DGAT (diacylglycerolacyltransferase), essential for VLDL lipidation, did not differ between lean and ob/ob mice (Fig. 3C).
Insulin signaling.
Hepatic mRNA levels of the IR (Ir) and IRS isoforms (Irs1 and Irs2) were decreased in ob/ob mice (Fig. 5A). Also, phosphorylation of IRβ, IRS-1, and IRS-2 proteins was reduced in ob/ob mice liver (Fig. 5B), indicating decreased hepatic insulin signaling. IRS-3 phosphorylation was slightly increased in the ob/ob mouse liver (Fig. 5B).
DISCUSSION
The primary defect in the ob/ob mouse model is the absence of leptin, resulting in an obese and diabetic phenotype (5). The intracellular signal transduction of leptin is similar to that of class 1 cytokine receptors and involves JAK-STAT signaling. Some of these receptors and possibly also leptin signaling can be linked to mitogen-activated protein kinase and phosphatidylinositol 3-kinase (PI3K) pathways (29,30). However, the metabolic relevance of leptin signaling via these pathways on hepatic fat and cholesterol metabolism and on VLDL production is currently not known. In this study, ob/ob mice showed increased plasma FFA levels, a 10-fold increase in hepatic DNL, and a severe perivenously localized hepatic steatosis. Despite these diabetic characteristics, hepatic VLDL production was not increased under fasting conditions, as is the case in humans with insulin resistance or type 2 diabetes (31). The absence of a simultaneous upregulation of hepatic cholesterol synthesis, recognized as a crucial factor in control of VLDL production rates (32–34), might contribute to this discordant phenotype. We further demonstrate that the VLDL production process in ob/ob mice was insensitive to the suppressive effects of insulin. This disturbance in the control of VLDL production is likely a result of impairment in the transduction pathway(s) of insulin; in livers of ob/ob mice, IRβ, IRS-1, and IRS-2 gene expression levels and protein phosphorylation were clearly decreased. Similar decreases of hepatic and muscle IRS phosphorylation after insulin stimulation in vivo were observed in other studies (35–37). It is well established that insulin-mediated suppression of VLDL-apoB secretion in rodent liver cells requires PI3K activation (38,39). Phosphatases might play a role in the sequence of events. The phosphotyrosine phosphatase-1B has been associated with insulin signaling in different models (40,41). Whether this phosphatase is involved in defective insulin signaling in ob/ob mouse liver remains to be determined.
Female ob/ob mice have been used extensively in metabolic studies concerning hepatic (and muscle) insulin sensitivity (8,42). To be able to compare our data with published work, we chose to use female ob/ob mice for the current experiments. Importantly, studies by Li and colleagues (13,18) indicate that there are no differences between male and female ob/ob mice regarding hepatic lipid deposition or VLDL-TG production.
DNL, suggested as a regulator of VLDL production, was increased 10-fold in ob/ob liver. Expression of enzymes involved in lipogenesis are under control of at least three transcription factors (SREBP-1c, LXR, and CHREBP) (43–45). Interestingly, hepatic SREBP-1c expression levels were increased in ob/ob mice, whereas those of LXR and CHREBP were decreased, indicating that SREBP-1c is independently able to induce DNL. Because SREBP-1c expression is influenced by insulin (46) and insulin levels are elevated in ob/ob mice, insulin may continuously induce expression of SREBP-1c and, thereby, of its target genes (43). Thus, insulin resistance may not involve all branches of insulin signaling. Alternatively, leptin has been shown to be able to downregulate SREBP-1c expression and protein levels and expression of its target gene (Fas) in ob/ob adipocytes (47) and in wild-type mouse liver (48). IRS-2−/− mice, like ob/ob mice, have increased hepatic SREBP-1c levels, which normalize upon leptin treatment (49). Irrespective of the underlying mechanism, however, our results indicate that upregulated DNL per se is not a regulator of hepatic VLDL production by mouse liver.
In our clamp experiments, plasma insulin levels were similar in both groups. However, both groups were clamped at their basal glucose levels, resulting in higher glucose levels in ob/ob mice. The question arises to which extent this may have influenced our results with respect to insulin sensitivity of VLDL-TG production. It could be argued that hyperglycemia may directly promote VLDL production. This effect might counteract inhibitory effects of insulin on VLDL-TG secretion. However, in the basal state, VLDL production was not increased despite hyperglycemia in ob/ob mice, indicating that hyperglycemia is not an independent driving force for VLDL secretion in these animals. Therefore, it is rather unlikely that hyperglycemia per se underlies the profound insulin resistance of VLDL production.
Theoretically, it may be that TG and apoB, required for VLDL assembly, are functionally separated in the ob/ob liver. Using in situ hybridization, we found that Apob mRNA is present in all cells in the liver lobule but with the highest intensity in periportal hepatocytes of control mice. Funahashi et al. (50) reported a uniform distribution of Apob mRNA in rat liver, suggesting the existence of species differences in this respect. In any case, our results suggest that perivenously localized TG in the ob/ob mouse liver may be less available for VLDL production.
Substrate availability has been proposed to regulate hepatic VLDL output (32,34). Because the availability of plasma FFA, de novo synthesized FA, and hepatic TG was increased in ob/ob mice, it is unlikely that the supply of TG is rate-controlling in this respect. The availability of newly synthesized cholesterol may also influence VLDL formation, as has been shown in rats (32), rabbits (33), and humans (34). Total hepatic cholesterol content in ob/ob mouse liver was increased, but the absolute cholesterol synthesis rate, as determined by MIDA, was not different from that in lean mice. Cellular cholesterol homeostasis is controlled by SREBP-2. Sterol depletion induces cleavage of membrane-bound SREBP-2, allowing its translocation to the nucleus to induce expression levels of genes involved in cholesterol synthesis and uptake. Increased hepatic cholesterol levels were associated with decreased SREBP-2 expression in ob/ob mice; however, this did not lead to alterations in expression levels of HMG-CoA reductase (Fig. 3) or absolute cholesterol synthesis rates (Table 3). We are aware that our studies were performed under specific experimental conditions. Yet, we tend to hypothesize that limited availability of newly synthesized cholesterol may compromise the ability of the ob/ob mouse liver to remove excess TG in the form of VLDL under basal conditions. The resistance of the VLDL assembly/secretion process to the suppressive effects of insulin, however, must result from decreased insulin signaling. According to current knowledge (38,39), it is highly likely that the PI3K pathway is defective in ob/ob mouse liver.
In conclusion, DNL is clearly increased in ob/ob mice, probably related to increased SREBP-1c expression levels and despite downregulation of LXR and CHREBP expression. The insufficient supply of newly synthesized cholesterol may become rate controlling for VLDL production in ob/ob mice in a situation in which the supply of FA from plasma and DNL is excessive. Metabolic zonation of TG accumulation and apoB production may contribute in this respect. The inability to induce VLDL production under these conditions, in combination with impaired hepatic β-oxidation, contributes to development of hepatic steatosis. Insulin signaling is clearly impaired in fatty livers of ob/ob mice, resulting in a decreased ability of insulin to suppress VLDL production.
VLDL-TG production rate (A), apoB100 production rate (B), and apoB100/B48 band-density-scan-ratio (C) during a hyperinsulinemic clamp in lean and ob/ob mice receiving saline (□) or insulin (▪). Triglyceride accumulation curve (D) and a representative Western blot of apoB (E) during hyperinsulinemic clamp in lean and ob/ob mice are also shown. n = 4 per group. *P < 0.05, insulin effect; †P < 0.05, mouse strain effect, Mann-Whitney U test.
VLDL-TG production rate (A), apoB100 production rate (B), and apoB100/B48 band-density-scan-ratio (C) during a hyperinsulinemic clamp in lean and ob/ob mice receiving saline (□) or insulin (▪). Triglyceride accumulation curve (D) and a representative Western blot of apoB (E) during hyperinsulinemic clamp in lean and ob/ob mice are also shown. n = 4 per group. *P < 0.05, insulin effect; †P < 0.05, mouse strain effect, Mann-Whitney U test.
Hematoxylin-eosin (HE) staining (A and B) and apoB in situ hybridization in liver slides from lean (C and E) and ob/ob mice (D and F), respectively. p, portal area (zone 1); c, central area (zone 3).
Hematoxylin-eosin (HE) staining (A and B) and apoB in situ hybridization in liver slides from lean (C and E) and ob/ob mice (D and F), respectively. p, portal area (zone 1); c, central area (zone 3).
Relative hepatic expression levels of several transcription factors (A); their target genes in FA synthesis, cholesterol synthesis, and FA oxidation (B); and genes involved in VLDL assembly (C) in lean (□) and ob/ob (▪) mice. Genes are relative to the hepatic β-actin expression level. *P < 0.05, Mann-Whitney U test.
Relative hepatic expression levels of several transcription factors (A); their target genes in FA synthesis, cholesterol synthesis, and FA oxidation (B); and genes involved in VLDL assembly (C) in lean (□) and ob/ob (▪) mice. Genes are relative to the hepatic β-actin expression level. *P < 0.05, Mann-Whitney U test.
Plasma glucose (A), insulin (B), and FFA (C) levels during a hyperinsulinemic clamp in lean mice receiving saline (▪) or insulin (□) and in ob/ob mice receiving saline (•) or insulin (○). n = 4 per group. *P < 0.05, insulin effect; †P < 0.05, mouse strain effect, Mann-Whitney U test.
Plasma glucose (A), insulin (B), and FFA (C) levels during a hyperinsulinemic clamp in lean mice receiving saline (▪) or insulin (□) and in ob/ob mice receiving saline (•) or insulin (○). n = 4 per group. *P < 0.05, insulin effect; †P < 0.05, mouse strain effect, Mann-Whitney U test.
Hepatic gene expression levels (A) and phosphorylation (B) of the IR and IRS isoforms in lean and ob/ob mice. *P < 0.05, Mann-Whitney U test.
Hepatic gene expression levels (A) and phosphorylation (B) of the IR and IRS isoforms in lean and ob/ob mice. *P < 0.05, Mann-Whitney U test.
List of sequences of primers and probes used
Standard . | Type . | Sequence . | GenBank number . |
---|---|---|---|
β-Actin | Forward | AGC CAT GTA CGT AGC CAT CCA | NM_007393 |
Reverse | TCT CCG GAG TCC ATC ACA ATG | ||
Probe | TGT CCC TGT ATG CCT CTG GTC GTA CCAC | ||
Transcription factors | |||
Srebp-1c | Forward | GGA GCC ATG GAT TGC ACA TT | BI656094 |
Reverse | CCT GTC TCA CCC CCA GCA TA | ||
Probe | CAG CTC ATC AAC AAC CAA GAC AGT GAC TTC C | ||
Srebp-2 | Forward | CTG CAG CCT CAA GTG CAA AG | AF374267 |
Reverse | CAG TGT GCC ATT GGC TGT CT | ||
Probe | CCA TCC AGC AGC AGG TGC AGA CG | ||
Ppar-α | Forward | TAT TCG GCT GAA GCT GGT GTA C | X57638 |
Reverse | CTG GCA TTT GTT CCG GTT CT | ||
Probe | CTG AAT CTT GCA GCT CCG ATC ACA CTT G | ||
Ppar-γ | Forward | CAC AAT GCC ATC AGG TTT GG | X57638 |
Reverse | GCT GGT CGA TAT CAC TGG AGA TC | ||
probe | CCA ACA GCT TCT CCT TCT CGG CCT G | ||
Lxr | Forward | GCT CTG CTC ATT GCC ATC AG | AF085745 |
Reverse | TGT TGC AGC CTC TCT ACT TGG A | ||
Probe | TCT GCA GAC CGG CCC AAC GTG | ||
Chrebp | Forward | GAT GGT GCG AAC AGC TCT TCT | AF156604 |
Reverse | CTG GGC TGT GTC ATG GTG AA | ||
Probe | CCA GGC TCC TCC TCG GAG CCC | ||
DNL, cholesterogenisis, and β-oxidation | |||
Fas | Forward | GGC ATC ATT GGG CAC TCC TT | AF127033 |
Reverse | GCT GCA AGC ACA GCC TCT CT | ||
Probe | CCA TCT GCA TAG CCA CAG GCA ACC TC | ||
Acc | Forward | GCC ATT GGT ATT GGG GCT TAC | AF374170 |
Reverse | CCC GAC CAA GGA CTT TGT TG | ||
Probe | CTC AAC CTG GAT GGT TCT TTG TCC CAG C | ||
Hmgr | Forward | CCG GCA ACA ACA AGA TCT GTG | BB664708 |
Reverse | ATG TAC AGG ATG GCG ATG CA | ||
Probe | TGT CGC TGC TCA GCA CGT CCT CTT C | ||
CptIa | Forward | CTC AGT GGG AGC GAC TCT TCA | AF017175 |
Reverse | GGC CTC TGT GGT ACA CGA CAA | ||
Probe | CCT GGG GAG GAG ACA GAC ACC ATC CAA C | ||
Mcad | Forward | GCA GCC AAT GAT GTG TGC TTA C | NM_007382 |
Reverse | CAC CCT TCT TCT CTG CTT TGG T | ||
Probe | CCC TCC GCA GGC TCT GAT GTG G | ||
Hmgs | Forward | TGG TGG ATG GGA AGC TGT CTA | U12790 |
Reverse | TTC TTG CGG TAG GCT GCA TAG | ||
Probe | CCA AGG CCC GCA GGT AGC ACT G | ||
VLDL metabolism | |||
Apob | Forward | GCC CAT TGT GGA CAA GTT GAT C | AW012827 |
Reverse | CCA GGA CTT GGA GGT CTT GGA | ||
Probe | AAG CCA GGG CCT ATC TCC GCA TCC | ||
Apobec-I | Forward | TCG TCC GAA CAC CAG ATG CT | NM_031159 |
Reverse | GGT GTC GGC TCA GAA ACT CTG T | ||
Probe | CCT GGT TCC TGT CCT GGA GTC CCT G | ||
Apoe | Forward | CCT GAA CCG CTT CTG GGA TT | NM_009696 |
Reverse | GCT CTT CCT GGA CCT GGT CA | ||
Probe | AAA GCG TCT GCA CCC AGC GCA GG | ||
Dgat-I | Forward | GGT GCC GTG ACA GAG CAG AT | NM_010046 |
Reverse | CAG TAA GGC CAC AGC TGC TG | ||
Probe | CTG CTG CTA CAT GTG GTT AAC CTG GCC A | ||
Mttp | Forward | CAA GCT CAC GTA CTC CAC TGA AG | NM_008642 |
Reverse | TCA TCA TCA CCA TCA GGA TTC CT | ||
Probe | ACG GCA AGA CAG CGT GGG CTA CA | ||
Insulin signaling | |||
Ir | Forward | TGA GTC AGC CAG TCT TCG AGA A | NM_010568 |
Reverse | ACT ACC AGC ATT GGC TGT CCT T | ||
Probe | CTG CCA TCA TGT GGT CCH CCT TCT | ||
Irs-1 | Forward | AGC ACC TGG TGG CTC TCT ACA | NM_010570 |
Reverse | CAG CTG CAG AAG AGC CTG GTA | ||
Probe | CTC GCT ATC CGC GGC AAT GGC | ||
Irs-2 | Forward | AGT CCC ACA TCG GGC TTG AAG | AF090738 |
Reverse | GGT CTG CAC GGA TGA CCT TAG | ||
Probe | CCT TCA AGT CAG CCA GCC CCC TG |
Standard . | Type . | Sequence . | GenBank number . |
---|---|---|---|
β-Actin | Forward | AGC CAT GTA CGT AGC CAT CCA | NM_007393 |
Reverse | TCT CCG GAG TCC ATC ACA ATG | ||
Probe | TGT CCC TGT ATG CCT CTG GTC GTA CCAC | ||
Transcription factors | |||
Srebp-1c | Forward | GGA GCC ATG GAT TGC ACA TT | BI656094 |
Reverse | CCT GTC TCA CCC CCA GCA TA | ||
Probe | CAG CTC ATC AAC AAC CAA GAC AGT GAC TTC C | ||
Srebp-2 | Forward | CTG CAG CCT CAA GTG CAA AG | AF374267 |
Reverse | CAG TGT GCC ATT GGC TGT CT | ||
Probe | CCA TCC AGC AGC AGG TGC AGA CG | ||
Ppar-α | Forward | TAT TCG GCT GAA GCT GGT GTA C | X57638 |
Reverse | CTG GCA TTT GTT CCG GTT CT | ||
Probe | CTG AAT CTT GCA GCT CCG ATC ACA CTT G | ||
Ppar-γ | Forward | CAC AAT GCC ATC AGG TTT GG | X57638 |
Reverse | GCT GGT CGA TAT CAC TGG AGA TC | ||
probe | CCA ACA GCT TCT CCT TCT CGG CCT G | ||
Lxr | Forward | GCT CTG CTC ATT GCC ATC AG | AF085745 |
Reverse | TGT TGC AGC CTC TCT ACT TGG A | ||
Probe | TCT GCA GAC CGG CCC AAC GTG | ||
Chrebp | Forward | GAT GGT GCG AAC AGC TCT TCT | AF156604 |
Reverse | CTG GGC TGT GTC ATG GTG AA | ||
Probe | CCA GGC TCC TCC TCG GAG CCC | ||
DNL, cholesterogenisis, and β-oxidation | |||
Fas | Forward | GGC ATC ATT GGG CAC TCC TT | AF127033 |
Reverse | GCT GCA AGC ACA GCC TCT CT | ||
Probe | CCA TCT GCA TAG CCA CAG GCA ACC TC | ||
Acc | Forward | GCC ATT GGT ATT GGG GCT TAC | AF374170 |
Reverse | CCC GAC CAA GGA CTT TGT TG | ||
Probe | CTC AAC CTG GAT GGT TCT TTG TCC CAG C | ||
Hmgr | Forward | CCG GCA ACA ACA AGA TCT GTG | BB664708 |
Reverse | ATG TAC AGG ATG GCG ATG CA | ||
Probe | TGT CGC TGC TCA GCA CGT CCT CTT C | ||
CptIa | Forward | CTC AGT GGG AGC GAC TCT TCA | AF017175 |
Reverse | GGC CTC TGT GGT ACA CGA CAA | ||
Probe | CCT GGG GAG GAG ACA GAC ACC ATC CAA C | ||
Mcad | Forward | GCA GCC AAT GAT GTG TGC TTA C | NM_007382 |
Reverse | CAC CCT TCT TCT CTG CTT TGG T | ||
Probe | CCC TCC GCA GGC TCT GAT GTG G | ||
Hmgs | Forward | TGG TGG ATG GGA AGC TGT CTA | U12790 |
Reverse | TTC TTG CGG TAG GCT GCA TAG | ||
Probe | CCA AGG CCC GCA GGT AGC ACT G | ||
VLDL metabolism | |||
Apob | Forward | GCC CAT TGT GGA CAA GTT GAT C | AW012827 |
Reverse | CCA GGA CTT GGA GGT CTT GGA | ||
Probe | AAG CCA GGG CCT ATC TCC GCA TCC | ||
Apobec-I | Forward | TCG TCC GAA CAC CAG ATG CT | NM_031159 |
Reverse | GGT GTC GGC TCA GAA ACT CTG T | ||
Probe | CCT GGT TCC TGT CCT GGA GTC CCT G | ||
Apoe | Forward | CCT GAA CCG CTT CTG GGA TT | NM_009696 |
Reverse | GCT CTT CCT GGA CCT GGT CA | ||
Probe | AAA GCG TCT GCA CCC AGC GCA GG | ||
Dgat-I | Forward | GGT GCC GTG ACA GAG CAG AT | NM_010046 |
Reverse | CAG TAA GGC CAC AGC TGC TG | ||
Probe | CTG CTG CTA CAT GTG GTT AAC CTG GCC A | ||
Mttp | Forward | CAA GCT CAC GTA CTC CAC TGA AG | NM_008642 |
Reverse | TCA TCA TCA CCA TCA GGA TTC CT | ||
Probe | ACG GCA AGA CAG CGT GGG CTA CA | ||
Insulin signaling | |||
Ir | Forward | TGA GTC AGC CAG TCT TCG AGA A | NM_010568 |
Reverse | ACT ACC AGC ATT GGC TGT CCT T | ||
Probe | CTG CCA TCA TGT GGT CCH CCT TCT | ||
Irs-1 | Forward | AGC ACC TGG TGG CTC TCT ACA | NM_010570 |
Reverse | CAG CTG CAG AAG AGC CTG GTA | ||
Probe | CTC GCT ATC CGC GGC AAT GGC | ||
Irs-2 | Forward | AGT CCC ACA TCG GGC TTG AAG | AF090738 |
Reverse | GGT CTG CAC GGA TGA CCT TAG | ||
Probe | CCT TCA AGT CAG CCA GCC CCC TG |
Plasma and hepatic parameters after a 4-h fast in lean and ob/ob mice
Plasma . | Lean . | ob/ob . |
---|---|---|
Glucose (mmol/l) | 8.6 ± 3.0 | 16.9 ± 4.6* |
Insulin (ng/ml) | 1.3 ± 0.9 | 6.9 ± 0.6* |
TG (mmol/l) | 0.3 ± 0.03 | 0.7 ± 0.1* |
Cholesterol (mmol/l) | 1.7 ± 0.4 | 4.1 ± 0.3* |
FFA (mmol/l) | 0.7 ± 0.1 | 1.3 ± 0.1* |
Liver | ||
Liver weight (g) | 1.3 ± 0.1 | 2.7 ± 0.3* |
TG (μmol/g liver) | 1.2 ± 0.4 | 2.7 ± 0.5* |
Total cholesterol (μmol/g liver) | 1.3 ± 0.4 | 1.6 ± 0.3* |
Free cholesterol (μmol/g liver) | 0.8 ± 0.1 | 1.0 ± 0.1 |
Cholesteryl esters (μmol/g liver) | 0.1 ± 0.01 | 0.5 ± 0.3* |
Phospholipids (μmol/g liver) | 0.3 ± 0.1 | 0.2 ± 0.3 |
Glucose-6-phosphate (μmol/g liver) | 0.2 ± 0.03 | 0.1 ± 0.1 |
Glycogen (μmol/g liver) | 212 ± 11 | 183 ± 17 |
Plasma . | Lean . | ob/ob . |
---|---|---|
Glucose (mmol/l) | 8.6 ± 3.0 | 16.9 ± 4.6* |
Insulin (ng/ml) | 1.3 ± 0.9 | 6.9 ± 0.6* |
TG (mmol/l) | 0.3 ± 0.03 | 0.7 ± 0.1* |
Cholesterol (mmol/l) | 1.7 ± 0.4 | 4.1 ± 0.3* |
FFA (mmol/l) | 0.7 ± 0.1 | 1.3 ± 0.1* |
Liver | ||
Liver weight (g) | 1.3 ± 0.1 | 2.7 ± 0.3* |
TG (μmol/g liver) | 1.2 ± 0.4 | 2.7 ± 0.5* |
Total cholesterol (μmol/g liver) | 1.3 ± 0.4 | 1.6 ± 0.3* |
Free cholesterol (μmol/g liver) | 0.8 ± 0.1 | 1.0 ± 0.1 |
Cholesteryl esters (μmol/g liver) | 0.1 ± 0.01 | 0.5 ± 0.3* |
Phospholipids (μmol/g liver) | 0.3 ± 0.1 | 0.2 ± 0.3 |
Glucose-6-phosphate (μmol/g liver) | 0.2 ± 0.03 | 0.1 ± 0.1 |
Glycogen (μmol/g liver) | 212 ± 11 | 183 ± 17 |
Data are means ± SD. n = 5 per group.
P < 0.05, Mann-Whitney U test.
Acetyl-CoA pool enrichment and fractional palmitate synthesis values of adipose tissue and liver and hepatic acetyl-CoA pool enrichment and fractional cholesterol synthesis values in [1-13C]-acetate–enriched diet–fed lean and ob/ob mice
. | Lean . | ob/ob . |
---|---|---|
DNL | ||
Liver | ||
Acetyl-CoA pool enrichment (%) | 6.3 ± 0.3 | 5.9 ± 0.8 |
Fractional hepatic palmitate (%) | 31.0 ± 6.1 | 53.1 ± 4.7* |
Hepatic palmitate (μmol/liver) | 1.1 ± 0.4 | 6.3 ± 0.9* |
Newly synthesized hepatic palmitate (μmol/liver) | 0.3 ± 0.1 | 3.4 ± 0.6* |
Adipose tissue | ||
Acetyl-CoA pool enrichment (%) | 1.6 ± 0.5 | ND |
Fractional synthesis adipose tissue palmitate (%) | 21.5 ± 8.7 | ND |
Cholesterol synthesis | ||
Acetyl-CoA pool enrichment (%) | 6.3 ± 0.1 | 5.5 ± 0.5 |
Fractional synthesis hepatic free cholesterol (%) | 17.8 ± 4.0 | 12.6 ± 3.4* |
Hepatic free cholesterol (μmol/liver) | 1.7 ± 0.1 | 2.7 ± 0.5* |
Newly synthesized hepatic cholesterol (μmol/liver) | 0.3 ± 0.1 | 0.3 ± 0.2 |
. | Lean . | ob/ob . |
---|---|---|
DNL | ||
Liver | ||
Acetyl-CoA pool enrichment (%) | 6.3 ± 0.3 | 5.9 ± 0.8 |
Fractional hepatic palmitate (%) | 31.0 ± 6.1 | 53.1 ± 4.7* |
Hepatic palmitate (μmol/liver) | 1.1 ± 0.4 | 6.3 ± 0.9* |
Newly synthesized hepatic palmitate (μmol/liver) | 0.3 ± 0.1 | 3.4 ± 0.6* |
Adipose tissue | ||
Acetyl-CoA pool enrichment (%) | 1.6 ± 0.5 | ND |
Fractional synthesis adipose tissue palmitate (%) | 21.5 ± 8.7 | ND |
Cholesterol synthesis | ||
Acetyl-CoA pool enrichment (%) | 6.3 ± 0.1 | 5.5 ± 0.5 |
Fractional synthesis hepatic free cholesterol (%) | 17.8 ± 4.0 | 12.6 ± 3.4* |
Hepatic free cholesterol (μmol/liver) | 1.7 ± 0.1 | 2.7 ± 0.5* |
Newly synthesized hepatic cholesterol (μmol/liver) | 0.3 ± 0.1 | 0.3 ± 0.2 |
Data are means ± SD. n = 5 lean and 5 ob/ob mice.
P < 0.05, Mann-Whitney U test. ND, not detectable.
Article Information
This work was supported by the Netherlands Diabetes Foundation (grant 96.604).
We thank Vincent Bloks and Juul Baller for excellent technical assistance.
REFERENCES
Address correspondence and reprint requests to Folkert Kuipers, Department of Pediatrics, Room Y2115 CMCIV, Hanzeplein 1, P.O. Box 30.001, 9700 RB, Groningen, The Netherlands. E-mail: [email protected].
Received for publication 16 June 2002 and accepted in revised form 29 January 2003.
C.H.W. and R.H.J.B. contributed equally to this work.
apo, apolipoprotein; CHREBP, carbohydrate responsive element–binding protein; DNL, de novo lipogenesis; FA, fatty acid; FFA, free fatty acid; IDL, intermediate-density lipoprotein; IR, insulin receptor; IRS, insulin receptor substrate; MIDA, mass isotopomer distribution analysis; PI3K, phosphatidylinositol 3-kinase; PPAR, peroxisome proliferator–activated receptor; SREBP, sterol regulatory element–binding protein; TG, triglyceride.