Insulin-Independent Induction of Sterol Regulatory Element-Binding Protein-1c Expression in the Livers of Streptozotocin-Treated Mice

Sugars were obtained from Wako Pure Chemical Industries. STZ was purchased from Sigma and insulin from Eli Lilly. Enhanced chemiluminescence Western blot detection kit and redivue [α-³²P]dCTP (6,000 Ci/mmol) were purchased from Amersham Pharmacia.

Seven-week-old male C57BL/6J mice (21–23 g) were purchased from Clea (Tokyo, Japan) and adapted to the environment for 1 week before initiation of the study. All mice were housed in a controlled environment with a 12-h light/dark cycle and free access to water and standard laboratory diet. Diabetes was induced by two intraperitoneal injections of STZ (100 mg/kg body wt; Sigma) with 1-day intervals. STZ was dissolved in 50 mmol/l sodium citrate buffer (pH 4.5) immediately before administration. Animals were considered diabetic when blood glucose levels exceeded 350 mg/dl, usually within 7 days from the first injection. For dietary studies, control and STZ-administered mice were fasted for 24 h and refed glucose, fructose, or sucrose for 3, 6, 9, or 12 h. STZ-administered mice injected with insulin received a combination of human regular insulin (3 units; Eli Lilly) subcutaneously and human NPH insulin (0.6 units; Eli Lilly) intraperitoneally. The mice in a control group received PBS intraperitoneally and subcutaneously.

Male Wistar rats (250–300 g; Clea) were housed in a controlled environment with a 12-h light/dark cycle and free access to water and standard laboratory diet. After the animals were fasted for 24 h, rats in the STZ group were injected intravenously with 60 mg/kg body wt STZ in 50 mmol/l sodium citrate buffer (pH4.5), and rats in the control group were injected with 50 mmol/l sodium citrate solution. Tail vein blood glucose levels were measured, and animals were killed 1 week after the induction of diabetes. For dietary studies, control and STZ-administered rats were fasted for 48 h and refed either glucose, fructose, or sucrose for 24 h.

The concentration of glucose, insulin, cholesterol, triglycerides, and nonesterified fatty acids in plasma and of cholesterol and triglycerides in liver were measured as previously described (37). Plasma leptin was determined by the mouse leptin enzyme-linked immunosorbent assay kit (Morinaga).

Nuclear extracts and membrane fractions of mouse livers were prepared as previously described (38). A small portion of each sample was removed and analyzed for protein content by a bicinchoninic acid reaction (Pierce). Aliquots of nuclear (25 μg) and membrane (30 μg) proteins were subjected to immunoblot analysis with rabbit IgG against mouse SREBP-1 (11), SREBP-1c (15), and SREBP-2 (39), followed by horseradish peroxidase-linked goat IgG against rabbit IgG and the enhanced chemiluminescence kit (Amersham Pharmacia Biotech) (11,15).

Total RNA from liver was isolated with Trizol reagent (Invitrogen), and 10-μg RNA samples pooled equally from three mice of each group were run on a 1% agarose gel containing formaldehyde and transferred to Hybond-N membranes (Amersham Pharmacia Biotech). The cDNA probes for aldolase A, aldolase B, fructokinase, carbohydrate response element-binding protein, and GLUT2 were generated by using the sequence from GenBank with accession numbers Y00516, X53402, BC013464, AF156604, and BC034675, respectively. The other cDNA probes used were cloned as previously described (11,13,35). The probes were labeled with [α-³²P]dCTP using the Megaprime DNA labeling system kit (Amersham Biosciences). Membranes were hybridized with the radiolabeled probes in Rapid-hyb Buffer (Amersham Pharmacia Biotech) and washed in 0.1× SSC, 0.1% SDS, at 65°C. Blots were exposed to Eastman Kodak XAR-5 film, and the BAS imaging plate for the BAS2000 Bio Imaging Analyzer (Fuji Photo Film). The quantification results obtained with the BAS2000 system were normalized to the signal generated from 36B4 (acidic ribosomal phosphoprotein P0) mRNA.

Formalin-fixed pancreata were embedded in paraffin, and 4-μm-thick sections were used for insulin staining. After being dewaxed through xylene and absolute ethanol, the slides were rehydrated through decreasing concentrations of ethanol (100, 90, 80, and 70%) and rinsed in distilled water. The slides were washed thoroughly in Tris-buffered saline containing 1 mmol/l CaCl₂ (pH 7.6) (TBS-Ca) and incubated with 0.03% trypsin in TBS-Ca for 30 min. After blocking with TBS-Ca containing 5% skim milk for 1 h, the slides were exposed to the primary antibody, a polyclonal guinea pig anti-swine insulin (1:4; Dako) at 4°C overnight. After being washed in TBS-Ca, the slides were incubated with the secondary antibody, polyclonal sheep anti-guinea pig IgG conjugated to fluorescein isothiocyanate (1:100, PARIS) for 30 min at room temperature. After washing three times with TBS-Ca, the slides were photographed with fluorescence microscopy (Leica Microsystems).

To examine the kinetics of effects of different sugars on the expression of hepatic SREBP-1 and lipogenic enzymes, mice were fasted for 24 h and refed glucose, fructose, or sucrose for 6, 9, or 12 h, and the amount of mRNA was measured by Northern blotting (Fig. 1). Robust induction of SREBP-1 mRNA level was observed by all of the sugars but exhibited a considerable difference in their time-course patterns. When mice were refed with glucose, hepatic SREBP-1 gene expression was highest at 6 h and declined at 9 and 12 h postfeeding. In the livers of fructose-refed mice, the SREBP-1 mRNA level was at its peak at 12 h, exhibiting a delay in induction response compared with glucose. Livers refed with sucrose, which contains both glucose and fructose, showed a pattern between those with glucose and fructose. Plasma glucose and insulin levels were also compared. Glucose, the most potent insulin secretagogue, caused a higher response of both plasma glucose and insulin levels than fructose and sucrose refeeding, as expected. The overall induction patterns of SREBP-1 mRNA with each sugar did not necessarily follow the plasma insulin levels except in the case of glucose refeeding. Induction of FAS expression, a major target of SREBP-1, followed the changes in SREBP-1. Unexpectedly, glucose was a relatively weak activator of FAS mRNA compared with fructose and sucrose. In addition, the activation of hepatic glucokinase was highly insulin dependent.

Previous studies have focused primarily on insulin-induced expression of SREBP-1. To determine whether the carbohydrate regulation of the SREBP-1 was insulin dependent in vivo, STZ-administered mice were fasted for 24 h and refed either glucose or fructose for 12 h. Hepatic membrane-bound precursor forms and mature nuclear forms of SREBP-1 and -2 were compared and measured by immunoblot analysis (Fig. 2). When mice were fasted for 24 h, both membrane and nuclear forms of SREBP-1 were almost undetectable from either STZ-administered mice or nontreated controls. Refeeding with glucose or fructose for 12 h caused induction of SREBP-1 precursor protein in control mice, leading to incremental increases of nuclear proteins, a well-established observation in mice refed with a high-sucrose diet (13–15). Unexpectedly, in insulin-depleted mice, there was an even more profound increase in precursor SREBP-1 after glucose or fructose refeeding. Nuclear SREBP-1 protein was equally induced in STZ-treated mice. The induced isoform was SREBP-1c as depicted by a specific antibody, which is consistent with previous studies (14). There was no significant change in the precursor for SREBP-2. These results indicated that dietary glucose and fructose can highly induce active SREBP-1c protein independent of insulin.

Figure 3 shows time-course changes of plasma glucose and insulin levels in STZ-administered and nonadministered mice. All STZ-administered mice had undetectable levels of plasma insulin in any time point (1, 3, 6, and 12 h refeeding), confirming insulin depletion (Fig. 3B). Hyperglycemia was marked at a very early time point (1 h) and was sustained during the refeeding measured with glucose, sucrose, and fructose feeding, in this order of magnitude (Fig. 3A). Immunohistological examination of the pancreas showed that immunoreactive insulin in islets essentially disappeared after STZ administration, supporting complete depletion of plasma insulin (Fig. 3C). To assess a residual secretary function of β-cells in STZ-administered mice, we intraperitoneally administered glibenclamide to control and STZ-administered mice at the beginning of refeeding (1 h). The animals were further refed with glucose for 6 h for estimation of hepatic SREBP-1 and glucokinase mRNA levels (Fig. 3D). In control mice, glibenclamide treatment increased hepatic glucokinase expression as compared with nontreated mice. However, in STZ-administered mice, plasma insulin levels were still undetectable and hepatic glucokinase expression was not changed by glibenclamide treatment. These results strongly support that our STZ-administered mice had no residual secretory capacity of insulin and completely lost the effects of insulin.

Table 1 represents other parametric measurements of mice with STZ administration and of controls without treatment. STZ administration caused a significant decrease in body weight, with disappearance of fat, and a marked increase in plasma glucose. Plasma leptin levels were also very low, reflecting loss of adipose tissue. STZ administration also reduced plasma and hepatic triglycerides.

To determine the scope of insulin effects on gene expression patterns of carbohydrate/lipid metabolism, Northern blot analysis was performed on a variety of related genes in the livers of STZ-administered mice refed with each sugar at 6 and 12 h (Figs. 4 and 5). In concert with protein levels, the amount of hepatic SREBP-1 mRNA in STZ-administered mice was equally or even more highly induced by any sugar refeeding than in nontreated controls. The time course of changes in SREBP-1 mRNA levels by each sugar refeeding was similar between STZ-administered and control mice (Fig. 4). In contrast, there was no significant change in expression of SREBP-2 by fasting/refeeding or by STZ administration. As shown in Fig. 5A, expression of SREBP-1-sensitive lipogenic genes, such as ACC, FAS, SCD, ATP citrate lyase, S14, glucose-6-phosphate dehydrogenase, and 6-phosphoglucanate dehydrogenase, were steadily induced in a time-dependent manner by refeeding with any sugar, with the highest levels by sucrose seen at 12 h. Overall, STZ administration enhanced the induction of these genes. As for genes involved in glucose/fructose metabolism, the mRNA encoding glucokinase was reduced by STZ administration, especially in a fed glucose state (Fig. 5B). In contrast, the gluconeogenic enzyme mRNAs for PEPCK and glucose 6 phosphatase (G6Pase) were slightly increased by STZ administration at 6 and 12 h refeeding. In the glycolytic pathway, aldolase A did not show significant changes. Fructokinase, aldolase B, and liver-type pyruvate kinase, all of which are involved in lipogenesis from fructose, were not affected by STZ but showed a tendency of induction in a late stage of refeeding. The mRNA for IRS-2 increased slightly during fasting and dramatically after STZ administration, whereas there was no significant change in the mRNAs for the insulin receptor and IRS-1 (Fig. 5C). GLUT2 expression was increased in refed STZ-administered mice, assuring increased glucose uptake by hepatocytes in hyperglycemia. The mRNA for the cholesterologenic enzyme, 3-hydroxy-3-methylglutaryl CoA synthase, was also increased by STZ administration. 36B4 was used as a loading control.

To more precisely confirm insulin deficiency, we measured mRNA levels of glucokinase, PEPCK, SREBP-1, and IRS-2 at 3-h refeeding with glucose. At this early time point, SREBP-1 was not yet induced (Fig. 5D). Hepatic glucokinase gene expression in control mice was already highly induced but was almost completely suppressed by STZ administration. Conversely, high expression of PEPCK was already suppressed at 1-h refeeding, and this suppression was blunted in STZ-administered mice. Thus, at this time point, the expression of genes involved in glucose metabolism was under the control of insulin, as expected.

To determine whether any changes in SREBP-1 and lipogenic gene expression could be reversed, the STZ-administered mice were supplemented with insulin during glucose refeeding. As shown in Fig. 6, exogenous insulin completely normalized the plasma glucose levels of STZ-administered mice, but had no effect on the SREBP-1 mRNA level. Mature forms of SREBP-1c did not show significant changes by STZ administration or subsequent insulin supplementation, as estimated by immunoblot analysis. The mRNA levels of lipogenic enzymes, including ACC, FAS, SCD, and S14, were significantly elevated by STZ administration and, in mice given insulin, declined to the levels of non-STZ-administered controls. The predicted effects of insulin on hepatic gene expression were observed in induction of glucokinase and suppression of G6Pase. Plasma leptin levels were suppressed by STZ administration and partially restored by insulin supplementation (3.42 ± 0.40, 0.085 ± 0.026, and 1.92 ± 1.12 ng/ml for control, STZ, and STZ + insulin, respectively).

Since insulin dependence of SREBP-1 expression has been proposed from rat primary hepatocyte or STZ-administered rats (26), it is possible that the discrepancy in our data from mice could be due to a difference between species. To examine whether the regulation of the SREBP-1 gene expression in our protocol was insulin dependent in rats, we also performed Northern blot analysis in the livers of control and STZ-administered rats fasted for 48 h and refed either glucose, fructose, or sucrose for 24 h. As shown in Fig. 7, SREBP-1 expression was significantly decreased in the livers of STZ-administered rats, suggesting more insulin dependency in rat SREBP-1 expression. Expression of the lipogenic genes FAS and S14 was also repressed in the livers of STZ-administered rats. Thus, there is a considerable species difference in the dependency of hepatic SREBP-1 on insulin.

While previous reports have shown that insulin appears to induce SREBP-1 expression, our current data suggest that dietary activation of SREBP-1c in a refed state could be insulin independent in mouse livers at least. In fully refed states (12-h refeeding after 24 h fasting), SREBP-1c and its target lipogenic genes were highly induced by any sugar tested in insulin-depleted conditions by STZ administration. The similar induction of SREBP-1c and FAS expression was also observed in the STZ-administered mice refed with a regular diet (data not shown). Fructose can be efficiently taken up by hepatocytes without insulin (40). Therefore, fructose activation of SREBP-1c in normal mice is also supportive evidence for the insulin independence of SREBP-1c induction. Previous studies (29) suggesting that insulin regulates SREBP-1c expression were performed at short fasting times, when insulin dominates the metabolic fate of glucose. Lipogenesis is activated after glycogen synthesis is fully stored and excess energy is ready for lipid deposition. Insulin would be expected to facilitate this process. However, when expression of lipogenic genes was fully activated in glucose refeeding, plasma insulin levels were already decreasing (Fig. 1). Furthermore, previous studies suggest that chronic insulin exposure to hepatocytes causes secondary insulin resistance through IRS-2 repression, which is observed in a refed state (35,41). Considering this time lag in the process of these nutritional events, it is unlikely that the same insulin signal that mediates acute glucose metabolism, such as glycogen synthesis, also regulates SREBP-1c and lipogenic gene expression in a late phase of nutritional process. The insulin dependency of SREBP-1c expression in previous reports was all based on rats or rat primary hepatocytes. It is also possible that the discrepancy could be a species difference between rats and mice. We also estimated the hepatic SREBP-1c expression of refed STZ-administered rats. STZ administration significantly decreased SREBP-1c expression, suggesting more insulin dependency in rat SREBP-1c expression (Fig. 7).

We have shown here that SREBP-1c induction was correlated with plasma glucose levels rather than plasma insulin levels in both STZ-administered and nonadministered mice. Uptake of glucose by hepatocytes is mediated through GLUT2, which does not require insulin for activation (42), and was interestingly increased by STZ. Thus, cellular glucose availability is correlated with plasma glucose level. We and others (24,34) have postulated that glucose could induce hepatic SREBP-1c expression with the permissive action of insulin. The current data are consistent with this notion and, at least in a late fed phase of mice, insulin is not required for induction of hepatic SREBP-1 and lipogenic enzymes.

SREBP-1c has been suggested to be involved in metabolic derangements in the liver as a part of the pathogenesis of the metabolic syndrome (43). Beside insulin and glucose, a diet rich in fructose has been known to induce hyperinsulinemia, hyperlipidemia, and fatty liver (44–46). Furthermore, a high-fructose diet was reported to induce hepatic SREBP-1c mRNA (47). After uptake by hepatocytes through GLUT2, fructose is converted to fructose-1-phosphate by fructokinase and thereafter enters into the glycolytic pathway as triose phosphates catalyzed by aldolase B, bypassing an insulin-regulated step by phosphofructokinase-1. Fructokinase and aldolase B are not direct SREBP targets because they are not induced in the livers of SREBP transgenic mice (data not shown). However, both genes showed time-dependent induction at refeeding irrespective of STZ administration, which is evidence for the activity of this pathway in the refed condition. It is conceivable that in a refed state with fructose as the only sugar source, the triose phosphates could be used for both lipogenesis and glucose production to maintain blood glucose.

The activation of hepatic SREBP-1c and lipogenic genes by refeeding with pure glucose or fructose feeding illustrates two major metabolic pathways from these monosaccharides to lipogenesis, as is shown in Fig. 8. Both SREBP-1-activated pathways could bypass the phosphofructokinase-1-catalyzed step, which is a rate-limiting and insulin-regulated step under feedback regulation. The subsequent rate-limiting step, pyruvate kinase, although not an SREBP-1 target, was also highly induced in a refed state without insulin, presumably supported by concomitant induction of carbohydrate response element-binding protein, which was recently identified as a transcription factor for l-pyruvate kinase (48). These data indicate that glucose and fructose ingestion can very efficiently lead to the induction of lipogenic genes via these pathways without insulin.

Although the roles of SREBP-1c in the control of lipogenic enzymes have been well established, the current studies also support the presence of SREBP-1-independent control of lipogenic enzymes. STZ administration did not change nuclear SREBP-1 protein but still increased lipogenic enzymes such as FAS, ACC, SCD, and S14 (Fig. 6). Insulin supplementation in turn decreased expression of these enzymes to the levels of controls. These data suggest that changes in lipogenic enzymes related to insulin action may be caused by some factor other than SREBP-1.

Because fructose feeding gradually induced SREBP-1c expression, it could be postulated that glucose and fructose need to be metabolized by some common substance for activation of SREBP-1c in a cumulative fashion. Insulin could enhance this conversion from glucose, which explains why refeeding normal mice with glucose causes the most rapid and potent activation of SREBP-1c, with a corresponding insulin increment (Fig. 1). But even in insulin depletion, severe hyperglycemia in STZ-administered mice presumably compensates for the lack of this effect of insulin. The identity of the parameter or signaling molecule for SREBP-1c induction is currently an enigma. It could be some metabolite in the final common pathway, or this metabolic parameter could be related to an energy-sensing molecular ratio, such as ATP-to-AMP. Further investigations are needed for comprehensive elucidation of the regulation of the carbohydrate/lipid metabolism. It is also of great clinical importance as this energy-sensing mechanism is likely to be related to obesity and insulin resistance.

This work was supported by grants-in-aid from the Ministry of Science, Education, Culture, and Technology of Japan.