Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

New Zealand obese (NZO) mice exhibit a polygenic syndrome of obesity, insulin resistance, dyslipidemia, and hypertension that is similar to the human metabolic syndrome (Ortlepp et al., 2000). Like other rodents with morbid obesity, the strain exhibits impaired glucose tolerance and eventually develops a type 2 diabetes-like hyperglycemia and hypoinsulinemia. Thus, the NZO mouse is a suitable model for the identification of obesity and diabetes genes and for the characterization of their interaction (Leiter et al., 1998; Plum et al., 2000).

We have recently established a backcross model of NZO mice with the lean and atherosclerosis-resistant SJL strain (Ortlepp et al., 2000). The male NZO × F1 (SJL × NZO) backcross population (referred to as NSZO) is heterogeneous and includes normoglycemic/normoinsulinemic animals, insulin resistant animals with a compensatory hyperinsulinemia, and diabetic animals with hyperglycemia/hypoinsulinemia. This heterogeneity defines different stages in the development of diabetes and reflects the different genetic burden of the subgroups. A striking characteristic of the backcross is that only a few female mice develop hypoinsulinemia, despite a marked obesity and hyperinsulinemia. In a genome-wide scan of the NSZO backcross population, we identified a susceptibility locus for obesity/hyperinsulinemia (Kluge et al., 2000) and a separate locus for hyperglycemia/hypoinsulinemia that was contributed by the SJL genome (Plum et al., 2000); together, these loci are responsible for approximately 90% of the prevalence of diabetes in the backcross.

In addition to the genome-wide search for disease susceptibility loci, we have used the NSZO backcross population for identification of differences in hepatic gene expression that may be related to the metabolic abnormalities. With the use of cDNA arrays, we identified transcripts that were differentially expressed in liver of normoglycemic, normoglycemic/hyperinsulinemic, and diabetic NSZO mice. Among these, several cytochrome P450 (P450) and glutathione S-transferase (GST) isoforms were identified (W. Becker, R. Kluge, T. Kantmer, K. Linnartz, M. Korn, G. Tschank, L. Plum, K. Giesen, and H.-G. Joost, unpublished results). This finding is particularly interesting, because members of these families are involved in the extramitochondrial oxidation of fatty acids and in the metabolism of steroids. Fatty acids and/or steroids have been thought to be involved in the pathogenesis of insulin resistance and diabetes (Kahn and Flier, 2000; Masuzaki et al., 2001).

For more than a decade, experimental, type 1-like diabetes has been known to cause significant alterations in the expression of individual P450 isoforms in the rat (Favreau and Schenkman, 1988; Barnett et al., 1990, 1993; Donahue and Morgan, 1990; Cheng and Morgan, 2001). Streptozotocin-induced hypoinsulinemia has been reported to induce hepatic expression of CYP2B, CYP2E1, and CYP1A2 (Thummel and Schenkman, 1990; Barnett et al., 1993) and to suppress CYP2C11, CYP2C13, CYP2A2, and CYP3A2 (Thummel and Schenkman, 1990). In addition, induction of CYP4A isoforms in streptozotocin-induced diabetic male rats, presumably mediated by the peroxisome proliferator-activated receptor- (PPAR), has been reported (Kroetz et al., 1998).

In contrast to the large amount of information available on the effect of hypoinsulinemia on P450 expression, few studies have investigated the P450 expression in animals with a type 2 diabetes-like syndrome, and these studies have reported, in part, contradictory results (Cheng and Morgan, 2001). Barnett and coworkers (1993) found that type 2 diabetes had no effect on the microsomal activity of CYP1, CYP2B, CYP2E, CYP3, and CYP4A in the obese-hyperglycemic (ob/ob) mouse. In contrast, Enriquez and coworkers (1999) reported a decrease in CYP2E1 activity and an up-regulation in Cyp4a11 and Cyp4a14 mRNA in the same mouse model.

It seems reasonable to assume that these seemingly conflicting results are due to a varying expression of the enzymes at different stages of the development of insulin resistance and diabetes. Thus, we felt that a comprehensive comparison of the P450 and GST expression in mouse populations with different, well-defined metabolic abnormalities of glucose homeostasis was necessary. The objective of this study, therefore, was to compare hepatic P450 and GST expression in the three groups of NSZO mice that are normoglycemic/normolinsulinemic, hyperinsulinemic, or hyperglycemic/hypoinsulinemic. In addition, comparisons in expression were made between male and female NSZO mice in an attempt to find associations between gender-specific gene expression and the development of the type 2 diabetes-like syndrome.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. The research was approved by the Animal Experimentation Ethics Committee at the Regierungspräsidium Cologne, Germany. SJL and NZO mice were obtained from Bomholtgard (Ry, Denmark). Housing and generation of F1 hybrids (SJL × NZO) and back-crosses (NZO × F1) were as described previously by this group (Ortlepp et al., 2000; Plum et al., 2000). In brief, after weaning (3 weeks of age), mice were maintained on either a standard rodent food (1314) or a fat-enriched diet (C1057; Altromin, Lage, Germany). The high-fat diet contained 16% fat, 46.8% carbohydrates, 17.1% protein, and 15.4 kJ/g digestible energy. The standard diet contained 5% fat, 48% carbohydrates, 22.5% protein, and 12.5 kJ/g digestible energy. Mice were killed at the age of 22 weeks in isoflurane anesthesia by exsanguination. Blood was then collected, and livers were immediately excised and snap frozen in liquid nitrogen. Samples were stored at 70°C until required.

Serum Parameters and Plasma Insulin. Blood glucose, serum cholesterol, and serum triglycerides were measured by AutoAnalyzer (Johnson & Johnson, Neckargemünd, Germany). Plasma insulin was determined in duplicate by radioimmunoassay (Amersham Biosciences, Freiburg, Germany) with anti-rat insulin antiserum and ¹²⁵I-labeled rat insulin as tracer. Free and bound radioactivity were separated with an anti-IgG antibody and samples were measured for activity using gamma scintillation counting. Plasma free fatty acid (FFA) content was measured using a FFA determination kit (Roche Diagnostics, Mannheim, Germany) (Table 1).

Northern- and Dot-Blot Analysis. Total hepatic RNA was extracted according to the method of Chomczynski and Sacchi (1987). For Northern-blot analysis, RNA from individuals in each group were pooled according to total RNA concentration. Total RNA for each pooled group (15 µg/lane) was then separated by denaturing formaldehyde electrophoresis on 1% agarose gels and transferred by capillary blot to nylon membranes (Hybond N⁺; Amersham Biosciences) in 10× standard saline citrate (SSC). Equal loading of RNA was ascertained by ethidium bromide staining. For dot-blot analysis, 10 µg RNA from each individual animal was loaded directly onto nylon membranes with the aid of a commercial dot-blot apparatus (Invitrogen, Carlsbad, CA). After application of RNA, blots were then washed with 2× SSC and allowed to dry at room temperature for 1 h. For Northern and dot blots, the RNA was cross-linked to the membrane by irradiation under UV at 0.4 J/cm².

cDNA Probes. IMAGE cDNA clones (Lennon et al., 1996) were purchased from the Resource Center/Primary Database (Berlin, Germany). IMAGE clone ID and GenBank accession numbers for the probes used are shown in Table 2. Identity of all clones was verified by nucleotide sequencing. Plasmids were digested with appropriate restriction enzymes, and inserts were isolated from agarose gels with the help of a commercial kit (QIAGEN, Hilden, Germany). Probes were labeled with [-³²P]dCTP (Amersham Biosciences) by random oligonucleotide priming.

Preparation of Hepatic Cytosol and Microsomal Fractions. All procedures were carried out at 0 to 4°C. Livers from individual mice were homogenized in 0.1 M phosphate buffer (pH 7.4) with 1.15% KCl using a Potter-Elvehjem homogenizer and centrifuged at 9000g for 20 min. The supernatant was then centrifuged at 108,000g for 60 min. The resulting cytosolic fraction (supernatant) was distributed into individual eppendorf tubes, snap frozen in liquid nitrogen, and stored at 70°C for later use. The microsomal pellet was resuspended in the same buffer and centrifuged at 108,000g for 60 min. Phosphate buffer (0.1 M, pH 7.4) with 20% glycerol was used to resuspend the final microsomal pellet, and this was also snap frozen in liquid nitrogen and stored at 70°C for later use. Protein concentration of microsomal and cytosolic fractions was estimated using the method of Bradford (1976). Microsomes and cytosol from animals in the different groups were then pooled relative to their protein concentration for use in Western-blot analysis.

Western-Blot Analysis. SDS-polyacrylamide gel electrophoresis was performed using either a 10% (CYP4A, CYP2C, and CYP2B) or 12% (GSTµ) separating gel. Pooled samples (15 µg of protein/lane) were solubilized and heated at 95°C for 5 min. The proteins were electroblotted from the gel to a nitrocellulose membrane and were then visualized by Ponceau Red to confirm transfer and to ensure equal loading of proteins for each pooled sample. Membranes were blocked in 3% bovine serum albumin in Tris-buffered saline/Tween 20 for 30 min at room temperature. Immunoblot analysis was performed by incubation (3 h) with appropriate dilutions of primary antibody (1:1000, CYP2C and CYP4A; 1:4000, CYP2B; and 1:100, GSTµ) followed by incubation (45 min) with polyclonal horseradish peroxidase-conjugated anti-sheep antibody (CYP2B and CYP2C) or anti-rabbit antibody (CYP4A and GSTµ). Polyclonal antibodies to rat CYP4A1/4A2/4A3, CYP2B1/2B2, CYP2C12 (Chemicon International Inc., Temecula, CA), and GSTµ (DPC Biermann, Bad Nauheim, Germany) were used to detect mouse CYP4A, CYP2B, CYP2C, and GSTµ isoforms, respectively, in the pooled hepatic microsomal (P450) and cytosolic (GST) fractions for each mouse group. Bands were visualized by enhanced chemiluminescence.

Statistical Analysis. For mRNA species, the mean value obtained from the normoglycemic male high-fat animals was taken to be 100%. The mean values obtained from all other groups were then expressed relative to this. All data are expressed as mean ± standard error (S.E.). Overall differences between mouse groups for each transcript were determined by one-way ANOVA. When the analysis indicated an overall difference between the groups for each transcript, the data were corrected using Bonferonni`s post hoc analysis for multiple comparisons. Pair wise comparisons among glycemic status, sex, and diet were carried out. A P value of <0.05 was considered significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Characteristics of the Animals

The characterization of this population of NSZO backcross mice that had been fed on a fat-enriched diet will be reported elsewhere by our group (Plum et al., 2002). For the purpose of this study, three groups (normoglycemic, hyperinsulinemic, and diabetic) of selected animals were assembled according to their metabolic parameters. In addition, NSZO mice were fed on a standard diet, and one group of both males and females that were of normal blood glucose and plasma insulin were selected to test for diet effects. Table 1 summarizes the body weight and metabolic parameters measured in the NSZO mice that where selected for the present study.

The concentration of plasma FFAs, which are known inducers of certain P450 isoforms (e.g., CYP4A subfamily) (Tollet et al., 1994; Zangar and Novak, 1997), is also shown in Table 1. As expected, a significant increase in plasma FFAs was observed with a fat-enriched diet in both sexes (males, p = 0.004; females, p = 0.001). Hyperinsulinemia significantly reduced the concentration of FFAs in the plasma in males (p = 0.018) but failed to do so in female mice (p = 0.213). The level of FFAs in the plasma was the same between males and females when maintained on a fat-enriched diet. On a standard diet, the level of FFAs in the plasma of male mice was greater than in females.

Expression of Cytochrome P450 Isoforms

Effects of Hyperinsulinemia and Diabetes. The data obtained from hybridization of P450 probes to Northern blots of pooled RNA and dot blots of RNA from individuals from each group are presented in Table 3 and Figs. 1 and 2. From these results, it was possible to distinguish three groups of the investigated P450 isoforms according to their response to diabetes or hyperinsulinemia. The first group showed an increase in P450 expression in the liver of diabetic NSZO backcross mice. It also showed a slight increase, or no change, in expression in the liver of hyperinsulinemic animals. This pattern of expression was found for the female-specific isoforms Cyp2b9 and Cyp3a16 and also for Cyp4a14, which is involved in fatty acid oxidation and known to be induced by PPAR (Kroetz et al., 1998; Enriquez et al., 1999) (Table 3; Fig. 1).

View larger version (41K): [in this window] [in a new window]   Fig. 1.   Analysis of P450 expression in normoglycemic (N), diabetic (D), and hyperglycemic (H) NSZO mice fed on a fat-enriched or standard diet. Animals maintained on standard diet were of normal glucose status. Northern blots of pooled hepatic RNA from individuals from each group and bar diagrams from dot-blot analyses of individual animals from each group are shown. Data are expressed as a percentage relative to the male high-fat normoglycemic group (mean ± S.E.).

View larger version (47K): [in this window] [in a new window]   Fig. 2.   Analysis of Cyp2c expression in normoglycemic (N), diabetic (D), and hyperglycemic (H) NSZO mice fed on a fat-enriched or standard diet. Animals maintained on standard diet were of normal glucose status. Northern blots of pooled hepatic RNA from individuals from each group and bar diagrams from dot-blot analyses of individual animals from each group are shown. Data are expressed as a percentage relative to the male high-fat normoglycemic group (mean ± S.E.).

Sexual Dimorphic Expression and Dietary Effects. The expression of Cyp2b9, Cyp2c29, Cyp3a16, and Cyp7b1 was sexually dimorphic. With the exception of Cyp7b1, expression was lower in male mice compared with females. The dimorphism for Cyp2b9 could also be extended to the effect of physiological state, with the level of this transcript increased in hyperinsulinemic males yet unchanged in hyperinsulinemic females. Similarly, although the level of expression of the putative Cyp2c22 was the same in normoglycemic males and females, a decrease in Cyp2c22 mRNA levels was observed with hyperinsulinemia in male mice only. Diet also influenced the expression of Cyp1a2, Cyp3a16, and Cyp4a14 isoforms. Cyp4a14 mRNA levels were higher in both males and females receiving a fat-enriched diet. On the other hand, the levels of Cyp1a2 (males and females) and Cyp3a16 (males) mRNA were decreased by a fat-enriched diet compared with a standard diet (Fig. 1).

Glutathione S-Transferase Expression

The expression of both Gstm3 and Gstm6 mRNA in NSZO mice followed the third pattern of expression described in the previous section, with a reduction in mRNA levels for these two isoforms in diabetic animals. Hyperinsulinemia seemed to reduce this expression to an even greater extent (Fig. 3).

View larger version (52K): [in this window] [in a new window]   Fig. 3.   Analysis of Gst expression in normoglycemic (N), diabetic (D), and hyperglycemic (H) NSZO mice fed on a fat-enriched or standard diet. Animals maintained on standard diet were of normal glucose status. Northern blots of pooled hepatic RNA from individuals from each group and bar diagrams from dot-blot analysis of individuals from each group are shown. Data expressed as a percentage relative to the male high-fat normoglycemic group (mean ± S.E.).

The level of expression of Gstt2 was low in all animal groups. However, the pattern of expression observed for this transcript did not conform with any of the three patterns identified above. This isoform was up-regulated in diabetic male animals and down-regulated with hyperinsulinemia for both male and females (Fig. 3).

No change in Gsta2 mRNA was detected with the onset of diabetes or hyperinsulinemia. No differences between males and females were found for all of the GST isoforms studied. A fat-enriched diet seemed to cause some down-regulation for Gsta2 (males only).

Western-Blot Analysis

Western-blot analysis of microsomal (P450) and cytosolic (GST) protein was carried out in an attempt to correlate protein concentration to those results obtained at the mRNA level. Immunoblots using polyclonal antibodies to rat CYP2B1/2B2 and CYP4A1/4A2/4A3 supported the results obtained from the hybridization of Cyp2b9 and Cyp4a14 cDNA, respectively, to Northern and dot blots of RNA. Two CYP2B-related immunoreactive bands were detected, and the intensity of the bands was in accordance with that found at the mRNA level (Fig. 4). A clear sexual dimorphism was apparent, with the level of CYP2B-related protein considerably higher in females compared with males, however, changes were only observed in males with the onset of diabetes and hyperinsulinemia. Similarly, two CYP4A-related immunoreactive bands, one major and one minor, were detected in both males and females (Fig. 4), and the bands were of greater intensity in microsomes from animals with diabetes. On the other hand, although Cyp4a14 mRNA levels were higher in both males and females receiving a fat-enriched diet (Fig. 1), this result was not supported at the protein level (Fig. 4).

View larger version (84K): [in this window] [in a new window]   Fig. 4.   Hepatic protein levels of CYP2B, CYP2C, CYP4A, and GSTµ in normoglycemic (N), diabetic (D), and hyperglycemic (H) NSZO mice fed on a fat-enriched or standard diet. Western blots of pooled hepatic microsomes (P450) or cytosol (GST) were probed with antibodies directed against rat CYP4A, CYP2B, CYP2C, and GSTµ.

Western-blot analysis using a CYP2C antibody showed a decrease in CYP2C-related immunoreactive protein in liver microsomes from diabetic mice (Fig. 4), however, this change did not reflect the highly significant decrease found in mRNA for Cyp2c22. There are several known murine isoforms belonging to this subfamily (at least six). In this study, the mRNA levels of three classified Cyp2c genes (Cyp2c29, Cyp2c39, and Cyp2c40) and two novel Cyp2c isoforms (Cyp2c22 and Cyp2c23) were examined, and a considerable amount of variability was found in the expression between each transcript. Therefore, because of the unspecificity of the polyclonal rat CYP2C antibody used, it is likely that a number of CYP2C isoforms would have contributed to the one immunoreactive band.

Western-blot analysis showed two GSTµ-related immunoreactive bands. Compared with the changes in mRNA levels, only minor differences in intensity were found between each group (Fig. 4). There are four known isoforms belonging to the GSTµ subfamily in the mouse (GSTµ1, GSTµ2, GSTµ3, and GSTµ6). Therefore, as for CYP2C, it is likely that these other GSTµ isoforms would have contributed to the two bands detected.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study provides the first comprehensive examination of P450 and GST expression in a polygenic mouse model of type 2 diabetes. Our findings indicate that hyperinsulinemia and type 2 diabetes-like hyperglycemia greatly affect the expression of a number of P450 and GST isoforms. Specifically, expression of Cyp2b9, Cyp3a16, Cyp4a14, and Gstt2 was increased, whereas expression of Cyp1a2, Cyp2c22, Cyp2c29, Cyp2c40, Gstm3, and Gstm6 was suppressed in diabetes. Hyperinsulinemia altered expression of Cyp3a16, Cyp2b9, Gstm3, and Gstm6 in male mice. Although there are a number of studies in the literature reporting changes in selected P450 and GST isoforms with diabetes in the leptin deficient ob/ob mouse or fa/fa rats (Barnett et al., 1992; Enriquez et al., 1999; Roe et al., 1999; Liang and Tall, 2001), this is the first study to show that a broad range of P450 and Gst genes are either up-regulated or down-regulated by the pathophysiological alterations associated with diabetes and insulin resistance.

Reduced Expression of CYP2C Isoforms in Diabetes. One of the most notable and novel findings to emerge from this study was the degree of down-regulation of the putative mouse Cyp2c22 mRNA in diabetic animals. This isoform has not yet been described in the mouse. The cDNA clone used here corresponds to a mouse transcript (UniGene entry Mm.29119; www.ncbi.nlm.nih.gov/UniGene/) that shows 87% identity to the rat Cyp2c22 nucleotide sequence. Emi et al. (1990) first described this isoform in primary cultures of rat hepatocytes and found that it was highly induced after plating hepatocytes on collagen-coated culture dishes. They speculated that CYP2C22 may be involved in the metabolism of steroid hormones because of its similarity with sex-specific isoforms of CYP2C (e.g., CYP2C11 and CYP2C12). To date, no further studies have reported on the regulation of this isoform in the rat. In NSZO mice, Cyp2c22 showed no sexually dimorphic pattern of expression, whereas the related Cyp2c genes, Cyp2c29 and Cyp2c40, were predominantly expressed in females. Transcript levels of these genes were also down-regulated in diabetes and insulin resistance, but to a much lower extent than that found for Cyp2c22.

Role of Fatty Acids and PPAR in the Expression of P450 Isoforms. The CYP4A subfamily includes isoforms that are important fatty acid -hydroxylases and are well known to be regulated through the activation of PPAR (Lee et al., 1995; Barclay et al., 1999). Our finding that Cyp4a14 was dramatically induced with diabetes and by high-fat diet consequently was expected and has been demonstrated previously in rats and mice (Barnett et al., 1990; Kroetz et al., 1998; Enriquez et al., 1999). Interestingly, induction of Cyp2b9 and Cyp3a16, which are not considered to be regulated by PPAR, was also found. Zangar and Novak (1997) reported induction of CYP2B and CYP4A isoforms by straight-chained saturated fatty acids. Therefore, as for Cyp4a14, the induction of Cyp2b9 and Cyp3a16 observed in this study may be due to the increased FFA.

Sexually Dimorphic Expression of P450 Isoforms in Hyperinsulinemia and Diabetes. It has long been known that the expression of many P450 isoforms, including Cyp2b9 and Cyp3a16, are sexually dimorphic in mice, with expression generally greater in females compared with males (Shapiro et al., 1995). The reasons for these dimorphisms are not fully understood, although sex-dependent patterns of growth hormone secretion play an important role (Davey et al., 1999). Basal levels of Cyp2b9 and Cyp3a16 mRNA in male mice were a fraction of that observed in females, and it was interesting to find that expression of these two isoforms was greater in diabetic and hyperinsulinemic male mice only. Western-blot analysis also showed an increase in protein levels of CYP2B-related isoforms in males, in support of that found at the mRNA level. Sakuma and coworkers (2001) observed this male-only induction of Cyp2b9 in streptozotocin-induced diabetic mice. They speculated that induction of Cyp2b9 was due to changes of growth hormone secretion. No other studies have reported on this phenomenon for Cyp3a16. It may be that common sexually dimorphic characteristics which are altered in diabetes and hyperinsulinemia, such as growth hormone secretion or sex hormone levels, are responsible for the regulation of these two isoforms.

Alterations in the Expression of GST Isoforms. In addition to P450 genes, a range of GST isoforms were examined due to the importance of these enzymes in the elimination of xenobiotics and in their role as deactivators of reactive intermediates produced in phase I (P450) metabolism. Alterations in GST activity and protein levels have been reported in streptozotocin-induced diabetic rats (Raza et al., 2000), with levels generally reduced with diabetes. No studies have investigated the effect of diabetes and hyperinsulinemia on the expression of Gst genes in mice. Barnett et al. (1992) did measure the overall GST activity, however, the range of GST isoforms detected with these assays is somewhat broad. From the results presented in this study, differential regulation of individual GST isoforms was evident, with a reduction in the level of Gstm mRNA, an increase in Gstt2 mRNA level, and no change in the amount of Gsta mRNA with the onset of diabetes. Moreover, expression of GSTµ3 and GSTµ6 was also reduced in hyperinsulinemic males, the same pattern of changes as found for CYP1A2 and CYP7B1. Because this expression pattern is not correlated with serum levels of FFA and insulin, it seems to reflect an early alteration in the course of the disease and may be related to the progression of the syndrome from insulin resistance to the type 2-like diabetes.