The Journal of Nutrition Vol. 128 No. 11 November 1998, pp. 1984-1988

The Role of Lipogenesis in the Development of Obesity and Diabetes in Israeli Sand Rats (Psammomys obesus)^1,2

Paul A. Lewandowski, David Cameron-Smith, Colleen J. Jackson, Eva R. Kultys, and Greg R. Collier³

School of Nutrition and Public Health, Deakin University, Geelong, Victoria 3217, Australia

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Obesity and diabetes in Israeli sand rats, Psammomys obesus, occur with the sequential transition of animals from normal insulin sensitivity to impaired insulin sensitivity, accompanied by increased adiposity, prior to insulin resistance and obesity, in a manner similar to susceptible human populations. The current study was designed to examine the role of de novo lipid synthesis in the development of excessive weight gain in P. obesus. Sand rats were classified at 12 wk of age into three groups: A, normoglycemic normoinsulinemic; B, normoglycemic hyperinsulinemic; C, hyperglycemic hyperinsulinemic, based on glucose and insulin responses in fed sand rats. Body weight, liver weight, white adipose tissue (WAT) mass and food intake were significantly elevated in Group C compared to Group A (P < 0.05). Lipogenic rate was measured by the amount of ³H incorporated into subscapular brown adipose tissue (BAT), epidiymal WAT and liver per hour, from sand rats with and without access to food. No difference in lipogenic rate was found between the groups in BAT, indicating that this tissue is of minor importance in whole body lipogenesis in P. obesus. In the WAT there was a greater lipogenic rate with the development of obesity and hyperinsulinemia (Group B vs. Group A) but no difference in the liver. However, the onset of hyperglycemia (Group C) further stimulated WAT lipogenesis and initiated increased hepatic lipogenesis, both of which contributed to the pre-existing obesity. This study suggests that elevated lipogenesis is not the primary cause of obesity in P. obesus, as lipogenic rate only markedly increases after obesity is already present in hyperglycemic animals.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Elevated lipogenesis was demonstrated in several animal models of obesity and non-insulin dependent diabetes mellitus (NIDDM)⁴. Simple genetic rodent models of obesity and NIDDM, including ob/ob mice and db/db rats, demonstrate excessive rates of de novo lipid synthesis, which provide an important pathway for the synthesis and subsequent excessive accumulation of lipid from nonlipid precursors in these rodents (Mercer and Trayhurn 1983, Trayhurn and Wusteman 1990). Psammomys obesus (Israeli sand rat) is a herbaceous gerbil, common to semiarid regions of the Middle East. When maintained in captivity and given free access to a nonpurified diet in the laboratory, P. obesus displays heterogeneous glucose and insulin levels, ranging from healthy animals with normoglycemia and normoinsulinemia to obese diabetic animals with persistent hyperglycemia and hyperinsulinemia (Barnett et al. 1995, Shafrir 1992).

Elevated activities of enzymes central in the lipogenic pathway, including acetyl-CoA carboxylase (ACC), fatty acid synthetase (FAS), NADP-malate dehydrogenase (NADP-MD) and pyruvate dehydrogenase in normoglycemic P. obesus, were compared with albino rats (Kalderon et al. 1983), implicating increased lipogenesis as a contributing factor to the increased propensity for excessive weight gain in this rodent model. Further studies in hyperglycemic and hyperinsulinemic P. obesus showed that the rate of the lipogenic enzymes ACC and NADP-MD is increased compared with normoglycemic normoinsulinemic animals (Kalderon et al. 1986). However it is unclear from these studies whether both hyperinsulinemia and hyperglycemia are necessary for the demonstration of increased lipogenesis.

The development of diabetes and obesity in P. obesus, like that of susceptible human populations, demonstrates a progression from normoinsulinemia and normoglycemia in a percentage of the animals to a stage of impaired insulin sensitivity characterized by hyperinsulinemia and the maintenance of normoglycemia. A further smaller percentage of the animals progresses to frank insulin resistance with hyperinsulinemia and hyperglycemia. These studies will determine whether elevated lipogenesis is present only in those animals with hyperinsulinemia and hyperglycemia or whether elevated lipogenesis is a contributing factor in the development of obesity and insulin resistance in P. obesus demonstrating impaired insulin sensitivity alone.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

Animals.  Male P. obesus were obtained at weaning (4 wk) from an outbred colony, maintained at Deakin University (Geelong, Victoria, Australia) (Barnett et al. 1994) with free access to food and water for 12 wk; the diet consisted of 57.3% of carbohydrate, 15.0% of fat and 27.6% of protein, expressed as a percentage of total energy (ARM Cubes, Clark, King and Co., NSW, Australia). The high-energy density of the diet, relative to the herbivorous diet that P. obesus subsists on in its native habitat, induced weight gain and glucose intolerance in a proportion of animals (Barnett et al. 1995).

Classification.  At age 12 wk, P. obesus were classified as described previously (Barnett et al. 1994). Group A (n = 12) were normoglycemic and normoinsulinemic (glucose <7.0 mmol/L, insulin <20 pmol/L), Group B (n = 14) were normoglycemic and hyperinsulinemic (Glucose <7.0 mmol/L, insulin 20 pmol/L) and Group C (n = 18) were hyperglycemic and hyperinsulinemic (glucose 7.0 mmol/L, insulin 20 pmol/L).

Lipogenesis.  The rate of lipogenesis was measured in vivo as described previously (Hems et al. 1975, Stansbie et al. 1976). Food was removed from all animals 6 h prior to commencement of the experiment. Psammomys obesus were then randomly assigned to two groups, with the first group receiving ³H₂O i.p. (37 MBq/100 g) (Dupont NEN Products, Boston, MA) and housed without food for another hour. The second group were supplied free access to a nonpurified diet for 2 h and food intake estimated. After 1 h, access to the food was interrupted briefly with the administration of ³H₂O (37 MBq/100 g) i.p., prior to animals being returned to the nonpurified diet. In all groups a final blood sample was taken immediately prior to killing to determine blood glucose, insulin, triglyceride concentrations and the specific activity of ³H in plasma (Hems et al. 1975). All animals were then killed by cervical dislocation and the subscapular brown adipose tissue (BAT), epididymal white adipose tissue (WAT) and liver were excised. These tissues were frozen in liquid nitrogen and stored at 20°C until required for analysis. The amount of plasma ³H and the quantity of ³H incorporated into the tissues were determined by liquid scintillation counting. Rates of lipogenesis were calculated as µg-atoms of ³H incorporated per hour per gram of tissue (µg ³H incorporated · h¹ · g tissue¹) and mg-atoms of ³H incorporated per hour per total tissue (µg ³H inc.h¹) (Hems et al. 1975).

Biochemical analyses.  Blood was collected from the tail vein in heparinized microtubes, and blood glucose was immediately determined by an enzymatic glucose analyzer (2300 Stat Plus; Yellow Springs Instruments, Yellow Spring, OH). Plasma insulin levels were determined by radioimmunoassay using a human primary antibody (Phadeseph; Kabi Pharmacia Diagnostics, Sweden). Plasma triglyceride levels were measured colormetrically (Boehringer-Mannheim, Mannheim, Germany).

Statistical analysis.  All results are expressed as mean ± SEM. Plasma glucose and plasma insulin concentrations were analyzed by one-way analysis of variance ANOVA. The lipogenic rates were analyzed by a two-way ANOVA. No diabetic or feeding status interaction existed. For the other variables, the differences were analyzed by one-way ANOVA followed by a Tukey's test for multiple comparisons (Snedecor and Cochran 1980). Simple linear regression analysis was used to determine if glucose or insulin concentrations were associated with lipogenesis rate in either the food withheld or fed conditions. A P-value of <0.05 was considered statistically significant. All analyses were performed using SPSS version 6.1 software (SPSS Inc., Chicago, IL).

	RESULTS

Abstract Introduction Methods Results Discussion References

Whole blood glucose concentrations were significantly greater in Group C than in Groups A and B (P < 0.001, Fig. 1). Plasma insulin levels were significantly greater in Groups B and C than in Group A (P < 0.001, Fig. 1). Body weights and liver weights of Groups B and C were significantly greater than Group A. WAT mass and relative food intake were significantly higher in Group C than Group A (P < 0.05, Table 1).

View larger version (18K): [in this window] [in a new window]   Fig 1. Blood glucose and plasma insulin concentrations, in fed sand rats that were normoglycemic and normoinsulinemic (glucose < 7.0 mmol/L, insulin < 20 pmol/L) (A), n = 6 normoglycemic and hyperinsulinemic (glucose < 7.0 mmol/L, insulin  20 pmol/L), (B) n = 7, OV hyperglycemic and hyperinsulinemic (glucose  7.0 mmol/L, insulin  20 pmol/L) (C) n = 9 at 12 wk of age. * Significantly different from Group A (P < 0.001), Significantly different from Group B (P < 0.001). All values are expressed as means ± SEM.

Lipogenic rate in all tissues studied was greater in P. obesus with free access to food 2 h prior to lipogenic determinations, suggesting that postprandial insulin secretion or hyperglycemia stimulated lipogenesis in all animals. (Tables 2-4)

Lipogenesis in BAT did not differ among groups (Table 2). However in epididymal WAT the three groups differed from one another, irrespective of feeding status (Table 3). Lipogenesis in WAT was greater in Group C than in Group B in Group B compared to Group C. In the fed P. obesus, lipogenic rates in WAT were positively correlated with circulating glucose concentrations when expressed as lipogenic rate per gram and/or as the absolute rate (P < 0.005, Table 5). Lipogenic rate in the fed sand rats was also positively correlated with circulating insulin concentration for total WAT (P < 0.05, Table 5). The positive relationship between lipogenic rate and insulin was found to be present also in sand rats that were not fed (P < 0.01, Table 5).

In liver, Group C that had food withheld had a significantly greater lipogenic rate when expressed per gram or whole tissue than the other two groups (P < 0.05, Table 4). These differences in lipogenic rate per whole tissue significantly correlated with circulating insulin levels (P < 0.05, Table 5). When hepatic lipogenic rate was measured in sand rats that had free access to food, Group C was significantly greater than Group A (P < 0.05, Table 4). These differences in lipogenic rate correlated with glucose when expressed per gram and with insulin when expressed per total liver (P < 0.05, Table 5).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Consistent with our previous studies, obese diabetic P. obesus had elevated body weight, liver mass, WAT weight and more food consumption compared with lean healthy controls. To examine whether elevated lipogenesis was a major contributing factor in the excessive body fat accumulation in the obese diabetic P. obesus, lipogenesis rates in various tissues were measured in vivo in fed and food-deprived animals. BAT mass and the rate of lipogenesis in this tissue did not differ among the groups, suggesting that elevated BAT lipogenesis is unlikely to explain the development of obesity in this animal model. However, lipogenic rates in epididymal WAT in Group B (obese) and Group C (obese diabetic) P. obesus were elevated, which correlated with plasma glucose concentrations.

The rate of lipogenesis was significantly greater than Group A only in Group C, obese-diabetic animals in which hyperinsulinemia and hyperglycemia are evident. Impaired insulin sensitivity in Group B animals, in which the hyperinsulinemia is sufficient to maintain normoglycemia, did not demonstrate elevated liver lipogenesis in either the fed or food-deprived states when compared to the lean normoinsulinemic Group A P. obesus. Yet these insulin-resistant animals had greater body weight and fat mass than normoinsulinemic P. obesus suggesting that increased liver lipogenesis is not necessary for excessive weight gain.

Previous in vitro studies showed the lipogenic rate of isolated hepatocytes from ventromedial hypothalumus lesioned mice is stimulated by the presence of glucose (Katz et al. 1977). A similar effect was demonstrated in isolated hepatocytes (Clarke et al. 1990) and in cultured adipocytes (Foufelle et al. 1992) where glucose alone was shown to stimulate the gene expression of the key lipogenic enzymes, FAS and ACC. In contrast, other investigators reported that in both isolated adipocytes (Pape et al. 1988) and hepatocytes (Fukuda and Iritani 1983) simultaneous additions of glucose and insulin to the incubation media are necessary to stimulate the expression of FAS and ACC. The current study provides evidence that in vivo, hyperinsulinemia can increase lipogenic rate only in WAT, which is further elevated by the onset of hyperglycemia. In this respect liver lipogenesis differed from that in WAT in that hyperinsulinemia was insufficient to stimulate lipogenesis. In this obese and diabetic rodent model, elevated hepatic lipogenesis was not evident until significant increases in both plasma insulin and blood glucose were evident.

The requirement for jointly elevated insulin and glucose concentrations has important implications for our understanding of the etiology of obesity in this polygenetic rodent model. Elevated fat pad weight and greater whole body weight, together with a slightly greater food consumption (11.5 vs. 9.8 mg/g body weight, P > 0.05), are apparent in the sand rats with impaired insulin sensitivity, characterized by hyperinsulinemia and normoglycemia (Group B) (Barnett et al. 1995). Yet despite the greater weight gain, the animals with impaired insulin sensitivity had significantly elevated rates of lipogenesis in WAT but not in liver. Earlier studies in P. obesus suggested that this increase in WAT lipogenesis is of secondary importance to the changes in hepatic lipogenic rate (Kalderon et al. 1986), which implies that increased de novo synthesis of lipids from nonlipid precursors is not the only factor involved in the development of obesity in this rodent model. However, with the development of insulin resistance, characterized by hyperglycemia and hyperinsulinemia, the expression of lipogenic enzymes is elevated in the liver and further stimulated in WAT, increasing lipogenesis, which acts to further exacerbate the accumulation of excess body fat.

In the obese, hyperinsulinemic and hyperglycemic P. obesus, lipogenesis was elevated in both the liver and WAT, which is consistent with previous studies in a variety of monogenetic and chemically-induced rodent models of obesity and diabetes that have reported increased lipogenesis in liver and WAT of mature animals (Cooney et al. 1989, Mercer and Trayhurn 1983, Trayhurn and Wusteman 1990). A transitory increase in BAT was demonstrated at weaning or immediately after chemical inducement in these other animal models (Cooney et al. 1989, Mercer and Trayhurn 1983, Trayhurn and Wusteman 1990). This transitory increase in lipogenesis may be present in P. obesus at weaning, although the current study measurement was not made in sand rats of this age. However, at maturity, P. obesus have markedly lower rates of liver lipogenesis compared with other rodents (Cooney et al. 1989, Mercer and Trayhurn 1983, Trayhurn and Wusteman 1990), while still demonstrating a normal although blunted activation in response to food. Lipogenic rate approximately doubled in fed compared to food-deprived sand rats, a difference that is markedly less than what was reported in rodents (Cooney 1989). These findings further demonstrate that elevated lipogenesis in P. obesus is not the only contributing factor in the development of obesity and diabetes.

In summary, lipogenic rate was measured in BAT, WAT and liver in lean, obese and obese diabetic P. obesus. No difference in lipogenic rate was found in BAT, indicating that it is unlikely to influence the nonobese or obese state in P. obesus. In the WAT there was an increase in lipogenic rate with the development of hyperinsulinemia; the onset of hyperglycemia stimulated a further increase in lipogenesis. In liver, lipogenesis increased only after the onset of hyperglycemia with preexisting hyperinsulinemia and pronounced weight gain. This increase in lipogenesis was not present with hyperinsulinemia alone and modest weight gain. This study demonstrates that hyperlipogenesis may not be the only cause of obesity in P. obesus. It only occurs after obesity is already present and merely accentuates a pre-existing condition.

	FOOTNOTES

Manuscript received 3 July 1997. Initial reviews completed 27 December 1997. Revision accepted 10 August 1998.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Barnett M., Collier G. R., Collier F. M., Zimmet P., O'Dea K. A cross-sectional and short-term longitudinal characterisation of NIDDM in Psammomys obesus. Diabetologia 1994; 37:671-676[Medline]
• Barnett, M., Collier, G. R., Zimmet, P. & O'Dea, K. (1995) Energy intake with respect to the development of diabetes mellitus in Psammomys obesus. Diabetes, Nutr. Metab. 8:
• Clarke S. D., Armstrong M. K., Jump D. B. Nutritional control of rat liver fatty acid synthase and S14 m RNA abundance. J. Nutr. 1990; 120:218-224
• Cooney G. J., Vanner M. A., Nicks J. L., Williams P. F., Caterson I. D. Changes in lipogenic response to feeding of liver, white adipose tissue and brown adipose tissue during the development of obesity in the gold thioglucose injected mouse. Biochem. J. 1989; 259:651-657[Medline]
• Foufelle F., Gouhot B., Pegorier J., Perdereau D., Girard J., Ferre P. Glucose stimulation of lipogenic enzyme gene expression in cultured adipose tissue. J. Biol. Chem. 1992; 267:20543-20546[Abstract/Free Full Text]
• Fukuda H., Iritani N. Effect of insulin-treatment on age-dependent decrease of lipogenic enzyme induction. BBA 1983; 760:149-153
• Hems D., Rath E., Verrinder T. Fatty acid synthesis in liver and adipose tissue of normal and genetically obese (ob/ob) mice during the 24-hour cycle. Biochem. J. 1975; 150:167-173[Medline]
• Kalderon B., Adler J. H., Levy E., Gutman A. Lipogenesis in the sand rat (Psammomys obesus). Am. J. Physiol. (Endocrinol. Metab.) 1983; 244:E480-E486[Abstract/Free Full Text]
• Kalderon B., Gutman A., Levy E., Shafrir E., Adler J. H. Characterization of stages of development of obesity-diabetes syndrome in sand rat (Psammomys obesus). Diabetes 1986; 35:717-724[Abstract]
• Katz J., Wals P., Golden S., Goldman J., Bernardis L. Lipogenesis by hepatocytes of rats with hypothalamic obesity. Horm. Metab. Res. 1977; 9:59-63[Medline]
• Mercer S. W., Trayhurn P. Developmental changes in fatty acid synthesis in intrascapular brown adipose tissue of lean and genetically obese (ob/ob) mice. Biochem. J. 1983; 212:393-398[Medline]
• Pape M., Lopez-Casillas F., Kim K. Physiological regulation of acetyl-CoA carboxylase gene expression: effects of diet, diabetes and lactation on acetyl-CoA carboxylase mRNA. Arch. Biochem. Biophys. 1988; 267:104-109[Medline]
• Shafrir E. Animal models of noninsulin dependent diabetes. Diabetes/Metab. Rev. 1992; 8:179-208
• Snedecor, G. & Cochran, W. (1980). Statistical Methods, 7th ed., Iowa State University Press, Ames, IA.
• Stansbie D., Brownsey R. W., Crettaz M., Denton R. M. Acute effects in vivo of anti-insulin serum on rates of fatty acid synthesis and activities of acetyl-coenzyme A carboxylase and pyruvate dehydrogenase in liver and epididymal adipose tissue of fed rats. Biochem. J. 1976; 160:413-416[Medline]
• Trayhurn P., Wusteman M. C. Lipogenesis in genetically diabetic (db/db) mice: developmental changes in brown adipose tissue and the liver. Biochim. Biophys. Acta 1990; 1047:168-174[Medline]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences