Transgenic Overexpression of Leptin Rescues Insulin Resistance and Diabetes in a Mouse Model of Lipoatrophic Diabetes

We crossed male A-ZIP/F-1 mice (A-ZIPTg/+) on the FVB/N background (22) with female transgenic skinny mice (LepTg/+) on the C57BL/6J background (9,12) to obtain F1 mice of four genotypes; +/+, LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ on a mixed FVB/N and C57BL/6J background. Genotypes were determined by polymerase chain reaction on tail genomic DNA with transgene-specific primers (9,22). The A-ZIP/F-1 mice on the FVB/N background were used for leptin injection experiments. Mice were housed in a temperature-, humidity-, and light-controlled room (12-h light/12-h dark cycle) with free access to water and standard rat diet (352 kcal/100 g) (CE-2; Japan CLEA, Tokyo). Male mice were used in the studies reported here. Animal care and all experiments were conducted in accordance with the institutional guideline of the Kyoto University Graduate School of Medicine.

Body weight was measured daily, beginning at 3 weeks of age. Food intake was measured daily for 7 days in 15-week-old animals.

Northern blot analysis using total liver RNA was performed using a mouse leptin cDNA probe, as previously described (9).

Plasma was obtained from the retro-orbital sinus of fed mice at 9:00 a.m. using EDTA as the anticoagulant. Plasma leptin concentrations were determined by radioimmunoassay (RIA) (Linco Research Immunoassay, St. Louis, MO). Concentrations of plasma glucose, insulin, triglycerides, and FFAs were determined by, respectively, the glucose oxidase method with a reflectance glucometer (One Touch II; Lifescan, Milpitas, CA), RIA with rat insulin standards (Linco Research Immunoassay), and enzymatic kits (Triglyceride G-test and NEFA C-test; Wako Pure Chemical, Osaka, Japan).

Mice aged 15 weeks were killed and skinned, and interscapular brown adipose tissue (BAT), liver, and pancreas were preserved in Bouin’s fixative. Paraffin-embedded sections 5-μm thick were stained with hematoxylin and eosin and examined by light microscopy.

Glucose tolerance tests were performed after an overnight fast by intraperitoneal injection of 1.5 mg/g glucose. For insulin tolerance tests, fed animals were injected intraperitoneally with 1.5 mU/g human regular insulin (Novolin R; Novo Nordisk, Copenhagen). Tail-vein blood samples were obtained before and 15, 30, 60, 90, and 120 min after the injection.

A-ZIPTg/+ mice were given the amount of food (divided in two portions) consumed by ad libitum–fed LepTg/+:A-ZIPTg/+ mice on the previous day. Pair-feeding was performed after weaning for 12 weeks. The epididymal white adipose tissue (WAT), interscapular BAT, liver, and pancreas were obtained from 15-week-old LepTg/+:A-ZIPTg/+ and pair-fed A-ZIPTg/+ mice and were examined histologically.

On the first day of the experiment, infusion with recombinant mouse leptin (Genzyme/Techne, Minneapolis, MN) (5 μg/g per day) or vehicle (saline) was begun in 15-week-old A-ZIP/F-1 mice using two osmotic minipumps (Alzet model 2001; Alza, Palo Alto, CA) per mouse (200-μl pump × 2) delivering 2 μl/h. Mice were anesthetized with 50 mg/kg i.p. pentobarbital (Nembtal), and two osmotic minipumps were implanted subcutaneously in the midscapular region. During the 6-day infusion experiment, body weight, food intake, and plasma parameters were measured. On the sixth day of the experiment, mice were killed and livers were preserved for microscopic examinations.

Data were expressed as the means ± SE. Comparison among genotypes was assessed by analysis of variance (ANOVA) and completed by Fisher’s probable least-significant difference test, as required. Comparison among genotypes and treatments were assessed by two-way ANOVA. If significant interactions among genotypes and treatments were found, one-way ANOVA among all groups was performed.

Figure 1A shows the appearance of +/+, LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice at 15 weeks of age. As reported (9), LepTg/+ mice were thinner than +/+ mice, and the A-ZIPTg/+ animals had distended abdomens (22). The gross appearance of the LepTg/+:A-ZIPTg/+ mice was similar to that of the LepTg/+ mice and not the A-ZIPTg/+ animals. The body length of A-ZIPTg/+ mice (88.1 ± 0.05 mm, n = 10) was longer than that of LepTg/+ or LepTg/+:A-ZIPTg/+ mice (83.4 ± 0.04 mm and 84.0 ± 0.18 mm, n = 10, respectively). After removal of the skin, no subcutaneous WAT was visible in LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice, in contrast to the ample WAT visible in the +/+ controls. The amount of interscapular BAT was reduced in LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice (Fig. 1B).

Northern blot analysis of liver mRNA revealed a significant amount of leptin message in the LepTg/+ and LepTg/+:A-ZIPTg/+ mice but not in the +/+ and A-ZIPTg/+ animals (Fig. 1C). Plasma leptin concentrations were significantly elevated in the LepTg/+ and LepTg/+:A-ZIPTg/+ mice (31 and 52 ng/ml, respectively) relative to the +/+ animals (7.6 ng/ml) (Fig. 1D). In contrast, leptin concentrations were significantly reduced in A-ZIPTg/+ mice (1.8 ng/ml).

Figure 1E shows the growth curves for the transgenic mice. At 3 weeks of age, there was no significant difference in body weight among the four genotypes. However, by 4 weeks of age and thereafter, the LepTg/+ were significantly smaller than the +/+ controls. In contrast, starting at week 8, the A-ZIPTg/+ mice were heavier than the +/+ controls. The body weight of the LepTg/+:A-ZIPTg/+ mice was comparable to that of LepTg/+ mice. Food intake was significantly reduced in the LepTg/+ mice and increased in the A-ZIPTg/+ mice compared with the +/+ controls (Fig. 1F). Like the LepTg/+ mice, the LepTg/+:A-ZIPTg/+ mice ate less than the +/+ controls. Thus, both the body weight and food intake of the LepTg/+:A-ZIPTg/+ mice are indistinguishable from those of the LepTg/+ mice and different from the A-ZIPTg/+ mice. This suggests that the A-ZIPTg/+ phenotype is suppressed by the overexpression of leptin.

Table 1 presents the organ weights of 15-week-old +/+, LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice. The epididymal WAT and interscapular BAT fat pads were markedly smaller in the LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice compared with the +/+ mice (P < 0.05, n = 5). The liver was smaller in the LepTg/+ and LepTg/+:A-ZIPTg/+ mice (P < 0.05) but enlarged in A-ZIPTg/+ mice (P < 0.05) relative to +/+ controls. The liver color was lighter in the A-ZIPTg/+ mice but darker in the LepTg/+ mice, as compared with the +/+ controls. The LepTg/+:A-ZIPTg/+ liver color was indistinguishable from that of +/+ livers (data not shown).

Microscopic examination revealed no subcutaneous WAT in the LepTg/+, A-ZIPTg/+, and LepTg/+:A-ZIPTg/+ mice (Fig. 2A). The interscapular BAT from the A-ZIPTg/+ mice resembled WAT, with large lipid droplets in the adipocytes (Fig. 2B). In contrast, the BAT from LepTg/+ mice was more eosinophilic than the +/+ BAT and had smaller lipid droplets. The appearance of the BAT from the LepTg/+:A-ZIPTg/+ mice was intermediate between that of A-ZIPTg/+ and +/+ mice. Livers from A-ZIPTg/+ mice showed marked hepatic steatosis (Fig. 2C). Lipid accumulation in the LepTg/+:A-ZIPTg/+ liver was markedly reduced, though not completely normal. The size of the pancreatic islets was increased in A-ZIPTg/+ mice but reduced in LepTg/+ and LepTg/+:A-ZIPTg/+ mice (Fig. 2D). No appreciable difference in islet size was observed between LepTg/+ and LepTg/+:A-ZIPTg/+ mice.

Fed plasma glucose concentrations were 98 mg/dl in the LepTg/+ mice, which was significantly lower than the 139 mg/dl of the +/+ mice (Fig. 3A). The A-ZIPTg/+ mice were hyperglycemic (302 mg/dl). However, transgenic overexpression of leptin completely normalized the A-ZIPTg/+ glucose levels (98 mg/dl in the LepTg/+:A-ZIPTg/+ mice). Plasma insulin concentrations were lower in the LepTg/+ mice (2.2 pmol/l) than in +/+ mice (15.6 pmol/l) (Fig. 3B). Insulin concentrations were markedly elevated in A-ZIPTg/+ mice (604.0 pmol/l), but completely reduced by leptin overexpression (8.2 pmol/l in LepTg/+:A-ZIPTg/+ mice).

Plasma triglyceride and FFA concentrations were significantly reduced in LepTg/+ mice and elevated in A-ZIPTg/+ mice relative to +/+ controls (Figs. 3C and D). Leptin overexpression reduced the triglyceride and FFA concentration in A-ZIPTg/+ mice to the level of +/+ mice but not quite to the level of the LepTg/+ mice.

Glucose and insulin tolerance tests revealed increased insulin sensitivity in LepTg/+ mice and insulin resistance in A-ZIPTg/+ mice (Figs. 3E and F). Leptin overexpression in the LepTg/+:A-ZIPTg/+ mice caused the insulin sensitivity of the A-ZIPTg/+ group to increase to the level of the LepTg/+ mice.

Taken together, these results demonstrate that hepatic leptin overexpression effectively compensates for the lack of adipose tissue of the A-ZIPTg/+ mice, preventing hyperglycemia, hyperinsulinemia, hypertriglyceridemia, and increased FFA levels.

We next investigated whether leptin’s action to decrease food intake is the reason for its efficacy in preventing diabetes. To do this, we pair-fed A-ZIPTg/+ mice to the amount of food consumed by LepTg/+:A-ZIPTg/+ mice for 12 weeks. The body weights of LepTg/+:A-ZIPTg/+ and pair-fed A-ZIPTg/+ mice were similar and ∼30% less, respectively, than that of the ad libitum–fed A-ZIPTg/+ mice (Table 1). No histological differences in the WAT, BAT, liver, or pancreas were found between pair-fed A-ZIPTg/+ and ad libitum–fed A-ZIPTg/+ mice (Figs. 2A–D). In addition, pair-feeding did not improve the plasma glucose, insulin, triglyceride, or FFA concentrations of the A-ZIPTg/+ mice (Figs. 4A–D). Finally, pair-feeding did not markedly improve either the glucose or insulin tolerance tests of the A-ZIPTg/+ mice (Figs. 4E and F). Thus, the improved phenotype of the LepTg/+:A-ZIPTg/+ mice caused by leptin is not simply due to the reduced food intake.

We explored whether exogenous administration of leptin that elevates plasma leptin concentrations to those of LepTg/+:A-ZIPTg/+ mice can improve lipoatrophic diabetes of A-ZIP/F-1 mice. To do this, we examined the metabolic effect of continuous leptin administration in A-ZIP/F-1 mice. At the third and sixth days of the experiment, continuous leptin administration through two osmotic minipumps increased plasma leptin concentrations in A-ZIP/F-1 mice to 57 and 54 ng/ml, respectively, roughly comparable to those in LepTg/+:A-ZIPTg/+ mice (Fig. 5A). Body weight reduction in leptin-treated A-ZIP/F-1 mice was significantly greater than that in vehicle-treated group after the 6-day infusion (−2.30 ± 0.53 vs. 1.93 ± 0.10 g, n = 5, P < 0.01). Food intake of leptin-treated A-ZIP/F-1 mice was also decreased significantly relative to the vehicle-treated group throughout the experiment (3.48 ± 0.18 vs. 5.85 ± 0.11 g/day, n = 5, P < 0.01). Liver weight of leptin-treated A-ZIP/F-1 mice was markedly reduced relative to that of the vehicle-treated group at the end of experiment (1.52 ± 0.09 vs. 3.00 ± 0.54 g, n = 5, P < 0.01). Figure 5B shows the representative histological sections of livers from leptin- and vehicle-treated A-ZIP/F-1 mice. Lipid accumulation in the A-ZIP/F-1 mice liver was markedly reduced by leptin treatment, although the lipid droplets had not completely disappeared after the 6-day infusion. Hyperglycemia of A-ZIP/F-1 mice was improved by continuous leptin treatment but not by vehicle treatment (203 and 423 mg/dl in the leptin- and vehicle-treated groups, respectively) (Fig. 5C). Plasma insulin concentrations of A-ZIP/F-1 mice were normalized by leptin treatment (4.3 and 667.7 pmol/l in the leptin- and vehicle-treated groups, respectively) (Fig. 5D). Plasma triglyceride and FFA concentrations were also significantly reduced in the leptin-treated group relative to the vehicle-treated group (Figs. 5E and F).

Taken together, these results demonstrate that continuous leptin administration that elevates plasma leptin concentrations of A-ZIP/F-1 mice to those of LepTg/+:A-ZIPTg/+ mice can effectively improve the disorders of glucose and lipid metabolism in A-ZIP/F-1 mice.

Deficiency of adipose tissue causes insulin resistance, diabetes, and dyslipidemia, although the underlying mechanisms are not known. In this study, we have bred A-ZIP/F-1 with LepTg/+ mice to study the effects of chronic hyperleptinemia in the near-total absence of adipose tissue. In LepTg/+:A-ZIPTg/+ mice, plasma leptin concentrations become elevated soon after birth and remain markedly elevated (∼50 ng/ml) throughout the animal’s life. Transgenic overexpression of leptin rescues insulin resistance, diabetes, and much of the hepatic steatosis in A-ZIP/F-1 mice. This lack of insulin resistance, diabetes, and hepatic steatosis in LepTg/+:A-ZIPTg/+ is observed in both sexes and at up to 6 months of age (K.E., Y.O., H.M., F.M., M.S., M.A.-A., and K.N., unpublished data). This study demonstrates that leptin is useful for treatment of severe lipoatrophic diabetes and provides the first evidence for leptin as a long-term therapeutic agent for lipoatrophic diabetes.

Using the aP2-nSREBP-1c mouse model of lipoatrophic diabetes, Shimomura et al. (24) showed that leptin infusion for 12 days that achieved a concentration of ∼3 ng/ml effectively treated the insulin resistance, diabetes, and hepatic steatosis of the mice (Table 2). In contrast, when leptin was infused into A-ZIP/F-1 mice, only moderate improvement in glucose and insulin levels occurred, despite increasing the plasma leptin concentrations by 5 ng/ml (25) (Table 2). The difference between these two studies could be attributable to the smaller amount of remaining fat in A-ZIP/F-1 mice (25). Indeed, surgical implantation of normal adipose tissue reverses the insulin resistance, diabetes, and hepatic steatosis of A-ZIP/F-1 mice while achieving plasma leptin concentrations of only ∼3 ng/ml (26) (Table 2). Fat transplantation also decreases hepatic and muscular triglyceride content and restores the insulin stimulation of glucose utilization and the suppression of hepatic glucose production (27). Taken together, these findings suggest that an adequate mass of adipose tissue is required for the antidiabetic effects of physiological doses of leptin. With no substantial amount of leptin, fat transplantation may not be sufficient for recovery from insulin resistance and diabetes in A-ZIP/F-1 mice or lipoatrophic patients, given that leptin-deficient ob/ob mice are profoundly diabetic, though less severely so than A-ZIP/F-1 mice (22,25,28) (Table 2). This study demonstrates that even with minimal residual adipose tissue, either transgenic overexpression of leptin or exogenous leptin administration that achieves a plasma leptin concentration of ∼50 ng/ml can rescue insulin resistance and diabetes in A-ZIP/F-1 mice, suggesting that adipose tissue itself is not required for the antidiabetic effects of chronic hyperleptinemia. Clinically, it is important to note that the efficacy of leptin treatment or fat transplantation in lipoatrophic diabetes may depend on how much adipose tissue remains or how much leptin is supplied.

The mechanisms through which leptin exerts its antidiabetic actions are not known with certainty. The data suggest that leptin acts primarily through the hypothalamus, and thus most of its effects on insulin sensitivity are indirect (13,14,15). High doses of leptin cause a dramatic loss of visible adipose tissue, presumably by increasing energy expenditure and decreasing food intake. However, pair-feeding showed that decreased food intake does not explain the antidiabetic effects of leptin in LepTg/+:A-ZIPTg/+ mice. One possible explanation for the insulin sensitization by leptin is that the reduced levels of triglycerides and FFAs, both in the circulation and in body tissues (e.g., liver and muscle), is etiologic (29,30). Tissue triglyceride levels correlate with insulin resistance, being elevated in both lipoatrophic diabetes and obesity. It is not known whether the tissue triglycerides themselves cause insulin resistance or whether their levels serve as a surrogate for other molecules (e.g., fatty acyl-CoAs or malonyl-CoA) that may be causing insulin resistance. However, in LepTg/+:A-ZIPTg/+ mice, the improvement in hepatic steatosis and circulating triglyceride and FFA concentrations is moderate relative to the reversal of insulin resistance and diabetes. This is consistent with findings of a relatively modest improvement in triglyceride and FFA concentrations but an impressive improvement in insulin resistance after fat transplantation into A-ZIP/F-1 mice (26). These observations suggest that additional mechanisms besides fat depletion of tissues may be involved in the antidiabetic effect of leptin in lipoatrophic diabetes.

In conclusion, this study demonstrates that transgenic overexpression of leptin can rescue insulin resistance and diabetes in a mouse model of severe lipoatrophic diabetes. Therefore, LepTg/+:A-ZIPTg/+ mice provide the first experimental model demonstrating that leptin can be useful as a long-term therapeutic agent for severe lipoatrophic diabetes.

We thank Shigeo Yura for discussion and Mayumi Nagamoto for technical assistance. The authors also acknowledge Yoshiko Isa and Yuko Nakajima for secretarial assistance. This work was supported in part by research grants from the Japanese Ministry of Education, Science, Sports, and Culture; the Japanese Ministry of Health and Welfare; the Yamanouchi Foundation for Research on Metabolic Disorders; the Inamori Foundation; the Naito Foundation; the Uehara Memorial Foundation; the Kato Memorial Trust for Nanbyo Research; the ONO Medical Research foundation; and “Research for the Future (RFTF)” of the Japanese Society for the Promotion of Science (JSPS-RFTF 96100204 and 98 l00801).