Transgenic Mice Overexpressing Neuropeptide Y in Noradrenergic Neurons: A Novel Model of Increased Adiposity and Impaired Glucose Tolerance

Total RNA from a mouse hypothalamus was purified by the GenElute Mammalian Total RNA kit (Sigma Diagnostics, St. Louis, MO) and synthesized into cDNA (M-MLV RT RNase kit; Promega, Madison WI). To isolate the 300-bp preproNPY coding cDNA, exons 2 and 3 and part of exon 4 were amplified by PCR using two primer pairs: forward-1, 5′-CATGTTATGCATATGCTAGGTAACAAGCGA-3′, with reverse-1, 5′-TCCTCTGCCGGCGCGTC-3, and forward-2, 5′-GACGCGCCGGCAGAGGA-3′, with reverse-2, 5′-CAAGTAATGCATTCACCACATGGAAGGGTC-3′ (Fig. 1A). In the third PCR run, exons were joined by using the primers forward-1 and reverse-2. A plasmid containing a viral 580-bp internal ribosomal entry site (IRES) followed by a 3.7-kb reporter gene LacZ encoding β-galactosidase was linearized, and the preproNPY cDNA was ligated upstream of the IRES fragment. IRES enables cap-independent translation of preproNPY and LacZ into single transgenic mRNA, which results in coexpression of both genes from the same promoter (13). The bicistronic NPY-IRES-LacZ construct was inserted downstream of the previously described 5.8-kb promoter of the human DβH gene (12), and the final 10.6-kb transgene construct was excised from the vector with NotI (Fig. 1B). Thus, DβH promoter drives the expression of transgenic NPY and LacZ, and positive β-galactosidase staining can be used as a marker of transgene expression. The construct was sequenced gene by gene to verify the correct alignment of base pairs before it was microinjected into fertilized FVB/n eggs. The oocytes were transferred into pseudopregnant FVB/n females to generate hemizygous offspring.

Genomic DNA was isolated from tail biopsies from weaned pups with a commercial kit (Gentra, Minneapolis, MN). Genotypes were determined by PCR using forward-3 primer 5′-AACAAACGAATGGGGCTGTGT-3′ and reverse-3 primer 5′-GTGATGAGATTGATGTAGTGTCG-3′ located in exon 2 of the preproNPY gene (Fig. 1A). The PCR products were digested with NaeI. Wild-type DNA showed a single 190-bp band and transgenic DNA 190-, 120-, and 70-bp bands (Fig. 1C). Eight transgene DNA-positive founder mice (F1) were produced and bred with FVB/n mice to produce F2 hemizygous transgenic animals. β-Galactosidase expression showed most consistent staining in line number 32, which was chosen for further studies and backcrossed to C57BL/6 for six generations.

Groups (n = 7–15) of 3- and 6-month-old male and 6-month-old female mice (C57BL/6) were used unless stated otherwise. The mice were kept in an animal room maintained at 21 ± 1°C with a fixed 12-h light/12-h dark cycle. Standard rodent chow (SDS, Essex, U.K.) and water were available ad libitum. Wild-type littermates were used as control mice. We obtained terminal blood samples from vena cava (plasma) or trunk blood (serum) after a 4-h fast between 1000 and 1400 h to analyze levels of NPY, leptin, resistin, corticosterone, ghrelin, and triglyceride. All samples were stored at −70°C and treated in duplicate. To access vena cava, the mice were quickly anesthetized with ketamine (Ketalar 75 mg/kg i.p.) and medetomidine (Domitor 1 mg/kg i.p.). Otherwise, mice were killed by decapitation. Samples for glucose and insulin analyses were obtained from tail vein after a 4-h fast. Experimental procedures were approved by the local animal ethics committee (The Lab-Animal Care and Use Committee at the University of Turku).

Coronary brain (80-μm) and adrenal (35-μm) frozen sections were mounted on gelatin-coated slides and stained for Escherichia coli β-galactosidase as previously described (14).

Brains and adrenal glands were immersed in a 4% paraformaldehyde fixative overnight and cryoprotected with 30% sucrose overnight before freezing. Frozen sections of coronary brain and adrenal glands (35 μm thick) were cut with a cryostat and placed in PBS with 0.01% Triton-X. Free-floating sections were washed in PBS and preincubated with blocking buffer (5% goat normal serum and 0.01% Triton-X in PBS) for 30 min at room temperature. The sections were incubated overnight at 4°C with rabbit polyclonal anti-NPY antibody (Affiniti Research Products, Exeter, U.K.) diluted 1:1,500 followed by incubation biotinylated goat anti-rabbit antibody (1:3,000) (Molecular Probes, Eugene, OR) for 1 h at room temperature. A Standard Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA) was used in conjunction with the biotinylated antibody, and subsequent horseradish peroxidase activity was developed using 3,3′-diaminobenzidene as a substrate (Sigma Fast DAB tablets; Sigma Diagnostics). The immunostaining was visualized using light microscopy.

Blood was collected from 3-month-old male and female mice (n = 8/group). The brain and both adrenal glands were removed and washed with cold saline. Medial basal hypothalamus was isolated with a mouse brain block using a 2-mm section caudal to the optic nerve chiasma, excluding hippocampal and cortical areas. The brainstem section extended 3 mm caudal from the hypothalamic section, including locus coeruleus and A5 noradrenergic nuclei, but the cerebellum and cerebral cortex were discarded. In adrenal glands, the majority of NPY is localized within the medulla (15). Therefore, whole adrenal glands were used in this study. Tissues were homogenized with 500 μl 0.1 N HCl and placed immediately on ice. The tissue homogenates were sonicated, and NPY concentrations were determined with a EURIA-NPY radioimmunoassay (RIA) kit (Euro-Diagnostica, Malmö, Sweden) with some modifications to the instructions for the kit: Tissues were diluted to 0.1 mol/l Na₂HPO₄, buffer pH 8.3, and the standard curve standards in the manufacturer's diluent. NPY concentrations were adjusted to the tissue protein levels determined with a BCA Protein Assay Reagent kit (Pierce Biotechnology, Milwaukee, WI) using phosphate buffer as a diluent. Plasma NPY concentrations were determined with the RIA kit by diluting plasma 1:4 to the assay buffer with 5% BSA.

All mice used in the studies were weighed before killing the animals. Additionally, a set of adult males and females were separated into single cages and allowed to adjust to the change for 1 week. The mice were weighed weekly for 3–8 weeks. Food was freely available but carefully weighed every 2 days to calculate the average consumption and spillage of chow per day per mouse.

Five-month-old male mice (n = 8–11/group) were placed individually into transparent polypropylene locomotor activity cages housed in a photo-beam recording system (San Diego Instruments, San Diego, CA). The locomotor activity was measured over 10-min intervals for 24 h. The measurements started at 1200 h after 24-h habituation.

Rectal body temperatures were measured with a digital thermometer (Ellab, Roedovre, Denmark) from 3-month-old male and female mice (n = 7–10/group) between 1000 and 1200 h.

The nose-anal lengths and femur lengths of the 3-month-old male mice (n = 6/group) were measured. Femurs were imaged with a Skyscan 1072 Micro-CT (Skyscan, Aartselaar, Belgium). The Micro-CT system was calibrated with two hydroxyapatite phantoms with known densities of 250 and 750 mg/cm³. For each specimen, a volume of interest containing trabecular bone and bone marrow was manually defined, and the trabecular bone was separated from the bone marrow using a global thresholding technique.

WAT weight was determined by collecting the subcutaneous, epididymal/gonadal, and retroperitoneal fat pads at death. The fat cell sizes were determined from a total of eight gonadal fat pads from 3-month-old female mice as previously described (16). Cells were analyzed and imaged under a light microscope, and diameters were calculated using Cell*A imaging software (Soft Imaging System, Münster, Germany).

Liver lipid contents were isolated and purified with the Folch method (17). Plasma, serum, and tissue triglycerides were quantified with Free Glycerol Reagent (F6428) and Triglyceride Reagent (T2449; Sigma Diagnostics). Serum total cholesterol from 6-month-old females was measured with the BioVision Cholesterol Quantitation kit (BioVision, Mountain View, CA).

Circulating concentrations of insulin (ELISA kit; Mercodia, Uppsala, Sweden), corticosterone (RIA kit; MP Biomedicals, Orangeburg, NY), ghrelin (RIA kit; Linco Research, St. Charles, MO), leptin, and resistin (LINCOplex Adipokines kit; Linco Research) were measured using commercial kits.

The left lateral lobes and gastrocnemius from male mice were dissected, placed in liquid nitrogen, and stored at −70°C. Frozen sections (5 μm) were cut and, based on the results from a pilot study, stained for hematoxylin-eosin (H-E), oil red O, and sudan black (liver) or for H-E and ATP 3.4 (muscle).

Mice were fasted from 600 to 1000 h and administered intraperitoneally with glucose (5% [wt/vol], 1 g/kg body wt). Tail vein blood glucose was measured at 0, 20, 40, 60, and 90 min with the Precision Xtra Glucose Monitoring Device (Abbott Diabetes Care, Abbott Park, IL). Areas under the resultant glucose curves were calculated with the trapezoidal method in GraphPad Prism 4.3 software (GraphPad Software, San Diego, CA).

The results were compared with unpaired Student's t test between the genotypes. Logarithmic transformations were used if data were not normally distributed (D'Agostino and Pearson omnibus normality test). The nonparametric Mann-Whitney test was used if the distribution remained skewed after the transformation. Comparisons between the two genotypes and both sexes were analyzed with two-way ANOVA with Bonferroni post hoc test. Intraperitoneal glucose tolerance test (IGTT) and locomotor activity data were analyzed with repeated-measures two-way ANOVA. Correlation analyses were performed using linear regression to compare body weight or WAT weight with other variables (area under the curve [AUC] in IGTT, fasting insulin, and leptin) and glucose (AUC) with ghrelin. Statistical analyses were carried out using GraphPad Prism 4.3. Data are presented as means ± SEM, and the results were considered statistically significant at P < 0.05.

NPY- and LacZ-overexpressing mice under DβH promoter were generated, resulting in eight founders from which a line of transgenic mice was established. Genotype distributions matched expected Mendelian ratios. NPY overexpression does not result in embryonic lethality, and the mice produce pups normally. Transgene expression was visualized with β-galactosidase analysis, which showed strong and specific LacZ staining in adrenal medulla and the locus coeruleus and A5 nuclei of the brainstem (Fig. 2), which are areas known to contain dense noradrenergic innervation. Staining was faint in the hypothalamus, the cortex, the medial habenular nuclei, and the paraventricular thalamic nucleus. Blue color was observed only in the OE-NPY^DβH mice, and no ectopic staining was observed in any of the sections. Examination of coronal brain sections and adrenal glands using immunohistochemistry with a specific NPY antibody revealed that in the OE-NPY^DβH mice, NPY was detected in the same areas as those observed in wild-type littermates. These areas included adrenal medulla, the arcuate, the ventro-/dorsomedial and the periventricular nuclei of the hypothalamus, the lateral hypothalamic area, the hippocampus, the amygdala, the thalamus, and the locus coeruleus and A5 neurons of the brainstem. No site of ectopic NPY expression was observed in transgenic animals (data not shown).

The NPY concentrations in adrenal, brainstem, and hypothalamus homogenates and plasma were measured with RIA. The OE-NPY^DβH mice had a significantly higher NPY concentration in the adrenals and in the brainstem compared with wild-type mice (genotype, P < 0.05 and 0.01, respectively; sex, NS; genotype × sex interaction, NS; Fig. 3A and B). NPY concentration in the hypothalamus did not differ between the genotypes but was increased in females in comparison with males (genotype, NS; sex, P < 0.01; genotype × sex interaction, NS; Fig. 3C). NPY concentration in plasma did not differ between the genotypes. However, NPY levels were significantly higher in male than in female mice (genotype, NS; sex, P < 0.001; genotype × sex interaction, NS; Fig. 3D).

Body weights were recorded from mice at 3 and 6 months of age, and no difference between the genotypes was observed (Fig. 4A). During a weekly follow-up, the mice retained the body weights steadily, and no difference in the amount of food consumed between the wild-type and transgenic mice was observed (males, 4.12 ± 0.27 and 4.19 ± 0.30 g/day; females, 3.59 ± 0.53 and 3.64 ± 0.48 g/day, respectively; NS). The mice displayed similar body and femur lengths at 3 months (wild type, 9.51 ± 0.14 and 1.46 ± 0.04 cm; OE-NPY^DβH, 9.48 ± 0.15 and 1.46 ± 0.04 cm, respectively; NS). Bone mineral density (BMD) did not differ between the genotypes (wild type, 1,119 ± 13.7; OE-NPY^DβH, 1,105 ± 27.5 mg/cm³; NS).

The OE-NPY^DβH mice had significantly more WAT relative to controls at 3 and 6 months of age (Figs. 4B and C and 5A). Mean fat cell diameter was significantly smaller in the OE-NPY^DβH fat pads (Fig. 5B). The triglyceride content of the OE-NPY^DβH livers was increased already at 3 months (Table 1), but the livers were not enlarged and did not show severe fatty liver morphology. No differences in circulating triglyceride and cholesterol levels were observed (Table 1). Basic morphology of the skeletal muscle was similar between the genotypes (data not shown).

No differences in overall activity or number of rearings between the genotypes were observed (Fig. 6). Also, no difference in rectal body temperature between wild-type and OE-NPY^DβH mice was observed (males, 37.16 ± 0.29 vs. 37.07 ± 0.46°C; females, 37.47 ± 0.31 vs. 36.84 ± 0.23°C, respectively; NS).

Six-month-old males had increased plasma insulin in comparison with wild-type males (Table 1), and the insulin values correlated positively with body weights in both genotypes (wild type, r = 0.60, P < 0.05; OE-NPY^DβH, r = 0.86, P < 0.001). Insulin levels of 3-month-old mice remained under the detection level of the insulin kit in both genotypes (Table 1). Leptin levels did not differ between the genotypes, although a tendency to increased leptin was seen in 6-month-old males (Table 1). There was no difference in corticosterone or resistin levels between the genotypes (Table 1). Ghrelin levels were significantly increased in the OE-NPY^DβH mice over wild-type values (Table 1). In females, we did not observe any differences between the genotypes in plasma insulin, leptin, resistin, or corticosterone levels (data not shown).

At 3 months, the glucose tolerance was normal in males and females. Six-month-old male OE-NPY^DβH mice showed impaired glucose tolerance, although fasting glucose remained normal (Fig. 7). In mutant mice, the AUC in IGTT was significantly elevated compared with that in wild-type mice (Fig. 7), and it correlated positively with body weights (wild type, r = 0.04, P = NS; OE-NPY^DβH, r = 0.64, P < 0.05). A strong correlation between glucose AUC and ghrelin was also observed in both genotypes (wild type, r = 0.57, P < 0.05; OE-NPY^DβH, r = 0.63, P < 0.05).

In this novel mouse model, we show that overexpression of NPY in noradrenergic neurons of the central and peripheral nervous systems leads to increased WAT mass and triglyceride accumulation in the liver already at an early age, leading to hyperinsulinemia and impaired glucose tolerance later in life. These changes occur without hyperphagia, increased body weight, hypercorticosteronemia, decreased BMD, or changes in locomotor activity or core temperature. Transgene expression is driven by the human DβH promoter, which has been used to generate transgenic mice in several other studies, and is shown to drive the transgene expression to noradrenergic neurons with very little ectopic expression (12,18,19). IRES sequence and LacZ reporter gene were inserted downstream of preproNPY, enabling the use of β-galactosidase staining as a marker of transgene expression. Because transgenic preproNPY-IRES-LacZ gene construct is transcribed to a single mRNA, detection of LacZ expression proves that preproNPY mRNA is also expressed in the same cells. β-Galactosidase staining showed that the transgene is strongly expressed in adrenal medulla and the locus coeruleus of the brainstem and faintly expressed in the hypothalamus, the thalamus, and the cortex, which fits with the previously described DβH expression pattern (12). NPY protein detection in tissue homogenates showed 1.3-fold increase in NPY in the adrenal glands and 1.8-fold in the brainstem dissection, including locus coeruleus and A5 noradrenergic nuclei, whereas there was no difference in the medial basal hypothalamus NPY in the OE-NPY^DβH compared with wild-type mice. These data confirm that transgenic preproNPY mRNA is expressed in noradrenergic neurons and that total NPY peptide level is increased in areas of noradrenergic innervation.

The OE-NPY^DβH mice have 20–25% increased body WAT depot mass compared with wild-type controls, which is observed already at the age of 3 months in both sexes. Fat cells from the transgenic mice are significantly smaller and, due to the heavier WAT weight, probably more numerous (hyperplasia). The finding is interesting because NPY has previously been linked to fat-cell hypertrophy, as central NPY administration promotes lipogenesis by stimulating lipoprotein lipase activity in WAT (5,20). In addition to increased WAT mass, triglyceride content in the liver is increased in 3- and 6-month-old OE-NPY^DβH mice. Central administration of NPY or its Y5 receptor agonist have been shown to increase circulating levels of triglycerides (3,21), and chronic intracerebroventricular administration of NPY has been shown to stimulate de novo lipogenesis in the liver by increasing the hepatic acetyl CoA carboxylase activity (3). Hepatic steatosis in OE-NPY^DβH mice could be a result of enhanced hepatic lipogenesis, or it may be secondary to increased fatty acid flux from the enlarged adipose tissue. In contrast, triglyceride accumulation in the liver is not due to hyperinsulinemia, which has been shown to induce steatosis (22), because 3-month-old mice are not hyperinsulinemic but show hepatic steatosis.

OE-NPY^DβH male mice are normoinsulinemic and glucose tolerant at the age of 3 months, but by the age of 6 months, they develop hyperinsulinemia and impaired glucose tolerance. These data show that chronically increased SNS NPY does not induce insulin release or affect glucose tolerance per se but augments hyperinsulinemia and impaired glucose tolerance associated with increased age and body weight. In contrast to male transgenic mice, glucose tolerance remains normal in 6-month-old females. The male and female mice differ in the degree of fat accumulation over age. Mean WAT weight (corrected with body weight) in 6-month-old males is twofold higher than in 3-month-old males compared with 1.4-fold increase in females both in wild-type and OE-NPY^DβH mice (Fig. 4C). The sex difference in impairment of glucose tolerance may arise from the difference in development of excess adiposity. Thus, increased adiposity and liver triglyceride content seem to be the primary effects of NPY overexpression, and impaired glucose metabolism arises as an interaction between augmented fat accumulation and excess NPY in the OE-NPY^DβH mice.

Serum ghrelin was significantly increased in the OE-NPY^DβH mice over wild-type controls. Ghrelin is a gastric hormone associated with increased food intake and obesity. Central administration of ghrelin increases food intake, body weight, and adiposity (23–27), and this effect is mediated via stimulation of NPYergic neurons in the hypothalamus (24). In contrast, mouse models of hyperghrelinemia have failed to show hyperphagia or obesity (28,29). These results fit with the phenotype of our OE-NPY^DβH mice, i.e., hyperghrelinemia without hyperphagia. Furthermore, ghrelin has been implicated in the control of glucose metabolism. Acutely, ghrelin attenuates glucose-stimulated insulin release (30), and genetic ablation of ghrelin augments insulin release and stimulates glucose clearance after glucose injection (31). In our study, serum ghrelin strongly correlates with glucose AUC in IGTT, suggesting that increased circulating ghrelin contributes to impaired glucose tolerance in the OE-NPY^DβH mice.

Increased adiposity in NPY-overexpressing mice is in line with previous studies in rats with chronic central infusion of NPY and Y2 receptor (Y2R)-null mice. In contrast to OE-NPY^DβH mice, weight gain in these models is mainly due to hypothalamically mediated increase in food intake and decrease in energy expenditure (1,3,32). The effects of extrahypothalamic NPY on energy homeostasis have been scarcely addressed (33) until a recent paper by Kuo et al. (34), which shows stimulated NPY release from SNS by chronic mild stress to increase diet-induced obesity in mice. The effects of stress-induced NPY release on energy homeostasis fit the phenotype of the OE-NPY^DβH mouse. The primary finding is increased WAT mass without change in food intake or body weight after 2 weeks of daily stress and high-fat diet. Prolonged stress and diet lead to liver steatosis, impaired glucose tolerance, and obesity, which are attenuated by local Y2R antagonist administration and fat-targeted Y2-gene knockdown procedure. Thus, SNS NPY seems to increase adipose tissue mass through local, Y2R-mediated mechanisms. Kuo et al. (34) show that NPY directly stimulates adipogenesis, i.e., proliferation and differentiation of preadipocytes in vitro, which correlates with the hyperplastic adipose tissue in our OE-NPY^DβH model. Therefore, the phenotype of the OE-NPY^DβH mouse closely resembles the effects of stress-induced NPY release from SNS on energy homeostasis, which supports that the phenotype is caused by overexpression of NPY in noradrenergic neurons.

In humans, association studies of a polymorphism in the NPY gene have linked NPY to metabolic disturbances. The L7P polymorphism, first described by our group, is associated with earlier onset of type 2 diabetes in obese carriers of L7P (35) and with other traits of the metabolic syndrome (36–39). Clinical studies have also shown that L7P carriers have altered insulin and ghrelin responses (40,41). L7P polymorphism is located in the signal peptide of the human NPY gene (36) and is functional leading to increased plasma levels of NPY during sympathetic activation in subjects carrying L7P allele (42). Thus, we have hypothesized that increased NPY causes the metabolic changes in L7P carriers. This hypothesis is now supported by the findings in OE-NPY^DβH mice showing that even a modest chronic excess of NPY can lead to significant alterations in lipid and glucose metabolism.

In conclusion, moderate overexpression of NPY in SNS and brain noradrenergic neurons causes increased adiposity and liver triglyceride accumulation, which leads to hyperinsulinemia and impaired glucose tolerance with age. The effect seems to be nonhypothalamic because no difference is seen in the hypothalamic NPY content between the OE-NPY^DβH mice and wild-type controls or in feeding and other hypothalamically induced NPY effects. These findings suggest that NPY exerts significant metabolic effects directly on peripheral target tissues and/or by modulation of autonomic nervous system activity. Furthermore, the results suggest that NPY may play an important role in the pathogenesis of disturbances in glucose and lipid metabolism. Thus, further analysis of the OE-NPY^DβH mouse may provide an important tool for studies concerning the development of metabolic disorders and the role of NPY in the process.

This study was supported by the Academy of Finland, the Finnish Funding Agency for Technology and Innovation, the Turku University Foundation, and the Finnish Cultural Foundation.

We thank Dr. Richard D. Palmiter (University of Washington) for providing the DβH promoter. Dr. Matti Poutanen and his staff (Transgenic Core Unit, University of Turku) are acknowledged for the microinjections. We thank Raija Kaartosalmi, Pipsa Lappalainen, Pirkko Huuskonen, Francisco Lopez-Picon, and Heikki Hiekkanen for technical and scientific assistance.