An animal model of spontaneous metabolic syndrome: Nile grass rat

Abstract

Metabolic syndrome (MetS) is a prevalent and complex disease, characterized by the variable coexistence of obesity, dyslipidemia, hyperinsulinaemia, and hypertension. The alarming rise in the prevalence of metabolic disorders makes it imperative to innovate preventive or therapeutic measures for MetS and its complications. However, the elucidation of the pathogenesis of MetS has been hampered by the lack of realistic models. For example, the existing animal models of MetS, i.e., genetically engineered rodents, imitate certain aspects of the disease, while lacking other important components. Defining the natural course of MetS in a spontaneous animal model of the disease would be desirable. Here, we introduce the Nile grass rat (NGR), Arvicanthis niloticus, as a novel model of MetS. Studies of over 1100 NGRs in captivity, fed normal chow, revealed that most of these animals spontaneously develop dyslipidemia (P<0.01), and hyperglycemia (P<0.01) by 1 yr of age. Further characterization showed that the diabetic rats develop liver steatosis, abdominal fat accumulation, nephropathy, atrophy of pancreatic islets of Langerhans, fatty streaks in the aorta, and hypertension (P<0.01). Diabetic NGRs in the early phase of the disease develop hyperinsulinemia, and show a strong inverse correlation between plasma adiponectin and HbA1c levels (P<0.01). These data indicate that the NGR is a valuable, spontaneous model for exploring the etiology and pathophysiology of MetS as well as its various complications.—Noda, K., Melhorn, M. I., Zandi, S., Frimmel, S., Tayyari, F., Hisatomi, T., Almulki, L., Pronczuk, A., Hayes, K. C., Hafezi-Moghadam, A. An animal model of spontaneous metabolic syndrome: Nile grass rat.

Metabolic syndrome (MetS) currently affects over a quarter of the population in developed countries, and its prevalence is rapidly rising (1). It is characterized by the variable coexistence of obesity, dyslipidemia, hyperinsulinemia or type 2 diabetes, and hypertension (2,3,4). MetS is associated with a marked increase in the risk for cardiovascular disease and atherosclerosis, both of which pose a major burden to the public health. However, the etiology of MetS is poorly understood.

Adiponectin, a key hormone in MetS, is produced by adipocytes and secreted into the bloodstream. It has insulin-sensitizing, antiinflammatory, and antiatherogenic properties (5, 6). Levels of adiponectin inversely correlate with body fat percentage (7). It protects against type 2 diabetes, atherosclerosis, and MetS. Decreases in circulating adiponectin are associated with increased leukocyte-endothelial interactions, possibly due to increased endothelial adhesion molecule expression (7). Another important hormone produced by adipocytes is leptin (8), a 16-kDa peptide that is a key regulator of energy balance, in part, by decreasing food intake (9) and by enhancing fatty acid catabolism in muscle, liver, and adipose tissues (10). Leptin deficiency both in rodents and humans results in severe obesity (11).

The study of MetS has been hampered by the lack of appropriate models. Genetically engineered rodents have been generated to resemble various aspects of MetS (12,13,14). Such models may allow investigation of links between individual genes and certain components of MetS. However, the genetic manipulations in these models may be causally unrelated to the natural roots of the human disease. To investigate the etiology and pathophysiology of MetS, a naturally occurring animal model would be invaluable.

The Nile grass rat (NGR), Arvicanthis niloticus, is a herbivorous murine rodent inhabiting dry savanna, woodlands, and grasslands in Africa (15). The animals live underground in burrows constructed in a central area with surface runways radiating outward. As opposed to other rodent species that are nocturnal, NGRs are primarily diurnal (15). This unique behavior pattern has made them a valuable tool in the study of circadian rhythm (16,17,18,19). Furthermore, NGRs primarily express cones in the retina (20).

This report describes the systemic pathology in NGRs that matches all aspects of human MetS. Fed a conventional lab diet, NGRs spontaneously develop obesity, hyperglycemia, and hypertension. The disease is manifest in various organs, such as lipid deposition in the liver, advanced glycation end product (AGE) deposits in the kidney, and β-cell failure coupled with early stages of atherosclerosis in the aorta.

MATERIALS AND METHODS

Animals

NGRs were fed standard laboratory chow (Lab Diet 5020; PMI Nutrition, St. Louis, MO, USA) and provided with water ad libitum. The animals were bred and raised in a 12-h light-dark cycle and housed in plastic cages in the climate-controlled Brandeis University animal facility. All animal experiments were conducted in accordance with protocols approved by the Animal Care Committees of the Massachusetts Eye and Ear Infirmary and Brandeis University.

Perfusion prior to tissue harvest

Prior to fixation of selected tissues in 4% paraformaldehyde, rats were perfused with PBS [500 ml/kg body weight (BW)], as described previously (21). Under deep anesthesia, the chest cavity was opened, and the rat was perfused via the left ventricle, with a nick in the right atrium to allow drainage of the perfusate.

HbA1c, blood glucose (BG), triglyceride (TG), and cholesterol measurements

BG levels in rats were measured from tail blood samples, taken randomly or after 16 h overnight food deprivation, using an Elite Glucometer (Bayer, Elkhart, IN, USA). Animals with repeated BG > 150 mg/dl were considered diabetic. After 3–5 d, the animals were deeply anesthetized, and blood was collected by cardiac puncture, using EDTA-treated syringes. Glycated hemoglobin (GHb) was measured by a Glyco-Tek affinity column kit (Helena Laboratories, Beaumont, TX, USA). HbA1c was calculated using the equation provided by the company. Animals with HbA1c ≥ 6.5% were considered diabetic, in accord with the diabetes management guidelines of the American College of Endocrinology for human patients (22, 23). Subsequently, blood was centrifuged at 12,000 g for 10 min at 4°C, and plasma was collected. Plasma level of total cholesterol (TC) and TGs were measured by enzymatic assay using Thermo Infinity™ kits (Thermo Electron, Pittsburgh, PA, USA). Ketone bodies in blood and urine were detected with DiaScreen reagent strips for urinalysis (HypoGurad; Minneapolis, MN, USA). The use of HbA1c as a determinant of diabetes in patients has recently been critically discussed (24), however, in our experimental setting, it appears a sufficiently strong indicator of disease, especially in combination with random BG testing.

Oil-Red-O staining

After perfusion of the animals with PBS, liver, and kidneys were excised and placed in 4% paraformaldehyde (PFA) for fixation. Frozen sections (15 μm) were prepared and stained with Oil-Red-O (Sigma, St. Louis, MO, USA), which detects hydrophobic lipids, including esterified cholesterol. The sections were counterstained with Mayer’s hematoxylin (Sigma).

Histomorphology of atherosclerosis

Animals were perfused with PBS and 4% PFA to obtain an initial fixation. The aorta was dissected from the aortic valve to the iliac bifurcation, opened longitudinally, and pinned flat on a black wax surface. After overnight fixation with 4% PFA and a 12 h rinse in PBS, the aortas were stained with Oil-Red-O. Specimens were examined by stereomicroscopy, and micrographs of the lesions were generated, using a color charge-coupled device (CCD) camera (Dage, Fremont, CA, USA).

Immunohistochemistry for insulin in pancreas

Using pancreas fixed with 4% PFA, 10-μm frozen sections were generated and incubated with blocking solution (Invitrogen, Carlsbad, CA, USA). Sections were then incubated with guinea pig polyclonal antibody against human insulin (Abcam, Cambridge, MA, USA), which also has cross-reactivity with nonhuman species, such as mice and rats (ab7842, Abcam). Thereafter, the sections were incubated for 30 min at room temperature with Texas Red-conjugated secondary antibodies (ab6906, Abcam). Sections were mounted with Vectashield mounting medium containing 4′,6-diamino-2-phenylindole (DAPI; Vector Laboratories, Burlingame, CA, USA).

Transmission electron microscopy (TEM)

Samples were fixed in 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M cacodylate buffer with 0.08 M CaCl₂ at 4°C. Samples were then postfixed for 1.5 h in 2% aqueous OsO₄, dehydrated in ethanol and water, and embedded in Epon. Ultrathin sections were cut from blocks and stained with saturated, aqueous uranyl acetate and Sato’s lead stain. The specimens were examined with a Philips CM10 electron microscope(Philips, Amsterdam, The Netherlands).

Enzyme-linked immunosorbent assay (ELISA) for insulin, adiponectin, and leptin

Plasma samples obtained from the animals were used to measure the levels of insulin, adiponectin, and leptin. Concentrations in plasma were determined with the ELISA kits for mouse/rat insulin (Linco Research, St. Charles, MO, USA), adiponectin (Phoenix Pharmaceuticals, Inc., Belmont, CA, USA) and leptin (Phoenix Pharmaceuticals) according to the manufacturers’ protocols.

Blood pressure (BP) measurement

Rats were briefly anesthetized by exposure for 10–40 s to 50% CO₂/50% O₂ (Air Gas East Item X02CD50B2005053) in a clear plastic restrainer (IITC Inc., Woodland Hills, CA, USA). The restrainer containing the animal was then placed in a temperature controlled chamber (IITC) and kept at 29°C to ensure physiological blood flow conditions. Animals were kept at this temperature for 8–10 min for acclimatization. A cuff (i.d. 1/4 or 5/8 inch; IITC) was placed around the animal’s tail. The BP measurement device’s output was digitalized using an analog/digital converter (Powerlab 8/30; ADInstruments, Colorado Springs, CO, USA), and recorded on an Apple computer (Apple, Cupertino, CA, USA) using Chart software (ADInstruments). The cuff was inflated to a pressure of 200–300 mmHg, followed by deflation, using a linear deflation piece. Three measurements were made in each rat over a 20-min period. Rats with BG > 150 mg/dl were considered diabetic for these measurements.

Statistical analysis

All results are expressed as means ± se; n as indicated. Student’s t test was used for statistical comparison between groups. Differences between means were considered statistically significant at values of P < 0.05.

RESULTS

Metabolic abnormality in NGRs

Fed with normal chow, NGRs grow at different rates, with males growing faster and larger than females (Fig. 1). By 1 yr of age, ∼90% of all males and 50% of all females developed elevated BG (nonfed values > 150 mg/dl). The propensity to develop diabetes in NGRs is illustrated by two representative groups of male rats, fed normal chow until 7 and 11 mo of age (Table 1). Adult males typically reached BW ∼ 125 g. More important, faster-growing males develop diabetes sooner (5/10 by 7 mo). Nearly all males became diabetic by 1 yr (7/8 at 11 mo). The BG values suggest that early diabetes (7 mo) is detected by random BG levels, whereas nonfed BG of moderate diabetes 125–250 mg/dl, does not always manifest until 8–12 mo of age. Diabetic rats at 7 mo had a 50% higher liver weight, which continued to increase in advanced diabetic animals up to 11 mo of age, which suggests a high association between liver weight and diabetes onset. Kidney weight followed a similar trend; however, it was less predictable than either BW or liver weight. Remarkably, total fat (as measured by the 3 largest fat pools) was unrelated to the onset of diabetes; however, the intra-abdominal fat pad (represented by the perirenal/retroperitoneal pad) was significantly larger in diabetic animals at 7 mo than in nondiabetic controls. In animals with advanced diabetes (11 mo), the total fat and perirenal pool were greatly diminished as ketosis progressed and fat became the major source of energy. Ketone bodies were detected routinely in blood and urine of rats with advanced diabetes. Blood lipids, similar to nonfed BG, tended to be elevated in diabetic rats, while both TC and TG were dramatically elevated in advanced diabetes at 11 mo.

Figure 1.

Growth characteristics of NGRs. Male and female NGRs fed normal chow and water were weighed individually at different ages. Animals experienced a growth spurt during the first 20 wk and a smaller weight gain until wk 50. Adult females were on average 10–20 g lighter than males at each time point. Few animals lived beyond wk 52, in most cases due to diabetes-associated complications. Observations n at each time point: males, 3–67; females, 2–70. Data are from 206 males and 200 females.

TABLE 1.

BW, organ weights, BG, and plasma lipids in 7- and 11-mo-old male NGRs with different susceptibilities to diabetes

Characteristic	7 mo		11 mo
	Nondiabetic, n = 5	Diabetic, n = 5	Nondiabetic, n = 3	Diabetic, n = 7
BW (g)	112 ± 7	127 ± 13*	126 ± 10	125 ± 13
BG (mg/dl)
Random	75 ± 43	266 ± 81*	98 ± 43	448 ± 39*†
Nonfed (15 h)	43 ± 15	67 ± 25	79 ± 64	210 ± 99†
Organ weight (g)
Liver	3.22 ± 0.55	5.59 ± 1.77*	3.32 ± 0.18	8.39 ± 1.54*†
Kidney	0.87 ± 0.21	1.28 ± 0.36	1.02 ± 0.11	1.70 ± 0.34*
Adipose tissue
Perirenal	1.86 ± 0.72	2.87 ± 0.60*	2.56 ± 1.38	0.87 ± 0.38*†
Epididymal	4.59 ± 0.59	3.64 ± 0.99	4.11 ± 1.24	3.22 ± 1.61
Brown fat	2.13 ± 0.60	2.55 ± 0.42	2.47 ± 0.28	1.83 ± 1.08
Total fat	8.57 ± 1.36	9.07 ± 1.40	9.15 ± 2.41	5.92 ± 2.41†
Plasma lipids
TC (mg/dl)	141 ± 21	173 ± 43	122 ± 26	391 ± 140*†
TG (mg/dl)	49 ± 16	124 ± 86	70 ± 16	659 ± 626

Rats with BG > 150 mg/dl were considered diabetic. Random BG typically measured between 11:00 AM and 3:00 PM. Perirenal data combine the weights of perirenal and retroperitoneal fat pads. Values are means ± sd. 

^*P < 0.05 vs. age-matched nondiabetic NGRs; 

^†P < 0.05 vs. 7-mo diabetic NGRs. 

Two age-matched female rats of different BW illustrate development of obesity in some NGRs (Fig. 2A). However, disease development, as depicted by nonfed BG, did not strongly correlate with BW (n=50, 12–24 wk old, r²=0.07) (Fig. 2B). Plasma samples from rats with chronic high BG and advanced diabetes had a dramatic milky appearance, suggesting high blood lipid values, while samples from rats with low BG appeared transparent and normal (Fig. 2C). To assess the chronicity of elevated BG and diabetic status, we measured HbA1c in a subset of 43 rats and found a biphasic pattern, with peaks at ∼5 and 10%, suggesting the presence of a nondiabetic and a diabetic population, respectively (Fig. 3A). Of these rats, 46% exhibited HbA1c values ≥ 6.5% (9.5±0.4%, n=20) and were considered diabetic, while rats with HbA1c values < 6.5% (5.6±0.1%, n=23) were considered normal at the time of examination.

Figure 2.

Mild obesity and pronounced hyperlipidemia in NGRs. A) Photo of diabetic (left) and nondiabetic (right) female NGRs. Diabetic NGR (106 g) had high nonfed BG (384 mg/dl) and HbA1c (10.1%), whereas corresponding values for the nonobese age-matched NGR (68 g) were normal (BG, 54 mg/dl; HbA1c, 5.4%). B) Correlation between BW and nonfed BG in NGRs (12–24 wk of age). Weak correlation (r²=0.07) suggests that obesity in NGRs is not a defining factor for MetS development. Male and female, n = 50. C) Representative blood samples from diabetic (left) and normal (right) rats. Milky appearance of plasma of the diabetic animal indicates pronounced dyslipidemia.

Figure 3.

BG and plasma lipids in NGRs. A) Distribution of HbA1c. B) Nonfed BG. C) Plasma TG. D) Plasma TC. Values are means ± se; n = 43. *P < 0.01; **P < 0.001. E) Correlation between plasma TG and HbA1c in NGRs. y = 0.016x + 4.98; n = 43; P < 10⁻⁸.

In line with the HbA1c values, the nonfed BG in the diabetic rats (250±26 mg/dl, n=23) was significantly higher than those in the nondiabetic rats (78±11 mg/dl, n=20, P<0.001, Fig. 3B).

Furthermore, plasma TG was significantly higher in the diabetic rats (222±28 mg/dl; n=20; P<10⁻⁴) than in the nondiabetic rats (87±10 mg/dl; n=23) (Fig. 3C). Similarly, TC was significantly higher in the diabetic rats (195±41 mg/dl; n=20) compared to the nondiabetic controls (89±8 mg/dl; n=23; P<0.01; Fig. 3D). Plasma TG positively correlated with the HbA1c values (r=0.77; n=43; P<0.001; Fig. 3E). Interestingly HbA1c did not correlate with BW (data not shown), suggesting that NGRs are not primarily an obesity model of metabolic disarray.

Influence of age and gender on the metabolic dysfunction in NGRs

To investigate the influence of age and gender on the development of MetS in NGRs, a cohort of 62 (38 female and 24 male) animals were divided into 4 groups. Rats < 10 mo were considered young (female, 5.7±0.1 mo, n=20; male, 6.1±0.3 mo, n=13), while rats ≥ 10 mo were considered aged (female, 12.7±0.5 mo, n=18; male, 12.6±0.3 mo, n=11). The BW of the young male NGRs (98±6 g; n=13) was similar to that of the age-matched females (97±3 g; n=20; P=0.8). The older males (126±5 g; n=11) weighed slightly more than the age-matched females (115±4 g; n=18); however, the difference was not statistically significant (P=0.1).

Nonfed BG in the older males (250±22 mg/dl; n=11) was significantly higher than in older females (163±31 mg/dl; n=18; P<0.05; Fig. 4A). Furthermore, nonfed BG was significantly higher in the older males than in younger males (157±38 mg/dl; n=13; P<0.05; Fig. 4A). Older female rats had higher nonfed BG than younger females, but the difference was not significant (P=0.19; Fig. 4A). In line with the BG levels, HbA1c was significantly higher in the older male animals (11.0±0.8%; n=7) than in the age-matched female rats (7.8±0.8%; n=9; P<0.05; Fig. 4B). HbA1c in older male NGRs was significantly higher than in the younger males (7.8±0.8%; n=9; P<0.05), while HbA1c in the older (7.8±0.8%) and younger females (6.3±0.3%) was not statistically different (P=0.1) and tended to be lower than males (Fig. 4B).

Figure 4.

Effects of age and gender on metabolic profile in young and aged, male and female NGRs. A) Nonfed BG. B) Plasma HbA1c. C) Plasma TG. D) Plasma cholesterol. Values are means ± se. *P < 0.01; **P < 0.001.

To investigate lipid metabolism, plasma TG and TC were measured in the various groups. Plasma TG was significantly higher in young male NGRs (322±59 mg/dl; n=13) compared with the age-matched females (86±10 mg/dl; n=20; P<0.01; Fig. 4C), while the aged males (391±105 g/dl; n=11) had significantly higher plasma TG than females (149±20 mg/dl; n=18; P<0.05; Fig. 4C). Aged female rats had significantly higher plasma TG levels than young females (P<0.01; Fig. 4C). Plasma TC in young male rats (212±60 mg/dl; n=13) was significantly higher than in age-matched females (76±3 mg/dl; n=20; P<0.05; Fig. 4D). Aged male rats had higher TC (306±56 mg/dl; n=11) compared to age-matched females (163±19 mg/dl; n=18; P<0.05; Fig. 4D). Aged females had significantly higher TC than young females (P<0.01; Fig. 4D).

Pathological changes in liver, kidney and pancreas of diabetic NGRs

To investigate potential organ manifestations of the metabolic changes in NGRs, the visceral organs of normal and diabetic animals were examined macroscopically and histologically. Livers from nondiabetic rats appeared normal by both measures (Fig. 5A). By contrast, livers from diabetic NGRs were typically enlarged, with a pale, marble-like appearance (Fig. 5A and Table 1). When examined histologically, intracellular microvesicular fatty depositions were noted in the pericentral and midzonal areas of the hepatic lobules (Fig. 5B). The lipid particles varied in size but appeared to be larger in the central areas compared to those in the midzonal or peripheral areas, suggesting a gradual confluence of the lipid droplets during the course of the disease. Electron microscopy revealed marked cytoplasmic lipid droplets, characteristic of microvesicular hepatic steatosis (Fig. 5C).

Figure 5.

Macroscopic, histological and ultrastructual images of liver tissues obtained from nondiabetic and diabetic NGRs. A) Gross appearance of representative livers of a normal (left panel; nonfed BG=49 mg/dl; plasma TG=35 mg/dl) and a diabetic NGR (right panel; nonfed BG=270 mg/dl; plasma TG=431 mg/dl). In contrast to the normal appearance of liver in nondiabetic NGRs, liver of diabetic rats was enlarged and showed yellowish markings, typical for steatosis. B) Oil-Red-O staining of frozen sections of livers obtained from nondiabetic (left panel) and diabetic animals (right panel). Mayer’s hematoxylin was used for counterstaining. Scale bar = 100 μm. C) Representative electron microscopic visualization of the cytoplasm of a normal (left panel) and a diabetic NGR (right panel). Arrow, typical lipid microdroplets that are indicative of microvesicular steatosis.

Diabetic NGRs also showed pathological changes of the kidneys, consistent with progressive chronic nephropathy. A representative kidney from a young male diabetic rat (5 mo old; nonfed BG=251 mg/dl; plasma TG=101 mg/dl) suggests early-stage dramatic enlargement during diuresis compared to the organ of an age-matched control (5 mo old; nonfed BG=67 mg/dl; plasma TG=43 mg/dl). However, the surface of the enlarged kidney of the young diabetic rat was smooth, similar to the kidney from an age-matched nondiabetic control (Fig. 6A). In contrast, a representative kidney of an aged diabetic NGR (14 mo old; nonfed BG=225 mg/dl; plasma TG=160 mg/dl) was markedly reduced in size compared with the kidney obtained from the young diabetic rat, suggesting atrophic changes in the renal tissue with chronic disease (Fig. 6A). Furthermore, the kidney of the aged diabetic animal also showed an irregular, pitted surface, consistent with late stage nephrosclerosis of diabetic nephropathy.

Figure 6.

Macroscopic, histological and ultrastructual images of kidney tissues obtained from nondiabetic and diabetic NGRs. A) Gross appearance of representative kidneys of a normal (left; 5 mo old; nonfed BG=67 mg/dl; plasma TG=43 mg/dl), a young diabetic (middle; 5 mo old; nonfed BG=251 mg/dl; plasma TG=101 mg/dl), and an aged diabetic NGR (right; 14 mo old; nonfed BG=225 mg/dl; plasma TG=160 mg/dl) NGRs. Compared with kidney of a normal young animal, kidney of the age-matched diabetic animal was enlarged; however, it exhibited a smooth surface, which suggests the existence of early stages of diabetic nephropathy. Kidney of aged diabetic rats was relative to those from young diabetic NGRs smaller in size and showed nodular surface, which suggests the presence of atrophic changes. B) Oil-Red-O staining of frozen sections of kidneys obtained from nondiabetic (left panel) and diabetic animals (right panel). Lipid depositions were prominent in the glomeruli (arrow) and proximal tubule epithelial cells (arrowhead) of young diabetic animals (right panel) but not found in normal controls (left panel). Mayer’s hematoxylin was used for counterstaining. Scale bar = 100 μm. C) Representative electron microscopic visualization of the kidney of a normal (left panel) and a diabetic NGR (right panel). Asterisk indicates amorphous deposits.

Light microscopy revealed lipid accumulation in glomeruli and tubular cells on Oil-red-O staining (Fig. 6B). EM of renal tissues from normal and diabetic rats at representative ages revealed effacement of the podocyte foot processes, thickened glomerular basement membrane, and amorphous AGE deposits in the glomeruli, while none of these changes were observed in age-matched nondiabetic animals (Fig. 6C).

To examine β-cell structure, immunofluorescence staining for insulin in normal and diabetic NGRs was performed. Whereas nondiabetic NGRs (5 mo old; nonfed BG=72 mg/dl; plasma TG=49 mg/dl) showed prominent staining of insulin-producing cells in the islets of Langerhans, the amount of insulin positive staining was dramatically reduced in pancreas tissues of diabetic rats (Fig. 7A). Azur-stained histological sections showed dramatically decreased size of the islets of Langerhans in NGRs with diabetes (Fig. 7B).

Figure 7.

Macroscopic, histological, and ultrastructual images of pancreas tissues obtained from nondiabetic and diabetic NGRs. A) Fluorescent micrograph of pancreas tissues immunostained for insulin (red); nuclei counterstained with DAPI (blue). Scale bar = 100 μm. B) Azur staining. In normal NGRs, islets show well-organized endocrine cells, while in the diabetic animals, islets appear degenerated and smaller in size.

Adipocytokine levels in NGRs

To investigate the metabolic correlates underlying the diabetes in NGRs, non-food-deprived levels of plasma adiponectin and leptin in fed rats were assessed. Plasma adiponectin inversely correlated with BW (r=−0.52; n=43; P<0.001; Fig. 8A) and HbA1c (r=−0.64; n=43; P<0.001; Fig. 8B). Non-food-deprived plasma insulin tended to be increased in moderate diabetes related to plasma adiponectin (6.5%≤HbA1c≤9.0%; Fig. 8C). In animals with advanced diabetes (HbA1c≥9.0%), in particular, rats with plasma adiponectin < 10 μg/ml, there was a paucity of plasma insulin (Fig. 8C). Non-food-deprived plasma insulin positively correlated with plasma leptin levels (r=0.73; n=43; P<10⁻⁷; Fig. 8D).

Figure 8.

Adipocytokines in NGRs. A) Correlation between plasma adiponectin and BW in NGRs, with equation of regression y = −0.25x + 44.54; n = 43; P < 10⁻³. B) Correlation between plasma adiponectin and HbA1c in NGRs, with equation of regression y = −0.16x + 10.27; n = 43; P < 10⁻⁵. C) Relationship between plasma level of adiponection and non-food-deprived insulin. Animals were divided into 3 groups, based on level of HbA1c (%), nondiabetic (open circle; HbA1c<6.5%), mild (plus symbol; 6.5% <HbA1c<9.0%), and severe (solid circle; HbA1c>9.0%); n = 43. D) Correlation between plasma leptin levels and non-food-deprived insulin in NGRs, with equation of regression y = 0.16x + 0.75; n = 43; P < 10⁻⁷.

Atherosclerotic changes in diabetic and hyperlipidemic NGRs

To investigate potential vascular manifestations of the hyperlipidemia and hyperglycemia in the NGRs, the aortas of normal and diabetic animals were examined for atherosclerotic lipid depositions. In young or aged normal animals, no significant abnormality was observed (Fig. 9A). In contrast, diabetic NGRs (n=7) with an average nonfed BG of 306.5 ± 42.7 mg/dl showed macroscopically detectable atherosclerotic lesions (12.9±6.4 mm²) total lesion size/aorta. Notably, most lesions were found in the area of the aortic arch and around bifurcation of aortic branches (i.e., at the brachiocephalic trunk, common carotid artery, and subclavian artery) (Fig. 9A). Of the male NGRs (n=15), 73% showed atherosclerosis development, compared to 29% in female NGRs (n=7; P=0.047).

Figure 9.

Atherosclerotic changes in NGRs. A) Morphology of atherosclerotic lesions. Oil-Red-O stained excised aortas of a normal (nonfed BG=93; male; 19 mo of age) and a diabetic NGR (BG=464; male; 14 mo of age). Lesions frequently found in branching areas of the aortic arch and bifurcations. B) Intima hyperplasia, structural damage in media, and lipid depositions as micrscopic correlates of fatty streaks in diabetic NGR. Toluidine blue-stained 1-μm plastic embedded sections at ×40.

Toluidine blue-stained 1-μm plastic embedded sections showed the organized nature of the aortic wall in the control animal, with layers of smooth muscle cells between bands of elastin (Fig. 9B, left panel). In contrast, in the diabetic animal, some of the elastin layers in the media are disrupted by invading epithelial and mononuclear cells (Fig. 9B, right panel).

Hypertension in the NGRs

To investigate whether NGRs develop risk factors of cardiovascular disease aside from metabolic changes, we measured the systemic BP in these animals. Rats ≤ 10 mo old with BW < 120 g had significantly lower BP (128±4.9; n=31) than those with average BW > 120 g (148±10.1; n=14; P=0.048), suggesting an influence of body mass on BP in these animals (Fig. 10A). Diabetic rats (BG>150 mg/dl) showed significantly higher systemic BP values (147±7.3, n=21) compared to normal controls (BG<150 mg/dl; 123±5.3; n=24; P=0.0093; Fig. 10B). Interestingly, the BP values in diabetic animals >10 mo of age (131±3.27; n=29) did not significantly differ from age-matched controls (124±6.1; n=9; P=0.33), which suggests that in severe diabetes, dehydration through increased urine production might counter the elevated BP.

Figure 10.

Hypertension in NGRs. Systolic BP was measured using the tail-cuff technique. A) Contribution of body mass to BP in NGRs (≤10 mo of age). Animals with average BW < 120 g (n=31) had significantly lower BP than those with average BW > 120 g (n=14). B) BP in normal and diabetic animals at different ages, as indicated. Diabetic rats (BG>150 mg/dl; n=21) showed significantly higher systemic BP values compared to normal controls (BG<150 mg/dl; n=24). BP values in diabetic animals >10 mo of age (n=29) did not significantly differ from age-matched controls (n=9).

DISCUSSION

The industrialized countries face a growing epidemic of obesity. With the current rate of weight gain, by 2015 ∼75% of U.S. adults will be overweight and 41% obese (25). The consequences of obesity are debilitating to the public health. Obesity is a risk factor for an array of diseases, a group of which is known as MetS. A preventive or therapeutic intervention for MetS is urgently needed, and understanding the etiology of MetS is critical for the development of an effective treatment. However, the study of MetS is complicated despite a number of available animal models. For instance, the existing experimental rodent models for MetS (12,13,14, 26, 27) may not mimic all aspects of the complex human condition. However, among the advantages of the existing mouse models are availability of reagents and the existence of probes for genetic manipulations.

In this study, we introduce and characterize a novel spontaneously occurring disease model of MetS in NGRs that includes most of the traits observed in the naturally occurring human disease. Our Western blot and ELISA experiments indicate substantial cross-reactivity between mice, rats, and NGRs. Furthermore, should the NGR genome become known in the future, molecular biological techniques would become applicable to this useful animal model.

The growth data show that the majority of NGRs, especially males, kept under normal laboratory environment and fed with normal rodent lab chow develop diabetes by 1 yr of age. The examined rats in this study constitute a representative subsample from a total of 1100 rats bred thus far. Examination of the plasma TG and TC showed that these values became progressively elevated with time in the animals with the disease.

Total body fat, especially intra-abdominal fat, disappears as ketosis evolves in advanced disease. The link between intra-abdominal fat and diabetes is well documented in humans with MetS (28), as well as an inverse relationship between adiponectin and intra-abdominal fat. Matching the human condition, we detected a decline in adiponectin with chronic diabetes (HbA1c) in this model, indicating the opportunity to define the sequence of events involved in the MetS scenario in male NGRs.

In the diabetic animals, a strong correlation was found between plasma TG levels and HbA1c, which suggests a link between the severity of the diabetic status and aberrant lipid metabolism, as is the case in humans with MetS and diabetes. Furthermore, our results suggest an age-related worsening of the metabolic features, measured as nonfed BG and HbA1c, plasma TG, and TC, especially in males. In the diabetic animals, we find macroscopic and microscopic manifestation of the metabolic dysfunction in the liver and kidneys.

An inverse relation between plasma adiponectin and insulin values suggest 3 overall stages of diabetic progression. During the first or prediabetic stage, animals display high adiponectin and normal insulin and BG levels. In the subsequent early diabetic stage, the adiponectin values decline, while insulin levels become elevated with rising BG. These changes are reminiscent of a hyperinsulinemic phase found in human patients with type 2 diabetes. In severely diabetic NGRs, both the adiponectin and insulin values were found to be very low, which suggests a loss of insulin production and adipose tissue, associated with ketosis and wasting. Androgens depress adiponectin in diabetes-susceptible humans (29, 30), which may explain why male NGRs are more prone to diabetes. These changes in adiponectin and insulin appear to represent the various phases of type 2 diabetes in human patients. Histological and ultrastructural examination of the pancreatic β-cells underscores these functional findings, with nearly total absence of detectable insulin-producing cells in the later stages of the disease in our NGRs.

The significant correlation between leptin and plasma insulin levels in the rats suggest an attempt to counterregulate disease progression, for instance through reduction of food intake (data not shown). The fact that in most animals the disease progresses unhindered might suggest a functional leptin resistance or loss of adiponectin, which enhances insulin sensitivity. A central defect may be an inability to produce enough adiponectin (29, 30) or altered leptin transport to the hypothalamus and/or impaired leptin signaling (31). Alternatively, a peripheral defect could exist at the level of the expanding adipocytes, with the consequence of reduced fatty acid oxidation (32).

Another prominent element of MetS is cardiovascular disease, as a consequence of atherosclerosis. In diabetic animals, we found various stages of atherosclerosis formation in the aorta in conjunction with high plasma TC and TG and blood HbA1c. Frequently found colocalization of the atherosclerotic lesions in and around vascular bifurcations suggests an important role for hemodynamic forces. This notion was fostered by observations of older diabetic rats having higher BP values than lighter-weight, younger controls. Rodents fed a normal diet do not develop atherosclerosis. Genetically engineered mice, deficient for apolipoprotein E (apoE) or LDL receptor, develop atherosclerotic lesions and have been a major advancement. However, lesion development in these animals may differ from the normal course in humans. Our finding of atherosclerosis in NGRs fed a normal diet might therefore become useful in studies of vascular aging as it occurs in humans. Future studies will address the plasma distribution of lipoproteins in NGRs and their contribution to atherogenesis in these animals.

An intriguing finding in NGRs is that adiposity per se does not correlate with BG; however, the heavier rats do show the highest BG. The most sensitive indicator for non-food-deprived BG (much better than nonfed BG) is total BW and liver weight, presumably through fat accumulation (steatosis). This finding suggests that the disease in NGRs may start with the inability of the liver to handle the BG and a shifting toward fat synthesis and storage, equivalent to the hepatic steatosis of type 2 patients with MetS. Surprisingly, total fat (i.e., obesity assessed as sum of the three largest fat pads) has almost no correlation to non-food-deprived BG. It is interesting that the perirenal fat pad (that might be interpreted as the abdominal fat equivalent in humans) does appear related. This finding corresponds to the correlation between abdominal obesity and type 2 diabetes in humans.

In sum, we describe a new model of spontaneously occurring metabolic syndrome in NGRs with all pertinent characteristics of the human condition. This model provides a powerful tool for studies of the mechanisms of the MetS.