Early-Life Stress and the Development of Obesity and Insulin Resistance in Juvenile Bonnet Macaques

The colony consists of ∼250 bonnet macaques (Macaca radiata), born and reared in a social, seminaturalistic environment (11). Our monkeys begin eating solids around 2 months of age, with weaning beginning at 7–8 months. Regardless of rearing condition, monkeys are separated from their mothers and socially housed with peers at the end of the 1st year of life. Males reach puberty at ∼4–4.5 and females at 3–3.5 years; full maturity is attained at 7–8 years of age. The control rearing condition entails ad libitum access to a commercial laboratory diet (Monkey Diet 5038; LabDiet, Richmond, IN; 4 kcal/g, 69% carbohydrate, 18% protein, and 13% fat), regularly supplemented with fresh fruit. All monkeys are treated ethically, under careful veterinary supervision in accordance with the State University of New York–Downstate Medical Center–Institutional Animal Care and Use Committee and Association for Assessment and Accreditation of Laboratory Animal Care and National Institutes of Health guidelines.

The primary sample consisted of 48 peripubertal bonnet monkeys (30 VFD and 18 controls; 25 males and 23 females; aged 3–4 years). A subset of 15 male macaques matched for weight and BMI (8 VFD and 7 controls) underwent euglycemic-hyperinsulinemic clamp studies.

Mother-infant dyads are randomly assigned to control or treatment rearing conditions shortly after birth. Beginning when their infants are ∼3–5 months old, mothers of the VFD-reared subjects confront ∼4 months of circumstances in which the time and effort required to obtain food are either relatively brief and easy (low foraging demand [LFD]; essentially ad libitum access) or more lengthy and difficult (high foraging demand [HFD]). Alternating periods of LFD and HFD in 2-week blocks is termed VFD. Animals always have ample food, and frequent weight and health checks confirm normal growth and development in all subjects. After the experimental period and at the time of testing, all offspring are on ad libitum feeding. Further details of the rearing procedure are provided by Rosenblum and Smiley (3).

There were no differences between the randomly selected mothers undergoing VFD versus the controls with respect to parity or phenotypic characteristics related to body size, insulin resistance, fat distribution, obesity, or blood lipids (D.K., M.A.B., I.S., E.L.P.S., J.D.C., L.A.R., J.G.K., unpublished data). Furthermore, baseline levels of plasma and csf cortisol and CRF were similar in both groups of mothers. Similarly, there were no recorded phenotypic differences between mothers with different timing of the VFD experience (“early” vs. “late” postpartum) (15).

After an overnight fast (food was withdrawn at 1600 h, while water remained available ad libitum), between 0800 and 1100 h, monkeys were individually captured in carrying cages, placed in single-animal squeeze cages, and anesthetized with ketamine (10–15 mg/kg). Venous blood was drawn from antecubital or femoral veins and was immediately placed on ice, centrifuged for plasma separation, and stored at −80°C within 1 h in plain nonheparinized tubes, as described previously (11). Samples drawn during clamp studies were handled similarly. Samples were analyzed by routine laboratory procedures for glucose; TGs; and total, HDL, and LDL cholesterol at the University Hospital of Brooklyn Clinical Chemistry Laboratory (Brooklyn, NY). Insulin, C-peptide, glucagon-like peptide-1 (GLP-1), adiponectin, leptin, and ghrelin levels were analyzed using the Human Endocrine Lincoplex kit (Linco Diagnostic Services). As simple measures of insulin resistance, IGR and the HOMA (16) were calculated.

During anesthesia for blood sampling, weight in kilograms, crown-rump length (CRL), and abdominal circumference were measured. CRL is the length in centimeters from the vertex of the head to the base of the tail. Abdominal circumference is the largest distance in centimeters around the abdomen at the level of the iliac crests. Measurements were performed by the same team of investigators blinded to rearing condition. A modified BMI was calculated as mass in kilograms divided by the square of the CRL in meters (17).

All euglycemic-hyperinsulinemic clamp studies were conducted by a team of two trained investigators at 0900 h after a 17-h period of food deprivation following procedures described by Bodkin et al. (18) in rhesus monkeys. In the afternoon before the day of the clamp, monkeys were captured in carrying cages and transferred to individual squeeze cages for overnight fasting. To decrease the stress of dehabituation and social isolation, experimental animals were kept in a pen along with other similarly aged individually caged monkeys with food in their cages. To reduce muscle fasciculation associated with ketamine, a muscle relaxant, xylazine, was added to the anesthesia. On the morning of the procedure, an intramuscular injection of a 9:1 mix of ketamine (100 mg/ml) and xylazine (20 mg/ml) solutions was used for induction of anesthesia and readministered as needed to achieve complete sedation throughout the procedure. The monkey was placed in a neonatal incubator with an ambient controlled temperature of 25–30°C. Heart rate, sphygmomanometric blood pressure, and core temperature measurements were continuously monitored. Bilateral antecubital or saphenous veins were cannulated, with one line used for infusates and the other for blood sampling. To maintain line patency, 1 cc heparinized saline (500 units heparin sodium in 60 cc 0.9% normal saline) was infused in the sampling line with each blood draw.

After baseline fasting plasma glucose levels were taken, a priming infusion of porcine insulin was administered followed by a continuous infusion (>40 mU/m² body surface area per min) for 120 min. Body surface area was calculated using a modified DuBois formula (19). Simultaneously, a variable-rate solution of 20% glucose was infused to maintain a euglycemic state (∼85 mg/dl) determined by blood samples (0.05–0.1 cc) taken every 5 min from the sampling line, centrifuged for plasma separation, and analyzed by a Beckman Glucometer 2 (Beckman Instruments, Brea, CA). Glucose disposal is determined as the mean glucose infusion rate in milligrams per kilogram per minute during the 0- to 30-, 30- to 60-, 60- to 90-, and 90- to 120-min intervals (M rates). An additional 3 cc blood was drawn at 0, 30, 60, 90, and 120 min for a related study. Upon completion of the procedure, the cannulas were removed, and the animal was placed in a specially prepared cage and allowed to recover completely. Monkeys were returned to their home pens in late afternoon or early morning.

Means ± SE are reported throughout. Sex distribution comparisons used χ² analysis. Simple between-group comparisons used two-tailed Student's t tests, ANCOVA covaried for age and sex, as indicated, and repeated-measures ANCOVA covaried for age. Partial correlations were controlled for age. Not all monkeys had values for every parameter examined. Missing data led to exclusion on a measure-by-measure basis. Outlier values (>3 SD from the mean) for weight and BMI in one VFD female; for GLP-1 in one VFD female; for insulin, IGR, and HOMA in one VFD female; and for glucose, IGR, HOMA, and TGs in another control female led to exclusion from respective ANCOVAs and correlations. All statistical analyses were made using SPSS for Windows (release 10.0.1; SPSS, Chicago, IL) with P < 0.05 as the criterion for significance.

VFD juveniles had significantly greater weight [F_(1,43) = 7.42; P = 0.009], BMI [F_(1,43) = 10.25; P = 0.003], abdominal circumference [F_(1,44) = 4.81; P = 0.034], and VLDL [F_(1,44) = 7.37; P = 0.009] than controls when covaried for age and sex (Table 1). No rearing group differences were noted for insulin, C-peptide, glucose, IGR, HOMA, adiponectin, ghrelin, or leptin. Leptin values were detected in only seven VFD and six control animals. GLP-1 levels were greater in VFD juveniles [F_(1,40) = 5.14; P = 0.029; Fig. 1], with no sex or sex by rearing group interactions. GLP-1 levels were inversely related to BMI (r = −0.49; P = 0.011).

Females across both rearing groups had higher levels of insulin [F_(1,31) = 5.32; P = 0.028], IGR [F_(1,30) = 4.65; P = 0.039], HOMA [F_(1,30) = 4.22; P = 0.049], and TGs [F_(1,44) = 12.32; P = 0.001] but weighed less than males [F_(1,44) = 10.81; P = 0.002] when covaried for age. Females also had greater adiponectin levels [F_(1,45) = 6.15; P = 0.017], with no rearing or sex by rearing interactions and had marginally higher C-peptide [F_(1,45) = 3.8; P = 0.058] and lower ghrelin levels [F_(1,37) = 4.06; P = 0.051; data not shown]. HDL, LDL, and total cholesterol did not differ by sex or rearing group status.

Blood pressure during anesthesia was highly variable and did not exhibit any statistically significant differences, consistent with our findings across ages, sex, rearing conditions, and methods of anesthesia (ketamine alone vs. ketamine plus xylazine) and of blood pressure measurement.

Among the male monkeys who underwent clamps, no differences in age-adjusted fasting plasma glucose measures were noted between control and VFD juveniles [VFD = 83.8 ± 2.74 vs. control = 83.2 ± 2.74; F_(1,11) = 0.018; P = 0.896]. Similarly, no rearing group differences were noted for fasting insulin, IGR, or HOMA (all P > 0.4). VFD M rates, however, were lower at all 30-min clamp time intervals, with an overall significant difference between VFD and controls [repeated-measures ANCOVA; F_(1,12) = 5.5; P = 0.037; Fig. 2]. Additionally, across both rearing groups, BMI was negatively correlated with M rates at the 0- to 30-min (r = −0.69; P = 0.003), 30- to 60-min (r = −0.51; P = 0.044), 60- to 90-min (r = −0.72; P = 0.002), and 90- to 120-min (r = −0.66; P = 0.006) intervals, whereas GLP-1 was negatively correlated with M rates at the 30- to 60-min (r = −0.68; P = 0.007), 60- to 90-min (r = −0.59; P = 0.026), and 90- to 120-min (r = −0.55; P = 0.042) intervals.

Several experimental models demonstrate distinct relationships between disturbances of early rearing and persistent behavioral alterations (20,21), modulation of neuroendocrine and immunological systems (22–24), and permanent changes in neuroanatomical architecture (25). To our knowledge, this is the first study to report differences in metabolic parameters relating to obesity and insulin resistance in differentially reared nonhuman primates.

VFD rearing resulted in higher weight, BMI, and abdominal circumferences. In light of the chronic behavioral and neuroendocrinological effects of VFD rearing, further studies of consumption parameters and their relationship to measures of central stress may determine whether increased ingestion of “comfort food” (26) led to this difference when compared with controls. Strong negative correlations between BMI and M rates demonstrate the close clinical association between obesity and insulin resistance in this species. As studies of obese adult rhesus macaques of Hansen and Bodkin (27) have shown, pancreatic β-cell hyperresponsiveness is the earliest stage in the progression of insulin resistance to overt diabetes. In our colony, hyperinsulinemia is present in juvenile monkeys with the metabolic syndrome (11), yet VFD juveniles, although more obese than controls, did not have different levels of insulin, glucose, IGR, or HOMA. They did, however, exhibit lower M rates by hyperinsulinemic-euglycemic clamps. Unlike simple indicators, the clamp procedure measures in vivo glucose disposal as an index of whole-body tissue sensitivity to insulin.

GLP-1, a “brain-gut” peptide (28), is the most potent intestinal incretin hormone, which has been shown to enhance pancreatic β-cell maturation and proliferation, in fact promoting glucose disposal (28,29). In VFD juveniles in our stress-related model of insulin resistance, early increases in GLP-1 with robust negative correlations with M rates imply that circulating GLP-1 is an early marker of stress, contemporaneous with adrenergic inhibition of insulin secretion. Longitudinal studies are required to ascertain whether GLP-1 participates in a neuro-adaptive response involved in the pathogenesis of type 2 diabetes.

Consistent with our previous experience, compared with males, peripubertal females showed general trends toward insulin resistance and the metabolic syndrome, including higher insulin, IGR, HOMA, and TGs, despite weighing less. This could be due to the earlier appearance of puberty in females of this species (30), although it remains to be determined whether the puberty-related insulin resistance normally seen in humans (31) is present in our macaques.

Adiponectin, an adipose-specific peptide with activity in the brain (32), was elevated in the early stages and decreased in the later stages of progression to insulin resistance and diabetes in adult rhesus macaques (33). As with the increased GLP-1 levels in VFD juveniles, elevations of this peptide with clear protective anti-atherogenic properties and strong correlations with insulin sensitivity (33), may represent a compensatory response to the earliest emergence of insulin resistance in peripubertal females.

Most experimental models of obesity and insulin resistance use gestational and postnatal nutritional and endocrinological manipulations (34,35,36). These studies underscore the importance of both plasticity and vulnerability of susceptible neural circuits in establishing and defending adult body weight and energy homeostasis (37). To our knowledge, ours is the only model of stress-related insulin resistance in juvenile nonhuman primates, without conventional factors that lead to obesity, such as dietary manipulation or restricted physical activity. The early decreases of in vivo insulin action may suggest that maternal food insecurity during nursing, a perceived threat to nutrition serving as an allostatic load (38), may trigger in the offspring an incretin response enabling glucose utilization, which becomes pathological through progression into older age. In separate analyses of previously published maternal plasma cortisol levels and csf CRF elevations after VFD, we found positive correlations with juvenile offspring BMI levels in both sexes but more evident among males (15). Furthermore, this percentage elevation in maternal csf CRF was positively correlated with insulin resistance (HOMA) in the young offspring. Interestingly, offspring GLP-1 levels were positively related to maternal subordinance (D.K., M.A.B., I.S., E.L.P.S., J.D.C., L.A.R., J.G.K., unpublished data), supporting the concept that GLP-1 is an early stress-linked response. We conclude that our validated experimental model of stress-related to adverse rearing predicts the early appearance of obesity and insulin resistance.

This work has received support from a grant from the Smart Family Foundation and from the Dean's Research Initiative Award at State University of New York Downstate Medical Center.

We thank Dr. Mustafa Tiewala, Dr. Mohammed Arif, Dr. Bruce Scharf, Shirn Baptiste, and Douglas Rosenblum for their contributions to this study.