Transient prenatal androgen exposure produces metabolic syndrome in adult female rats

Animal treatments and diet.

All animal procedures were approved by the Animal Care and Use Committee at Northwestern University (Evanston, IL). Rats were housed at 23–24°C on a 10:14-h light-dark cycle.

Time-pregnant female Sprague-Dawley rats were obtained from Charles River (Portage, WI) at day 14 of gestation and treated from embryonic day (E) 16 to E19 with daily subcutaneous (sc) injections of 5 mg of free testosterone (T-1500; Sigma, St. Louis, MO) dissolved in 500 μl sesame oil (S3547; Sigma)/benzyl benzoate (B6630; Sigma) (prenatally androgenized: PA n = 8) or of 500 μl sesame oil vehicle as a control (C, n = 12). This hormonal paradigm mimics the fetal testosterone surge that is observed in male rats (17). In preliminary experiments, we found no significant difference in either the female-to-male offspring ratio, number of pups per litter, or birth weights between PA and C animals (unpublished observations). Pups were culled from control litters to equalize group sizes. All litters were weaned, and females were separated from males at 21 days of age and fed regular chow diet (RD) (3.30 kcal/g: kcal from fat 15%, carbohydrate 62%, and protein 23%; 7912 Harlan-Teklad, Madison, WI) and water ad libitum. Body weights were monitored weekly from the day of weaning to tissue harvest at age 60 days.

Study protocols.

At 35 days of age, female rats were either kept on RD (n = 37) or switched to the high-fat diet (HFD; n = 19) (5.24 kcal/g: kcal from fat 60%, carbohydrate 20%, and protein 20%, D12492 Test Diet; Purina Mills, Richmond, IN). There were 20 C RD, 17 PA RD, 11 C HFD, and 8 PA HFD female rats. Adult female rats (aged 60–63 days) on their respective diets were fasted overnight, anesthetized with CO₂, weighed, and decapitated. To examine insulin signaling, one-half of the animals in all experimental groups were injected intraperitoneally with 5 U of regular human insulin (Novo Nordisk, Princeton, NJ) in 100 μl 0.9% sodium chloride (Abraxis, Schaumburg, IL), and half with 100 μl 0.9% sodium chloride 10 min before death. Trunk blood was collected and centrifuged, and serum was stored at −80°C. Glucose, insulin, cholesterol, and triglyceride (TG) levels were determined in samples from saline-treated animals only (10 C RD, 8 PA RD, 6 C HFD, 4 PA HFD). Liver, adipose tissue depots (parametrial and sc), and soleus muscles were excised, weighed, frozen, and stored at −80°C for further analyses. Tissues from both saline- and insulin-treated rats were used for analyses of body composition and hepatic TG content (15 C RD, 13 PA RD, 11 C HFD, 8 PA HFD), since administration of insulin would not acutely alter these end points (Fig. 1).

Dynamic studies.

Dynamic studies of glucose homeostasis were performed in separate groups of animals at age 60 days after a 6-h fast with jugular catheters implanted 1 day before testing. Intraperitoneal (ip) glucose tolerance test (IPGTT) was performed in 5 C RD and 6 PA RD rats. A baseline blood sample was followed by intraperitoneal injection of 2 g/kg body wt dextrose and blood sampling at 2, 5, 10, 15, 30, 60, 90, and 120 min. Intraperitoneal insulin tolerance test (IPITT) was performed in a separate group of 4 C RD and 6 PA RD rats. A baseline blood sample was obtained followed by ip injection of insulin (5 U/rat) with blood sampling at 15, 30, 60, 90 and 120 min.

Hormone assays and hepatic TG content.

Whole blood glucose levels were measured using a Prestige Smart System Glucose Monitor (Home Diagnostics, Ft. Lauderdale, FL). Insulin levels were measured by a rat insulin ELISA kit (Crystal Chem, Downers Grove, IL). Circulating TG and cholesterol levels and hepatic TG content were assessed by spectrophotometric assay (Infinity Triglyceride Reagent Kit; Thermo Electron, Pittsburgh, PA). Hepatic TG content was expressed as percent of protein content as described previously (41). Frozen livers were stained with Oil red-O for lipids (19).

Western blotting.

Liver and soleus muscle lysates (100 μg protein) were resolved using SDS-PAGE on 7.5% gels and incubated with specific antibodies to insulin receptor substrate (IRS)-1 (Upstate Biotechnology, Lake Placid, NY), IRS-2 (gift of Dr. M. White; Joslin Diabetes Center, Boston, MA), protein kinase B (Akt), phospho-Akt (Ser⁴⁷³), extracellular signal-regulated kinase (ERK)1/2, phospho-ERK1/2 (Thr²⁰²/Tyr²⁰⁴) (Cell Signaling, Beverly, MA), insulin receptor β-subunit (IR-β; BD Transduction Laboratories, San Jose, CA), and appropriate secondary antibodies (goat anti-rabbit and -mouse horseradish peroxidase conjugates; Cell Signaling), with PA and C samples run together on the same gels. Bands were visualized and analyzed as described previously (10). Phosphorylation was expressed as the ratio of phosphorylated to total protein.

Statistical analysis.

Repeated-measures two-way ANOVA with time and treatment as factors was applied to body weight, IPITT, and IPGTT with a Bonferroni's post hoc test to determine which groups differed significantly. For all other analyses, comparisons of interest, 1) C RD vs PA RD, 2) C RD vs C HFD, 3) C HFD vs PA HFD, and 4) PA RD vs PA HFD, were made separately using the appropriate parametric (Student's t-test) or nonparametric (Mann-Whitney) tests according to the sample size (n ≤ 6, nonparametric test used) and/or the normality of the data. A Bonferroni correction for multiple testing was used to adjust the threshold for statistical significance to P < 0.0125 for these four comparisons. The data were transformed when necessary to achieve homogeneity of variance. Data are reported as means ± SE. All analyses were performed using GraphPad Software (GraphPad Software, San Diego, CA).

Baseline features.

PA treatment did not result in changes in body weight at weaning (Fig. 2). Beginning at 42 days, body weight began to increase in PA compared with C rats (Fig. 2). PA rats had significantly increased body weight compared with C rats at 60 days (P < 0.01, Fig. 2). The parametrial and subcutaneous fat depots were increased by 30% at 60 days in PA compared with C rats (P < 0.0125, Fig. 2). HFD produced similar changes in C rats in body and fat pad weights to those observed in PA rats with RD and did not further change these parameters in PA rats (Fig. 2). PA rats exhibited ∼12-fold higher fasting insulin levels compared with C rats at 60 days (P < 0.005, Fig. 3). However, fasting glucose levels at 60 days did not differ in PA compared with C rats (Fig. 3). HFD produced a significant ∼2.5-fold increase in fasting insulin levels in C rats (C RD vs C HFD, P < 0.0125). In PA rats, HFD did not further increase circulating insulin levels (Fig. 3).

There was an increase in both fasting TG and cholesterol levels in PA compared with C rats (P < 0.0125, Fig. 4). HFD increased TG levels in C rats and did not further influence them in PA rats (Fig. 4). Cholesterol levels in C and PA rats on HFD did not differ from those in C RD rats. Hepatic TG content was increased in PA compared with C rats (P < 0.0125), and the increased lipid content was also visualized with oil red-O stain (Fig. 4). HFD produced greater increases in hepatic TG content in both PA and C rats than those seen with PA alone (Fig. 4).

Dynamic testing.

Following IPGTT, there were no differences in glucose or insulin responses between PA and C groups (Fig. 5). There were no differences in insulin sensitivity assessed by IPITT in PA and C rats (Fig. 5).

Insulin signaling.

There were no significant changes in IRS-1 or IRS-2 protein abundance in PA compared with C rats in liver or skeletal muscle with either diet (Fig. 6). There were no differences in the abundance of IR-β in the livers of PA compared with C rats on either diet (data not shown). There were also no differences in insulin-stimulated activation of Akt (Fig. 7) or of ERK1/2 (not shown) in PA compared with C rat skeletal muscle or liver with either diet.