Animals and treatments

Female rats of the Sprague Dawley strain were housed in group cages under controlled conditions of light (12 h light, 12 h dark) and temperature (23–25°C). Animals received food and water ad libitum. Virgin female rats were mated with fertile males of the same strain. Day 1 of pregnancy was defined as the morning in which spermatozoa were observed in the vaginal fluid. Between day 16 to day 19 of pregnancy, rats were hyperandrogenized as described before [8]. Briefly, pregnant rats (N = 120) received subcutaneous injections of either 2 mg or 5 mg of free testosterone (T-1500; Sigma, St. Louis, MO) dissolved in 100 µl sesame oil from day 16 to day 19 of pregnancy whereas the control group received only 100 µl of sesame oil. The doses of T2 of free testosterone used result in circulating testosterone levels that are similar to male rats [8], [39]. Under the conditions of our animal facilities, spontaneous term labor occurs on day 22 of gestation. The treatments described did not modify the spontaneous term labor, the female-to-male offspring ratio or the number of pups per litter. Pups were culled from litters to equalize group sizes (ten pups/each mother). Females were separated from males at 21 days of age and randomly chosen. Animals were allowed free access to Purina rat chow and water. All procedures involving animals were conducted in accordance with the Animal Care and Use Committee of Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) 1996. The Ethic Committee of Facultad de Medicina (UBA) approved the present study. The group of female offspring prenatally treated with 2 mg testosterone was designated as T2, whereas that corresponding to those prenatally treated with 5 mg was designed as group T5 and that without treatment as control group.

Study protocols

To examine whether prenatal hyperandrogenization altered the intra-uterine development, offspring from animals of the three groups were weighed at 21 and 60 days of age. The offspring were not weight at birth because at that time it was almost impossible to determine reliably the sex of the rats. In addition, the weight at birth has a significant higher error than at 21 days and methodologically it is difficult since manipulation of the offspring could conduce to the rejection of the mother. To establish if prenatal hyperandrogenization induced defeminization, the uro-genital distance was determined at 60 days of age. In order to study whether treatments altered the estrous cycle, the estrous cycle was determined by vaginal smears taken daily from 45 to 60 days of age (day of sacrifice). At 60 days of age, sixty female Sprague Dawley rats per group of treatment were anesthetized with carbon dioxide and killed by decapitation. Trunk blood was collected and serum was separated by centrifugation at 1000 g for 15 min and stored at −80°C until estradiol, progesterone and testosterone were determined by radioimmunoassay (RIA). Ovaries were immediately removed from each group (control, T2 and T5) and divided as follows: 10 were immediately fixed in 4% (w/v) paraformaldehyde for histological studies whereas 40 were immediately frozen at −80°C. Then, 10 of the latter were used for PGE determination by RIA whereas the other 30 used to determine the ovarian oxidant-antioxidant balance (10 ovaries were used to determine the lipid peroxidation index, 10 to determine the nitric oxide synthase activity and 10 ovaries to determine the total glutathione). Other 10 ovaries per group were immediately homogenized in western blotting buffer and stored at −20°C until protein expression of cyclooxygenase 2 (COX2), the limiting enzyme of PGs synthesis, steroidogenic acute regulatory protein (StAR) and PPAR gamma.

Intraperitoneal glucose tolerance test (IPGTT)

As a measurement of glucose homeostasis and insulin resistance, the intraperitoneal glucose tolerance test (IPGTT) was performed in separate groups of ten animals per group (control, T2 and T5) at 60 days of age after 8 h fast [8]. Briefly, a baseline blood sample was obtained followed by intraperitoneal injection of 2 g/kg body weight dextrose with blood sampling at 30, 60, 90 and 120 min.

Fasting blood glucose was determined by using the Haemoglukotest (Roche) test strips for visual determination in the range of 20–800 mg/100 ml (1–44 mmol/l). The test principle uses the glucose-oxidase/peroxidase reaction. Results are expressed in mg glucose/dl.

Morphological studies

To study the effect of prenatal hyperandrogenization on the ovarian function, ten ovaries from each group were fixed as described above, were embedded in paraffin wax and consecutively cut. To prevent counting the same follicle twice, 4-µm step sections were mounted at 50-µm intervals onto microscope slides according to the method described by Woodruff et al. [40]. To count the number of different stages of follicles per ovary section, a set of slides was stained with hematoxylin and eosin. An atretic follicle was defined as the follicle that presented more than 10 pycnotic nuclei per follicle; in the smallest follicles, the criterion for atresia was a degenerate oocyte, precocious antrum formation, or both [41], [42]. For morphological analysis, the sections were chosen as follows: five from each extreme and five from the middle of each ovary. Five ovaries from each group were observed for three different researchers.

Progesterone and estradiol radioimmunoassay (RIA)

Serum progesterone and estradiol levels were determined by specific RIA as described before [19]. Briefly, serum samples from ten rats per treatment were extracted with the same volume of diethyl ether three times. The extracts were collected and evaporated in a vacuum oven and saved at −80°C until the RIA was performed. The antibodies from progesterone and estradiol were provided by Dr Niswender (Colorado State University, Fort Collins, CO, USA). Both sensitivities were 5–10 pg/tube, 2–5 µl of serum. Results are expressed as ng progesterone or estradiol/ml serum.

Testosterone radioimmunoassay (RIA)

Testosterone was quantified by RIA as previously described [43], [44]. Briefly, serum samples were extracted as described for progesterone and estradiol. The utility range of the assay was 25–1600 pg. The intra-assay and inter-assay variations were 7.5 and 15.1%, respectively. Results are expressed as pg testosterone/ml serum.

Prostaglandin radioimmunoassay (RIA)

To study whether prenatal hyperandrogenism was able to induce a pro-inflamatory state in ovarian tissue, PGE was determined by RIA as previously reported [19]. Briefly, one ovary per point (ten points per treatment) was weighed and homogenized three times in ethylic alcohol 1∶5 (weight of tissue: volume of alcohol) at room temperature. The extracts were collected and evaporated in a vacuum oven and saved at −80°C until the RIA was performed. PGE was quantified by using rabbit antiserum from Sigma Chemical Co. Sensitivity was 10 pg/tube and cross-reactivity with other PGs was <0.1%. Results are expressed as pg/mg tissue.

Oxidative stress-related parameters

Ovarian nitric oxide synthase activity

Nitric oxide synthase (NOS) activity was evaluated as a measure of NO produced by ovarian tissue. NOS was quantified by monitoring the production of [L-¹⁴C] citrulline from [L-¹⁴C] arginine as described previously [19]. Briefly, the frozen ovarian tissue (one ovary per point and ten points per treatment) was homogenized (Tissuemizer Tekmar; Thomas Scientific, Swedesboro, NJ, USA) at 0°C in three volumes of 50 mM Hepes, 1 mM DL-dithiothreitol, 1 mM NADPH and 50 mM L-valine, pH = 7.5. Samples were incubated at 37°C for 15 min with 10 microM [¹⁴C] arginine (0.3 microCi; 1 Ci = 37 GBq). The samples were centrifuged for 10 min at 1000 g and then applied to 1 ml of DOWEX AG50W-X8 (Na+ form; Bio-Rad, Hercules, CA, USA) resin. The radioactivity was measured by liquid scintillation counting. Results are expressed as pmol/g.min.

Glutathione (GSH) content

The antioxidant metabolite GSH was quantified as previously described [23]. The reduced form of GSH comprises the bulk of cellular protein sulphydryl groups. Thus, measurement of acid-soluble thiol is used to estimate the GSH content in tissue extracts. Briefly, one ovary per point and ten points for group was homogenized in homogenization buffer (EDTA (1 mM), KCl (150 mM), beta Mercoptoethanol (1 mM), Trizma base (20 mM) and sacarose (500 mM), pH = 7.6) and centrifuged at 800 g for 10 minutes at 4°C. Then, supernatants (50 microl/point) were incubated with 800 microl of 1.5 M Tris buffer (pH 7.4) containing 50 microl of 5.10⁻³ M NADPH and 6 IU of GSH reductase. The reaction involves the enzymatic reduction of the oxidized form (GSSG) to GSH. When Ellman's reagent (a sulphydryl reagent 5, 5-dithiobis-2 nitrobenzoic acid; Sigma and Co, St Louis, MO, USA) is added to the incubation medium, the chromophoric product resulting from this reaction develops a molar absorption at 412 nm that is linear to the first beyond 6 min; after this, the reaction remains constant. Results are expressed as microM GSH/mg protein.

Western blotting

Ovarian tissue were lysed for 20 min at 4°C in lysis buffer (20 mM Tris–HCl, pH = 8.0, 137 mM NaCl, 1% Nonidet P-40 and 10% glycerol) supplemented with protease inhibitors (0.5 mM PMSF, 0.025 mM N-CBZ-L-phenylalanine chloromethyl ketone, 0.025 mM N′-p-tosyl-lysine chloromethyl ketone and 0.025 mM L-1-tosylamide-2-phenyl-ethylchloromethyl ketone). The lysate was centrifuged at 4°C for 10 min at 10,000 g and the pellet discarded. Protein concentrations in the supernatant were measured by the Bradford assay (Bio-Rad). After boiling for 5 min, 90 microg of protein from each sample was applied to an SDS-polyacrylamide gel (10%) and electrophoresis was performed at 100 Volts for 1.5 h. The separated proteins were transferred onto nitrocellulose membranes in transfer buffer (20% methanol, vol/vol; 0.19 M glycine; 0.025 M Tris-Base, pH = 8.3) for 1 h at 4°C. Blots were blocked for 1.5 h in TBS (4 mM Tris–HCl, pH = 7.5, 100 mM NaCl) containing bovine serum albumin (0.1%) at room temperature. Rabbit polyclonal anti-COX2 (Santa Cruz Biotechnology, Inc., USA) (1∶200 overnight), rabbit polyclonal anti-StAR (Cayman, Ann Arbor, MI, USA) (1∶2000 overnight) or rabbit polyclonal anti-PPAR gamma (Cayman, Ann Arbor, MI, USA) (1∶2500 overnight) were used as primary antibodies. Rainbow-colored protein mass markers (14.3–200 kDa, Bio-Rad) were applied to samples to determine the bands of COX2 (72 kDa), StAR (30 kDa) and PPAR gamma (67 kDa). Protein bands were visualized by incubating the blots with biotin-conjugated secondary anti-rabbit IgG (1∶5000, 1 h) followed by streptavidin–peroxidase complex and diaminobenzidine solution. Consistency of protein loading was evaluated by staining the membranes with Ponceau-S and applying the protein beta actin (43 kDa) (Sigma Co, USA). The intensities (area×density) of the individual bands on western blots were quantified by densitometry (Model GS-700, Imaging Densitometer, Bio-Rad). The experiment was independently repeated three times. Results are expressed in arbitrary units.

Protein concentration

Ovarian protein concentration was determined as described by Bradford [45].

Statistical analysis

Statistical analyses were carried out by using the Instant program (GraphPad software, San Diego, CA, USA). ANOVA followed by Newman- Keuls test were used to compare all pairs of columns. A Bonferroni correction for multiple testing was used to adjust the threshold for statistical significance to P<0.05.

Prenatal hyperandrogenization and body weight

In order to determine whether prenatal hyperandrogenization affected fetal development, we determined the body weight at 21 and 60 days of age. We found that hyperandrogenization induces an adverse intrauterine condition since it diminished the body weight at 21 days (Fig. 1A, a vs b P<0.01; b vs c P<0.0001; a vs c P<0.001) as compared with controls. Therefore, the higher dose of androgen injected caused a significant decrease in body weight (Fig. 1; T2 vs T5 P<0.0001). This adverse effect of prenatal hyperandrogenization was compensated when the animals were 60 days of age since no significant differences were found between groups (Fig. 1B). This mechanism of compensation in the body weight was clearly manifested by the slope of the growth curve (Fig. 1C). We also found that the “compensatory mechanism” was directly related to the dose of testosterone prenatally injected, thus, the slope of the growth curve (slope of control group = 3.717; slope of T2 = 3.974; slope of T5 = 4.160) increased (r = 0.99, p = 0.0001) with the dose of testosterone (r = 0.66, p = 0.0001).

Prenatal hyperandrogenization and glucose homeostasis

The IPGTT used to evaluate homeostasis of glucose. Figure 2 showed that prenatal hyperandrogenization induced increased levels of circulating glucose. We also found that this effect was directly related to the dose of testosterone prenatally injected. The area under the curve (control = 16557±200; T2 = 18225±150; T5 = 19638±130 arbitrary units) of glucose concentration clearly shows that circulating glucose was significantly higher in T5 than in T2 (P = 0.0001, r = 0.999).

Glucose homeostasis and body weight

Increased circulating glucose after IPGTT directly correlated (r = 0.99, p = 0.0001) with the slope of the growth curve (r = 0.66, p = 0.0001).

Prenatal hyperandrogenization and estrous cycle

The uro-genital distance (UGD) determined at 60 days of age showed that prenatal hyperandrogenization induced defeminization since the UGD was significantly increased in T2 and T5 (control = 1.36±0.15; T2 = 1.66±0.20; T5 = 1.75±0.08 cm; T2 vs control P<0.05; T2 vs T5 P<0.001). With respect to the sexual cycle, we found that 80/80 (100%) of control rats showed a regular estrous cycle (4–6 days). In contrast, only 16/80 (20%) of rats from T2, showed a regular estrous cycle (4–6 days) and T5 rats showed vaginal opening atresia. Vaginal smears from rats from T2 group which did not ovulate showed that these rats stayed at diestrous stage.

Prenatal hyperandrogenization and Ovarian Steroidogenesis

Since hormonal levels change through the estrous cycle, we determined both serum progesterone and estradiol levels in two well differentiated stages of the cycle: proestrus and diestrus. We chose diestrus because rats from T2 which did not ovulate stayed at that stage and we chose proestrus because it is the stage of follicular development which is one of the altered conditions in PCOS. We found that T2 rats showed increased serum progesterone levels as compared with controls in both proestrus and diestrus stages without significant differences between them (Fig. 3A; a vs b; a vs c and b vs c P<0.0001). Similarly, T5 rats showed significantly increased serum progesterone levels as compared to controls. No differences were found between rats from T2 and T5 (Fig. 3A). As expected, serum levels of estradiol were lower in control rats in diestrus than in control rats in proestrus (Fig. 3B; P<0.0001). T2 and T5 showed decreased levels of serum estradiol as compared with control rats in proestrus (Fig. 3B; P<0.0001). Serum testosterone levels were increased in both T2 and T5 as compared with controls (Fig. 3C; P<0.0001). Rats from T5 showed higher serum testosterone levels than rats from T2 (Fig. 3C; P<0.0001).

Prenatal hyperandrogenization and ovarian morphology

Ovaries from T2 rats (Fig. 4A–K) and ovaries from T5 rats (Fig. 4L–O) were smaller than controls (control = 55±9 mg vs T2 = 38±5 mg; P<0.005; control = 55±9 mg vs T5 = 42±6 mg, P<0.005). Ovaries from T2 rats had a larger number of developing follicles in early stages (primary and secondary follicles: PF and SF respectively) than controls. These data were quantified in Fig. 5: b vs a and e vs d; P<0.0001. Ovaries from T2 rats showed smaller number of antral follicles (Fig. 4G) than ovaries from control rats (Fig. 4A). These data were quantified in Fig. 5: h vs g P<0.0001. In contrast with the oocytes from control rats (Fig. 4C and 4D), oocytes from T2 rats suffered atresia (Fig. 4H) that it was quantified in Fig. 5: k vs j P<0.0001, and displayed follicular cysts (three cysts/ovary) (Fig. 4I) with altered distribution of granulosa (GC) and theca (TC) cells (Fig. 4 J) as compared with control ovaries (Fig. 4B and 4D). Finally, ovaries from T2 rats displayed abnormal hyper-luteinization (Fig. 4K) characterized by enhanced pyknotic cells with compacted chromatin (Fig. 4K) as compared with controls (Fig. 4E and 4F).

Ovaries from T5 rats showed an increased number of atretic follicles (Fig. 4L: AtF) as compared with controls (Fig. 4A). Since rats from T5 group showed vaginal opening atresia, we used the ovarian morphology to compare ovarian structures with the other groups. Ovarian morphology of rats from T5 group (Fig. 4L) was similar to that of rats from the T2 group which did not ovulate and the two groups stayed in constant diestrus. These data were quantified in Fig. 5: l vs j P<0.0001, and follicular cysts (three cysts per ovary) (Fig. 4M: FC: follicular cysts). Oocytes from ovaries from T5 rats showed atresia (Fig. 4N: AO) and a disorganized distribution of the surrounding layers of granulosa (GC) and theca (TC) cells (Fig. 4N). The number of corpus luteum per ovary between groups did not differ (five corpus luteum/ovary) (data not shown).

Prenatal hyperandrogenization and ovarian inflammatory status

To assess whether prenatal hyperandrogenization was able to induce a pro-inflammatory status in the ovarian tissue, we determined the PGE content. In order to establish whether the estrous cycle is able to modify the PGE content, we quantified PGE from ovarian tissue in proestrus and diestrus (Fig. 6; P<0.0001). We found that prenatal T2 hyperandrogenization did not modify PGE content in either stage but that prenatal T5 hyperandrogenization significantly increased PGE content as compared with controls and T2 rats (Fig. 6; P<0.0001).

Prenatal hyperandrogenization and ovarian oxidant-antioxidant balance

To determine whether prenatal hyperandrogenization was able to induce ovarian oxidative stress, lipid peroxidation index, nitric oxide synthase activity and the antioxidant metabolite; GSH were evaluated in the ovarian tissue from control, T2 and T5 rats. Since the content of MDA (the main product which results of lipid peroxidation) was not modified either in ovarian tissue from T2 or T5 rats, we inferred no damage in the ovarian membrane (Fig. 7A). To evaluate the possible accumulation of ROS, we quantified NOS activity. We found that NOS activity was not modified in T2 or T5 (Fig. 7B). However, the ovarian content of the antioxidant metabolite GSH was increased in ovarian tissues from T2 and T5 rats (Fig. 7C; P<0.0001).

Mechanisms involved in the prenatal hyperandrogenization in ovarian tissue

To clarify some of the molecular mechanisms involved in the altered ovarian steroidogenesis induced by prenatal hyperandrogenization, the protein expression of the limiting protein that regulates cholesterol availability, the steroidogenic acute regulatory protein: StAR and the protein expression of the transcriptional factor PPAR gamma were evaluated by Western blotting (Fig. 8; P<0.0001).

We found that the protein expression of StAR was increased in ovarian tissue from T5 rats as compared to both controls and T2 rats (Fig. 8A–C; P<0.0001). We also found that the protein expression of PPAR gamma was increased in ovarian tissue from T5 rats as compared to both controls and T2 rats (Fig. 8D–F; P<0.0001).

In order to determine whether protein expression of COX2, the limiting enzyme of PG synthesis, was affected by prenatal hyperandrogenization, we evaluated the protein expression of COX2. We found that protein expression of COX2 was increased in ovarian tissue from T5 rats as compared with controls (Fig. 8G–I; P<0.0001). No significant differences were found between T2 rats and controls (Fig. 8G–I; P<0.0001).

A direct correlation was found between COX2 (T5 vs T2 and control, P = 0.0001, r = 0.999) and PPAR gamma (T5 vs T2 and control, P = 0.01; r = 0.950).