Maternal androgen excess reduces placental and fetal ...€¦ · Administration of testosterone during pregnancy in rats or sheep does not affect the food intake or weight gain of

Maternal androgen excess reduces placental and fetal weights, increasesplacental steroidogenesis, and leads to long-term health effects in their femaleoffspring

Miao Sun,1,2 Manuel Maliqueo,1 Anna Benrick,1 Julia Johansson,1 Ruijin Shao,1 Lihui Hou,2

Thomas Jansson,3 Xiaoke Wu,2 and Elisabet Stener-Victorin1,2

1Institute of Neuroscience and Physiology, Department of Physiology, Sahlgrenska Academy, University of Gothenburg,Gothenburg, Sweden; 2Department of Obstetrics and Gynecology, First Affiliated Hospital, Heilongjiang University ofChinese Medicine, Harbin, China; and 3Center for Pregnancy and Newborn Research, Department of Obstetrics andGynecology, University of Texas Health Science Center, San Antonio, Texas

Submitted 15 August 2012; accepted in final form 30 September 2012

Sun M, Maliqueo M, Benrick A, Johansson J, Shao R, Hou L,Jansson T, Wu X, Stener-Victorin E. Maternal androgen excess re-duces placental and fetal weights, increases placental steroidogenesis, andleads to long-term health effects in their female offspring. Am J PhysiolEndocrinol Metab 303: E1373–E1385, 2012. First published October 9,2012; doi:10.1152/ajpendo.00421.2012.—Here, we tested the hypothe-sis that excess maternal androgen in late pregnancy reduces placentaland fetal growth, increases placental steroidogenesis, and adverselyaffects glucose and lipid metabolism in adult female offspring. Preg-nant Wistar rats were randomly assigned to treatment with testoster-one (daily injections of 5 mg of free testosterone from gestational days16 to 19) or vehicle alone. In experiment 1, fetal and placentalweights, circulating maternal testosterone, estradiol, and corticoste-rone levels, and placental protein expression and distribution ofestrogen receptor-� and -�, androgen receptor, and 17�-hydroxys-teroid dehydrogenase 2 were determined. In experiment 2, birthweights, postnatal growth rates, circulating testosterone, estradiol, andcorticosterone levels, insulin sensitivity, adipocyte size, lipid profiles,and the presence of nonalcoholic fatty liver were assessed in femaleadult offspring. Treatment with testosterone reduced placental andfetal weights and increased placental expression of all four proteins.The offspring of testosterone-treated dams were born with intrauterinegrowth restriction; however, at 6 wk of age there was no difference inbody weight between the offspring of testosterone- and control-treatedrats. At 10–11 wk of age, the offspring of the testosterone-treateddams had less fat mass and smaller adipocyte size than those born tocontrol rats and had no difference in insulin sensitivity. Circulatingtriglyceride levels were higher in the offspring of testosterone-treateddams, and they developed nonalcoholic fatty liver as adults. Wedemonstrate for the first time that prenatal testosterone exposure altersplacental steroidogenesis and leads to dysregulation of lipid metabo-lism in their adult female offspring.

testosterone; prenatal; maternal; placenta; polycystic ovary syndrome;insulin sensitivity; steroidogenesis; estrogen receptor; androgen re-ceptor

THE MATERNAL ENVIRONMENT may influence epigenetic processesduring placental and fetal development that have long-lastingeffects and lead to diseases such as hypertension, obesity, type2 diabetes, and endocrine and reproductive dysfunction in adultoffspring (6, 24). Polycystic ovary syndrome (PCOS) is themost common endocrine disorder in women of reproductiveage and is associated with hyperandrogenism, oligo/anovula-

tion (infertility), and polycystic ovaries (5, 26). PCOS is alsoassociated with metabolic disturbances such as hyperinsulin-emia and type 2 diabetes and dysfunctional lipid profile,symptoms that are aggravated by obesity (5). Women withPCOS are at a higher risk of delivering prematurely, develop-ing gestational diabetes and preeclampsia (33), and havingboth small-for-gestational-age (40) and large-for-gestational-age infants (33). These findings suggest a complex relationshipbetween the altered maternal environment and fetal growth inPCOS. The mechanisms leading to the development of PCOSare largely unknown, but the pathogenesis of PCOS is likely toinclude a combination of genetic and epigenetic factors (51) aswell as environmental influences (2).

Prenatal androgenization is associated with development ofa PCOS-like phenotype in female mammals, including rhesusmonkeys, sheep, rats, and mice, suggesting that the intrauterineenvironment plays a role in the etiology of PCOS (1). It alsoleads to a dose-dependent reduction in birth weight in nonpri-mate species (25, 35, 36, 50). Despite the recognition thatexcess androgen during pregnancy affects fetal developmentand can program for future development of adult metabolicdiseases, there are few studies examining the underlying mech-anisms of fetal origins of PCOS.

Administration of testosterone during pregnancy in rats orsheep does not affect the food intake or weight gain of the damor the maternal levels of insulin, estrogen, progesterone, glu-cose, or lipids (36, 47). These findings indicate that the effectsof prenatal androgenization on the placenta, fetus, and femaleoffspring are not caused by changes in maternal metabolism.On the other hand, in monkeys, maternal androgen adminis-tration during pregnancy does increase weight gain and altersinsulin and glucose responses to glucose and increases conju-gated estrogen of the dam, indicating that effects in femaleoffspring may be caused by altered maternal metabolism (2).The placenta normally acts as a barrier that protects the fetusfrom excessive maternal androgens. However, because testos-terone is lipophilic, it may diffuse across the placenta and exertdirect effects on fetal growth and/or energy homeostasis (28).Alternatively, testosterone may affect placental developmentand function by modulating amino acid transporters (35, 42) orby regulating the expression of enzymes and/or androgen/estrogen receptors, as demonstrated in human placentas inpregnancies complicated by preeclampsia (21, 34). Whetherplacental steroidogenesis is dysregulated due to excess prenatal

Address for reprint requests and other correspondence: E. Stener-Victorin, Instituteof Neuroscience and Physiology, Dept. of Physiology, Univ. of Gothenburg, Box 434,SE-405 30, Gothenburg, Sweden (e-mail: [email protected]).

Am J Physiol Endocrinol Metab 303: E1373–E1385, 2012.First published October 9, 2012; doi:10.1152/ajpendo.00421.2012.

0193-1849/12 Copyright © 2012 the American Physiological Societyhttp://www.ajpendo.org E1373

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

androgen exposure has not, to our knowledge, been investi-gated.

Because PCOS is associated with excessive circulating an-drogen that further increases over the course of pregnancy (41),an increased risk of adverse pregnancy outcomes (33), andmetabolic disturbances during adulthood, we tested the hypoth-esis that excess maternal androgen late in pregnancy reducesplacental and fetal growth, increases placental steroidogenesis,and adversely affects glucose and lipid metabolism in adultfemale offspring.

MATERIALS AND METHODS

Ethics Statement

Animals were cared for in accordance with the principles of theGuide to the Care and Use of Experimental Animals (www.sjv.se).The study was approved by the Animal Ethics Committee of theUniversity of Gothenburg (Dnr: 70-2011).

Animals

Timed-pregnant Wistar rats purchased from Charles River Labo-ratories (Sulzfeld, Germany) arrived on day 7 of gestation and werehoused one per cage under controlled conditions (21–22°C, 55–65%humidity, and a 12:12-h light/dark cycle). All rats were fed HarlanTeklad Global Diet (16% protein rodent diet no. 2016; Harlan Win-kelmann, Harlan, Germany) and tap water ad libitum. Pregnant ratswere treated from embryonic days 16 to 19 with daily injections of 5mg of free testosterone (T-1500; Sigma) dissolved in 0.5 ml of a 1:1mixture of sesame oil (S3547; Sigma) and benzyl benzoate (B6630;Sigma) or with 0.5 ml of sesame oil-benzyl benzoate vehicle as acontrol. The dose was selected to mimic the fetal testosterone surgethat is observed in male rats (45, 48). Maternal weight gain and foodintake were recorded from gestational days 16 to 21.

Experimental Design and Methods

Experiment 1. PLACENTAL SAMPLES. Seven control and eighttestosterone-treated dams were anesthetized with thiobutabarbital so-dium (130 mg/kg ip, Inactin; Sigma) and euthanized on gestationalday 21. Maternal blood was collected by cardiac puncture and cen-trifuged, and the serum was stored at �80°C for further analysis.Laparotomy was performed, and the fetuses and placentas werecollected and quickly dried on blotting paper to remove any remainingfetal membranes and weighed. Two placentas from each litter werefixed in Histofix containing 6% formaldehyde (Histolab, Gothenburg,Sweden) before final storage in 70% ethanol. The remaining placentasin each litter were pooled and cut in smaller pieces. Some placentawere pieces immediately frozen and stored at �80°C until later geneexpression analysis, and the remaining tissue was quickly homoge-nized on ice in a Polytron in a buffer containing 10 mM Tris-HEPES,250 mM sucrose, 1 mM EDTA, 1.6 �M antipain, 0.7 �M pepstatin,and 0.5 �g/ml aproptinin. The homogenate was frozen in liquidnitrogen and stored at �80°C until subsequent protein expressionanalysis.

PLACENTA MORPHOLOGY. Paraffin-embedded histological placentasections (6 �m) were mounted on glass slides and stained withhematoxylin and eosin. Each section was converted to a virtual slidewith a Zeiss Mirax Desk scanning device (Zeiss, Oberkochen, Ger-many). In each section, the total placenta area, labyrinth zone area,and basal zone area were measured with MIRAX SCAN ControlSoftware, and the ratios of labyrinth zone to total area and basal zoneto total area were calculated.

IMMUNOHISTOCHEMISTRY. Immunohistochemical analysis wasperformed as described previously (17), with minor modifications.Paraffin sections were deparaffinized and rehydrated through a graded

alcohol series followed by antigen retrieval in 10 mM sodium citratebuffer (pH 6.0) for 18 min in a microwave oven at full power.Endogenous peroxidases and nonspecific binding were removed bypreincubation with 3% H2O2 for 10 min, 0.25% Triton X-100 for 30min at room temperature, and 2.5% normal horse serum for 1 h at37°C. Sections were then incubated with commercially availableprimary antibodies [estrogen receptor-� (ER�) 1:400, estrogen recep-tor-� (ER�) 1:200, androgen receptor (AR) 1:100, or 17�-hydroxys-teroid dehydrogenase 2 (17�-HSD2) 1:200] from Santa Cruz Biotech-nology (sc-542, sc-8974, sc-816, and sc-135042; Sigma) overnight at4°C. After washing, the sections were incubated with the appropriatebiotinylated secondary antibody for 1 h at 37°C. The Vectastain ABCkit (Vector Laboratories, Burlingame, CA) was used for the avidin-biotin peroxidase complex detection system according to the manu-facturer’s instructions. Antigens were visualized by reaction with thechromogen 3,3=-diaminobenzidine-tetrahydrochloride (Sigma) andhydrogen peroxidase for 1 min. Sections were viewed on an OlympusDP50 microscope using Image-Pro plus software or on an Axiovert200 confocal microscope (Carl Zeiss, Jena, Germany) equipped witha laser-scanning confocal imaging LSM 510 META system (CarlZeiss), and all sections were photomicrographed. Rat testis tissueserved as a positive control for AR.

WESTERN BLOT ANALYSIS. Protein expression levels of ER�, ER�,AR, and 17�-HSD2 were determined in placenta homogenates usinghorseradish peroxidase-conjugated anti-rabbit IgGs (A0545; Sigma).Western blot analysis was performed as described (22). In brief,20–50 �g of total proteins was separated by electrophoresis throughan SDS-PAGE gel and then transferred to a polyvinylidene difluoridemembrane. Membranes were rinsed in Tris-buffered saline with 1%Tween-20 (TBS-T), blocked in 3% BSA in TBS-T for 1 h at roomtemperature, and incubated with primary antibody overnight at 4°C.The blots were washed in TBS-T, incubated in secondary antibody for1 h at room temperature, and washed again in TBS-T. Protein bandswere developed with SuperSignal West Dura Extended DurationSubstrate (Pierce Biotechnology) and photographed with an LAS-1000 camera system (Fujifilm, Tokyo, Japan). The intensities of theprotein signals were quantified by densitometry with MultiGaugesoftware version 3.0. �-Actin (A1978; Sigma) was used as a loadingcontrol and for normalization. Values are expressed in arbitrarydensitometric units.

AR SEMIQUANTITATIVE PCR ANALYSIS. Total RNA from placentaltissues was extracted using the RNeasy Micro Kit (Qiagen) andRNase inhibitor (Applied Biosystems). Single-stranded cDNA wassynthesized from each sample (0.5 �g) with a High-Capacity cDNAReverse Transcription kit (Applied Biosystems). Oligo(dT)-primedcDNAs were synthesized from total RNA (2 �g) for 1 h at 55°C in areaction volume of 20 �l. The resulting cDNAs (2 �l) were amplifiedby PCR in a final volume of 25 �l containing 2.5 units of HotStar TaqDNA polymerase (Qiagen) and 0.2 �M each of the sense andantisense primers used to amplify specific nucleotide sequences pres-ent in AR and �-actin transcripts. The primer pair used for amplifi-cation of a 29-bp fragment of AR cDNA corresponding to theligand-binding domain of AR (NM_012502.1) was 5=-CCC ATCGAC TAT TAC TTC CCA CC-3= (sense) and 5=-TTC TCC TTC TTCCTG TAG TTT GA-3= (antisense). The primer pair for amplificationof a 454-bp fragment of �-actin (NM_031144.2) was 5=-CTG TGCCCA TCT ATG AGG GTT AC-3= (sense) and 5=-AAT CCA CACAGA GTA CTT GCG CT-3= (antisense). Results were calculated asthe ratio of the signal of the gene in each sample to its correspondinginternal control (�-actin). To ensure reliability, PCR analysis for eachgene was independently performed in duplicate for each tissuesample.

MATERNAL HORMONE LEVELS. Serum concentrations of cortico-sterone, 17�-estradiol, and testosterone were determined with double-antibody RIA kits (Siemens Healthcare Diagnostics, Los Angeles,CA). The intra- and interassay coefficients of variation were 6.4–10.6

E1374 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

and 5.9–11.9% for 17�-estradiol, 6.0–9.1 and 7.5–12.7% for testos-terone, and 4.0–12.2 and 4.8–14.9%, respectively, for corticosterone.The sensitivities of the assays were 7.2 pg/ml, 0.03 ng/ml, and 5.7ng/ml for 17�-estradiol, testosterone, and corticosterone, respectively.

Experiment 2. FEMALE OFFSPRING. Spontaneous delivery wasallowed for five control and five testosterone-treated dams. Four daysafter birth pups were counted and weighed, and the offspring weresubsequently weighed twice/wk until 70 days of age. After 21 days,the female offspring were separated from the dam and their malelittermates, and their weights, lengths, and anogenital distances weremeasured. Microchips (AVID, Norco, CA) with an identificationnumber were inserted subcutaneously in their necks under lightanesthesia (2% isoflurane in a 1:1 mixture of oxygen and air, IsobaVet; Schering-Plough). Food intake was recorded from day 21 until 70days of age.

VAGINAL SMEARS AND BLOOD SAMPLING. The estrous cycle stagesof the female offspring were determined by microscopic analysis ofthe predominant cell type in vaginal smears. The smears were takenstarting 1 wk prior to the collection of fasting tail blood samples at 10wk of age and were continued until the euglycemic hyperinsulinemicclamp at weeks 11 and 12 (27). None of the testosterone-treatedoffspring had vaginal openings, so blood was collected, and the clampwas performed independent of cycle day. In control offspring, bloodwas collected and the clamp performed during estrous phase.

BODY COMPOSITION BY DUAL-ENERGY X-RAY ABSORPTIOMETRY.A whole body dual-energy X-ray absorptiometry (DEXA) instrument(QDR-1000/W; Hologic, Waltham, MA) was used to assess the totalbody fat, lean body mass, and bone mineral content at 10 wk of age.Rats were lightly anesthetized with isoflurane during the procedure.

EUGLYCEMIC HYPERINSULINEMIC CLAMP AND SAMPLE COLLECTION.At 11–12 wk of age, rats were subjected to a euglycemic hyperinsu-linemic clamp as described (27). In brief, rats were anesthetized withthiobutabarbital sodium (130 mg/kg ip, Inactin; Sigma). Body tem-perature was maintained at 37°C with a heating pad throughout theclamp. Insulin (100 U/ml Actrapid; Novo Nordisk) together with 0.2ml of albumin and 10 ml of physiological saline was infused at 8mU·min�1·kg�1 during the clamp after an initial bolus dose. Theglucose infusion rate (GIR) was guided by blood glucose concentra-tion measurements taken every 5 min with an Accu-Chek CompactPlus glucometer (Roche Diagnostics, Indianapolis, IN), and 20%glucose in saline solution was continuously administered to maintainplasma glucose at a constant euglycemic level (6.0 mM). At steadystate (after 50–70 min), the mean GIR was normalized to bodyweight, and blood samples were taken to determine plasma insulinconcentrations.

After the clamp, the rats were decapitated, and their ovaries, uteri,hindlimb muscles (tibialis anterior, extensor digitorum longus, andsoleus), fat depots (inguinal, parametrial, retroperitoneal, and mesen-teric), and livers were dissected and weighed. Collected tissues weredivided into two halves, with one half being snap-frozen in liquidnitrogen and stored at �80°C until protein analysis and the other halfbeing fixed in Histofix containing 6% formaldehyde (Histolab) beforefinal storage in 70% ethanol.

OVARIAN MORPHOLOGY. Paraffin-embedded histological sections(6 �m) of the ovaries were mounted on glass slides and stained withhematoxylin and eosin. Each section was converted to a virtual slideusing a Zeiss Mirax Desk scanning device.

OIL RED O ANALYSIS OF THE LIVER. Frozen 10-�m sections fromliver samples in cryostat-embedding medium were cut with a LeicaCryostat microtome 3050S (Leica Microsystems Nussloch, Heidel-berger, Germany) at �20°C, mounted onto glass slides, and air-dried.The tissue samples were rinsed with 60% isopropanol for 2 min andoil red O dissolved in 98% isopropanol for 15 min, 60% isopropanolfor 30 s, and distilled water for 5 min. The tissue samples werecounterstained with hematoxylin for 30 s and washed thoroughly withdistilled water prior to mounting in glycerol gelatin. Unsaturated

hydrophobic lipids were identified microscopically by the presence ofa red stain and were subsequently classified into negative (�),possible early (�/�), and positive (�) fatty liver categories comparedwith predetermined reference sections (20).

COMPUTERIZED DETERMINATION OF ADIPOCYTE SIZE. The meanadipocyte size was determined by computerized image analysis ac-cording to Björnheden et al. (10). Adipocytes were obtained from�500 mg of inguinal or mesenteric adipose tissues. Small pieces oftissue were incubated with 0.2 U/ml type A collagenase (Roche) in 10ml of minimum essential medium (Invitrogen) for 50 min at 37°C ina shaking water bath. Adipocytes were washed three times andsuspended in fresh medium. The cell suspension was placed betweena siliconized glass slide and a coverslip and transferred to the micro-scope (DM6000B, �5 objective; Leica Microsystems). Twelve ran-dom visual fields were photographed with a charge-coupled devicecamera (DFC320; Leica Microsystems), and the adipocyte size wasmeasured using the Leica QWin software package version 3. Uniformmicrospheres 98 �m in diameter (Dynal; Invitrogen) served as areference.

HORMONE LEVELS AND BIOCHEMICAL ANALYSIS. Serum concen-trations of 17�-estradiol, testosterone, and corticosterone in the off-spring were determined as described above for the dams. All lipidanalyses were carried out at an accredited laboratory at the Wallen-berg Laboratory, Sahlgrenska University Hospital, Gothenburg, Swe-den, on a Konelab 20 autoanalyzer (Thermo Fisher Scientific), withinterassay coefficients of variation �3%. Plasma concentrations oftotal cholesterol, HDL cholesterol, LDL cholesterol, and triglycerides(TG) were determined enzymatically. HDL concentrations were de-termined after precipitation of apolipoprotein B-containing lipopro-teins with magnesium sulfate and dextran sulfate (Thermo FisherScientific). Human insulin levels given during the clamp were mea-sured in duplicate with an ELISA kit (Mercodia, Uppsala, Sweden),with intra- and interassay coefficients of variation of 3.4 and 3.0%,respectively, and a sensitivity of 1 mU/l.

Statistical Analysis

Data are reported as means SE. Body weight gains at each timepoint were analyzed by mixed between/within-subjects ANOVA fol-lowed by Student’s t-test. Remaining analyses were performed onlywith Student’s t-test. All statistical evaluations were performed withthe SPSS software package (version 19.0; SPSS, Chicago, IL). P �0.05 was considered significant.

RESULTS

Experiment 1

Maternal body weights were measured from gestation day(GD) 16 to GD 21. No significant differences were observedbetween testosterone-treated and control dams at any timepoint (Table 1).

Maternal hormone levels. Administration of testosterone topregnant rats from GD 16 to GD 19 resulted in a 2.5-fold

Table 1. Body weights (g) in control and T-treated dams

Gestational Day Control (n 12) T-treated (n 13) P Value

16 294.9 5.6 287.7 5.5 0.37217 311.4 5.6 300.4 5.8 0.18418 325.1 6.2 316.0 6.2 0.30919 339.6 6.7 329.6 6.3 0.29220 356.1 7.0 337.0 9.2 0.11721 373.0 7.8 357.1 6.9 0.138

Values are means SE. T, testosterone. P values were determined withStudent’s t-test.

E1375MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

increase (P 0.032) in circulating testosterone levels on GD21 (Table 2). There were no significant differences in circulat-ing corticosterone or 17�-estradiol levels in testosterone-treated dams compared with controls (Table 2).

Placental and fetal data. Maternal administration of testos-terone decreased fetal weight, fetal length, and placentalweight compared with controls (Table 3). No significant dif-ferences in litter size or in the fetal-to-placental weight ratiowere observed between testosterone-treated and control dams(Table 3).

Placenta morphology. The total placental area and the basalzone area were smaller in testosterone-treated dams comparedwith control dams at GD 21 (Table 3 and Fig. 1, A and D), butno significant differences were seen in the area of the labyrinthzone, the ratio of the labyrinth zone to total area, or the ratio ofthe basal zone to total area (Table 3). The labyrinth zones ofthe placentas from testosterone-treated dams had fewer redblood cells compared with controls (Fig. 1, B and E), and thebasal zones of the placentas from testosterone-treated dams hadlarger and more dispersed trophoblast giant cells than controls(Fig. 1, C and F). No significant differences in the appearanceof glycogen cells or spongiotrophoblasts in the basal zone wereobserved (Fig. 1, C and F).

Changes in placental ER, AR, and 17�-HSD2 protein ex-pression levels. The placental distribution patterns of ER�,ER�, AR, and 17�-HSD2 on GD 21 are shown in Fig. 2. Theexpression of all four proteins was higher in placentas fromtestosterone-treated dams compared with controls. ER� wasexpressed mainly in the cytoplasm of cells in the basal zone(Fig. 2, A and B), ER� was expressed mainly in the nucleiof cells in the placental labyrinth, basal, and decidual zones(Fig. 2, C and D), AR was expressed in the basal zone (Fig. 2,E and F), and 17�-HSD2 was expressed within the basal anddecidual zones (Fig. 2, G and H).

Western blot experiments were performed to confirm theimmunohistochemical results (Fig. 3, A–D). The placentas oftestosterone-treated dams showed 3.2-fold higher levels ofexpression of ER�, 2.5-fold higher ER� levels, and 2.3-fold higher 17�-HSD2 levels compared with control placentas(Fig. 3, A–D). Expression of the AR protein was very low andcould not be detected by Western blot. Semiquantitative PCRof AR confirmed its expression in the placenta (Fig. 3, E andF). For ER�, measurements were based on the 52-kDa bandsbecause the 66-kDa bands were too weak to be measuredaccurately (14).

Experiment 2

Postnatal growth. The offspring of dams treated with tes-tosterone late in pregnancy weighed less than controls frompostnatal days 4 to 35 (Fig. 4A). However, from postnatal day42, female offspring from testosterone-treated and control rats

did not differ in body weights (Fig. 4B), nor was there anydifference in food intake during the study period (Fig. 4C).

On postnatal day 21, sex was determined, and females wereseparated from males. The mean anogential distance was 26%longer in the female offspring of testosterone-treated damscompared with controls (11.5 0.5 vs. 9.0 1.5 mm, P �0.001). There was no significant difference in body lengthbetween the groups (data not shown). Inspection of femalesrevealed an absence of the vaginal opening in 100% of theoffspring of the testosterone-treated dams, and some diedduring the study, most likely due to a large, distended, fluid-filled uterus and upper vagina (hydrometrocolpos).

Female offspring hormone levels and biochemical analysis.At 9 wk of age, circulating testosterone levels were higher infemale offspring of the testosterone-treated dams comparedwith controls, whereas 17�-estradiol and corticosterone levelsdid not differ between the groups (Table 4). Fasting TG levelswere higher in female offspring of the testosterone-treateddams compared with controls (P � 0.01), whereas fastinglevels of cholesterol, HDL, and LDL did not differ between thegroups (Table 4).

Body composition measured by DEXA. At 10 wk of agethere was no difference in body weight between the two groups(Table 5). Despite this, the female offspring of the testosterone-treated dams had less body fat, both in terms of total weightand as a percentage of body weight, and a tendency for reducedbone mineral density compared with controls as measured byDEXA.

Insulin sensitivity and tissue weights. There was no differ-ence in mean GIR (Fig. 5A) or insulin sensitivity index (ratioof the mean GIR to the steady-state plasma insulin level; Fig.5B) between the two groups. When steady state was reached,the glucose level was �6 mmol/l in both groups, and theplasma insulin levels were 152.8 12.3 mU/l in the controloffspring and 151.6 13.6 mU/l in the testosterone-treatedoffspring.

After application of the clamp, the animals were euthanized,and various tissues were dissected and weighed (Table 6). Theweights of the tibial muscle and of the inguinal and retroper-itoneal fat depots were reduced in the offspring of the testos-terone-treated dams, whereas the ovarian, uterine, and livertissue weights did not differ between the two groups (Table 6).

Morphology of ovary and liver. The offspring of the testos-terone-treated dams displayed normal ovarian morphologiesand had antral follicles at different stages and fresh corpus

Table 3. Placental and fetal data from control T-treateddams

Control (n 12) T-treated (n 13) P Value

Litter size (n) 12.25 0.80 12.00 1.72 0.892Fetal weight, g 4.46 0.10 3.73 0.09 �0.001Fetal length, mm 37.86 0.36 36.35 0.46 0.013Placenta weight, g 0.560 0.008 0.467 0.011 �0.001Fetal/placental weight ratio, g/g 8.08 0.17 8.19 0.20 0.669Placental morphology (n 6)

Total area, mm2 34.9 1.2 30.9 1.1 0.035Basal zone area, mm2 7.3 0.5 5.4 0.5 0.025Labyrinth zone area, mm2 24.2 1.0 21.4 0.9 0.056Labyrinth zone/total area 0.69 0.02 0.69 0.01 0.889Basal zone/total area 0.21 0.02 0.18 0.02 0.174

Values are means SE. P values were determined with Student’s t-test.

Table 2. Maternal serum testosterone, 17�-estradiol, andcorticosterone concentrations in control and T-treated dams

Control (n 7) T-treated (n 8) P Value

Testosterone, ng/ml 0.04 0.01 0.10 0.02 0.03217�-Estradiol, pg/ml 55.0 16.9 73.7 12.3 0.380Corticosterone, ng/ml 229.4 26.7 221.0 33.3 0.855

Values are means SE. P values were determined with Student’s t-test.

E1376 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

luteum, just as in the control ovaries (Fig. 6A). However, therewere signs of early nonalcoholic fatty liver in 37.5% of the theoffspring of the testosterone-treated dams compared with10.0% in controls (P � 0.05), and there were signs of fattyliver in 50% of the offspring compared with 30% in controls(P � 0.01) (Fig. 6B).

Adipocyte size. Mean adipocyte size and size distributionwere determined in inguinal and mesenteric adipose tissues. Infemale offspring of the testosterone-treated dams, the sizedistribution curve of inguinal adipocytes was shifted to the left(Fig. 7, A–C), with smaller mean inguinal adipocytes than incontrols (61.9 2.4 vs. 70.9 7.6 �m, P � 0.05). In themesenteric fat depot, the mean adipocyte size and size distri-bution did not differ compared with controls (data not shown).

DISCUSSION

This study is, to our knowledge, the first to demonstrateenhanced placental ER, AR, and 17�-HSD2 protein expressionlevels in rats after maternal exposure to high levels of testos-terone during late pregnancy. The increase in placental 17�-

HSD2 protein expression suggests altered placental steroido-genesis, and the increased expression of ER and AR indicatesaltered estrogen and androgen activity that occurred togetherwith decreased placental weight, altered placental morphology,and decreased fetal weight. The female offspring of the testos-terone-treated dams were born with intrauterine growth restric-tion (IUGR), but by puberty no difference in body weight wasobserved between the two groups. Furthermore, female off-spring of testosterone-treated dams displayed normal insulinsensitivity but had less adipose tissue, smaller adipocytes, andhigh levels of circulating TG with clear signs of nonalcoholicfatty liver. These results indicate that prenatal androgen expo-sure in rats induces early signs of metabolic dysfunction infemale offspring as in human PCOS.

Effect of Prenatal Testosterone Exposure on the Placentaand Fetal Growth

Administration of testosterone to rats in late pregnancy didnot affect maternal weight gain, food intake, metabolic status,or circulating estradiol or corticosterone concentrations, and

Fig. 1. Placental morphology as indicated by hematox-ylin and eosin staining of placentas from control (A–C)and testosterone-treated (D–F) dams at gestation day 21.A and D: the total placental area was smaller in thetestosterone-treated dams compared with controls. Band E: the labyrinth zone of the placentas from thetestosterone-treated dams had fewer red blood cells(RB) compared with controls. L, labyrinth zone; Ba,basal zone; De, decidual zone; MS, maternal sinusoid;Gly C, glycogen cell; FC, fetal capillary; G, trophoblas-tic giant cell; ST, spongiotrophoblast. Magnification:100 �m.

E1377MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

these results are consistent with previous reports in rodents(36). This suggests that the effect of testosterone on placentaland fetal growth in rats is a direct effect of the testosterone andnot a secondary consequence of maternal malnutrition or al-terations in metabolic and steroid hormones. However, theseresults may not be extrapolated directly to humans sincesimilarly treated monkeys show altered maternal weight gain,metabolic status, and circulating conjugated estrogens (2).Testosterone is a lipophilic hormone and can diffuse from theamniotic fluid into the fetal circulation and across the placenta(16, 28, 44). However, previous studies have reported that asignificant increase in maternal testosterone level in rats is notassociated with a significant increase in testosterone in thefetus (35, 50), whereas in both sheep and monkeys it has beenshown that elevated maternal testosterone levels also elevatefetal testosterone levels in female offspring (2, 47). Also, arecent preliminary report indicates altered enzymatic regu-lation in PCOS placentas, as demonstrated by higher 3�-HSD1 placental expression and lower P450 aromatase ac-tivities compared with controls (11). Thus, maternal testos-terone may exert an indirect or direct effect on the fetus andalter critical placental functions that support fetal growth.One hypothesis is that a high maternal testosterone levelalters placental metabolism of testosterone, although thepathway by which this may occur is still unknown. It hasbeen reported recently that high maternal testosterone con-centrations do not directly cross the placenta to suppress

fetal growth but instead decrease amino acid nutrient deliv-ery to the fetus by downregulation of the activity andexpression of specific placental amino acid transporters(35). Although it is unclear whether the maternal environ-ment in women with PCOS directly influences the develop-ment of PCOS in their offspring, it has been demonstratedthat female offspring of mothers with PCOS are exposed totestosterone levels comparable with male levels in utero (7).Altered birth weight with both small-for-gestational-age(40) and large-for-gestational-age infants (33) has beenreported in women with PCOS. Whether the placentalweight or the placental metabolism of testosterone is alteredin women with PCOS has not been investigated to ourknowledge, but our finding that high maternal testosteroneduring late pregnancy in rats affects placenta weight, size,morphology, and steroidogenesis as well as androgen andestrogen activity may be relevant to human PCOS.

Excess maternal androgen levels in rats decreased placentalsize, increased circulating androgens, and tended to increasecirculating 17�-estradiol and the expression of AR and the ERsand altered placental morphology, and this may adverselyaffect the ability of the placenta to deliver nutrients to the fetus.Importantly, blockade of AR action has not been performed,and thus an alternative explanation may exist. Maternal stress(i.e., food restriction) and maternal glucocorticoid excess alsoinhibit placental and fetal development in rats (9). In thepresent study, it is clear that prenatal exposure to high levels of

Fig. 2. Immunohistochemistry of estrogen receptor(ER), androgen receptor (AR), and 17�-hydroxys-teroid dehydrogenase 2 (17�-HSD2) in the placentasof control (A, C, E, and G) and testosterone-treateddams (B, D, F, and H) at gestation day 21. Theplacentas of testosterone-treated dams displayed moreintense immunoreactivity for ER and 17�-HSD2 com-pared with control dams, but AR expression was weakin both groups. Magnification: 100 �m.

E1378 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

testosterone inhibits placenta growth, especially in the basalzone. Maternal food restriction is also known to cause areduction in placental weight, with associated hypotrophy inthe basal and labyrinth zones of the placenta (9). In thematernal stress model, maternal circulating corticosterone con-centrations are increased (9), but no such increase after mater-nal testosterone exposure was seen in our experiments. Thisindicates that the effects on placental and fetal developmentthat we have observed are not due simply to a stress response.

In contrast to the human placenta, the rat placenta normallysecretes only small amounts of progesterone and testosterone(46) and cannot synthesize estrogens since they do not expressaromatase, and therefore, it is dependent on ovarian estrogensince it cannot synthesize any estrogen itself (3, 43). High

estrogen concentrations are known to inhibit placental growth,whereas estrogen deficiency induces placental hypertrophy (8,12). Estrogen exerts its biological effects by binding to the ER,and the ER exists as two different isoforms, ER� and ER�. Inour present study, high maternal testosterone levels increasedboth ER� and ER� expression in the placental basal zoneduring late gestation. Furthermore, placental expression of17�-HSD2 was increased in response to maternal administra-tion of testosterone. In the rat placenta, 17�-HSD2 is expressedmainly in the decidual and basal zones (the functional zone),and this enzyme converts the active 17�-hydroxy forms ofestradiol and testosterone to their less active 17-keto forms (29,49). Mustonen et al. (29) found that the distribution of 17�-HSD2 expression in the placenta changes as gestation ad-

Fig. 3. ER, AR, and 17�-HSD2 protein expression in placentas of control (n 7) and testosterone-treated (n 8) dams. A: representative blots of ER�, ER�,AR, and 17�-HSD2 protein expression. B–D: densitometric analysis of ER� (B), ER� (C), and 17�-HSD2 protein expression (D). E: semiquantitative PCRanalysis. Values are means SE. *P � 0.05 vs. control; **P � 0.01 vs. control.

E1379MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

vances, and this enzyme is expressed mainly in both the giantcells at the interface of the chorioallantoic placenta and decid-ual zone as well as in the functional zone during gestation, andthis is consistent with our observations. It is well known thatthe amniotic fluid surrounding the fetus contains large amounts

of various steroids among other compounds (38). Hence, 17�-HSD2 in the fetus, together with its oxidative activity, suggeststhat this enzyme protects the embryo from excessive action ofthe 17-hydroxysteroids present in the amniotic fluid (29). Thismay explain why previous studies found less testosterone infetal serum than in testosterone-treated maternal serum (35,50). Based on these collective results, we speculate that theincreased expression of placental ER may contribute to theinhibition of placental growth in the current study. Further-more, we propose that the increased expression of 17�-HSD2in the placenta in response to maternal administration of

Table 4. Serum T, 17�-estradiol, and lipid profiles inT-exposed and C offspring

C Offspring (n 11) T Offspring (n 9) P Value

T, ng/ml 0.01 0.001 0.03 0.015 0.01217�-Estradiol, pg/ml 103.6 11.3 146.5 29.9 0.209Corticosterone, ng/ml 846.4 68.1 663.2 86.5 0.109Cholesterol, mmol/l 2.30 0.08 2.28 0.09 0.859Triglycerides, mmol/l 0.64 0.04 0.89 0.08 0.008HDL, mmol/l 1.68 0.06 1.70 0.05 0.815LDL, mmol/l 0.22 0.02 0.26 0.03 0.258

Values are means SE. C, control. P values were determined with Student’st-test.

Fig. 4. Body weight development in the offspring of control and testosterone-treated dams. A: body weight gain from days 4 to 35 after birth. B: body weightgain from days 4 to 70 after birth. C: food intake corrected by weight gain fromdays 21 to 70 after birth. Values are means SE. *P � 0.05 vs. control; **P �0.01 vs. control; ***P � 0.001 vs. control.

Table 5. Body composition as determined by DEXAin T-exposed and C offspring

C Offspring (n 11) T Offspring (n 9) P Value

Body weight, g 264.9 8.9 260.1 10.3 0.733Bone mineral content, g 6.8 0.1 6.4 0.2 0.120Bone mineral content,

%BW 2.59 0.03 2.49 0.05 0.130Bone mineral density,

g/cm2 0.136 0.002 0.132 0.001 0.052Fat mass, g 40.1 3.2 27.4 2.4 0.007Fat mass, %BW 15.3 1.1 10.9 1.0 0.008Lean body mass, g 215.8 8.1 218.4 9.0 0.833Lean body mass, %BW 81.5 1.4 84.0 1.0 0.180

Values are means SE. DEXA, dual-energy X-ray absorptiometry; %BW,%body weight. P values were determined with Student’s t-test.

Fig. 5. Glucose infusion rate (GIR) and insulin sensitivity index (ISI) in femaleoffspring from control (n 9) and testosterone (T)-treated (n 8) dams at11–12 wk of age. There was no significant difference in GIR or ISI betweenthe 2 groups. Values are means SE.

E1380 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

testosterone is consistent with increased 17�HSD2 activity thatinactivates testosterone and estrogen and thereby preventsexposure of the fetus to high testosterone levels and counter-acting effects of increased ER expression. However, it does notprevent decreased fetal growth in the testosterone-treatedgroup.

Although the AR protein was expressed only at very lowlevels, we observed increased immunoreactivity for AR proteinexpression in the basal zone of the placentas in the testoster-one-treated rats. Studies of AR gene expression confirmedplacental expression in controls and increased expression inresponse to maternal testosterone treatment, and AR levels areincreased in the preeclampsia placenta independent of fetal sex(34). The role of androgen signaling in the placenta is largelyunknown, but dysregulation may affect placental developmentand function. Furthermore, testosterone exposure in pregnantmonkeys, sheep, and rats has been shown to program forcardiovascular, metabolic, and endocrine dysfunction in theoffspring during adulthood (1, 4, 13, 15, 20, 23, 30, 37). Thuswe hypothesize that the overexpression of placental AR, to-gether with increased expression of ER and 17�-HSD2 protein,that is observed in response to high maternal testosterone levelsis involved in the development of adult diseases in the off-spring.

In a normal pregnancy, the red blood cell mass increasesduring the latter half of gestation and is associated withincreased erythropoietin production. Adequate blood flow tothe placenta is essential for the successful outcome of thepregnancy (31). Here, we observed fewer red blood cells in thelabyrinth zone in the testosterone-treated dams compared with

controls, which may indicate decrease of placental blood flowbut also a diminution of red blood cells. Thus, high maternaltestosterone may play an adverse role in the control of placen-tal blood flow.

Effect of Maternal Testosterone on Birth Weight, PostnatalGrowth, and Offspring Metabolism

It has been proposed that PCOS has its origin in fetal lifebased on reports that prenatal exposure to testosterone insheep, monkeys, and rodents causes endocrine and metabolicdysfunction in the offspring when they reach adulthood (1, 4,13, 15, 20, 23, 30, 37). The etiology of PCOS remains uncer-tain, but there is increasing evidence for a genetic basis. PCOSmanifests during adolescence along with maturation of thehypothalamic-pituitary-ovarian axis, but the genesis of thesyndrome may occure anywhere from very early fetal devel-opment of the ovary to the onset of puberty (19). However, atpresent, it is unclear whether the maternal environment directlyinfluences the development of PCOS in the offspring.

In this study, we found tissue-specific abnormalities in thefemale offspring of testosterone-treated dams. A novel findingis that these offspring had no vaginal openings, and it was notpossible to observe estrous cycle changes in these rats. Thisbirth defect, together with an increased anogenital distance,appears to be a direct effect of prenatal androgen exposure. Onthe other hand, inspection of ovarian morphology demon-strated clearly that the female offspring of the testosterone-treated dams have a normal estrous cycle independent of highercirculating testosterone levels. The uterus was large and fluidfilled, which also indicates that normal estrus cycles wereoccurring. Furthermore, the female offspring of testosterone-treated dams had less fat mass and smaller adipocyte sizetogether with normal insulin sensitivity. The decrease in fatmay reflect disturbances in water balance regulation as indi-cated by the DEXA results, since a 5% variation in fat-freemass hydration can change the DEXA-determined body fatpercentage by nearly 3% (32). Other possible factors are thelarge fluid-filled uteri, which also may reflect this differ-ence. Regarding smaller adipocytes, this may not always beconsidered to be healthy. It has been demonstrated recentlythat adipocytes taken on days 1 and 21 after birth from ratsthat were born small for their gestational age exhibitedenhanced adipogenesis and lipogenesis (52), and this mayincrease the risk of development of obesity and insulinresistance in adulthood.

The offspring of the testosterone-treated dams in the presentstudy were born intrauterine growth restricted, but their bodyweights increased until they were similar to the offspring of thecontrol group from 6 wk of age onward. However, the mech-anisms underlying fetal growth restriction in response to ma-ternal testosterone treatment remain to be established. Antena-tal exposure to glucocorticoids has been found to reduceoffspring birth weight, and this is followed by catchup growth(39). In our study, although circulating corticosterone levelsdid not differ in maternal serum, we cannot completely excludethat maternal exposure to testosterone activates the fetal hypo-thalamus-pituitary-adrenal axis, resulting in increased fetalcorticosterone levels. Testosterone may stimulate the placentaltransfer of corticosterone by modulating the expression of both11�-hydroxysteroid dehydrogenases, as has been shown in

Table 6. Dissected tissue weights in T-exposed and Coffspring

C Offspring (n 10) T Offspring (n 8) P Value

Body weight, g 285.7 10.1 287.9 10.0 0.877Ovary

g 0.135 0.008 0.155 0.013 0.192g/kg BW 0.047 0.002 0.055 0.006 0.172

Uterusg 0.533 0.019 0.908 0,399 0.274g/kg BW 0.188 0.007 0.311 0.128 0.264

Liverg 11.33 0.49 11.69 0.70 0.667g/kg BW 3.96 0.09 4.05 0.16 0.621

Muscles, gTibialis 1.095 0.052 1.022 0.032 0.280EDL 0.255 0.012 0.246 0.009 0.558Soleus 0.266 0.010 0.260 0.010 0.719

Muscles, g/kg BWTibialis 0.383 0.010 0.356 0.007 0.046EDL 0.089 0.003 0.085 0.002 0.325Soleus 0.093 0.002 0.091 0.002 0.453

Fat depots, gInguinal 1.34 0.08 1.03 0.10 0.023Parametrial 3.79 0.39 3.00 0.36 0.174Retroperitoneal 2.93 0.34 1.85 0.20 0.022Mesenteric 2.42 0.13 2.18 0.15 0.246

Fat depots, g/kg BWInguinal 0.48 0.03 0.36 0.04 0.047Parametrial 1.33 0.13 1.08 0.14 0.210Retroperitoneal 1.02 0.11 0.66 0.08 0.019Mesenteric 0.85 0.04 0.76 0.05 0.189

Values are means SE. EDL, extensor digitorum longus. P values weredetermined with Student’s t-test.

E1381MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

adipose tissue in children (53). However, the circulating cor-ticosterone level had a tendency to be lower in the femaleoffspring of the testosterone-treated offspring. In addition tohigher circulating testosterone, an indication of enhanced pla-cental steroidogenesis and reduced placental and fetal weightin testosterone-treated dams indicate that it is the high maternaltestosterone exposure rather than activation of the fetal stressaxis that causes the low birth weight and the catchup growth.

Interestingly, the offspring of the testosterone-treateddams had lower body fat, tended to have lower bone mass,had smaller adipocytes, and had higher circulating TG levelsthan the offspring of the control rats. The lower fat mass andadipocyte size may or may not indicate a more favorablemetabolic status, as discussed above. To elucidate the rea-son for increased circulating TG but no other lipids, weexamined the liver and found that the female offspring ofthe testosterone-treated dams showed signs of early nonal-coholic liver. A recent report showed an increased incidence

of nonalcoholic fatty liver independent of obesity in theoffspring of sheep that had been treated with testosterone(20). This indicates that adult liver metabolism and signal-ing could be altered by early exposure to testosterone andimplicates an epigenetic regulation of metabolic distur-bances.

The results in the present study are somewhat contradic-tory to what has been demonstrated previously in the femaleoffspring of testosterone-treated rats (13, 18, 37). In theseprevious studies, the offspring of pregnant rats treated withthe same dose of testosterone as used in our current studyexhibited metabolic disturbances at 9 wk of age, includingincreased body weight and fat mass, increased serum insu-lin, cholesterol, and TG levels, increased hepatic TG con-tent, and hypertension. They also showed signs of endo-crine/reproductive disturbances, including irregular estruscycles, polycystic ovaries, and neuroendocrine changes. Incontrast, the offspring of the testosterone-treated dams in

Fig. 6. A: morphology of ovarian changes in female offspring of control (n 10) and T-treated (n 8) dams. The morphologies did not differ significantlybetween the 2 groups, and both groups had follicles (F) at different stages and fresh corpus luteum (CL). Magnification: 500 �m. B: the morphology of the livers[oil red O and hematoxylin and eosin (H & E) staining] in the offspring of control (n 10) and T-treated dams (n 8). C: the offspring of the T-treated damsdisplayed signs of early nonalcoholic fatty liver compared with controls. *P � 0.05, and signs of fatty liver; **P � 0.01. Magnification: 50 �m. Representativestaining for lipids in the liver sections illustrating negative (�), early nonalcoholic fatty liver (�/�), or fatty liver (�) and associated analysis show an increasedfatty liver in the offspring of T-treated dams compared with controls.

E1382 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

the present study had normal insulin sensitivity, decreasedfat mass, smaller adipocytes, and normal estrus cycles. Theonly differences between the previous and current studiesare the rat strains used in the current study. Also, allmeasurements, including clamp studies in the testosterone-treated offspring, were performed independent of estruscycle day since we assumed that they were acyclic due to thelack of vaginal opening. These differences may lead to adifferent phenotype. The dose of prenatal testosterone ex-posure may also determine the extent of metabolic andendocrine alterations since it has been demonstrated re-cently that different doses of free testosterone cause differ-ent endocrine and metabolic phenotypes, with the dose offree testosterone (5 mg) used in the present study causing amore severe phenotype (4).

In conclusion, we demonstrate for the first time thatprenatal testosterone exposure in rats increases the placentalexpression of ER, 17�-HSD2, and AR proteins, which isindicative of enhanced placental steroidogenesis. The off-spring of the testosterone-treated dams were born withIUGR but showed catchup growth at puberty. These ratsalso developed early signs of metabolic dysfunction withincreased TG levels and nonalcoholic liver at adult agedespite less adiposity, smaller adipocytes, and normal insu-lin sensitivity. In addition to these metabolic effects, thefemale offspring of the testosterone-treated dams suffered

from severe birth defects and were born with a lack ofvaginal openings and an increased anogenital distance.

ACKNOWLEDGMENTS

We thank the Center for Mouse Physiology and Bio-Imaging of theUniversity of Gothenburg for technical assistance.

GRANTS

This study was financed by grants from the Swedish Research Council(Project No. K2012-55X-15276-08-3), the Jane and Dan Olsson Foundations,the Novo Nordisk Foundation, the Hjalmar Svensson Foundation, the Adler-bert Research Foundation, Wilhelm and Martina Lundgrens’s Science Fund,the Swedish federal government under the LUA/ALF agreement (ALFGBG-136481), and the Program for Longjiang Scholars and Innovative ResearchTeam in Heilongjiang Provincial Universities (2011TD006).

DISCLOSURES

The authors confirm that there are no conflicts of interest, financial orotherwise.

AUTHOR CONTRIBUTIONS

M.S., M.M., J.J., R.S., and E.S.-V. performed the experiments; M.S. andE.S.-V. analyzed the data; M.S., M.M., A.B., J.J., R.S., T.J., and E.S.-V.interpreted the results of experiments; M.S. and E.S.-V. prepared the figures;M.S. and E.S.-V. drafted the manuscript; M.S., M.M., A.B., J.J., R.S., L.H.,T.J., X.W., and E.S.-V. edited and revised the manuscript; M.S., M.M., A.B.,J.J., R.S., L.H., T.J., X.W., and E.S.-V. approved the final version of themanuscript; A.B., J.J., L.H., T.J., X.W., and E.S.-V. did the conception anddesign of the research.

0.00

5.00

10.00

15.00

20.00

25.00

%

Size Range (µm)

Adipocyte size Inguinal

A B

C

500µm

T-offspring

Control offspring

Fig. 7. The inguinal adipocyte distribution and size infemale offspring from control (n 6) and T-treated(n 6) dams. A and B: the mean inguinal adipocytesize was smaller than in controls (P � 0.05). C: thesize distribution was shifted to the left in the Toffspring.

E1383MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

REFERENCES

1. Abbott DH, Barnett DK, Bruns CM, Dumesic DA. Androgen excessfetal programming of female reproduction: a developmental aetiology forpolycystic ovary syndrome? Hum Reprod Update 11: 357–374, 2005.

2. Abbott DH, Bruns CR, Barnett DK, Dunaif A, Goodfriend TL,Dumesic DA, Tarantal AF. Experimentally induced gestational androgenexcess disrupts glucoregulation in rhesus monkey dams and their femaleoffspring. Am J Physiol Endocrinol Metab 299: E741–E751, 2010.

3. Al-Bader MD. Estrogen receptors alpha and beta in rat placenta: detectionby RT-PCR, real time PCR and Western blotting. Reprod Biol Endocrinol4: 13, 2006.

4. Amalfi S, Velez LM, Heber MF, Vighi S, Ferreira SR, Orozco AV,Pignataro O, Motta AB. Prenatal hyperandrogenization induces meta-bolic and endocrine alterations which depend on the levels of testosteroneexposure. PLoS One 7: e37658, 2012.

5. Azziz R, Carmina E, Dewailly D, Diamanti-Kandarakis E, Escobar-Morreale HF, Futterweit W, Janssen OE, Legro RS, Norman RJ,Taylor AE, Witchel SF; Task Force on the Phenotype of the PolycysticOvary Syndrome of The Androgen Excess and PCOS Society. TheAndrogen Excess and PCOS Society criteria for the polycystic ovarysyndrome: the complete task force report. Fertil Steril 91: 456–488, 2009.

6. Barker DJ. The developmental origins of adult disease. J Am Coll Nutr23: 588S–595S, 2004.

7. Barry JA, Kay AR, Navaratnarajah R, Iqbal S, Bamfo JE, David AL,Hines M, Hardiman PJ. Umbilical vein testosterone in female infantsborn to mothers with polycystic ovary syndrome is elevated to male levels.J Obstet Gynaecol 30: 444–446, 2010.

8. Bartholomeusz RK, Bruce NW, Lynch AM. Embryo survival, and fetaland placental growth following elevation of maternal estradiol bloodconcentrations in the rat. Biol Reprod 61: 46–50, 1999.

9. Belkacemi L, Jelks A, Chen CH, Ross MG, Desai M. Altered placentaldevelopment in undernourished rats: role of maternal glucocorticoids.Reprod Biol Endocrinol 9: 105, 2011.

10. Björnheden T, Jakubowicz B, Levin M, Oden B, Eden S, Sjöström L,Lönn M. Computerized determination of adipocyte size. Obes Res 12:95–105, 2004.

11. Chalder M, Wiles NJ, Campbell J, Hollinghurst SP, Searle A, HaaseAM, Taylor AH, Fox KR, Baxter H, Davis M, Thorp H, Winder R,Wright C, Calnan M, Lawlor DA, Peters TJ, Sharp DJ, Turner KM,Montgomery AA, Lewis G. A pragmatic randomised controlled trial toevaluate the cost-effectiveness of a physical activity intervention as atreatment for depression: the treating depression with physical activity(TREAD) trial. Health Technol Assess 16: 1–164, iii–iv, 2012.

12. Csapo A, Dray F, Erdos T. Letter: Oestradiol 17beta: inhibitor ofplacental growth. Lancet 2: 51–52, 1974.

13. Demissie M, Lazic M, Foecking EM, Aird F, Dunaif A, Levine JE.Transient prenatal androgen exposure produces metabolic syndrome inadult female rats. Am J Physiol Endocrinol Metab 295: E262–E268, 2008.

14. Dominguez R, Micevych P. Estradiol rapidly regulates membrane estro-gen receptor alpha levels in hypothalamic neurons. J Neurosci 30: 12589–12596, 2010.

15. Eisner JR, Dumesic DA, Kemnitz JW, Abbott DH. Timing of prenatalandrogen excess determines differential impairment in insulin secretionand action in adult female rhesus monkeys. J Clin Endocrinol Metab 85:1206–1210, 2000.

16. Even MD, Dhar MG, vom Saal FS. Transport of steroids between fetusesvia amniotic fluid in relation to the intrauterine position phenomenon inrats. J Reprod Fertil 96: 709–716, 1992.

17. Feng Y, Johansson J, Shao R, Mannerås L, Fernandez-Rodriguez J,Billig H, Stener-Victorin E. Hypothalamic neuroendocrine functions inrats with dihydrotestosterone-induced polycystic ovary syndrome: effectsof low-frequency electro-acupuncture. PLoS One 4: e6638, 2009.

18. Foecking EM, Szabo M, Schwartz NB, Levine JE. Neuroendocrineconsequences of prenatal androgen exposure in the female rat: absence ofluteinizing hormone surges, suppression of progesterone receptor geneexpression, and acceleration of the gonadotropin-releasing hormone pulsegenerator. Biol Reprod 72: 1475–1483, 2005.

19. Franks S, McCarthy MI, Hardy K. Development of polycystic ovarysyndrome: involvement of genetic and environmental factors. Int J Androl29: 278–285, 2006.

20. Hogg K, Wood C, McNeilly AS, Duncan WC. The in utero program-ming effect of increased maternal androgens and a direct fetal interventionon liver and metabolic function in adult sheep. PLoS One 6: e24877, 2011.

21. Hsu TY, Lan KC, Tsai CC, Ou CY, Cheng BH, Tsai MY, Kang HY,Tung YH, Wong YH, Huang KE. Expression of androgen receptor inhuman placentas from normal and preeclamptic pregnancies. Taiwan JObstet Gynecol 48: 262–267, 2009.

22. Johansson J, Feng Y, Shao R, Lönn M, Billig H, Stener-Victorin E.Intense electroacupuncture normalizes insulin sensitivity, increases mus-cle GLUT4 content, and improves lipid profile in a rat model of polycysticovary syndrome. Am J Physiol Endocrinol Metab 299: E551–E559, 2010.

23. King AJ, Olivier NB, Mohankumar PS, Lee JS, Padmanabhan V,Fink GD. Hypertension caused by prenatal testosterone excess in femalesheep. Am J Physiol Endocrinol Metab 292: E1837–E1841, 2007.

24. Lillycrop KA, Burdge GC. Epigenetic changes in early life and futurerisk of obesity. Int J Obes (Lond) 35: 72–83, 2011.

25. Manikkam M, Crespi EJ, Doop DD, Herkimer C, Lee JS, Yu S,Brown MB, Foster DL, Padmanabhan V. Fetal programming: prenataltestosterone excess leads to fetal growth retardation and postnatal catch-upgrowth in sheep. Endocrinology 145: 790–798, 2004.

26. March WA, Moore VM, Willson KJ, Phillips DI, Norman RJ, DaviesMJ. The prevalence of polycystic ovary syndrome in a community sampleassessed under contrasting diagnostic criteria. Hum Reprod 25: 544–551,2010.

27. Marcondes FK, Bianchi FJ, Tanno AP. Determination of the estrouscycle phases of rats: some helpful considerations. Braz J Biol 62: 609–614, 2002.

28. Meulenberg PM, Hofman JA. Maternal testosterone and fetal sex. JSteroid Biochem 39: 51–54, 1991.

29. Mustonen M, Poutanen M, Chotteau-Lelievre A, de Launoit Y, Iso-maa V, Vainio S, Vihko R, Vihko P. Ontogeny of 17beta-hydroxysteroiddehydrogenase type 2 mRNA expression in the developing mouse placentaand fetus. Mol Cell Endocrinol 134: 33–40, 1997.

30. Padmanabhan V, Veiga-Lopez A, Abbott DH, Recabarren SE, Her-kimer C. Developmental programming: impact of prenatal testosteroneexcess and postnatal weight gain on insulin sensitivity index and transferof traits to offspring of overweight females. Endocrinology 151: 595–605,2010.

31. Poston L. The control of blood flow to the placenta. Exp Physiol 82:377–387, 1997.

32. Prior BM, Cureton KJ, Modlesky CM, Evans EM, Sloniger MA,Saunders M, Lewis RD. In vivo validation of whole body compositionestimates from dual-energy X-ray absorptiometry. J Appl Physiol 83:623–630, 1997.

33. Roos N, Kieler H, Sahlin L, Ekman-Ordeberg G, Falconer H, Stepha-nsson O. Risk of adverse pregnancy outcomes in women with polycysticovary syndrome: population based cohort study. BMJ 343: d6309, 2011.

34. Sathishkumar K, Balakrishnan M, Chinnathambi V, Chauhan M,Hankins GD, Yallampalli C. Fetal sex-related dysregulation in testos-terone production and their receptor expression in the human placenta withpreeclampsia. J Perinatol 32: 328–335, 2012.

35. Sathishkumar K, Elkins R, Chinnathambi V, Gao H, Hankins GD,Yallampalli C. Prenatal testosterone-induced fetal growth restriction isassociated with down-regulation of rat placental amino acid transport.Reprod Biol Endocrinol 9: 110, 2011.

36. Sathishkumar K, Elkins R, Yallampalli U, Balakrishnan M, Yallam-palli C. Fetal programming of adult hypertension in female rat offspringexposed to androgens in utero. Early Hum Dev 87: 407–414, 2011.

37. Sathishkumar K, Yallampalli U, Elkins R, Yallampalli C. Fetal pro-gramming of adult hypertension in female rat offspring exposed toandrogens in utero. Reprod Sci 18: 219a–220a, 2011.

38. Schindler AE. Hormones in human amniotic fluid. Monogr Endocrinol21: 1–158, 1982.

39. Seckl JR. Glucocorticoid programming of the fetus; adult phenotypes andmolecular mechanisms. Mol Cell Endocrinol 185: 61–71, 2001.

40. Sir-Petermann T, Hitchsfeld C, Maliqueo M, Codner E, Echiburu B,Gazitua R, Recabarren S, Cassorla F. Birth weight in offspring ofmothers with polycystic ovarian syndrome. Hum Reprod 20: 2122–2126,2005.

41. Sir-Petermann T, Maliqueo M, Angel B, Lara HE, Perez-Bravo F,Recabarren SE. Maternal serum androgens in pregnant women withpolycystic ovarian syndrome: possible implications in prenatal androgeni-zation. Hum Reprod 17: 2573–2579, 2002.

42. Slob AK, den Hamer R, Woutersen PJ, van der Werff ten Bosch JJ.Prenatal testosterone propionate and postnatal ovarian activity in the rat.Acta Endocrinol (Copenh) 103: 420–427, 1983.

E1384 MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from

43. Sridaran R, Basuray R, Gibori G. Source and regulation of testosteronesecretion in pregnant and pseudopregnant rats. Endocrinology 108: 855–861, 1981.

44. Wang YC, Su HY, Liu JY, Chang FW, Chen CH. Maternal and femalefetal virilization caused by pregnancy luteomas. Fertil Steril 84: 509, 2005.

45. Ward IL, Ward OB, Affuso JD, Long WD 3rd, French JA, HendricksSE. Fetal testosterone surge: specific modulations induced in male rats bymaternal stress and/or alcohol consumption. Horm Behav 43: 531–539, 2003.

46. Warshaw ML, Johnson DC, Khan I, Eckstein B, Gibori G. Placentalsecretion of androgens in the rat. Endocrinology 119: 2642–2648, 1986.

47. Veiga-Lopez A, Steckler TL, Abbott DH, Welch KB, MohanKumarPS, Phillips DJ, Refsal K, Padmanabhan V. Developmental program-ming: impact of excess prenatal testosterone on intrauterine fetal endo-crine milieu and growth in sheep. Biol Reprod 84: 87–96, 2011.

48. Weisz J, Ward IL. Plasma testosterone and progesterone titers of preg-nant rats, their male and female fetuses, and neonatal offspring. Endocri-nology 106: 306–316, 1980.

49. Wetzel M, Marchais-Oberwinkler S, Perspicace E, Möller G, Adam-ski J, Hartmann RW. Introduction of an electron withdrawing group on

the hydroxyphenylnaphthol scaffold improves the potency of 17�-hydrox-ysteroid dehydrogenase type 2 (17�-HSD2) inhibitors. J Med Chem 54:7547–7557, 2011.

50. Wolf CJ, Hotchkiss A, Ostby JS, LeBlanc GA, Gray LE Jr. Effects ofprenatal testosterone propionate on the sexual development of male andfemale rats: a dose-response study. Toxicol Sci 65: 71–86, 2002.

51. Xu N, Kwon S, Abbott DH, Geller DH, Dumesic DA, Azziz R, Guo X,Goodarzi MO. Epigenetic mechanism underlying the development ofpolycystic ovary syndrome (PCOS)-like phenotypes in prenatally andro-genized rhesus monkeys. PLoS One 6: e27286, 2011.

52. Yee JK, Lee WN, Ross MG, Lane RH, Han G, Vega J, Desai M.Peroxisome proliferator-activated receptor gamma modulation and lipo-genic response in adipocytes of small-for-gestational age offspring. NutrMetab (Lond) 9: 62, 2012.

53. Zhu L, Hou M, Sun B, Buren J, Zhang L, Yi J, Hernell O, Li X.Testosterone stimulates adipose tissue 11beta-hydroxysteroid dehydroge-nase type 1 expression in a depot-specific manner in children. J ClinEndocrinol Metab 95: 3300–3308, 2010.

E1385MATERNAL ANDROGENS AND PLACENTAL AND OFFSPRING DEVELOPMENT

AJP-Endocrinol Metab • doi:10.1152/ajpendo.00421.2012 • www.ajpendo.org

by 10.220.33.6 on April 1, 2017

http://ajpendo.physiology.org/D

ownloaded from