Gestational Exposure to Ethane Dimethanesulfonate Permanently ...

BIOLOGY OF REPRODUCTION 69, 959–967 (2003) Published online before print 14 May 2003. DOI 10.1095/biolreprod.103.017343

Gestational Exposure to Ethane Dimethanesulfonate Permanently Alters Reproductive Competence in the CD-1 Mouse1 Dana K. Tarka-Leeds,3 Juan D. Suarez,4 Naomi L. Roberts,4 John M. Rogers,4 Matthew P. Hardy,5 and Gary R. Klinefelter2,4 Curriculum in Toxicology,3 University of North Carolina at Chapel Hill, Chapel Hill, North Carolina 27599-7270 Reproductive Toxicology Division,4 National Health and Environmental Effects Research Laboratory, United States Environmental Protection Agency, Research Triangle Park, North Carolina 27711 Population Council,5 New York, New York 10021 ABSTRACT

INTRODUCTION

Although the adult mouse Leydig cell (LC) has been considered refractory to cytotoxic destruction by ethane dimethanesulfonate (EDS), the potential consequences of exposure during reproductive development in this species are unknown. Herein pregnant CD-1 mice were treated with 160 mg/kg on Gestation Days 11–17, and reproductive development in male offspring was evaluated. Prenatal administration of EDS compromised fetal testosterone (T) levels, compared with controls. EDS-exposed pups recovered their steroidogenic capacities after birth because T production by hCG-stimulated testis parenchyma from prepubertal male offspring was unchanged. However, prepubertal testes from prenatally exposed males contained seminiferous tubules (STs) devoid of germ cells, indicating a delay in spermatogenesis. In adults, some STs in exposed males still contained incomplete germ cell associations corroborating observed reductions in epididymal sperm reserves, fertility ratios, and litter size. Morphometry revealed an EDS-induced increase in interstitial area and a concomitant decrease in ST area, but stereology revealed an unexpected decrease in the number and size of the LCs per testis in exposed males. Paradoxically, there was an increase in both serum LH and T production by adult testis parenchyma, indicating that the LCs were hyperstimulated. These data demonstrate permanent lesions in LC development and spermatogenesis caused by prenatal exposure in mice. Thus, although adult mouse LCs are insensitive to EDS, EDS appears to have direct action on fetal LCs, resulting in abnormal testis development.

In 1964 Harold Jackson [1] first described the antifertility effects of ethane dimethanesulfonate (EDS), an alkylating sulfonic ester first exploited for its potential in cancer chemotherapy [2]. EDS has since been used as a prototypic agent to study the androgen dependency of spermatogenesis and Leydig cell (LC) ontogeny [3]. EDS, unlike other compounds in its class that directly affect the spermatogenic epithelium, selectively and rapidly destroys LCs in the adult rat testis with a single injection of EDS (75 mg/kg) [4]. Serum testosterone (T) levels begin to decrease as early as 6 h after injection and are undetectable 9–10 days later [5]. Without androgen stimulation, spermatogenesis is suppressed and fertility is subsequently compromised; these alterations persist until LCs repopulate and resume androgen production [6, 7]. In the rat, the LC population is fully regenerated from undifferentiated progenitor cells by 49 days after injection and produces normal levels of T to restore and maintain spermatogenesis [5, 8–11]. Further studies of the toxicology of EDS have revealed temporal, organ, and species differences with respect to effects on LCs. For example, EDS was found to be cytotoxic to LCs in neonatal and suckling rats [12]. In neonatal rats dosed (Postnatal Day [PND] 5–16) s.c. with 50 mg/kg EDS, the fetal LC population is destroyed and the seminiferous tubules were permanently damaged either by direct cytotoxic effects prior to the formation of the blood-testis barrier or androgen deprivation resulting from fetal LC loss [13]. Furthermore, EDS was cytotoxic to adult LCs in the testis of the guinea pig [4], frog [14], rabbit [15], and lizard [16]. LCs in the adult Syrian hamster [17] and immature rat [18] exhibit significant effects of EDS exposure, although they are considered less sensitive than adult rat LCs based on the relative decrease in T production by LCs incubated in vitro with EDS. Additional effects of EDS, on targets other than LCs, have also been identified. For example, EDS is not cytotoxic to adult LCs in quail [19], mouse [4], and goat [20], although spermatogenesis is still disrupted by exposure in these species, creating a state of temporary infertility [4, 21]. In the adult mouse, EDS treatment induces partial spermatogenic disruption, evidenced by a loss of germ cells in the seminiferous tubules (STs) and expanded extracellular space between surviving germ cells, all without affecting LCs in the interstitium [4]. Furthermore, Klinefelter et al. [22] identified multiple effects of EDS exposure on the rat epididymis that are unrelated to androgen deprivation. The formation of sperm granulomas, morphological alterations of the epididymis, modification of sperm membrane pro-

Leydig cells, testis, testosterone, toxicology The investigation by D.K.T. was funded by the EPA/UNC Toxicology Research Program, Training Agreement CT 902908 and a predoctoral traineeship (National Research Service Award 5 T32 ES07126) from the National Institute of Environmental Health Sciences, National Institutes of Health, both with the Curriculum in Toxicology, University of North Carolina at Chapel Hill. The information in this document has been funded wholly (or in part) by the U.S. Environmental Protection Agency. It has been subjected to review by the National Health and Environmental Effects Research Laboratory and approved for publication. Approval does not signify that the contents reflect the views of the Agency, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. 2 Correspondence: Gary Klinefelter, Reproductive Toxicology Division MD 72, NHEERL, USEPA, Research Triangle Park, NC 27711. FAX: 919 541 4017; e-mail: [email protected] 1

Received: 19 March 2003. First decision: 11 April 2003. Accepted: 8 May 2003. Q 2003 by the Society for the Study of Reproduction, Inc. ISSN: 0006-3363. http://www.biolreprod.org

959

960

TARKA-LEEDS ET AL.

teins, and decreases in progressive motility and velocity of sperm could not be prevented with exogenous T [8, 22]. Furthermore, in vitro analysis revealed that EDS acts directly on the epididymal epithelium to mediate changes in sperm membrane proteins responsible for sperm motility [22]. Additionally, adverse functional effects on cultured Sertoli cells, including inhibition of the synthesis of the glycoprotein, transferrin [23], have been demonstrated at relevant sublethal in vitro doses of EDS (300–500 mM). Declining levels of this soluble iron transport protein can result in decreased numbers of germ cells and suppression of spermatogenesis [24]. The purpose of this study was to determine whether gestational exposure to EDS would adversely affect steroidogenesis in the developing mouse fetus and, if so, whether this results in persistent reproductive effects in the prepubertal and adult mouse. To our knowledge, this is the first study employing gestational exposure of EDS in any species. This study was designed to examine the effects of gestational EDS exposure on T biosynthesis throughout reproductive development, the attainment of puberty, and subsequent reproductive performance in the adult mouse. Here we provide the first demonstration that fetal mouse LCs are sensitive to EDS, as shown by decreases in the fetal T peak alterations persisting into adulthood including incomplete spermatogenesis, altered LC structure and function, and reduced fertility. MATERIALS AND METHODS

Animals Animal procedures in this study were approved by the Institute Animal Care and Use Committee at the National Health and Environmental Effects Research Laboratory (NHEERL) of the United States Environmental Protection Agency. Timed pregnant CD-1 mice were mated at Charles River Laboratories, Inc. (Raleigh, NC) and delivered to the NHEERL animal facility on Gestation Day (GD) 6. Dams were housed five per cage on laboratory grade pine shavings (Northeastern Products, Warrensburg, NY) in clear polycarbonate boxes until GD 17, when they were individually housed. They were housed under controlled temperature (228C 6 18C), relative humidity (50% 6 10%) and light (12L:12D, lights out at 1800 h). They were provided an ad libitum diet of rodent chow (RMH3000; PMI Nutrition International, Inc., Bentwood, MO) and tap water. For fertility studies, male CD-1 offspring (PND 72) previously exposed in utero were housed individually, and virgin female mice (8–10 wk) were initially housed 10 per cage. During the fertility study, two female mice were cohabited with one male mouse.

Treatment Administration Timed pregnant mice were randomly assigned to one of two treatment groups: vehicle (dimethyl sulfoxide [DMSO]: deionized water; 3:7, v/v) or EDS (gift from Sterling Research Group, Rensselaer, NY) (160 mg/kg body weight [BW] per day, in a 30% DMSO solution, i.p.) from GD 11 through GD 17; GD 0 5 day of postcoital plug. A preliminary study in which pregnant dams were given 160 mg/kg EDS daily from GD 11–17 revealed no overt maternal toxicity, but male offspring had significantly reduced fetal T and persistent decreases in body weight, testis weight, and fertility. Therefore, the current study focused on this same dosage of EDS. The volume administered was equivalent to 0.2 ml per 25 g of maternal BW.

Handling and Offspring For all end points, the litter became the experimental unit; therefore, litter mean data were analyzed and reported here. For endogenous wholebody T determination, both male and female fetuses and neonates were recovered via cesarean section on consecutive days from GD 15 to PND 4. For treatment-related effects on whole body T, time pregnant CD-1 dams were dosed from GD 11 to GD 17 with 160 mg/kg EDS. Fetal whole-body T production in male mice was evaluated from GD 15, when morphological identification of the male gonad could first be achieved,

through GD 18 just prior to birth. Dosing was terminated the day prior to cesarean section. Thus, a dam killed on GD 16 was dosed from GD 11 to GD 15. Fetuses were killed with CO2 followed by decapitation. They were weighed, placed in 12 3 75 mm borosilicate disposable glass culture tubes (Fisher Scientific, Pittsburgh, PA) (for fetuses aged GD 11–17) or 13 3 100 mm tubes (for fetuses and neonates aged GD 18 to PND 4), and frozen at 2808C until steroid extraction. On the day of birth, PND 1, live pups were counted, weighed, and examined for overt signs of toxicity (e.g., tremors, lethargy, malformations). Anogenital distance (AGD) was measured in all offspring on PND 1 and PND 82 from the base of the genital papilla to the proximal end of the anal opening with calipers calibrated to 0.1 mm: distances of 1.5 mm and longer on PND 1 were designated as male. On PND 4, litters were standardized to four males and four females. Male pups were individually identified with a s.c. injection of Permanent Tattoo Pigment 242 (Aims, Inc., Piscataway, NJ) with a 26-gauge needle and tuberculin syringe on the back of alternating paws [25]. At weaning, PND 21, dams and female offspring were killed via CO2 asphyxiation. Male littermates were housed together. On PND 24, two males per litter were removed for prepubertal necropsy. Remaining males were examined for preputial separation from PND 25 through PND 32 until puberty was attained. Thereafter, mice were housed individually until maturity (PND 70) for a 10-d fertility analysis (PND 70–80) and subsequent necropsy on PND 82.

Postnatal Necropsy of Male Offspring On PND 1, one male pup per litter was killed via CO2 followed by decapitation. The internal morphology of the urogenital tract and the positioning of the testes were noted. The right testis was removed and placed in a 5-ml glass scintillation vial containing 5% glutaraldehyde in 0.05 M collidine buffer with 0.1 M sucrose for histological analysis. The left testis was removed, placed in a microcentrifuge tube, and frozen at 2808C for proteomic analysis. On PND 24 and 82, male mice were killed via CO2 followed by cervical dislocation. Extravasated blood was collected from the neck region following decapitation and transferred to 1 ml serum separator tubes, placed on ice, and centrifuged (30 min, 3000 3 g, 48C). The serum was decanted and frozen at 2808C until serum T and LH could be assayed. Postmortem analysis was conducted on two prepubertal males per litter on PND 24 and two sexually mature males per litter on PND 82. Body and organ weights (testes, epididymides, and seminal vesicles trimmed free of fat and fascia) were recorded. The left testis was placed in a 20ml scintillation vial containing 5% glutaraldehyde in 0.05 M collidine buffer with 0.1 M sucrose for histological analysis. The right testis (PND 24) or 50-mg pieces of testis parenchyma (PND 82) were removed and placed into a microcentifuge tube with 1 ml Medium 199 (M199; #4001100, Gibco Laboratories, Grand Island, NY) for ex vivo determination of T production. The right cauda epididymis was removed, placed in a microcentrifuge tube, and frozen at 2808C for subsequent sperm enumeration. For immunohistochemical localization of P450 side chain cleavage enzyme in the LCs, adult testes (PND 53) were removed and immediately immersed in Bouin fixative (50 3 the volume of the tissue). A 26-gauge needle was used to puncture the tunica albuginea (two nicks on each side of the testis) to facilitate penetration of fixative into the testis. The testes were fixed overnight at 48C. The following morning, the testes were placed into cassettes and rinsed three times in 70% ethanol for 30 min on a stir plate at 48C. Following a fourth overnight rinse, the tissue remained in 70% ethanol until paraffin embedding.

Whole-Body T Extraction Whole-body T extraction was performed as described by Parks et al. [26]. Briefly, 1 ml double-distilled deionized water (ddH20) was added to each 12 3 75 mm culture tube or 2 ml ddH20 to each 13 3 100 mm tube, each tube containing a single fetus, and the contents were homogenized with a Polytron homogenizer (Brinkman Instruments, Westbury, NY). Samples were double extracted with diethyl ether, and the ether phase was evaporated in a fume hood and stored at room temperature for no more than 2 wk until RIA analysis for whole-body T was performed.

Sperm Counts The cauda epididymis of prepubertal (PND 24) and adult (PND 82) mice was removed, trimmed free of fat, and weighed. Each sample was homogenized in a 1-ml handheld glass homogenizer for 30 sec with 0.25 ml homogenization buffer containing 10 mM Tris (#161-0719; BioRad,

GESTATIONAL EDS EXPOSURE ALTERS MOUSE REPRODUCTION Hercules, CA), 1 mM EDTA (#E-6758; Sigma, St. Louis, MO), 0.5% 3[3-cholamidopropyl-dimethyl-ammonio]-1 propane sulfonate (CHAPS; #75621-03-3; Calbiochem, La Jolla, CA), and 0.25% octyl-beta-D-glucopyranoside (OBG; #494460; Calbiochem) pH 7.2. The sample was decanted into a microcentrifuge tube and stored at 48C. An aliquot (0.1 ml) was diluted with 0.9 ml of buffer, and sperm were counted using a hemocytometer slide. Minimally, three hemocytometer chambers were counted per sample.

Radioimmunoassay of T and LH Testosterone levels from dried perinatal extracts were quantified via RIA using a Coat-a-Count Total T kit according to the manufacturer instructions (Diagnostic Products Corporation, Los Angeles, CA). Dried perinatal extracts were resuspended in 120 ml of zero standard, vortexed for 30 sec, and a 50-ml aliquot assayed in duplicate for T [26]. Levels of T in both ex vivo media and blood serum were also quantified using a Coat-a-Count total T kit. Serum LH concentrations were measured using the method from Chandrashekar and Bartke [27]. Rat LH standards NIDDK-r-LH-19 and rat LH antibody NIDDK-anti-rLH-S-11 were obtained through the National Hormone and Pituitary Program. Radioactive 125I-rat LH was obtained from Covance Laboratories (Vienna, VA), and IgG antiserum was obtained from ICN Pharmaceuticals (Costa Mesa, CA).

Ex Vivo Testis Parenchyma Incubations Production of T by testis parenchyma following gestational exposure to EDS was determined by modification of the protocol used by Klinefelter et al. [28] and Parks [26]. Briefly, the right testis of each mouse was removed. The tunica was perforated (PND 24) or removed (PND 82), and the whole testis (PND 24) or 50-mg pieces of testis parenchyma (PND 82) were incubated (348C) while shaking for 30 min in 1 ml of M199 buffered with 0.71 g/L sodium bicarbonate, 2.21 g/L N-2-hydroxyethyl piperazine-N9-2 ethenesulfonic acid (Hepes; Sigma), 0.1% BSA (SchwarzMann, Cleveland, OH), and 25 mg/L soy bean trypsin inhibitor, pH 7.4 (Sigma). After this incubation, medium was removed and replaced with fresh medium, with or without maximal hCG stimulation (100 mIU/ml), for 2 h. Media were frozen at 2808C until RIA analysis.

Fertility Analysis The fertility analysis was based on a modification of two different mating protocols [29, 30]. Initially, two untreated virgin female mice (PND 65–75) were cohabited with an individual adult male offspring (PND 70– 80) that had been prenatally exposed to either 160 mg EDS/kg or 30% DMSO vehicle. Initially, two females were placed in a box with a single male at 1700 h (lights out at 1800 h). Each morning (between 0700 and 0900 h; lights on at 0600 h) of the 10-day study, the vagina of each of the two females was examined for the presence of a postcoital plug. In the event a plug was observed, the female was designated as GD 0, promptly removed from the box with the male, and group housed with other females (up to 10 per box) that had cohabited with a specific male. In the event that a plug was not observed, the female remained in the box with the male until she presented with a postcoital vaginal plug but not to exceed 4 days. If a female did not present with a vaginal postcoital plug after 4 days, she was removed from the box with the male on the morning of the fifth day of cohabitation and placed in the box with the other females that had cohabited with that specific male. This 4-day maximum was chosen to ensure that the female had completed one full estrous cycle while in the presence of a male. Each evening of the 10-day study (at 1700 h; lights out at 1800 h), the female mice that had been removed that morning were replaced with new virgin female mice, assuring that two females cohabited with each male every evening. At the completion of the 10-day study, the males were necropsied at PND 82 and the females that had cohabited with a male were group housed according to the male they cohabited with until GD 14. On GD 14, females were euthanized with CO2 followed by cervical dislocation and cesarean section. The following parameters were evaluated in pregnant dams: litter size, number of fetal resorptions, and observations of gross terata.

Testis Histology The method for testis histology is a modification of the protocol by Klinefelter et al. [22] and Parks et al. [26]. The testes from males on PND 1, 24, and 82 were removed and immersion fixed in 5% glutaraldehyde (#01909; Polysciences, Warrington, PA) in 0.05 M collidine buffer with

961

0.1 M sucrose overnight at 48C. The tissue was rinsed twice with 0.2 M collidine buffer for 5 min. Testes were postfixed in 1% aqueous osmium tetroxide (#223A; Polysciences) in 0.05 M collidine buffer on ice for 1 to 2 h followed by a 5-min rinse in 0.2 M collidine. Dehydration consisted of two 5-min washes in 70%, 80%, and 95% ethanol on ice, and finally three 20 min washes in 100% ethanol at room temperature. After two 10min rinses in propylene oxide, a 1-h incubation in propylene oxide:Epon (1:1), and an overnight incubation (48C) in 100% Epon, tissues were embedded in 100% fresh Epon at 608C for 48 h. Sections (1 mm thick) were cut with an RMS Ultramicrotome (MT-7; RMC Inc., Tucson, AZ) and stained with an aqueous solution of toluidine blue (#02205; Electron Microscopy Sciences, Fort Washington, PA). Images were captured with a Nikon Eclipse (E800, Nikon Instruments, Melville, NY) equipped with a digital SPOT camera (Southern Micro Instruments, Atlanta, GA) and corresponding SPOT RT version 3.2 software (Diagnostics Instruments, Sterling Heights, MI). For immunocytochemical/stereological studies, testes were removed from 70% ethanol, rinsed two times in 95% ethanol for 15 min, three times in 100% ethanol for 20 min, and two times in 100% xylene for 30 min. Tissue was incubated in a 1:1 xylene:paraffin (Paraplast Plus; Oxford Labwares, St. Louis, MO) bath at 608C, two times for 45 min, and then three times for 30 min in a 100% paraffin bath at 608C. Paraffin blocks were allowed to cool overnight and then stored at 48C until sections (4 to 6 mm) were obtained. After sectioning, the slides were held at room temperature overnight and then stored at 48C until immunocytochemical analysis.

Immunocytochemistry P450 side chain cleavage (P450scc) enzyme was immunolocalized to positively identify LCs [31]. Paraffin sections of the testis were deparaffinized via a graded ascending series of ethanol and xylene and microwaved at low power for 60 sec in 500 ml 0.1 M sodium citrate buffer (#2121; Polysciences) pH 6 for antigen retrieval. Sections were washed two times in PBS for 3 min, incubated for 20 min at room temperature in 30% H2O2 in methanol, and again washed two times in PBS. Sections were blocked for 1 h at 348C with sterile PBS containing 1.0% proteasefree BSA (#A-3294, Sigma), 10% normal goat serum (#S-1000, lot #I 1013; Vector Laboratories, Inc., Burlingame, CA), and 0.3% Triton X-100 (Surfact-Amps X-100; #46472; Pierce, Rockford, IL). Slides were incubated at room temperature for 30 min in rabbit anti-Cytochrome P450scc (#RDI-P450sccabr, Research Diagnostics, Inc.; Flanders, NJ), diluted 1: 200 in PBS containing 1.0% BSA, and then washed two times in PBS. Following a 1-h incubation at room temperature with biotinylated antirabbit IgG (H1L) (#BA-10000; Vector Laboratories) diluted 1:100 in PBS, sections were washed twice in PBS, incubated for 1 h with avidin-biotin peroxidase complex (Vectastain ABC reagent; Vector Laboratories), and washed twice with PBS. Bound antibody was localized with Vector DAB peroxidase substrate kit (#SK-4100; Vector Laboratories). Sections were washed in dH2O and counterstained with hematoxylin. Slides were rinsed in running tap water and dehydrated through a graded descending ethanol series before coverslipping with Vectamount (H-5000; Vector Laboratories).

Stereology Testis sections representing 9 and 10 testes from control and EDS exposed PND 53 testes were used for stereology. For each section, five distinct areas were selected in a uniform fashion beginning with the center and moving to each of the four corners. Images were captured with a digital SPOT camera (Southern Micro Instruments, Atlanta, GA) connected to a Nikon Eclipse E800 microscope equipped with a 340 objective. Images were analyzed with SPOT RT version 3.2 software. This software was used to determine diameters and areas of a traced image. For each of the five captured images, the number of immunostained cells (LCs) was counted and interstitial and tubular areas were delineated. We justified using the volume equation for a sphere to define the LC volume because the ratio of the longest to shortest diameter did not differ significantly (P , 0.05). Thus, roundness could be assumed. Briefly, the area and volume of an individual LC were approximated by taking the average of the longest and shortest diameter, dividing it in half to yield the average radius, and then applying the calculations for area (A 5 p 3 radius2) and volume (V 5 4/3[ p 3 radius3]). The LC volume density was calculated as the percentage of testis profile area that was LC area, and the total volume of LC per testis was estimated as the LC volume density 3 testis volume. The absolute number of LC per testis was estimated by dividing the total volume of LC per testis by the average volume of an individual LC [32].

962

TARKA-LEEDS ET AL.

FIG. 1. Effect of gestational EDS exposure on whole-body T levels in the fetal male mouse. The endogenous GD 16 T peak is significantly diminished with treatment. Values represent litter means 6 SEM; statistical significance (P # 0.05 5 *) is indicated; N 5 3–4 per time point.

Statistical Analysis For all end points, the litter means were analyzed by ANOVA with PROC GLM (SAS, version 8; SAS Institute Inc., Cary, NC); the numbers of litters represented are provided in the figure legends. Where statistically significant (P # 0.05) effects were seen with the overall ANOVA model, these results were compared with the least-square means (LSMEANS: a two-tailed t-test) of treatment groups to determine significant differences (P # 0.05) between treatments. Data on anogenital distance, preputial separation, and reproductive organ weights were analyzed with body weight as a covariate [37]. Unless stated otherwise, only statistically significant (P , 0.05) effects are discussed in the following section.

RESULTS

Administration of 160 mg/kg EDS during GD 11–17 did not affect overall maternal weight gain; however, it did decrease neonatal body weight on PND 1 by 15%, 1.72 vs. 1.47 g. This decrease was statistically significant and persisted in both the prepubertal and adult offspring. Both AGD and the onset of preputial separation, sensitive indicators of circulating androgens or variations in androgen responsiveness during sexual differentiation [37, 38], were evaluated in male offspring. On PND 1 and PND 82, 9–12 litters were represented per treatment group per time point. On both PND 1 and PND 82, AGD was decreased in EDSexposed males, 2.3 vs. 2.2 and 17.2 vs. 14.7 on PND 1 and PND 82, respectively. However, these differences were not significant after adjusting for the EDS-induced decreases in body weight by covariate analysis. Likewise, the observed delay in the onset of puberty (PND 26 vs. PND 30) was not delayed in males exposed prenatally to EDS when covariate analysis was run to correct for the decrease in body weight. Initial studies were performed to evaluate endogenous whole-body T levels in the fetal CD-1 mice utilized in this study. In control mice T began to increase beginning on GD 15, peak on GD 16 at 1.0 6 0.1 pg/mg of fetus, and subsequently decrease until PND 4 (data not shown). In male offspring prenatally exposed to EDS, the T levels on GD 15 and 16 were decreased, compared with controls (Fig.1). Litter means for T levels on GD 15 were 0.94 6 0.01 vs. 0.84 6 0.02 pg/mg of fetus and on GD 16 were 1.45 6 0.17 vs. 0.99 6 0.02 pg/mg of fetus for control and

FIG. 2. Effect of EDS exposure on reproductive organ weights and cauda epididymal sperm number. A) Effect of EDS on reproductive organ weights of adult (PND 82) male mice. Values represent litter means 6 SEM; statistical significance (P # 0.001 5 *) is indicated; N 5 12 and 11 for vehicle and EDS-exposed groups, respectively. B) Effect of EDS on cauda epididymal sperm numbers in adult (PND 82) male mice. Values represent litter means 6 SEM; statistical significance (P # 0.0001 5 *) is indicated; N 5 9.

EDS, respectively. By GD 17, T levels were comparable with control. Although seminal vesicle weight was not affected, the weights of other T-dependent organs, i.e., testis and epididymis, of adult mice were decreased in a body weight-independent manner (Fig. 2). Enumeration of cauda epididymal sperm from adult mice revealed a decrease in sperm; 17.9 3 106 6 9.0 3 105 vs. 2.7 3 106 6 6.7 3 105 cauda epididymal sperm for males exposed to control and EDS, respectively. The steroidogenic capacity of the LCs was not significantly affected in either unstimulated or hCG-stimulated prepubertal testis parenchyma from control and EDS exposed male offspring on PND 24, although there were trends toward increases in T production under both unstimulated and hCG-stimulated conditions (Fig. 3). Likewise, there was a trend toward increased unstimulated T production on PND 82. However, in these adults hCG-stimulated T production was more than doubled, 2.4 6 0.2 vs. 8.4 6 2.0 mg/mg testis in control and EDS-treated mice, respectively. Although hCG-stimulated T production was increased in

GESTATIONAL EDS EXPOSURE ALTERS MOUSE REPRODUCTION

FIG. 3. Treatment-related changes in ex vivo T production by testis parenchyma. A) T production (ng/mg of testis) of both vehicle and EDSexposed prepubertal (PND 24) CD-1 mouse testes incubated with or without maximal hCG-stimulation for 2 h. Note a trend toward increases in T production. Values represent litter means 6 SEM; N 5 7 and 8 for vehicle and EDS-exposed groups, respectively. B) T production (mg/mg of testis) of both vehicle and EDS-exposed adult (PND 82) CD-1 mouse testes incubated with or without maximal hCG-stimulation for 2 h. Values represent litter means 6 SEM; statistical significance (P # 0.01 5 *) is indicated; N 5 9 and 6 for vehicle- and EDS-exposed groups, respectively.

adult testis parenchyma ex vivo, serum T was unaffected. Serum LH was not affected by EDS treatment at the prepubertal time point. In adult EDS-exposed mice, despite the unchanged serum T levels, serum LH levels were increased 4-fold, 0.46 6 0.09 vs. 2.10 6 0.67 ng/ml for control and EDS-exposed mice, respectively (Fig. 4). Light microscopic examination of the neonatal (PND 1) testes from EDS-exposed mice revealed a seminiferous epithelium containing primarily Sertoli cells with only a limited organization of germ cells (Fig. 5). There was an abundance of LCs in the interstitium of both vehicle-treated and EDS exposed testis; however, the LCs tended to form aggregates of greater than 20 LCs in testes of EDS-exposed males. The testes of prepubertal (PND 24) control males revealed a maturing seminiferous epithelium with the presence of a lumen and evidence of active spermatogenesis. By contrast, the testes of the EDS-exposed prepubertal mice primarily contained Sertoli cells with abundant cytoplasm but few germ cells. Testes from adult (PND 82) control males revealed complete spermatogenesis. By contrast, although some seminiferous tubules in EDS-exposed mice

963

FIG. 4. Effect of EDS exposure on serum T (A) and LH (B) levels in the prepubertal (PND 24) and adult (PND 82) male mouse. Values represent litter means 6 SEM; statistical significance (P # 0.05 5 *); N 5 at least 5 in each group.

appeared normal, other tubule cross-sections continued to have incomplete spermatogenesis with Sertoli cell only tubules. Morphometric analysis revealed a decrease in the seminiferous tubule area, 57 516 6 535 vs. 52 835 6 1203.56 mm2, an increase in the corresponding interstitial area, 7728 6 535 vs. 12 409 6 1204 mm2. The smaller apparent tubule size in EDS-exposed testes coupled with a concomitant increase in interstitial space resulted in a perceptional increase in LC number. This apparent increase was analyzed using stereology to determine whether there were real alterations in the size of the LC population. Immunocytochemistry for P450scc enzyme was used to positively identify LCs in paraffin sections of adult testes (PND 52). Stereologic analysis of these tissue sections indicated that the total number of LC per testis was actually decreased in the EDS-exposed males, 5.7 3 106 6 4.6 3 105 vs. 4.2 3 106 6 2.8 3 105 LCs for control and EDS, respectively (Fig. 6). Analysis also revealed a decrease in the average volume of LCs, 1325.4 6 96.9 vs. 779.6 6 38.2 mm3 as well as an increase in the LC volume density, 6.2% vs. 9.3%. The average T production per testis from ex vivo incubations was divided by the average number of LCs per testis to estimate T production per LC. On either an unstimulated or hCG-stimulated basis, estimated T production per LC was increased in EDS-exposed males. The estimate for unstimulated T production was 9.3 3 1029 ng T per LC and 373.4 3 1029 ng T per LC for control and EDS, re-

964

TARKA-LEEDS ET AL.

FIG. 5. Light micrographs of testes from vehicle and EDS-exposed males. A) Cross section of the testis of a vehicle-exposed neonatal (PND 1) male. The germ cell layers have not yet formed and the LCs are dispersed uniformly throughout the interstitium. B) Testis from an EDS-exposed neonatal male. There are multiple areas of LC aggregates throughout the interstitium (arrow). C) Testis of a vehicle-exposed prepubertal (PND 24) male. Multiple germ cell layers have been formed and the lumen is visible. D) Testis from an EDS-exposed prepubertal mouse. The germ cell layers have not yet developed and the majority of tubule cross sections are Sertoli cell only (arrow). E) Testis from a vehicleexposed adult (PND 82) male mouse. Spermatogenesis is complete with the complete development of the germ cell layers. F) Testis from an EDS-exposed adult male mouse. Spermatogenesis is still incomplete as evidenced by few normal germ cell associations and abundant Sertoli cell cytoplasm with vacant spaces or vacuoles (arrow). It appears that the interstitial space is larger and that LCs are more abundant.

spectively. The estimate for stimulated T productions was 424.0 3 1029 ng T per LC and 1979.0 3 1029 ng T per LC for control and EDS, respectively. Adult male fertility was assessed for differences in mating ratio (the number of total observed copulatory plugs divided by the number of female mice that cohabited with each male), the fertility ratio (the number of actual litters divided by the number of observed copulatory plugs), and litter size (Fig. 7). The mating ratio was reduced, 0.74 6 0.04 vs. 0.56 6 0.08, as was the fertility ratio, 0.96 6 0.08 vs. 0.56 6 0.13, and the size of the F2 litters, 12.7 6 0.4 vs. 9.4 6 1.3 pups. DISCUSSION

Although indices of normal sexual development such as AGD and time to preputial separation were altered, these changes most likely were due, not to compromised fetal T production, but rather to the EDS-induced decrease in body weight of the male offspring. Fetal whole-body and serum T values in normal male mice were consistent with previous reports, showing an increase between GD 12.5 and 16.5, with a peak in plasma T levels at GD 16.5, just prior to a marked drop on the last day before parturition [39]. Gestational exposure to EDS from GD 11–17 resulted in a diminished fetal T peak in male offspring and associated alterations in testis development and reproductive compe-

tence in the adult. Our results are consistent with literature, suggesting that mouse LCs are less sensitive to the cytotoxicity of EDS than rat LCs [21] because fetal LCs were not destroyed following exposure as evidenced by their presence in histological sections. Moreover, these fetal LCs were still producing T, albeit at a reduced level. Although 160 mg/kg EDS was administered to the dams, we did not determine what level of EDS the fetus was exposed to; however, it may have been considerably less. The trend toward increases in both unstimulated and hCG-stimulated T production per milligram prepubertal testis in the EDS-exposed animals suggests that the hypothalamic-pituitary axis may compensate for the reduced levels of fetal T. The significant increase in the hCG-stimulated T production in adult testes from EDS exposed offspring further supports this notion. The increase in adult serum LH levels, despite the absence of significant changes in the serum T levels, suggests that there might be an alteration in the endocrine axis, i.e., changes in feedback regulation at the level of the hypothalamus/pituitary. Estimations of testosterone produced per adult LC in control and EDS-exposed males suggest that the LCs in EDS-exposed males are hyperstimulated. We estimated 40- and 5-fold increases in testosterone production per LC under unstimulated and hCG-stimulated conditions, respectively, in EDS-exposed males.

GESTATIONAL EDS EXPOSURE ALTERS MOUSE REPRODUCTION

965

FIG. 7. Assessment of adult (PND 70–80) male fertility. Values represent litter means 6 SEM; statistical significance (P # 0.05 5 *) is indicated; N 5 16 and 17 for vehicle- and EDS-exposed groups, respectively.

FIG. 6. Light micrographs of paraffin-embedded, P450scc-stained, 4- to 6-mm cross sections of vehicle- and EDS-exposed testes. A) Cross section of a vehicle-exposed testis with P450scc enzyme immunostaining to positively identify LCs; brown reaction product indicates LC staining. B) Testis from an EDS-exposed male. Spermatogenesis is incomplete, as evidenced by the paucity of germ cell associations and there is an abundance of small, stained LCs within the interstitium. C) Summary of the stereology of testes from vehicle and EDS-exposed male mice. Values represent litter means 6 SEM; statistical significance (P # 0.05 5 * and P # 0.001 5 **) is indicated; N 5 9 and 10 for vehicle- and EDS-exposed groups, respectively. Estimates of T production per LC are also provided.

It was possible that the elevated ex vivo T production in the adult merely resulted from an increase in number of LCs rather than an increase in LH stimulation per LC. Although it appeared that there was indeed an increase in LC numbers in the EDS-exposed males, it was necessary to perform stereologic analysis to discriminate an actual increase in LC numbers versus the appearances of LC hyperplasia because of seminiferous tubule damage, a historic controversy [40, 41]. Accordingly, LCs were immunostained for P450scc enzyme and enumerated. The observed increase in the interstitial area resulted from diminished tubule volume, and the apparent increase in the number of LCs was a consequence of this tubular alteration. It was

surprising to observe not only a decrease in the overall number of LCs but also a decrease in the volume of an individual LC. This alone indicates a permanent lesion in LC development; a defect that is presumably compensated by the observed increase in serum LH and the resulting increase in T production by these adult LCs. An evaluation of testicular histology in neonatal, prepubertal, and adult testes from EDS-exposed mice demonstrated that EDS exposure of the male fetus results in a persistent deficit in spermatogenesis. The maturation of the testis, i.e., the onset of spermatogenesis, was delayed as evidenced by the predominance of Sertoli cell only tubules throughout the development of the testes in EDS-exposed offspring. The fetal testes of EDS-exposed mice contained large LC aggregates within the interstitium; these aggregates were juxtaposed to tubules with disorganized seminiferous epithelium. A similar phenomenon has been observed in the fetal testes of rats exposed gestationally to DEHP, an antiandrogenic chemical that does not bind to the androgen receptor but does decrease fetal T production [26]. In the prepubertal mouse, seminiferous tubules in EDS-exposed offspring contained Sertoli cell only profiles. It is likely that the appearance of intraepithelial vacuoles is due partly to the absence of germ cells. Because the accumulation of vacuoles in conjunction with chemical disruption of spermatogenesis is a common, nonspecific response to a variety of insults on the testis, this may or may not be a direct consequence of EDS exposure [3]. At adulthood, EDS-exposed mice still manifest incomplete spermatogenesis and numerous tubule cross sections still lacked germ cells. Similar disruption of spermatogenesis was previously observed in adult EDS-exposed mice 3 wk after five daily treatments (25 mg EDS/100 g BW) in the absence of apparent effects on the interstitium or LCs [21]. It is well accepted that fertility end points in individuals are relatively insensitive measures of reproductive competence as mice with a 90% reduction in sperm counts have been shown to reproduce normally [42]. We evaluated fertility with a continuous 10-day mating regimen to enhance the sensitivity of fertility by natural mating. The observed decreases in the mating ratio suggest dysfunctional mating behavior in the male mice. This may be due to persistent effects of EDS on the brain, erectile responsiveness, or merely a consequence of decreased body weight and the resulting size incompatibility with the females. The observed decrease in fertility ratio, as well as the smaller litter sizes, is likely a direct result of the body weight-independent reduction in epididymal sperm reserves, a consequence of the incomplete spermatogenesis we observed by histological evaluation of the adult testis. The sustained effects of EDS may not be confined to LCs. Indeed, EDS may exert a direct effect on the semi-

966

TARKA-LEEDS ET AL.

niferous tubules because studies have demonstrated that EDS exerts direct effects on organs (i.e., epididymis) and cells (i.e., Sertoli cells) other than the LC [20, 22, 23]. Therefore, it is possible that EDS may exert direct effects on both LCs and the seminiferous tubule. Alternatively, the effects on the LCs may represent indirect action resulting from direct seminiferous tubule toxicity. Finally, the effects on the seminiferous tubules may represent an indirect action resulting from direct LC toxicity. Regardless, it is now clear that prenatal exposure to EDS results in reduced fetal T and persistent alterations in LC development. Whether the permanent alterations in the seminiferous tubule, i.e., spermatogenesis, manifest from the reduced fetal T peak, are independent of the LC, or the result of alterations in LC paracrine function remains unknown. Preliminary results (G. Klinefelter, unpublished) of a proteomic evaluation of PND 1 testes from control and EDS-exposed males reveal that three proteins potentially implicated in LC steroidogenesis are diminished by gestational exposure to EDS: apolipoprotein A-1, phosphatidylethanolamine, and thioredoxin-like protein. Future efforts will be made to follow proteomic changes at subsequent time points (i.e., PND 7 and 14) to determine whether and when putative paracrine changes might occur. Moreover, it will be important to compare changes in protein expression between isolated (i.e., laser capture and/or in vitro purification methods) fetal and progenitor LCs to changes that are observed in whole testis parenchyma. In summary, male mice exposed to EDS from GD 11– 17 manifest a significant reduction in the fetal T peak, body weight-independent decreases in testis and epididymal weights, decreased epididymal sperm reserves, altered steroidogenic potential of LC, incomplete spermatogenesis, and deficits in mating and fertility. Given that LCs are fewer and smaller in the testis of EDS-exposed males, the gestational EDS exposure clearly led to a permanent lesion in the development of the LC and spermatogenesis. ACKNOWLEDGMENTS The authors gratefully acknowledge the statistical support provided by Judith E. Schmid of the USEPA. We appreciate the time and technical histological expertise of Gail H. Grossman of the Cell and Developmental Biology Department at the University of North Carolina at Chapel Hill. Additionally, we would like to thank Dr. John Rockett for his advice in setting up our fertility study as well as Dr. Louise Parks and Christy Lambright for their expertise in the diethyl ether extraction of testosterone.

REFERENCES 1. Jackson H. Effects of alkylating agents on fertility. Br Med Bull 1964; 20:107–114. 2. Jackson H, Craig AW. Effects of alkylating chemicals on reproductive cells. Ann N Y Acad Sci 1969; 160:215–227. 3. Kerr JB, Savage GN, Millar M, Sharpe RM. Response of the seminiferous epithelium of the rat testis to withdrawal of androgen: evidence for direct effect upon intercellular spaces associated with Sertoli cell junctional complexes. Cell Tissue Res 1993; 274:153–161. 4. Kerr JB, Knell CM, Abbott M, Donachie K. Ultrastructural analysis of the effect of ethane dimethanesulphonate on the testis of the rat, guinea pig, hamster and mouse. Cell Tissue Res 1987; 249:451–457. 5. Edwards G, Lendon RG, Morris ID. Accelerated recovery of Leydig cells of the immature rat testis after administration of the cytotoxic ethylene dimethanesulphonate. J Endocrinol 1988; 119:475–482. 6. Gray LE Jr, Wolf C, Lambright C, Mann P, Price M, Cooper RL, Ostby J. Administration of potentially antiandrogenic pesticides (procymidone, linuron, iprodione, chlozolinate, p,p9-DDE, and ketoconazole) and toxic substances (dibutyl- and diethylhexyl phthalate, PCB 169, and ethane dimethane sulphonate) during sexual differentiation produces diverse profiles of reproductive malformations in the male rat. Toxicol Ind Health 1999; 15:94–118.

7. Sharpe RM. Toxicity of spermatogenesis and its detection. In: Korach KS (ed.), Reproductive and Developmental Toxicology. New York, NY: Marcel Dekker, Inc., 1998: 625–634. 8. Cooper ER, Jackson H. Comparative effects of methylene, ethylene and propylene dimethanesulphonates on the male rat reproductive system. J Reprod Fertil 1970; 23:103–108. 9. Jackson H. Chemical methods of male contraception. Am Sci 1973; 61:188–193. 10. Jackson AE, O’Leary PC, Ayers MM, de Kretser DM. The effects of ethylene dimethane sulphonate (EDS) on rat Leydig cells: evidence to support a connective tissue origin of Leydig cells. Biol Reprod 1986; 35:425–437. 11. Teerds KJ, de Rooij DG, Rommerts FF, Wensing CJ. Development of a new Leydig cell population after the destruction of existing Leydig cells by ethane dimethane sulphonate in rats: an autoradiographic study. J Endocrinol 1990; 126:229–236. 12. Smart JL, Massey RF, Lendon RG, Morris ID. Growth and development of male and female rats treated with the Leydig cell cytotoxin ethane dimethane sulphonate during the suckling period. Food Chem Toxicol 1990; 28:121–128. 13. Zaidi A, Lendon RG, Dixon JS, Morris ID. Abnormal development of the testis after administration of the Leydig cell cytotoxic ethylene1,2-dimethanesulphonate to the immature rat. J Reprod Fertil 1988; 82:381–392. 14. Minucci S, Fasano S, Di Matteo L, Chieffi Baccari G, Pierantoni R. Morphological and hormonal changes in the frog, Rana esculenta, testis after administration of ethane dimethane sulfonate. Gen Comp Endocrinol 1990; 79:335–345. 15. Laskey JW, Klinefelter GR, Kelce WR, Ewing LL. Effects of ethane dimethanesulfonate (EDS) on adult and immature rabbit Leydig cells: comparison with EDS-treated rat Leydig cells. Biol Reprod 1994; 50: 1151–1160. 16. Minucci S, Fasano S, Marmorino C, Chieffi P, Pierantoni R. Ethane 1,2-dimethane sulfonate effects on the testis of the lizard, Podarcis s. sicula Raf: morphological and hormonal changes. Gen Comp Endocrinol 1995; 97:273–282. 17. Gray LE, Klinefelter G, Kelce W, Laskey J, Ostby J, Ewing L. Hamster Leydig cells are less sensitive to ethane dimethanesulfonate when compared to rat Leydig cells both in vivo and in vitro. Toxicol Appl Pharmacol 1995; 130:248–256. 18. Kelce WR, Zirkin BR, Ewing LL. Immature rat Leydig cells are intrinsically less sensitive than adult Leydig cells to ethane dimethanesulfonate. Toxicol Appl Pharmacol 1991; 111:189–200. 19. Jones P, Kominkova E, Jackson H. Effects of antifertility substances on male Japanese quail. J Reprod Fertil 1972; 29:71–78. 20. Onyango DW, Wang EO, Oduor-Okelo D, Werner G. Early testicular response to intraperitoneal administration of ethane dimethanesulphonate (EDS) in the goat (Capra hircus). J Submicrosc Cytol Pathol 2001; 33:117–124. 21. Jackson CM, Jackson H. Comparative protective actions of gonadotrophins and testosterone against the antispermatogenic action of ethane dimethanesulphonate. J Reprod Fertil 1984; 71:393–401. 22. Klinefelter GR, Laskey JW, Roberts NR, Slott V, Suarez JD. Multiple effects of ethane dimethanesulfonate on the epididymis of adult rats. Toxicol Appl Pharmacol 1990; 105:271–287. 23. Roberts KP, Griswold MD. Ethylene dimethanesulfonate inhibits the function of Sertoli cells in culture at sublethal doses. Endocrinology 1990; 126:1618–1622. 24. Sylvester SR, Griswold MD. The testicular iron shuttle: a ‘‘nurse’’ function of the Sertoli cells. J Androl 1994; 15:381–385. 25. Bobbitt C. Foot Tattoo Identification for Rodent Pups. Research Triangle Park, NC: Chemical Industry Institute of Toxicology; 1997: 1–3. 26. Parks LG, Ostby JS, Lambright CR, Abbott BD, Klinefelter GR, Barlow NJ, Gray LE Jr. The plasticizer diethylhexyl phthalate induces malformations by decreasing fetal testosterone synthesis during sexual differentiation in the male rat. Toxicol Sci 2000; 58:339–349. 27. Chandrashekar V, Bartke A. Influence of endogenous prolactin on the luteinizing hormone stimulation of testicular steroidogenesis and the role of prolactin in adult male rats. Steroids 1988; 51:559–576. 28. Klinefelter GR, Laskey JW, Kelce WR, Ferrell J, Roberts NL, Suarez JD, Slott V. Chloroethylmethanesulfonate-induced effects on the epididymis seem unrelated to altered Leydig cell function. Biol Reprod 1994; 51:82–91. 29. Nagao T, Kuwagata M, Saito Y. Effects of prenatal exposure to 5fluoro-29-deoxyuridine on developing central nervous system and reproductive function in male offspring of mice. Teratog Carcinog Mutagen 1998; 18:73–92.

GESTATIONAL EDS EXPOSURE ALTERS MOUSE REPRODUCTION 30. Rockett JC, Mapp FL, Garges JB, Luft JC, Mori C, Dix DJ. Effects of hyperthermia on spermatogenesis, apoptosis, gene expression, and fertility in adult male mice. Biol Reprod 2001; 65:229–239. 31. Pudney J. Comparative cytology of the Leydig cell. In: Payne AH, Hardy MP, Russell LD (eds.), The Leydig Cell. Vienna: Cache River Press; 1996: 97–142. 32. Kaler LW, Neaves WB. Attrition of the human Leydig cell population with advancing age. Anat Rec 1978; 192:513–518. 33. Holmes M, Suarez JD, Roberts NL, Mole ML, Murr AS, Klinefelter GR. Dibromoacetic acid, a prevalent by-product of drinking water disinfection, compromises the synthesis of specific seminiferous tubule proteins following both in vivo and in vitro exposures. J Androl 2001; 22:878–890. 34. Klinefelter GR, Roberts NL, Suarez JD. Direct effects of ethane dimethanesulphonate on epididymal function in adult rats. An in vitro demonstration. J Androl 1992; 13:409–421. 35. Klinefelter GR, Laskey JW, Perreault SD, Ferrell J, Jeffay S, Suarez J, Roberts N. The ethane dimethanesulfonate-induced decrease in the fertilizing ability of cauda epididymal sperm is independent of the testis. J Androl 1994; 15:318–327. 36. Klinefelter GR, Strader LF, Suarez JD, Roberts NL. Bromochloroacetic acid exerts qualitative effects on rat sperm: implications for a novel biomarker. Toxicol Sci 2002; 68:164–173. 37. Gallavan RH Jr, Holson JF, Stump DG, Knapp JF, Reynolds VL. Interpreting the toxicologic significance of alterations in anogenital distance: potential for confounding effects of progeny body weights. Reprod Toxicol 1999; 13:383–390. 38. Theobald HM, Peterson RE. In utero and lactational exposure to 2,3,7,8-tetrachlorodibenzo-rho-dioxin: effects on development of the

39. 40. 41. 42.

43.

44.

45.

46.

967

male and female reproductive system of the mouse. Toxicol Appl Pharmacol 1997; 145:124–135. Pointis G, Latreille MT, Mignot TM, Janssens Y, Cedard L. Regulation of testosterone synthesis in the fetal mouse testis. J Steroid Biochem 1979; 11:1609–1612. Aoki A, Fawcett DW. Is there a local feedback from the seminiferous tubules affecting activity of the Leydig cells? Biol Reprod 1978; 19: 144–158. De Kretser DM, Kerr JB. The effect of testicular damage on Sertoli and Leydig cell function. Monogr Endocrinol 1983; 25:133–154. Gray LE Jr, Ostby J, Sigmon R, Ferrell J, Rehnberg G, Linder R, Cooper R, Goldman J, Laskey J. The development of a protocol to assess reproductive effects of toxicants in the rat. Reprod Toxicol 1988; 2:281–287. Plump AS, Erickson SK, Weng W, Partin JS, Breslow JL, Williams DL. Apolipoprotein A-I is required for cholesteryl ester accumulation in steroidogenic cells and for normal adrenal steroid production. J Clin Invest 1996; 97:2660–2671. Frayne J, McMillen A, Love S, Hall L. Expression of phosphatidylethanolamine-binding protein in the male reproductive tract: immunolocalisation and expression in prepubertal and adult rat testes and epididymides. Mol Reprod Dev 1998; 49:454–460. Lee KK, Murakawa M, Takahashi S, Tsubuki S, Kawashima S, Sakamaki K, Yonehara S. Purification, molecular cloning, and characterization of TRP32, a novel thioredoxin-related mammalian protein of 32 kDa. J Biol Chem 1998; 273:19160–19166. Klinefelter GR, Laskey JW, Roberts NL. In vitro/in vivo effects of ethane dimethanesulfonate on Leydig cells of adult rats. Toxicol Appl Pharmacol 1991; 107:460–471.