TOXICOLOGY AND APPLIED PHARMACOLOGY 142, 192–200 (1997)ARTICLE NO. TO967966

Vinclozolin and p,p*-DDE Alter Androgen-Dependent Gene Expression:In Vivo Confirmation of an Androgen Receptor-Mediated Mechanism1

WILLIAM R. KELCE,*,2 CHRISTY R. LAMBRIGHT,* L. EARL GRAY, JR.,* AND KENNETH P. ROBERTS†

*Reproductive Toxicology Division, Endocrinology Branch, National Health and Environmental Effects Research Laboratory,U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711; and †Department of

Urologic Surgery, University of Minnesota Medical School, Minneapolis, Minnesota 55455

Received April 25, 1996; accepted July 9, 1996

The study of environmental endocrine disruptors is a high-Vinclozolin and p,p*-DDE Alter Androgen-Dependent Gene Ex- visibility issue as these chemicals have the potential to ad-

pression: In Vivo Confirmation of an Androgen Receptor Mediated versely impact human reproductive health and because ad-Mechanism. KELCE, W. R., LAMBRIGHT, C. R., GRAY, L. E., JR.,

verse reproductive effects currently are being observed inAND ROBERTS, K. P. (1996). Toxicol. Appl. Pharmacol. 142,

wildlife populations such as the demasculinization of male192–200.alligators in Florida’s DDT-contaminated Lake Apopka (Gu-

Vinclozolin and p,p*-DDE induce antiandrogenic developmental illette et al., 1994). While much attention has focused oneffects in vivo and are potent inhibitors of androgen receptor (AR) the effects of environmental estrogens, we recently havebinding and AR-dependent gene expression in vitro. To determine described two environmental chemicals, vinclozolin andwhether this molecular mechanism is operative in vivo, the effects

p,p*-DDE, whose effects in vivo are likely mediated by inhi-of these compounds on two androgen-regulated prostatic mRNAsbition of androgen receptor action.were studied. Rats were sham operated or castrated and immedi-

Vinclozolin is a systemic dicarboximide fungicide usedately implanted with one or two empty 2.5-cm silastic capsules orin the United States and throughout Europe for the controlwith one (11) or two (21) 2.5-cm capsules containing testosteroneof fungal disease in grapes, fruits, vegetables, hops, orna-(T). T-implanted rats were treated by gavage for 4 days with vehiclemental plants, and turf (Rankin et al., 1989). Multigenera-(corn oil), vinclozolin (200 mg/kg/day), p,p*-DDE (200 mg/kg/day),

or the antiandrogen flutamide (100 mg/kg/day) as a positive control. tional studies suggest that developmental exposure to vin-Vinclozolin, p,p*-DDE, and flutamide all induced a reciprocal de- clozolin demasculinizes male rat pups (van Ravenzwaay,cline in seminal vesicle (p õ 0.01) and prostate (p õ 0.01) weight 1992), a finding which was later confirmed and extendedas well as a reduction in immunohistochemical staining of AR in in our laboratories (Gray et al., 1994). In vitro studies ofepididymal nuclei compared to vehicle-treated T-implanted con- the biochemical and molecular mechanism responsible fortrols. Specific AR antagonism was assessed by determining the abil- these adverse developmental effects were consistent withity of these chemicals to induce a testosterone-repressed prostatic

an inhibition of AR binding (Kelce et al., 1994a) andmessage (i.e., TRPM-2) and/or repress a testosterone-induced pros-subsequent inhibition of AR-dependent transcriptional ac-tatic message (i.e., prostatein subunit C3). Densitometry scans oftivation (Wong et al., 1995). In a similar series of experi-Northern blots indicated that vinclozolin, p,p*-DDE, and flutamidements, the persistent environmental contaminant p,p*-each induced TRPM-2 mRNA and repressed C3 mRNA comparedDDE, a metabolite of DDT, was found to exhibit littleto vehicle-treated T-implanted controls. These antiandrogenic ef-ability to bind the estrogen receptor (ER), but was a potentfects were competitively reduced in castrate rats implanted with

two 2.5-cm T capsules (21), where serum T levels were elevated inhibitor of androgen binding to the AR and subsequentmore than twofold above physiological levels. Taken together, these AR-dependent gene expression (Kelce et al., 1995). Whiledata indicate that vinclozolin and p,p*-DDE act as antiandrogens these studies provided compelling evidence that vinclo-in vivo by altering the expression of androgen-dependent genes. zolin and p,p*-DDE act to inhibit the expression of andro-q 1997 Academic Press gen-regulated genes in vitro, they did not address whether

alterations in AR-regulated gene expression also are mani-fest in vivo. Hence, the objective of the present study wasto bring these in vitro molecular findings back to the level1 This manuscript has been reviewed in accordance with the policy of

the National Health and Environmental Effects Research Laboratory, U.S. of the whole animal and determine whether these chemi-Environmental Protection Agency, and approved for publication. Approval cals alter steady-state levels of specific AR-regulated genedoes not signify that the contents necessarily reflect the views and policies transcripts in vivo. To this end, the ability of vinclozolinof the Agency, nor does mention of trade names or commercial products

and p,p*-DDE to repress a testosterone-induced prostaticconstitute endorsement or recommendation for use.message (i.e., prostatein C3) and/or induce a testosterone-2 To whom correspondence should be addressed. Fax: (919) 541-5138:

E-mail: Kelce@her145.herl.epa.gov. repressed prostatic message (i.e., TRPM-2) was assessed.

1920041-008X/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID TOX 7966 / 6h14h$$521 12-10-96 09:26:58 toxal AP: Tox

193VINCLOZOLIN AND p,p*-DDE ALTER GENE EXPRESSION

Prostate specific binding protein (i.e., prostatein) is the MATERIALS AND METHODSmajor androgen-dependent secretory protein of the rat

Chemicals and supplies. Radioinert testosterone (T) was purchasedventral prostate and was used in this study to examine thefrom Steraloids (Wilmington, NH) and successively recrystallized fromeffects of antiandrogen exposure on steady-state mRNAaqueous ethanol to constant melting point prior to use. Ethanol was pur-

levels encoding an androgen-induced protein. Prostatein chased from Midwest Grain Products (Weston, MO). Vinclozolin (analyticalis a tetrameric glycoprotein composed of three different grade, ú99% purity) was purchased from Crescent Chemical Co. (Haup-

pauge, NY). p,p*-DDE was purchased from Aldrich Chemical Co. (Milwau-polypeptides (C1, C2, and C3) arranged as two subunitskee, WI). Flutamide was purchased from Sigma Chemical Co. (St. Louis,(C1C3 and C2C3) (Heyns et al., 1978). The C3 polypep-MO). The purity of all treatment solutions was greater than 99% as assessedtide is encoded by a discrete mRNA (Parker et al., 1980, by reversed-phase HPLC with on-line uv absorbance detection (Kelce et

1982), the expression of which is regulated by androgen al., 1995). The Riboprobe Gemini System, IPTG, X-Gal, calf intestinalat the level of transcription and mRNA stability (Page and alkaline phosphatase, and restriction endonucleases were purchased from

Promega (Madison, WI) and used as supplied. SeaKem agarose was pur-Parker, 1982). AR-induced expression of the C3 gene ischased from FMC Bioproducts (Rockland, ME). Library efficiency (compe-reduced following castration (Page and Parker, 1982; Par-tant) DH5a cells, formamide, phenol, cesium chloride, ethidium bromide,

ker et al., 1978) and following in vivo exposure to antian- and 201 SSC buffer were obtained from Gibco-BRL (Grand Island, NY).drogenic pharmaceuticals such as cyproterone acetate and Magna nylon transfer membrane was obtained from Micron Separations

Inc. (Westboro, MA). All other chemicals and reagents were purchasedflutamide (Leger et al., 1988; Guenette et al., 1994). Whilefrom Sigma Chemical Co., Gibco-BRL, Promega, or Difco Laboratoriesthere are a number of potential regulatory sequences(Detroit, MI).within the C3 gene, specific AR–DNA binding and subse-

Animals, dosing, and necropsy procedures. Adult male Sprague–Daw-quent transcriptional activation appear to take place at anley rats (120 days of age) were purchased from Harlan–Sprague Dawley,

active androgen-responsive element (ARE) within a 0.5- Inc. (Indianapolis, IN). The rats were housed two per cage with laboratory-kb region of the first intron (Tan et al., 1992). In this grade pine shavings as bedding. Rats were maintained under controlled

temperature (227C), humidity (40–50%), and light (14L:10D) conditionsregard, we recently have demonstrated that vinclozolinand were given Purina Laboratory Rat Chow (No. 5001) and water adand its active metabolites, M1 and M2, inhibit AR bindinglibitum. Rats were weighed, weight ranked, and randomly assigned to treat-to this ARE sequence, an effect likely explaining the abil-ment groups in a manner that provided each treatment with equal means

ity of these chemicals to inhibit transcriptional activation and variances in body weight.of AR-dependent genes in vitro (Wong et al., 1995). Prior to treatment, rats were gavaged daily for 3 days with 2.5 ml of

water/g body wt to acclimate them to handling and the gavage procedure.Testosterone-repressed prostatic message-2 (TRPM-2, alsoIn the first study, rats (n Å 24) were sham operated (SHAM; n Å 4) orreferred to as clusterin or sulfated glycoprotein-2) is amongcastrated (n Å 20) and castrate rats were implanted with either one emptya group of proteins that are involved in the active involution 2.5-cm silastic capsule (CASTRATE; n Å 4) or one 2.5-cm capsule con-

of the prostate (i.e., apoptosis) following androgen withdrawal taining T (n Å 16). Castrate T-implanted rats were dosed daily by gavageand was used in these studies to examine the effects of antian- for 5 days with vehicle (2.5 ml of corn oil/g body wt; CASTRATE / T; n Å

4), 200 mg/kg/day p,p*-DDE (p,p*-DDE; nÅ 4), 200 mg/kg/day vinclozolindrogen exposure on steady-state mRNA levels encoding an(VINCLOZOLIN; n Å 4), or 100 mg/kg/day of the positive control flutam-androgen-repressed protein. TRPM-2 is a heterodimer com-ide (FLUTAMIDE; n Å 4). In the second study, rats (n Å 36) were sham

posed of two disulfide-linked subunits, a and b (35–40 kDa), operated (n Å 6) or castrated (n Å 30) and castrate rats were implantedcontaining 20–30% highly sulfated carbohydrate moieties with either two empty 2.5-cm silastic capsules (n Å 6) or two 2.5-cm

capsules containing T (n Å 24). Castrate T-implanted rats (n Å 6/group)(Jenne and Tschopp, 1992). TRPM-2 is the translation productwere dosed daily by gavage for 5 days with vehicle, p,p*-DDE, vinclozolin,of a single copy gene, the expression of which is associatedor the positive control flutamide exactly as described above. The doses ofwith the onset of cellular atrophy and death in several rodentvinclozolin and p,p*-DDE used in these studies are high-dose exposures

tissues including the regressing rat ventral prostate following that previously have been determined to be antiandrogenic in vivo (Grayandrogen withdrawl induced by castration and/or pharmaco- et al., 1994; Kelce et al., 1995). Castrate T-implanted rats were used in

these studies to maintain or clamp serum T levels at 11 or 21 physiologicallogical treatment with antiandrogens (Leger et al., 1988; But-levels in the presence of endocrine-disrupting antiandrogens. This 5-daytyan et al., 1989; Wong et al., 1993; Guenette et al., 1994).treatment was chosen based on previous studies demonstrating that optimalThere are a number of potential regulatory sequences within alterations in both TRPM-2 and C3 mRNA levels occur 4–6 days following

the TRPM-2 gene, including four ARE half-sites, raising the administration of the antiandrogenic pharmaceuticals cyproterone acetateand flutamide (Leger et al., 1988; Guenette et al., 1994).possibility that these latter regions alone or in combination are

Following the 5-day treatment, rats were necropsied and seminal vesicle,important in androgen-induced repression of TRPM-2 geneventral prostate, and liver weights were recorded. Blood was collectedexpression (Wong et al., 1993).for the determination of serum T and luteinizing hormone (LH) levels by

The results reported here indicate that vinclozolin and radioimmunoassay. The ventral prostates were immediately homogenizedp,p*-DDE act as antiandrogens in vivo by altering the expres- (10 ml/g tissue) in RNA homogenization buffer containing 4 M guanidinium

isothiocyanate, 25 mM sodium citrate, 0.5% sarcosyl, and 100 mM b-mer-sion of specific AR-regulated genes. These studies are criti-captoethanol and immediately frozen in liquid nitrogen for subsequent RNAcal to the final assessment of endocrine disruptor activity asisolation. Caput epididymides were harvested, cut into 1- to 2-mm tissuethey clearly establish that in vivo exposure to environmental cubes, immediately placed in OCT compound (Miles Laboratories, Elkhart,

antiandrogens inhibits AR-regulated cellular processes, in- IN), and frozen in isopentane precooled in liquid nitrogen for subsequentimmunohistochemical localization of AR protein.cluding AR-regulated gene expression, as is the case in vitro.

AID TOX 7966 / 6h14h$$522 12-10-96 09:26:58 toxal AP: Tox

194 KELCE ET AL.

Radioimmunoassay procedures. Testosterone concentrations were de-termined in 100-ml serum aliquots as previously described (Kelce et al.,1991). The percentage cross-reactivity of p,p*-DDE, flutamide, hydroxyflu-tamide, and vinclozolin and its major serum metabolites M1 and M2 (Kelceet al., 1994) at 50% bound was less than 0.1% in all cases. LH concentra-tions were determined in 200-ml serum aliquots using the following materi-als supplied by the National Hormone and Pituitary Agency: iodinationpreparation I-9, reference preparation RP-3, and antisera S-10 as previouslydescribed (Stoker et al., 1993). Iodination material was radiolabeled with125I (New England Nuclear, Boston, MA) using chloramine-T.

Immunohistochemistry. Nine-micrometer frozen sections of the caputepididymides were air-dried for 3–5 min on Fisherbrand Plus slides, fixedfor 10 min in 4% paraformaldehyde containing 10% sucrose, rinsed for 10min in pH 7.2 phosphate buffer, and blocked for 10 min with 5% normalgoat serum. Sections were incubated overnight (47C) with polyclonal rabbitanti-AR antibody (AR52; Dr. Elizabeth M. Wilson, University of NorthCarolina, Chapel Hill, NC) at 5 mg/ml in phosphate buffer containing 1%bovine serum albumin. After a 10-min wash in phosphate buffer, the slideswere treated for 1 hr with biotinylated goat anti-rabbit IgG secondary anti-body (Vector Laboratories, Burlingame, CA) followed by avidin–biotinperoxidase (ABC kit, Vector Laboratories) for an additional 1 hr as de-scribed previously (Sar et al., 1990). Slides were stained in 0.01 M Trisbuffer, pH 7.6, using diaminobenzadine (DAB) in 3% hydrogen peroxide,dehydrated in an ascending ethanol series, and coverslipped using Permount.Each slide contained one section that received no primary antibody to assessspecificity of the immunostaining procedure. Preabsorption of the primaryAR52 antibody with the uncoupled synthetic AR peptide used to generatethe antibody response reduced nuclear AR immunostaining to backgroundlevels. FIG. 1. Serum concentrations of testosterone (T) (top) and LH (bottom)

in sham-operated rats, castrate rats, and castrate rats implanted with oneRNA isolation and Northern blotting. Total RNA was isolated from2.5-cm T capsule and treated daily for 5 days with vehicle (castrate / T),the ventral prostate homogenates as per the method of Chomczynski andp,p*-DDE, vinclozolin, or the positive control flutamide (mean { SD, n ÅSacchi (1987). Aliquots (10 mg) of total RNA were electrophoresed through4). **p õ 0.01 vs castrate / T.a denaturing 1.2% agarose gel and subsequently blotted by capillary action

to a nylon filter (Sambrook et al., 1989). Loading efficiency was confirmedby ethidium bromide staining of the 18s and 28s ribosomal RNA bandsand was found to be equivalent in all lanes. Filters were hybridized (12 hr,

(i.e., castrate T-implanted rats receiving vehicle only) when the analysis of607C) with 32P-labeled cRNA probes, prepared from linearized plasmidsvariance first rejected the null hypothesis of equal means (Miller, 1981).(pGEM 4Z; Promega, Madison, WI). The cDNA for rat SGP-2 (i.e., TRPM-The level of significance was set at p õ 0.05.2), cloned into pSP64 (Collard and Griswold, 1987), was subcloned into

the HindIII/EcoRI restriction site of PGEM 4Z. The plasmid was linearizedwith HindIII and transcribed with SP6 RNA polymerase to yield a 829-bp RESULTScRNA probe. The cDNA for rat C3, originally cloned into the PstI restrictionsite of pBR322 (obtained from Dr. Glen Cunningham, Baylor College of

Serum T and LH. Testosterone levels in castrate adultMedicine, Houston, TX), was subcloned into the PstI site of PGEM 4Z.male rats implanted with a single 2.5-cm T capsule (Study 1)The plasmid was linearized with EcoRI and transcribed with T7 RNAwere not significantly different from those in sham-operatedpolymerase to yield a 400-bp cRNA probe. Incubation with 32P-labeled

sense TRPM-2 or C3 cRNA resulted in no detectable hybridization. controls (Fig. 1, top), while castrate rats implanted with twoFollowing low- and high-stringency washes, the blots were sealed in 2.5-cm silastic T capsules (Study 2) exhibited serum T levels

plastic wrap, exposed to a phosphor screen for 2 hr, and scanned for radioac-that were approximately twice those of sham-operated con-tivity emission using a Phosphorimager 400S (Molecular Dynamics, Sun-trols (Fig. 2, top; 6.16 ng T/ml vs 2.56 ng T/ml, respectively).nyvale, CA). ImageQuant software version 3.3 was used to backgroundThe fact that serum T levels in castrate T-implanted ratscorrect and quantify TRPM-2 and C3 band densities. For comparative pur-

pose, the blots were exposed to Kodak X-OMAT AR film overnight and were not different following treatment with vehicle, p,p*-the TRPM-2 and C3 bands were background corrected and scanned for DDE, vinclozolin, or the positive control flutamide indicatesoptical density using a Scanalytics image analyzer (Billerica, MA). Quanti-

that none of these chemicals increased T metabolism duringtations of relative band intensities within a single blot were virtually identi-these 5-day studies. Castration, on the other hand, reducedcal independent of the method chosen for analysis. Steady-state levels ofserum T levels to less than 0.4 ng T/ml, thus providing aTRPM-2 and C3 mRNA measured by Northern blot reflect the in situ

levels of mRNA resulting from alterations in the rate of transcription and/ good control for the effects of androgen withdrawal in bothor alterations in transcript stability, both of which are androgen-dependent studies.processes in the rat ventral prostate (Page and Parker, 1982; Leger et al.,

Castrate or castrate T-implanted rats treated with vinclo-1988).zolin or flutamide, but not p,p*-DDE, exhibited increasedStatistics. One-way analysis of variance was used to compare meansserum LH levels compared to the castrate T-implanted con-among treatment groups (Miller, 1981). Dunnett’s many-one t test was used

to compare differences between treatment groups and the single control trols (Fig. 1, bottom). The ability of vinclozolin and flutam-

AID TOX 7966 / 6h14h$$522 12-10-96 09:26:58 toxal AP: Tox

195VINCLOZOLIN AND p,p*-DDE ALTER GENE EXPRESSION

likely reflecting the reduced ability of these chemicals tocompete with higher levels of endogenous T for binding tothe AR. Flutamide was the most potent inhibitor in thisregard, as ventral prostate and seminal vesicle weights werereduced in both studies to near castrate levels.

AR immunohistochemistry. The effects of the varioustreatments on the immunohistochemical localization of theAR in epididymal epithelial cells are illustrated in Fig. 5.The specificity of the AR anti-peptide antibody in the immu-nohistochemical reaction was demonstrated by treating adja-cent tissue sections with nonimmune serum or with immuneIgG adsorbed by the uncoupled anti-peptide used to generatethe immune response; positive AR staining was not detect-able in either case (not shown). In addition, a control tissue(i.e., spleen), which does not express AR (Sar et al., 1990),lacked positive AR immunostaining (not shown). Positiveimmunostaining for AR was readily detectable in epididymalepithelial cells as illustrated in Fig. 5. The AR was primarilyimmunolocalized to the nuclei of epididymal epithelial cellsfrom sham control rats (Fig. 5A), an observation which couldbe markedly reduced by castration (Fig. 5B) and restoredby 21 androgen replacement (Fig. 5C). The intense immuno-staining of epididymal nuclei from castrate T-implanted con-trol rats (Fig. 5C) likely reflects the elevated levels of serumFIG. 2. Serum concentrations of testosterone (T) (top) and LH (bottom)

in sham-operated rats, castrate rats, and castrate rats implanted with two2.5-cm T capsules and treated daily for 5 days with vehicle (castrate / T),p,p*-DDE, vinclozolin, or the positive control flutamide (mean { SD, n Å6). **p õ 0.01 vs castrate / T.

ide to increase serum LH levels in castrate T-implanted ratswas less pronounced in the second study, where castrate ratsreceived two 2.5-cm T implants (Fig. 2, bottom). The ele-vated levels of LH in the sham-operated rats compared tocastrate T-implanted controls likely reflect the fact that se-rum T levels were less than half those in the T-implantedcontrol group. As in the first study, p,p*-DDE failed to alterserum LH levels compared to those in castrate T-implantedrats treated with vehicle only (Fig. 2, bottom).

Androgen-dependent organ weights. The weights of theventral prostate and seminal vesicles in castrate rats im-planted with a single 2.5-cm T capsule were equivalent tothose obtained from sham-operated controls (Fig. 3), un-doubtedly reflecting the fact that serum T levels were equiva-lent among these two groups. Castration or administrationof p,p*-DDE, vinclozolin, or flutamide to castrate T-im-planted rats resulted in a significant decline (p õ 0.01) inventral prostate and seminal vesicle weights compared tothose for the castrate T-implanted controls (Fig. 3). Despiteincreased serum T levels in Study 2, p,p*-DDE, vinclozolin,and the positive control flutamide all reduced ventral prostate

FIG. 3. Ventral prostate (top) and seminal vesicle (bottom) weights inand seminal weights compared to the castrate T-implantedsham-operated rats, castrate rats, and castrate rats implanted with one 2.5-

controls (p õ 0.01; Fig. 4). The magnitude of the effects of cm T capsule and treated daily for 5 days with vehicle (castrate / T), p,p*-p,p*-DDE and vinclozolin on ventral prostate and seminal DDE, vinclozolin, or the positive control flutamide (mean { SD, n Å 4).

**p õ 0.01 vs castrate / T.vesicle weight was less pronounced in the second study,

AID TOX 7966 / 6h14h$$523 12-10-96 09:26:58 toxal AP: Tox

196 KELCE ET AL.

illustrated for Study 2 in Figs. 8 and 9. The Northern blotsin Fig. 8 are representative of three of the six animals pertreatment group, while the optical density scans of theseNorthern blots (Fig. 9) are representative of all six animalsper treatment group. Castrate T-implanted rats treated withp,p*-DDE and vinclozolin exhibited a significant increase inTRPM-2 mRNA levels (p õ 0.05) and a significant declinein C3 mRNA levels (p õ 0.05). As anticipated, the magni-tude of the alterations induced by vinclozolin and p,p*-DDEwas less pronounced than in the first study, likely reflectingthe reduced ability of these chemicals to compete with higherlevels of endogenous T for binding to the AR.

DISCUSSION

Previously, we demonstrated that vinclozolin and p,p*-DDE are potent environmental antiandrogens acting to com-petitively inhibit AR androgen binding (Kelce et al., 1994a,1995) and subsequent expression of AR-dependent genes invitro (Wong et al., 1995; Kelce et al., 1995). Here, we extendthese studies to the level of the whole animal to demonstratethat vinclozolin and p,p*-DDE inhibit expression of two spe-cific androgen-regulated genes (i.e., TRPM-2 and C3) in therat ventral prostate following in vivo administration. These

FIG. 4. Ventral prostate (top) and seminal vesicle (bottom) weights instudies are critical in the final assessment of endocrine dis-sham-operated rats, castrate rats, and castrate rats implanted with two 2.5-ruptor activity as they indicate that this antiandrogenic mech-cm T capsules and treated daily for 5 days with vehicle (castrate / T),

p,p*-DDE, vinclozolin, or the positive control flutamide (mean { SD, n Å anism is operative in vivo and that sufficient levels of the6). **p õ 0.01 vs castrate / T. parent chemical and/or metabolites can be attained at the

level of the target tissue, and specifically at the level of theAR, for a sufficient time to alter AR-regulated processes.

T which bind AR and induce nuclear AR import and AR– While the assessment of TRPM-2 and C3 mRNA levels hasDNA binding, resulting in stabilization of the AR against been used for years to study the process of apoptosis in thecytoplasmic degradation. In contrast, exposure of castrate T- rat ventral prostate following androgen withdrawal (Parkerimplanted rats to p,p*-DDE (Fig. 5D), vinclozolin (Fig. 5E), et al., 1980; Page and Parker, 1982; Bettuzzi et al., 1992;or flutamide (Fig. 5F) for 5 days markedly reduced nuclear Celis et al., 1993; Wong et al., 1993) and following in vivoAR immunostaining compared to the castrate T-implanted administration of antiandrogenic pharmaceuticals (Leger etcontrol rats. al., 1988; Guenette et al., 1994), the exploitation of these

androgen-regulated mRNAs to study in vivo toxicologicalSteady-state levels of TRPM-2 and C3. Castrate rats im-planted with a single T capsule and treated with p,p*-DDE, mechanisms of environmental antiandrogens is new.

It is clear from the Northern analyses that antiandrogenvinclozolin, or the positive control flutamide exhibited asignificant increase in TRPM-2 mRNA levels (Figs. 6 and administration to castrate T-implanted rats is less effective

than castration at increasing steady-state levels of TRPM-27; p õ 0.01) and a significant decline in C3 mRNA levels(Figs. 6 and 7; p õ 0.01) compared to castrate T-implanted mRNA and reducing steady-state levels of C3 mRNA. These

disparate effects have been reported by others (Leger et al.,controls. It is clear from these results that androgen with-drawal via castration induces the expression of the androgen- 1988; Guennette et al., 1994) and are likely due to incom-

plete inhibition of the agonist effects of endogenous hor-repressed TRPM-2 message and reduces the expression ofthe androgen-induced C3 message relative to the castrate T- mone. However, TRPM-2 and C3 gene expressions are an-

drogen-dependent processes in the rat ventral prostate andimplanted controls. Note that the C3 Northern blot in Fig. 6was deliberately overexposed to illustrate that antiandrogen their regulation likely involves a number of other signaling

pathways, as evidenced by the complex transcriptional con-treatment did not totally eliminate C3 mRNA in the rat ven-tral prostate, but the levels were dramatically reduced com- trol sequences in the promoter regions of both TRPM-2

(Wong et al., 1993) and C3 (Tan et al., 1992; Celis et al.,pared to those from castrate T-implanted controls.The effect of antiandrogen exposure on steady-state levels 1993) genes. As these multiple signaling pathways likely

act in concert with androgens to regulate ventral prostateof TRPM-2 and C3 mRNA in the rat ventral prostate is

AID TOX 7966 / 6h14h$$523 12-10-96 09:26:58 toxal AP: Tox

197VINCLOZOLIN AND p,p*-DDE ALTER GENE EXPRESSION

FIG. 5. Immunohistochemical localization of AR in epididymides obtained from sham-operated rats (A), castrate rats (B), and castrate rats implantedwith two 2.5-cm T capsules and treated daily for 5 days with vehicle (C), p,p*-DDE (D), vinclozolin (E), or the positive control flutamide (F). Theintense staining of epididymal nuclei in castrate T-implanted rats treated with vehicle (C) compared to sham-operated rats (A) likely reflects the increasedserum T levels in these rats. It is clear that castration and treatment of castrate T-implanted rats with p,p*-DDE, vinclozolin, and flutamide markedlyreduce AR nuclear staining compared to the castrate T-implanted controls. Sections, 9 mm; magnification, 120.

functions, we cannot exclude the possibility that antiandro- pared to castrate T-implanted control rats, is in agreementwith biochemical studies demonstrating that 3H-R1881 bind-gens simply do not block all these other signaling pathways

to the same degree as androgen withdrawl via castration. ing to high-salt nuclear extracts also is reduced (i.e., approxi-mately 50%) 4 hr following in vivo exposure of adult maleThe ability of p,p*-DDE, vinclozolin, and flutamide to

inhibit immunostaining of the AR in epididymal nuclei, com- rats to 400 mg vinclozolin/kg body wt (Kelce et al., 1994b).

AID TOX 7966 / 6h14h$$523 12-10-96 09:26:58 toxal AP: Tox

198 KELCE ET AL.

of TRPM-2 mRNA levels, and induction of C3 mRNAlevels is consistent with competition at the level of theAR. p,p*-DDE, vinclozolin, and flutamide clearly act toinhibit AR-dependent processes in vivo, as predicted fromour in vitro studies.

Although p,p*-DDE, vinclozolin, and flutamide induced a20 to 35% increase in liver weight (data not shown), serumT levels in these animals were not different from castrateT-implanted rats treated with vehicle only. In this regard,p,p*-DDE has been reported to increase sterol hydoxylaseactivity resulting in increased T metabolism (Kupfer, 1975),and other antiandrogens, such as cyproterone acetate (Rob-erts et al., 1995) and Zanterone (Descotes et al., 1996),have been reported to increase liver weight, presumably via

FIG. 6. Steady-state levels of TRPM-2 and C3 mRNA in 10 mg of total increased mitogenic activity and/or inhibition of the normalRNA extracted from sham-operated rats, castrate rats, and castrate rats rate of hepatocyte apoptosis (Roberts et al., 1995). The pres-implanted with one 2.5-cm T capsule and treated daily for 5 days with ent results suggest that despite the increase in liver mass,vehicle (castrate / T), p,p*-DDE, vinclozolin, or the positive control flu-

p,p*-DDE, vinclozolin, and flutamide do not increase thetamide. Northern blots were hybridized with 32P-labeled cRNA probes spe-ability of the liver to metabolize T, at least during the coursecific for the 1.9-kb TRPM-2 (top) and the 0.5-kb C3 (bottom) transcripts.

Each lane is representative of mRNA extracted from a different rat; data of this 5-day study. The mechanism responsible for hepaticshown are scanned images from a Molecular Dynamics Phosphoimager and enlargement remains to be definitively identified; however,are representative of three of the four rats per treatment group in this study. it is clear from the present data that hepatic enlargement

does not always imply induction of hepatic monooxygenasesinvolved in T metabolism.

The reduction in nuclear immunostaining probably is notrelated to alterations in nuclear AR import as these chemicalsbind AR and are efficiently imported into the nucleus ofcells transiently transfected with the human AR (Wong etal., 1995). It seems possible that the decline in nuclear ARlocalization in these studies is due to increased cytoplasmicdegradation of the antiandrogen-bound AR, reflecting thefact that antiandrogens do not promote stable AR–DNAbinding (Wong et al., 1995). In this regard, other investiga-tors have demonstrated that the AR is rapidly degraded in theabsence of hormone and following antiandrogen exposure(Kemppainen et al., 1992; Zhou et al., 1995). We cannotexclude the possibility, however, that the antiandrogen-bound AR is more easily displaced during tissue preparationand/or fixation or that the loss of AR immunostaining follow-ing treatment with antiandrogens is due to reduced AR syn-thesis or to an altered AR conformation that prevents anti-body binding to its epitope.

Previous in vitro studies have demonstrated that p,p*-DDE and vinclozolin metabolites M1 and M2 are purecompetitive antagonists of AR androgen binding (Kelceet al., 1994a, 1995); once bound to the AR, these antian-drogenic chemicals are imported into the nucleus wherethey inhibit transcription of androgen-regulated genes(Kelce et al., 1995; Wong et al., 1995) by inhibiting theability of the antiandrogen-bound AR to bind androgen

FIG. 7. Optical densitometry scans of the Northern blots in Fig. 6.response element DNA (Wong et al., 1995). The abilityNorthern blots were exposed to a phosphor screen for 2 hr and scanned forof elevated (21) serum T concentrations to reduce theradioactivity emission using a Phosphorimager 400S (Molecular Dynamics).

magnitude of the antiandrogenic effects of p,p*-DDE, vin- ImageQuant software version 3.3 was used to background correct and quan-clozolin, and flutamide on AR-regulated processes such tify TRPM-2 and C3 band densities (mean { SD, n Å 4). **p õ 0.01 vs

castrate / T.as maintenance of accessory sex organ weight, repression

AID TOX 7966 / 6h14h$$523 12-10-96 09:26:58 toxal AP: Tox

199VINCLOZOLIN AND p,p*-DDE ALTER GENE EXPRESSION

Endocrine-disrupting chemicals often act through multipleendocrine pathways, emphasizing the need to demonstratethat the purported mechanism of action identified in vitroalso is operative in vivo. Just because an endocrine-disruptingchemical binds a steroid receptor in vitro does not mean thatany resulting reproductive toxicity is mediated through thatreceptor; an endocrine-disrupting chemical could bind the re-ceptor in vitro at concentrations that could never be attainedin vivo. These possibilities may be related to the inability ofp,p*-DDE to up-regulate serum LH levels via an endocrinefeedback mechanism. That is, the absence of an effect of p,p*-DDE on serum LH levels suggests that sufficient levels ofp,p*-DDE are not attained within the hypothalamo–pituitaryaxis, specifically at the level of the AR, to induce compensa-tory synthesis and/or release of LH. Alternatively, p,p*-DDEmay directly inhibit LH synthesis and/or release by a mecha-nism unrelated to its ability to bind AR. These target organ-specific responses emphasize the need to evaluate a numberof different in vivo end points for a comprehensive assessmentof endocrine disruptor activity.

Steroid hormone receptors control fundamental events inembryonic development and sex differentiation through theirfunction as ligand-inducible transcription factors that eitheractivate or repress transcription of target genes. The conse-

FIG. 9. Optical densitometry scans of the Northern blots in Fig. 8.quences of disrupting these processes can be especially pro- Northern blots were exposed to a phosphor screen for 2 hr and scanned forfound during development due to the crucial role hormones radioactivity emission using a Phosphorimager 400S (Molecular Dynamics).

ImageQuant software version 3.3 was used to background correct and quan-play in controlling transient and irreversible developmentaltify TRPM-2 and C3 band densities (mean { SD, n Å 6). *p õ 0.05 vscastrate / T; **p õ 0.01 vs castrate / T.

processes. To date our laboratories have focused on develop-mental effects of environmental endocrine disruptors actingto inhibit steroid hormone receptor action, as it is thesepermanent alterations that will most likely affect the riskassessment process. Rapid progress has been made in eluci-dating the in vitro biochemical and molecular mechanismsof several environmental antiandrogens (Kelce et al., 1994a,1995; Wong et al., 1995) and this study brings these resultsfull-circle to suggest that these mechanisms also are manifestin vivo. By exploiting the expression of two specific andro-gen-dependent genes (i.e., TRPM-2 and C3) in this castrateT-implanted rat model, antiandrogenic chemicals acting toinhibit T biosynthesis, increase T metabolism, or inhibit ARaction can be differentiated. We suggest that these two well-characterized, androgen-dependent, and oppositely regulated

FIG. 8. Steady-state levels of TRPM-2 and C3 mRNA in 10 mg of total transcripts in the rat ventral prostate are useful end pointsRNA extracted from sham-operated rats, castrate rats, and castrate rats to assess specific antiandrogenic cellular responses in vivo.implanted with two 2.5-cm T capsules and treated daily for 5 days withvehicle (castrate / T), p,p*-DDE, vinclozolin, or the positive control flu-

ACKNOWLEDGMENTStamide. Northern blots were hybridized with 32P-labeled cRNA probes spe-cific for the 1.9-kb TRPM-2 (top) or the 0.5-kb C3 (bottom) transcript.Each lane is representative of mRNA extracted from a different rat; data The authors thank Tammy Stoker for measurement of LH, Joe Ostby

and Ora Huey for assistance in the animal dosing procedures, Juan Suarezshown are scanned images from a Molecular Dynamics Phosphorimagerand are representative of three of the six rats per treatment group in the for assistance with the Northern blot image analysis, and John Tindall for

assistance with the initial Northern blot studies. The authors also greatfullystudy.

AID TOX 7966 / 6h14h$$523 12-10-96 09:26:58 toxal AP: Tox

200 KELCE ET AL.

acknowledge Drs. Glen R. Cunningham and Micheal D. Griswold for pro- Kemppainen, J. A., Lane, M. V., Sar, M., and Wilson, E. M. (1992). Andro-gen receptor phosphorylation, turnover, nuclear transport, and transcrip-viding the original C3 and TRPM-2 cDNAs, respectively.tional activation. J. Biol. Chem. 267, 968–974.

Kupfer, D. (1975). Effects of pesticides and related compounds on steroidREFERENCESmetabolism and function. CRC Crit. Rev. Toxicol. 4, 83–124.

Leger, J. G., Guellec, R. L., and Tenniswood, M. P. R. (1988). TreatmentBettuzzi, S., Zoli, M., Ferraguti, M. C., Ingletti, M. C., Agnati, L. F., andwith antiandrogens induces an androgen-repressed gene in the rat ventralCorti, A. (1992). Regional and cellular distribution within the rat prostateprostate. Prostate 13, 131–142.of two mRNA species undergoing opposite regulation by androgens. J.

Miller, R. G., Jr. (1981). Simultaneous Statistical Inference, 2nd ed.,Endocrinol. 132, 361–367.Springer Series in Statistics, Springer–Verlag, New York.Buttyan, R., Olsson, C. A., Pintar, J., Chang, C., Bandyk, M., Ng,

Page, M. J., and Parker, M. G. (1982). Effect of androgen on the transcrip-P.-Y., and Sawczuk, I. S., (1989). Induction of the TRPM-2 gene in cellstion of rat prostatic binding protein genes. Mol. Cell. Endocrinol. 27,undergoing programmed death. Mol. Cell. Biol. 9, 3473–3481.343–355.Celis, L., Claessens, F., Peeters, B., Heyns, W., Verhoeven, G., and Rom-

Parker, M., Needham, M., and White, R. (1982). Prostatic steroid bindingbauts, W. (1993). Proteins interacting with an androgen-responsive unitprotein: Gene duplication and steroid binding. Nature 298, 92–94.in the C3(1) gene intron. Mol. Cell.. Endocrinol. 94, 165–172.

Parker, M. G., White, R., and Williams, J. G. (1980). Cloning and character-Chomczynski, P., and Sacchi, N. (1987). Single-step method of RNA isola-ization of androgen-dependent mRNA from the rat ventral prostate. J.tion by acid guanidinium thiocyanate–phenol–chloroform extraction.Biol. Chem. 255, 6996–7001.Anal. Biochem. 162, 156–159.

Parker, M. G., Scrace, G. T., and Mainwaring, W. I. P. (1978). TestosteroneCollard, M. W., and Griswold, M. D. (1987). Biosynthesis and molecularregulates the synthesis of major proteins in rat ventral prostate. Biochem.cloning of sulfated glycoprotein 2 secreted by rat Sertoli cells. Biochemis-J. 170, 115–121.try 26, 3297–3303.

Rankin, G. O., Teets, V. J., Nicoll, D. W., and Brown, P. I. (1989). Compar-Descotes, G., Osman, H., Gallas, J.-F., Penacchio, E., and Moreau, C.

ative acute renal effects of three N-(3,5-dichlorophenyl)carboxide fungi-(1996). Functional, biochemical, and morphological hepatobiliary effects

cides: N-(3,5-Dichlorophenyl)-succinimide, vinclozolin and iprodione.in rats chronically exposed to a steroidal antiandrogen. Toxicol. Appl.

Toxicology 56, 263–272.Pharmacol. 137, 23–33.

Roberts, R. A., Soames, A. R., James, N. H., Gill, J. H., and Wheeldon,Gray, L. E., Ostby, J. M., and Kelce, W. R. (1994). Developmental effects

E. B. (1995). Dosing-induced stress causes hepatocyte apoptosis in ratsof an environmental antiandrogen: The fungicide vinclozolin alters sex

primed by the rodent nongenotoxic hepatocarcinogen cyproterone acetate.differentiation of the male rat. Toxicol. Appl. Pharmacol. 129, 46–52.

Toxicol. Appl. Phamacol. 135, 192–199.Guenette, R. S., Daehlin, L., Mooibroek, M., Wong, K., and Tenniswood, Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular Cloning—

M. (1994). Thanatogen expression during involution of the rat ventral A Laboratory Manual, Vol. 1, pp. 7.39–7.52, Cold Spring Harbor Labo-prostate after castration. J. Androl. 15, 200–211. ratory Press, Cold Spring Harbor, NY.

Guillette, L. J., Gross, T. S., Masson, G. R., Matter, J. M., Percival, H. F., Sar, M., Lubahn, D. B., French, F. S., and Wilson, E. M. (1990). Immuno-and Woodward, A. R., (1994). Developmental abnormalities of the gonad histochemical localization of the androgen receptor in rat and humanand abnormal sex hormone concentrations in juvenile alligators from tissues. Endocrinology 127, 3180–3186.contaminated and control lakes in Florida. Environ. Health Perspect. 102,

Stoker, T. E., Goldman, J. M., and Cooper, R. L. (1993). The dithiocarba-680–688.

mate fungicide thiram disrupts the hormonal control of ovulation in theHeyns, W., Peeters, B., Mous, J., Rombauts, W., and De Moor, P. (1978). female rat. Reprod. Toxicol. 7, 211–218.

Purification and characterization of prostatic binding protein and its sub-Tan, J., Marschke, K. B., Ho, K.-C., Perry, S. T., Wilson, E. M., and French,

units. Eur. J. Biochem. 89, 181–186.F. S. (1992). Response elements of the androgen-regulated C3 gene. J.

Jenne, D. E., and Tschopp, J. (1992). Clusterin: The intriguing guises of a Biol. Chem. 267, 4456–4466.widely expressed glycoprotein. Trends Biochem. Sci. 17, 154–159. van Ravenzwaay, B. (1992). Discussion of prenatal and reproduction toxic-

Kelce, W. R., Zirkin, B. R., and Ewing, L. L. (1991). Immature rat Leydig ity of Reg. No. 83-258 (Vinclozolin). Data submission to USEPA fromcells are intrinsically less sensitive than adult Leydig cells to ethane BASF Corporation, MRID 425813-02.dimethanesulfonate. Toxicol. Appl. Pharmacol. 111, 189–200. Wong, C.-I., Kelce, W. R., Sar, M., and Wilson, E. M. (1995). Androgen

Kelce, W. R., Monosson, E., Gamcsik, M. P., Laws, S. C., and Gray, L. E., receptor antagonist versus agonist activities of the fungicide vinclozolinJr. (1994a). Environmental hormone disruptors: Evidence that vinclozolin relative to hydroxyflutamide. J. Biol. Chem. 270, 19998–20003.developmental toxicity is mediated by antiandrogenic metabolites. Tox- Wong, P., Pineault, J., Lakins, J., Taillefer, D., Leger, J., Wang, C., andicol. Appl. Pharmacol. 126, 276–285. Tenniswood, M. (1993). Genomic organization and expression of the rat

Kelce, W. R., Monosson, E., and Gray, L. E., Jr. (1994b). In vitro/in vivo TRPM-2 (clusterin) gene, a gene implicated in apoptosis. J. Biol. Chem.evidence that vinclozolin (V) is an environmental antiandrogen. Biol. 268, 5021–5031.Reprod. 50(Suppl. 1), 189. Zhou, Z. X., Lane, M. V., Kemppainen, J. A., French, F. S., and Wilson,

E. M. (1995). Specificity of ligand-dependent androgen receptor stabiliza-Kelce, W. R., Stone, C. R., Laws, S. C., Gray, L. E., Jr., and Wilson, E. M.(1995). Persistent DDT metabolite p,p*-DDE is a potent androgen recep- tion: Receptor domain interactions influence ligand dissociation and re-

ceptor stability. Mol. Endocrinol. 9, 208–218.tor antagonist. Nature 375, 581–585.

AID TOX 7966 / 6h14h$$524 12-10-96 09:26:58 toxal AP: Tox