Reproductive Toxicology 17 (2003) 673–681

Effects of the mycotoxins�- and�-zearalenol on regulation ofprogesterone synthesis in cultured granulosa cells from porcine ovaries

U. Tiemanna,∗, W. Tomeka, F. Schneidera, J. Vanselowb

a Unit of Reproductive Biology, Research Institute for the Biology of Farm Animals, 18196 Dummerstorf, Wilhelm-Stahl-Allee 2, Germanyb Unit of Molecular Biology, Research Institute for the Biology of Farm Animals, 18196 Dummerstorf, Wilhelm-Stahl-Allee 2, Germany

Received 16 May 2003; received in revised form 3 July 2003; accepted 14 July 2003

Abstract

Mycotoxins as contaminants of animal food can impair fertility and can cause abnormal fetal development in farm animals. Therefore, thepresent study has investigated whether derivatives of the mycotoxin zearalenone,�-zearalenol (�-ZOL) and�-zearalenol (�-ZOL), influenceprogesterone synthesis via cytochrome P450 side chain cleavage enzyme (P450scc) and 3�-hydroxysteroid dehydrogenase/isomerase(3�-HSD) in cultured porcine granulosa cells. Both enzymes are essential for the conversion of cholesterol to progesterone. No differencesin basal progesterone levels and numbers of viable cell were observed between untreated granulosa cells and those treated with�- or�-ZOL (15 and 30�M). FSH (0.01�g/ml) or forskolin (10�M) enhanced the basal progesterone secretion in the absence of mycotoxins.The addition of�- or �-ZOL (7.5, 15 and 30�M) to cultures stimulated with FSH (0.01�g) or forskolin (10�M) reduced progesteronesynthesis and the levels of P450scc and 3�-HSD transcripts in a dose-dependent manner (P < 0.05). The enzymatic activity of 3�-HSD andthe abundance of P450scc protein were also reduced by these mycotoxins. In conclusion, effects of mycotoxins on FSH receptor-dependentand receptor-independent pathways indicate that adenylate cyclase activity and/or regulatory pathways further downstream are targets ofmycotoxin actions. The apparent dose-dependent reduction of P450scc and 3�-HSD transcripts implies an effect of�- and�-ZOL ontranscriptional regulation of these enzymes.© 2003 Elsevier Inc. All rights reserved.

Keywords: Zearalenol; Pig granulosa cells; Progesterone synthesis; Follicle-stimulating hormone; P450scc; 3�-HSD

1. Introduction

The mycotoxins�-zearalenol (�-ZOL) and�-zearalenol(�-ZOL) are derivatives of zearalenone that is synthesizedby fungalFusarium species. Both derivatives are mycotox-ins that frequently contaminate cereals and other plant prod-ucts[1]. Fusarium species are probably the most prevalenttoxin-producing fungi of the northern temperate regions andare commonly found in cereals grown in America, Europeand Asia.

Fusarium toxins have been shown to cause a variety oftoxic effects in experimental animals and livestock. Zear-alenone causes problems of the reproductive tract, impairedfertility, and abnormal fetal development in farm animals[2]. Pigs are quite sensitive to reproductive effects caused byhigh natural levels of zearalenone in their diet. Metabolismof zearalenone takes place in the liver where it is reducedto �- and�-ZOL. In experiments where we have tested the

∗ Corresponding author. Tel.:+49-38208-68-751;fax: +49-38208-68-752.E-mail address: tiemann@fbn-dummerstorf.de (U. Tiemann).

zearalenone derivatives,�- and�-ZOL, an examination ofexcretory products indicated a predominance of the� epimerin pig and man[3].

Most data on alterations in the female reproductive tractcaused by zearalenone exposure are based on in vivo in-vestigations[1]. In vivo assays provide information on neteffects in whole animals, where pollutants may act at multi-ple sites and organs, but are more difficult systems to studymechanisms at the molecular level. Therefore, cell and tis-sue cultures are of increasing importance to the evaluationof risks due to toxic compounds. The interaction of chemicalsubstances with the plasma membrane may alter membranestructure and function. Owing to the lack of toxicologicaldata for these mycotoxins, in vitro studies may contributeto risk assessments of toxins, which can impair fertility bychanging granulosa cell–oocyte interaction[4].

The purpose of the present study was to investigate theeffects of�- and�-ZOL on the regulation of progesteroneproduced by ovarian porcine granulosa cells cultured in aserum-free medium. Progesterone is an important ovarianhormone involved in preparing the reproductive tract forzygote implantation and the subsequent maintenance of the

0890-6238/$ – see front matter © 2003 Elsevier Inc. All rights reserved.doi:10.1016/j.reprotox.2003.07.001

674 U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681

pregnant stage[5]. The accumulation of progesterone resultsfrom an increased activity of the first enzyme of the steroido-genic pathway, the mitochondrial enzyme cytochrome P450cholesterol side-chain cleavage (P450scc), which catalyzesthe conversion of cholesterol to pregnenolone[6]. Thesubsequent conversion of pregnenolone to progesterone iscatalyzed by the microsomal enzyme 3�-hydroxysteroiddehydrogenase/isomerase (3�-HSD) [7]. FSH is of centralimportance as a stimulator of granulosa cell differentiation.Its actions are mediated by second-messenger signalingmechanisms through activation of adenylate cyclase[8].During the present study effects of�- and�-ZOL on pro-gesterone synthesis were analyzed by using an in vitromodel. In order to determine whether stimulation by areceptor-dependent or a receptor-independent mechanisminfluences the sensitivity of porcine granulosa cells to�-and �-ZOL, the effects of both zearalenone derivatives onthe expression of P450scc and 3�-HSD transcripts, and onthe abundance of the enzyme P450scc, were determined incells stimulated with FSH[9] or forskolin [10].

2. Material and methods

2.1. Materials

Tissue culture supplies, medium 199 (HEPES modifica-tion, TCM 199 medium), ABAM: penicillin (100 IU/ml),amphotericin B (250 ng/ml), streptomycin (100�g/ml),ITS: selenium (4 ng/ml), transferrin (2.5�g/ml), insulin(10 ng/ml), forskolin, pregnenolone, and the mycotoxins�- and �-ZOL were obtained from Sigma (Deisenhofen,Germany). FSH-preparat: Ovaset Lot No. 4782 T (contentof one ampulla= 39 mg pFSH according to NIH standard,39 mg FSH pure) was from CEVA (Libourne, France).Polyclonal P450scc antibody was from Chemicon Inter-national (Temecula, USA). For electrophoresis acrylamideand bisacrylamide were from Serva, and all other chemicalsfor electrophoresis and Western blotting were from Sigmaunless stated otherwise.

2.2. Cell cultures and experimental design

The collection and isolation of pig granulosa cells wascarried out as described by Tiemann et al.[11]. Ovariesfrom pigs at a commercial slaughterhouse were collectedinto PBS. Two hours after slaughter granulosa cells wereaspirated from medium to large non-atretic follicles (>3 mmin diameter) by means of a syringe and the granulosa cellswere flushed with PBS. To disperse cell clusters the cellaggregates were resuspended several times through a Pas-teur pipette. The cells were centrifuged at 200×g for 5 minand then resuspended in TCM199 medium containing 10%fetal calf serum (FCS), 1% ABAM. Viability, as measuredby trypan blue exclusion, was∼70%. Granulosa cells werealiquoted in 24-well plates at approximately 1× 105 cells

per well and incubated at 37◦C in a humidified 95% air–5%CO2 environment. After 24 h, the culture medium waschanged and the cells were cultured in complete media forfurther 24 h. After two days (50–70% confluency) the cellswere rinsed with complete Hank’s balanced salt solution(HBSS), then 1.0 ml serum-free medium was added whichwas supplemented with 1% ABAM, 1% ITS and differentadditives (see below). Control cultures were treated withmedium alone. Mycotoxins were dissolved in dimethylsul-foxid (DMSO) before addition to the medium (the finalconcentration of DMSO in the medium was 0.2%). ThisDMSO content was found to have no detectable effect onthe cells.

In the first experiment, monolayers of granulosa cellswere incubated with or without different concentrations of�- or �-ZOL (0, 15 and 30�M final concentrations). In ex-periment 2 the monolayers were incubated with or without0.01�g/ml FSH in the absence or presence of�- or �-ZOL(0, 5, 7.5, 15 and 30�M final concentrations). Experiment 3and 4 were carried out similarly except that 10�M forskolinor 2.5�M pregnenolone were used to stimulate granulosacell progesterone synthesis. After the incubation time thecells were centrifuged at 800× g for 10 min; media wereremoved and stored at−80◦C for subsequent progesteronedetermination. In all experiments, in the same wells mediawere harvested for progesterone analysis and viable cellswere counted as described below. For estimation of P450sccand 3�-HSD mRNA, and P450scc protein, the media wereremoved for progesterone detection and the cells were storedat −80◦C until use.

2.3. RIA

The progesterone concentrations were measured in cul-ture medium by specific radio-immunoassay (RIA) as pre-viously described[11]. Intra- and inter-assay precision were4.8 and 13.2% for progesterone. For the basal progesteronesynthesis (experiment 1) the progesterone synthesis was ex-pressed as ng/1×105 granulosa cells, which were plated perwell at the beginning of the experiment. For experiments 2,3 and 4 the progesterone synthesis was also calculated tothe cell number at the start of an experiment, but becauseof variation between the experiments the results were ex-pressed in relationship to the mean of the FSH-, forskolin-,or pregnenolone-control groups. For the determination ofmRNAs in cell monolayers the progesterone synthesis wasexpressed as ng progesterone/ml medium.

2.4. Cell viability and counting

2.4.1. Cell viabilityThe viability of adherent cells was estimated by

trypan-blue exclusion at the end of experiments 1–4, in thesame well in which medium for the measurement of pro-gesterone synthesis was collected. After removing mediumthe cells were washed twice with complete HBSS. Adherent

U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681 675

cells were then treated with trypan-blue (0.5% in PBS andfiltered prior to use). Since attached cells were not stainedwith trypan-blue they were regarded as viable cells.

2.4.2. Cell countingFor estimating the numbers of viable cells per well, the

monolayers were washed with HBSS (without Ca2+andMg2+) to remove the trypan-blue solution twice and in-cubated with 500�l trypsin–EDTA (0.02–0.05%) solutionper well for 30 min to loosen cells from the plate. Thenumber of detached cells was estimated by a cell counter(Coulter-Multisizer, Krefeld, Germany). For counting,100�l of cells were suspended in 10 ml, 0.9% NaCl solu-tion. Each sample measurement was repeated twice.

2.5. RNA preparation

Total RNA was prepared with the RNeasy mini kit (Qi-agen, Hilden, Germany). Briefly, after removing the culturemedium for progesterone radio-immunoassays cells werelysed and collected in 300�l of a guanidineiso-thiocyanatecontaining buffer (Lysis buffer) and subjected to homoge-nization by using QIAshredderTM Homogenizers (Qiagen).Subsequently, an equal volume of 70% ethanol was addedto the homogenized lysates and the RNA was extracted byadsorption to silica-gel spin columns. After three washesand elution in 30�l deionized, RNAse free water (washingbuffers and water provided by the kit) RNA was quantified ina GeneQuant II instrument (Pharmacia, Freiburg, Germany).Quality of RNA was monitored from randomly selected sam-ples by denaturing agarose (1%) gel electrophoresis.

2.6. Reverse transcription and primer design

Primers for reverse transcription (RT) and PCR were de-rived from different exons to avoid amplification of residualgenomic DNA. All primers were designed according to pub-lished sequences (P450scc: EMBL/GenBank accession no.X13768, RT-primer: 5′- CCAGGTCTTGGTCCTGAACA-GAC-3′; PCR-primers: 5′-TTTACAGGGAGAAGCTCGG-CAAC-3′ and 5′-TTACCTCCGTGTTCAGGACCAAC-3′;3�-HSD: EMBL/GenBank accession no. AF232699, RT-primer: 5′-CTATGCTGCTGGTGTGGATGAAG-3′; PCR-primers: 5′-AGGGTTTCTGGGTCAGAGGATC-3′ and5′-CGTTGACCACGTCGATGATAGAG-3′). For cDNAsynthesis, 0.15�g total RNA was reverse-transcribed ina 25�l reaction volume using MMLV reverse transcrip-tase, RNase H Minus, Point Mutant (Promega, Mannheim,Germany). Specifically primed reverse transcriptions wereperformed by simultaneously adding primers that bind toP450scc and 3�-HSD in order to avoid variations caused bydifferent RT reactions. The freshly synthesized cDNA sam-ples were cleaned with the High Pure PCR Product Purifica-tion Kit (Roche, Mannheim, Germany) and eluted in 50�lelution buffer. The identity of products generated with dif-ferent primer pairs had been controlled once by sequencing.

2.7. Real-time PCR

For real-time PCR, 0.3 and 0.15�l of purified cDNA sam-ples were amplified with the LightCycler-FastStart DNAMaster SYBR Green I Kit (Roche) in 10�l total reaction vol-ume. Amplification and quantification of generated productswere performed in a LightCycler instrument (Roche) underthe following cycling conditions: pre-incubation at 95◦Cfor 10 min, followed by 45 cycles denaturation at 95◦C for15 min, annealing at 60◦C for 10 min, extension at 72◦C for10 min and single point fluorescence acquisition at 83◦C for6 min in order to avoid quantification of primer artifacts.

The melting peaks of all samples were routinely deter-mined by melting curve analysis in order to ascertain thatonly the expected products had been generated. Additionally,molecular sizes of PCR products from randomly selectedsamples were monitored by agarose gel electrophoresisanalysis (3% agarose, ethidium bromide stained).

To generate external standards for each gene RT–PCRproducts of both genes were cloned into GEM-T plasmidvector (Promega). Routinely dilutions of standards covering5 orders of magnitude (5× 10−16 to 5× 10−12 �g DNA/�l)were co-amplified during each run. Fluorescence signals,which were recorded on-line during amplification, were sub-sequently analyzed using the “Second Derivative Maximum”method of the LightCycler Data Analysis software. Copynumbers were calculated relative to the amount of initiallytranscribed RNA. To normalize for variations between indi-vidual LightCycler runs one or two arbitrarily selected sam-ples were co-amplified during all runs for each gene.

2.8. SDS–PAGE and Western blotting for P450scc

To analyze granulosa cells (106) for their expressionof P450scc protein, cultured granulosa cells treated with0.01�g/ml FSH or 10�M forskolin with or without dif-ferent concentrations of�- or �-ZOL were analyzed byWestern blotting. Therefore, the cells were lysed in 10�lof SDS-sample buffer and denatured by boiling for 2 min.Then SDS–PAGE was performed according to Laemmlion 12.5% gels, where the acrylamid:bisacrylamide ratiowas 30:0.8. After electrophoresis proteins were transferredto PVDF membranes (Millipore, Schwalbach, Germany).Thereafter membranes were saturated with 5% fat free drymilk in Tween–Tris-buffered saline (TTBS) overnight at4◦C under constant agitation. After repeated washing withTTBS, the blots were incubated with a polyclonal antibodyto P450scc (Chemicon International, Hofheim, Germany)for three hours at room temperature in TTBS containing5% BSA. The dilution of the antibody was 1:1000. Thenblots were washed three times with TTBS and incubatedwith a HRP conjugated goat–anti-rabbit–antibody (SantaCruz, Heidelberg, Germany). This secondary antibody wasdiluted 1:2000 in TTBS. After a final washing in TTBS,the bands were visualized on X-ray films (Kodak, Stuttgart,Germany), using a chemiluminescent kit (Amersham,

676 U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681

Braunschweig, Germany) according to the manufacturer’sinstruction.

To elucidate the protein loading of each lane, blots wereanalyzed for the abundance of�-tubulin. Therefore, themembranes were stripped by incubation in 50 mM Tris–HCl,pH 6.7, 100 mM mercaptoethanol, 2% SDS at 50◦C for20 min. Thereafter the membranes were blocked again andreprobed with a monoclonal anti-�-tubulin antibody (Sigma)in a dilution of 1:2000 in TTBS as described above.

The optical density of the bands was determined usingscanalytics One-Dscan software (Fairfax, VA, USA). Therelative abundance of P450scc was noted, where the valueswere normalized to that of�-tubulin expression for eachlane (tubulin expression represents 100%).

After Western blotting membranes were stained with sil-ver (incubation for 5 min in 2% sodium citrate, 0.8% FeSO4and 0.2% AgNO3) to determine the total protein composi-tion of the samples.

2.9. Statistical analysis

The experiments were run in sets of three or four gran-ulosa cell preparations isolated from pig ovaries. Data areexpressed as the mean±S.E.M. of triplicate measurements.A one-way analysis of variance (ANOVA) with StudentNewman–Keuls post test was used to determine significantdifferences between treated and the corresponding controlswithout additives. AP value of less than 0.05 was consid-ered to be significant.

3. Results

3.1. Effect of α- and β-zearalenol on FSH- or forskolin-stimulated progesterone synthesis

In experiment 1 (n = 3) the basal level of progesteronesynthesis was 2.2±0.28 ng/1×105 granulosa cells. This levelwas not significantly affected after 48 h incubation with con-centrations (15 and 30�M) �- or �-ZOL (data not shown).

In the presence of FSH the progesterone synthesis in gran-ulosa cells was increased 2.7-fold (0.001 FSH), 5.6-fold(0.01�g FSH) and 5.3-fold (0.1�g FSH) versus basal lev-els (data not shown). We chose for the following experi-ment 0.01�g FSH. In experiment 2 (n = 4), coincubationwith increasing concentrations of both mycotoxins signif-icantly suppressed FSH-stimulated progesterone synthesis.Fig. 1A demonstrates that at a concentration of 15�M �-or �-ZOL, the FSH-stimulated progesterone synthesis was57.9 or 72.0%, respectively, compared to the unexposed con-trols (P < 0.05). At the highest dose used (30�M) of �- or�-ZOL, the FSH-stimulated progesterone synthesis was 26.3or 31.8%, respectively, compared to the unexposed controls(P < 0.05).

In experiment 3 (n = 4), a concentration of 10�Mforskolin increased progesterone secretion approximately10-fold versus basal levels produced by granulosa cells.

Fig. 1. Effects of�- or �-ZOL on FSH (A), forskolin (B), or preg-nenolone (C)-stimulated progesterone synthesis. Cell monolayers werecultured either with 0.01�g/ml FSH, 10�M forskolin or 2.5�M preg-nenolone and with or without different concentrations of�- or �-ZOL.After 2 days of incubation culture media were collected for progesteroneradio-immunoassay. Data are the means± S.E.M. of three to four in-dependent replications. Values are expressed relative to control for eachtreatment. Asterisks denote values that were significantly different fromrespective controls (P < 0.05).

Coincubation with�- or �-ZOL at concentration of 30�Mreduced the forskolin-stimulated progesterone synthesis to51% for �-ZOL or 61% for�-ZOL (P < 0.05) comparedto the unexposed controls (Fig. 1B).

3.2. Effect of α- and β-zearalenol on number of viable cells

After 96 h culture, the incubation of granulosa cells withmedium containing�- or �-ZOL, up to 30�M for 48 h,did not affect the number of adherent cells. In the presenceof FSH, 30�M of either zearalenol derivative caused a

U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681 677

Fig. 2. Effects of�- or �-ZOL on FSH (A), forskolin (B), or pregnenolone(C)-stimulated progesterone synthesis. Cell monolayers were cultured ei-ther with 0.01�g/ml FSH, 10�M forskolin or 2.5�M pregnenolone andwith or without different concentrations of�- or �-zearalenol. After 2days the number of viable cells (no trypan-blue stained monolayers) wasdetermined with a cell counter. Data are means± S.E.M. of three to fourindependent replications. Asterisks denote values that were significantlydifferent from respective controls (P < 0.05).

significant reduction of adherent cells to 60% for�-ZOLor 71% for �-ZOL (P < 0.05) compared to controls after48 h cultivation (Fig. 2A). The number of adherent cellswas not significantly decreased after treatment with differ-ent concentrations of both mycotoxins in combination withforskolin (Fig. 2B).

3.3. Effects of α- and β-zearalenol on FSH-stimulatedexpression of the P450scc and 3β-HSD genes

In a separate set of experiments we monitored the effectsof �- and�-ZOL on FSH-stimulated expression ofP450scc

Fig. 3. Effects of�-zearalenol (A) and�-zearalenol (B) on FSH-stimulatedexpression of the P450scc (SCC) and 3�-HSD (HSD) genes. Cells werecultured with 0.01�g/ml FSH with or without different concentrations of�- or �-zearalenol. After 2 days of incubation, culture media were col-lected for progesterone (P4) radio-immunoassay and the cells were pro-cessed for transcript quantification as described. Data are means±S.E.M.

of six independent replications. Concentrations of P450scc and 3�-HSDtranscripts are expressed as copies/�g RNA (left axes) and progesteroneconcentrations as ng/ml of medium (right axes). Means with differentletters are significantly different (P < 0.05).

and 3β-HSD. During these experiments transcript concen-tration and progesterone production were determined fromthe same samples. As shown for progesterone production(Fig. 1A) the transcript abundance of both genes increasedremarkably upon adding 0.01�g FSH.P450scc transcriptsincreased 11-fold and3β-HSD transcripts 25-fold (data notshown). FSH-stimulated transcripts and progesterone con-centrations were significantly reduced by 15�M �- and�-ZOL (Fig. 3). Slight differences may exist comparing ef-fects onP450scc and3β-HSD in that inhibitory effects on3β-HSD expression were observed already at lower zear-alenol concentrations.

3.4. Effects of α- and β-zearalenol on forskolin-stimulatedexpression of the P450scc and 3β-HSD genes

Effects of�- and�-ZOL on forskolin-stimulatedP450sccand3β-HSD gene expression were analyzed by adding in-creasing concentrations of zearalenol to cultures stimulatedwith 10�M forskolin. Transcript and progesterone profiles

678 U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681

Fig. 4. Effects of�-zearalenol (A) and�-zearalenol (B) on forskolin-stimulated expression of the P450scc (SCC) and 3�-HSD (HSD) genes. Cells werecultured with or without 10�M forskolin and different concentrations of�- or �-zearalenol. After 2 days of incubation, culture media were collectedfor progesterone (P4) radio-immunoassay and the cells were processed for transcript quantification as described. Data are means± S.E.M. of threeindependent replications. Concentrations of P450scc and 3�-HSD transcripts are expressed as copies/�g RNA (left axes) and progesterone concentrationsas ng/ml of medium (right axes). Means with different letters are significantly different (P < 0.05).

were determined from the same samples. In the absenceof zearalenol forskolin increased transcript concentrations8.6- and 3.8-fold in case ofP450scc and3β-HSD, respec-tively (Fig. 4). The addition of�- and �-ZOL decreasedforskolin-stimulated expression of both genes. Significantinhibition of both genes was observed at concentrations of30�M; however, reduced expression of the3β-HSD genewas apparent already at lower concentrations.

3.5. P450scc enzyme abundance and 3β-HSDenzyme activity

The decrease of P450scc transcript abundance in granu-losa cells that were co-cultured with either FSH or forskolinand either�- or �-ZOL was further evaluated by analyz-ing the abundance of P450scc protein as detected by West-ern blot analysis. On these blots, a single band of P450sccwas found with a size of 45 kDa (Fig. 5A–C). When the ex-pression of P450scc was normalized to�-tubulin as a load-ing control, the decrease of P450scc protein was apparentonly above 7.5�M �- or �-ZOL in FSH-stimulated cells,

and above 15�M during forskolin stimulation. This find-ing is in agreement with data obtained by analyzing theP450scc-mRNA (Figs. 3 and 4).

Addition of 2.5�M pregnenolone led to a 13.3-fold in-crease of progesterone synthesis at 48 h. This finding indi-rectly reflects the activity of 3�-HSD; however, co-culturewith �- or �-ZOL at a concentration of 30�M signifi-cantly inhibited pregnenolone-stimulated progesterone syn-thesis compared to the unexposed controls (experiment 4,n = 3; Fig. 1C) without a significant change in number ofadherent cells (Fig. 2C).

4. Discussion

Cultured primary porcine granulosa cells have been usedto test the toxic potential of xenobiotics on reproduction[12].The present study has investigated the influence that�- and�-ZOL had on the regulation of progesterone synthesis inprimary cultures of porcine granulosa cells. In these exper-iments, basal progesterone synthesis was very low and we

U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681 679

Fig. 5. Effects of different concentrations of�- or �-ZOL on the expression of cytochrome P450scc protein in porcine granulosa cells co-cultured with0.01�g/ml FSH (A, B) or 10�M forskolin (C). The cells were incubated with different additives at 37◦C for 2 days. Then the cells were collectedand analyzed by Western blotting for the abundance of P450scc. Equal cell numbers (106) were loaded on each lane. As a loading control, blots werereprobed with�-tubulin antibody and the total protein composition of the samples was determined by staining the membranes with silver. The opticaldensity of the obtained bands was measured with a scanalytics One-Dscan software. The expression of P450scc in each sample is expressed in percentnormalized for the abundance of�-tubulin.

did not find any significant inhibitory effect of both myco-toxins on unstimulated cells. Granulosa cells were culturedin serum-free medium and progesterone synthesis was stim-ulated with either FSH, forskolin or pregnenolone.

FSH-stimulation of progesterone synthesis in primary cul-ture of porcine granulosa cells is in agreement with the re-sults reported by Channing and Ledwitz-Rigby[13] and isfurther extended by the enhanced expression of cytochromeP450scc and 3�-HSD transcripts observed here. Expres-sion of 3�-HSD mRNA increased to a greater degree thanP450scc mRNA. This finding is consistent with results show-ing that the conversion of cholesterol to pregnenolone, a re-action catalyzed by P450scc, is critical for steroid hormonesynthesis and rate-limiting in the enzymatically-catalyzedsynthesis (steroidogenic acute regulatory protein, StAR) ofprogesterone[14]. Urban et al.[15] have shown that cAMPstimulates P450scc mRNA accumulation in porcine granu-losa cells. These results were supported by investigations ofPicton et al.[9]. In rat granulosa cells, the cAMP effect waspartly attributed to increased transcription of theP450sccgene[16]. In human[17] and bovineP450scc genes[18]regulatory elements have been identified which are respon-sible for cAMP-dependent transcription.

Evidently, hormone receptor binding in granulose cellsactivates a G protein that increases intracellular cAMP con-tent [19,20] and, in turn, stimulates cytochromeP450sccgene transcription and progesterone synthesis[21]. Our datasuggest that both mycotoxins at concentrations of 15 and

30�M (Fig. 1) inhibited the FSH-stimulated progesteronesynthesis. The inhibitory effect at a concentration of 30�Mseen with both mycotoxins could be caused by excessive celldeath because the numbers of adherent cells were signifi-cantly reduced. On the other hand, viability and cell numbersof samples treated with 15�M of either mycotoxin did notshow significant differences compared to controls. There-fore, we may conclude that the antisteriodogenic effect of15�M �- or �-ZOL was not due to general cytotoxicity andsuggest that the inhibition of FSH-stimulated progesteronesynthesis might be caused by different mechanisms.

One way that both mycotoxins may inhibit FSH-stimulatedprogesterone accumulation by binding and changingthe FSH receptor conformation and/or by interruptingG-protein–GTP-mediated pathways. We found that�- and�-ZOL inhibited the FSH-induced abundance of P450sccand 3�-HSD transcripts in a dose-dependent manner al-though the transcription of both genes was influenced toa slightly different extent (Fig. 3). These results indicatethat both mycotoxins altered the expression of P450scc and3�-HSD on the transcriptional level. Western blot analysisshowed decreased expression of P450scc protein in paral-lel with mRNA in response to both mycotoxin treatments(Fig. 5A). These data suggest that�- or �-ZOL inhibitobligatory and regulated steps of steroid synthesis in theovary as well.

The final step in progesterone synthesis is conversion ofpregnenolone to progesterone, a reaction catalyzed by the

680 U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681

microsomal enzyme 3�-HSD [5,22]. Addition of exoge-nous pregnenolone as a substrate caused a strong increaseof progesterone synthesis in granulosa cells. This suggestedthat cells express an excess of active 3�-HSD enzyme, aninterpretation consistent with the observation that 3�-HSDtranscript abundances were in excess of those for P450scc.However, 3�-HSD expression appeared more sensitive to�- or �-ZOL inhibition. This may reflect the importanceof P450scc as the rate limiting enzyme of progesteronesynthesis.

Taken together, these results suggest that mycotoxin-induced changes of progesterone synthesis in FSH-stimulated granulosa cells could be caused by a decreasein FSH receptor function and/or receptor-coupled adenylatecyclase activity.

To investigate whether both mycotoxins influenced pro-gesterone synthesis by post-receptor steps we use the adeny-late cyclase activator, forskolin. This substance increasesintracellular cAMP levels by nonreceptor-mediated mecha-nisms[23]. The application of forskolin to porcine granu-losa cells led to progesterone accumulation similar to thatreported previously[24]. Forskolin-stimulated progesteronesynthesis most likely corresponded with the increased ex-pression ofP450scc and3β-HSD gene transcripts.

Our study demonstrated that�- and�-ZOL significantlyinhibited forskolin-induced abundance of P450scc and3�-HSD mRNAs and progesterone secretion in approxi-mately the same manner. Because the viability of granulosacells was not significantly affected the observed inhibitionof forskolin-stimulated progesterone synthesis was not at-tributed to increased cell death. Thus, it can be explained bythe reduction of adenylate cyclase activity or the regulatorypathway further downstream from cAMP generation. Theinhibition of adenylate cyclase could be caused by an inter-action of both mycotoxins with the catalytic subunit of theenzyme. Seamon and Daly[25] reported a unique mecha-nism whereby forskolin directly interacts with the catalyticbut not with the regulatory subunit of adenylate cyclase.However, it cannot be excluded that reactions upstreamfrom progesterone synthesis may be targets of mycotoxinactions as well. Synthesis of cholesterol precursors cat-alyzed in the cytoplasm by HMG–CoA reductase[19], andcholesterol transport to the inner mitochondrial membranemediated by StAR[14], might both comprise direct targetsof the mycotoxins.

In summary, our results indicate that�- and �-ZOLinhibit the FSH-, forskolin- and pregnenolone-stimulatedprogesterone synthesis in porcine granulosa cells in vitro.�- and �-ZOL can impair FSH receptor-mediated andreceptor-independent stimulation of progesterone synthesisand expression of important enzymes involved in proges-terone synthesis, P450scc and 3�-HSD. This indicates thatadenylate cyclase and/or the regulatory pathways furtherdownstream are targets of mycotoxin actions. These resultssupport the notion that these mycotoxins influence geneexpression at a transcriptional level. Whether or not these

mycotoxins may also perturb other mechanisms in proges-terone synthesis, like HMG-Co reductase or StAR function,remains to be investigated.

Acknowledgments

We gratefully acknowledge the technical assistance ofMrs. Ch. Pöppel, M. Anders and G. Krüger. The study wassupported by a grant from the Deutsche Forschungsgemein-schaft (Ti 189/5-1).

References

[1] Kuiper-Goodman T, Scott PM, Watanabe H. Risk assessment of themycotoxin zearalenone. Regul Toxicol Pharmacol 1987;7:253–306.

[2] Rotter BA, Dan B, Pestka JJ, Prelusky JJ. Toxicology ofdeoxynivalenol (Vomitoxin). J Toxicol Environ Health 1996;48:1–34.

[3] Farnworth ER, Trenholm HL. The metabolism of the mycotoxinzearalenone and its effect on the reproductive tracts of young maleand female pigs. Can J Anim Sci 1983;63:967–75.

[4] Hirshfield AN. Overview of ovarian follicular development: consi-derations for the toxicologist. Environ Mol Mutagen 1997;29:10–5.

[5] Hadley ME. Hormones and female reproductive physiology. In:Hadley ME, editor. Endocrinology. New York: Prentice-Hall; 1995.p. 476–504.

[6] Lahav M, Garmey JC, Veldhuis JD. Paradoxical effect of 3-isobutyl-1-methylxanthine on cytochrome P450 cholesterol side-chain cleavage mRNA accumulation in porcine granulosa cells. MolCell Endocrinol 1996;117:203–10.

[7] Chedrese PJ, The VL, Labrie F, Juorio AV, Murphy BD. Evidencefor the regulation of 3�-hydroxysteroid dehydrogenase messengerRNA by human chorionic gonatrophin in luteined porcine granulosacells. Endocrinology 1990;126:2228–30.

[8] Hsueh AJW, Adashi EY, Jones PBC, Welsh Jr TH. Hormonalregulation of the differentiation of cultured ovarian granulosa cells.Endocr Rev 1984;5:76–127.

[9] Picton HM, Campbell BK, Hunter MG. Maintenance of oestradiolsynthesis and expression of cytochrome P450 aromatase enzymemRNA in long-term serum-free cultures of pig granulosa cells. JReprod Fertil 1999;115:67–77.

[10] Gu Y, Lin YC, Rikihisa Y. Inhibitory effect of gossypol onsteroidogenic pathways in cultured bovine luteal cells. Biol BiophysRes Commun 1990;169:455–61.

[11] Tiemann U, Pöhland R, Schneider F. Influence of organochlorinepesticides on physiological potency of cultured granulosa cells frompreovulatory follicles in bovine. Theriogenology 1996;46:253–65.

[12] Haney A, Hughes S, Hughes CJ. Screening of potential reproductivetoxicants by use of porcine granulosa cell cultures. Toxicology1984;30:227–41.

[13] Channing C, Ledwitz-Rigby F. Methods for assessing hormone-mediated differentiation of ovarian cells in culture and in short-termincubations. Methods Enzymol 1975;39:183–230.

[14] Stocco DM, Clark BJ. Role of the steroidogenic acute regulatoryprotein (StAR) in steroidogenesis. Biochem Pharmacol 1996;51:197–205.

[15] Urban RJ, Garmey JC, Shupnik MA, Veldhuis D. Follicle-stimulatinghormone increases concentrations of messenger ribonucleic acidencoding cytochrome P450 cholesterol side-chain cleavage enzymein primary cultures of porcine granulosa cells. Endocrinology1991;128:2000–7.

[16] Kurten RC, Richard JS. An adenosine 3′,5′-monophosphate-responsive deoxyribonucleic acid element confers forskolinsensitivity on gene expression by primary rat granulosa cells.Endocrinology 1989;125:1345–57.

U. Tiemann et al. / Reproductive Toxicology 17 (2003) 673–681 681

[17] Inoue H, Higashi Y, Morohashi K, Fujii-Kuriyama Y. The 5′-flanking region of the humanP-450(SCC) gene shows responsive-ness to cAMP-dependent regulation in transient gene-expressionsystem of Y-1 adrenal tumor cells. Eur J Biochem 1988;171:435–40.

[18] Ahlgren RA, Simpson ER, Waterman MR, Lund J. Characterizationof the promoter/regulatory region of the bovine CYPl1A(P-450scc)gene. J Biol Chem 1990;265:3313–9.

[19] Richards JS. Hormonal control of gene expression in the ovary.Endocr Rev 1994;15:725–51.

[20] Conti M. Minireview: specificity of the cyclic adenosine 3′,5′-monophosphate signal in granulosa cell function. Biol Reprod2002;67:1653–61.

[21] Brentano ST, Black SM, Lin D, Miller WL. cAMP post-transcrip-tionally diminishes the abundance of adrenodoxin reductase mRNA.Proc Natl Acad Sci USA 1992;89:4099–103.

[22] Chedrese PJ, Braileanu GT, Samon R. 3�-Hydroxy-5-ene steroiddehydrogenase gene expression regulation in procine granulosacells I: FSH- and LH-mediated transcriptional activation. Endocrine1995;3:195–9.

[23] Dean ED, Byrd JA, Williams JD, Hargis BM. Influence of follicularmaturation on inhibition of luteinizing hormone-, cyclic 3′,5′-adenosine monophosphate-, and forskolin-stimulated progesteroneproduction in chicken ovarian granulosa cells exposed to bursal anti-steroidogenic peptide. Biol Reprod 1995;52:771–5.

[24] Veldhuis JD, Rodgers RJ, Hewlett EL. Actions of cyclic adenosinemonophosphate on the cytodifferentiation of ovarian cells: studiesin cultured swine granulosa cells using a novel exogenous adenylatecyclase fromBordetella pertussis. Mol Endocrinol 1988;2:499–506.

[25] Seamon KB, Daly JW. Forskolin: an unique diterpene activator ofcyclic AMP-generating systems. Review. J Cyclic Nucleotide Res1981;7:201–24.