For some years, there has been growing con-cern about the existence of contaminants in ourfood and environment with the ability to modu-late or disrupt the human endocrine system
(Jensen et al., 1995; Chia, 2000). Reason for thisconcern is 1) an increase in the incidence of testiscancer (Møller, 1993), 2) a possible increase inhypospadias and cryptorchidism (Garcia-Ro-driguez et al., 1996; Weidner et al., 1998;Paulozzi, 1999) and 3) a decline in semen qualityobserved in some regions (Carlsen et al., 1992).Compounds such as vinclozolin, p,p�-DDE andprocymidone have been found to function as an-tiandrogens during sexual differentiation of males
C. Nellemann et al. / Toxicology 163 (2001) 29–3830
in animal studies resulting in e.g. hypospadia,cleft phallus, reduced anogenital distance, at-rophic seminal vesicles and ventral prostateglands as well as delayed puberty (Gray et al.,1994; Kelce et al., 1994, 1995; Ostby et al., 1999;Ashby & Lefevre, 2000).
After binding of ligand, the intracellular andro-gen receptor functions through regulation of tran-scription of specific genes by either increasing orsuppressing the expression (Chang et al., 1995).As examples of androgen-responsive genes wehave chosen to investigate the expression ofprostate specific binding protein polypeptide C3(PBP C3) (Bossyns et al., 1986) and testosterone-repressed prostatic message 2 (TRPM-2) (Mont-petit et al., 1986) in the ventral prostate of the rat.Prostate binding protein, PBP, constitutes ap-proximately 50% of the secreted protein from thenormal rat prostate and is thereby the most abun-dant androgen-regulated protein in this tissue.The androgen regulation of PBP C3 is pre-sumably caused by an androgen-responsive ele-ment in the first intron (Heyns et al., 1978; Page& Parker, 1982; Vercaeren et al., 1996). TRPM-2,also called clusterin or SGP-2, is not normallyexpressed in the rat ventral prostate but is inducedby castration (Leger et al., 1987). The expressionof TRPM-2 reaches a maximum approximately 4days after castration followed by a return tonormal levels (Montpetit et al., 1986; Bettuzzi etal., 1989). TRPM-2 has been suggested to beinvolved in the onset of apoptosis, but the latestdata suggest that TRPM-2 acts as a survivalfactor from apoptosis (Buttyan, 1991; Ho et al.,1998; Viard et al., 1999).
Several weak androgen receptor antagonistshave been identified in vitro (Vinggaard et al.,1999, 2000) causing a need for the establishmentof sensitive endpoints for antiandrogenic effects invivo. In vitro and in vivo screening have shownthat flutamide and the fungicide vinclozolin act asantiandrogens (Kelce et al., 1997; Ashby &Lefevre, 2000). Previously, the expressions ofTRPM-2 and PBP C3 have been shown to beaffected by these antiandrogens using a moretime-consuming technique involving NorthernBlotting (Kelce et al., 1997). We wanted to estab-lish a protocol for gene expression analysis in the
Hershberger assay (Hershberger et al., 1953) em-ploying real-time RT-PCR.
To study changes in gene expression, we em-ployed the LightCycler technology that uses rapidthermocycling and on-line detection of fluores-cence. By using specific hybridization probes, arapid quantification of initial amount of mRNAcan be achieved by correlation to the time-pointwhere the fluorescence signal rises above the back-ground and not by correlation to the plateauphase of the reaction as in normal PCR. Thegreat advantages of this method are: (1) that thePCR amplification and analysis take place in thesame capillary without risk of contamination (2)that the level of signal is standardized relative tothe expression of a housekeeping gene and (3)that the total analysis maximally endures 60 min(Emig et al., 1999). In addition, this method hasthe potential of investigating gene expressions invery small amounts of tissue from e.g. fetuses orneonates.
2. Materials and methods
2.1. Animals, chemicals and treatment
Forty adult male Wistar rats (60 days of age,approximately 200 g body weight) were purchasedfrom M&B, Eiby, Denmark, 1 week after castra-tion. The animals were maintained under stan-dardized conditions, weighed and thenrandomised into four different treatment groupswith ten rats in each group. The animals werecaged and maintained under the following condi-tions: relative humidity 55% � 5, 12 h day:nightcycle with light from fluorescent tubes from 0900to 2100 h, room temperature 21°C�1, acidifiedtap water and standard diet (Altromin, rat no.1324, Brogaarden, DK-2820 Gentofte, Denmark)ad libitum. Daily for 7 days, the rats in three ofthe groups were weighed and subcutaneously ad-ministered with 1 mg/kg/day testosteronepropi-onate (CAS no. 319491, lot no. 228351, Nomeco)with or without addition of either 20 mg/kg/dayflutamide (CAS no. F-9397, lot no. 87H1511,Sigma-Aldrich) or 50 mg/kg/day of vinclozolin(CAS no. 050471–44–8, lot no. 45705–9126X,
C. Nellemann et al. / Toxicology 163 (2001) 29–38 31
Riedel-de Haen). One group of castrated rats wasonly given vehicle (called the control group). Allsubstances were purchased as solids and solubi-lized easily in peanut oil purchased from TheSection of Pharmacy, The Royal Danish Veteri-nary and Agricultural University. Following the7-day treatment period, blood was collected forhormone analysis by heart puncture of animalsthat were anesthetized by inhalation of CO2/O2.The reproductive organs: ventral prostate, �esiculaseminalis and musculus le�ator ani/bulboca�ernosuswere isolated and the wet weight obtained. Theliver, combined kidneys and combined adrenalglands were also weighed. Immediately afterweighing, the ventral prostates were stored inRNAlater (Ambion) and kept at −20°C untiltotal RNA purification was performed.
2.2. RNA isolation and cDNA production
The ventral prostates were disrupted and ho-mogenised in RLT buffer from the RNeasy Midi-kit (QIAGEN) by an Ultra Turrax rotor-statorhomogenizer for approximately 30 s and subse-quent extraction of total RNA was performedusing the RNeasy Midi-kit (QIAGEN) followingthe manufacturer�s instructions. The quantity andquality of the purified RNA was evaluated byspectophotometry. cDNA was produced from0.5–1.0 �g of total RNA by using the display-THERMO-RT kit and the manufacturer’s in-structions (Display Systems Biotech, Kem-en-tec,Denmark).
2.3. Real-time RT-PCR
Real-time RT-PCR with online detection of thePCR reaction based on fluorescence monitoring
(LightCycler, Roche) was employed. We used hy-bridization probes (TIB MolBiol, Berlin, Ger-many) to monitor the amount of specific targetsequence produced. Hybridization probes hy-bridize adjacent on the PCR product and theprobe with the donor fluorphor (in this caseFL530 nm labelled in the 3�-end) transfers energyto the adjacent acceptor fluorphor (LC640 nm,labelled in the 5�-end). A fluorescence signal isemitted and detected in each PCR cycle. Quanti-tative results are obtained by the cycle thresholdvalue where a signal rises above background level.Expression of the genes coding for TRPM-2 andPBP C3 was compared to the steady expression of�-actin or 18S rRNA. The primers were selectedspanning an intron to avoid amplification ofDNA and by gel electrophoresis the correct sizeof the PCR product was checked. For the fourdifferent genes reported on in this paper, eachPCR-reaction was optimized with concern to theMgCl2 concentration, the amount of cDNA orRNA and the optimal concentration of primersand hybridization probes. Optimal reactions wereachieved with 4mM MgCl2 (except for 18SrRNAwhere 5mM MgCl2 was used), 5000 pg RNA orcDNA per reaction, 0.1 �M primers and probesfor �-actin, 0.5 �M primers and 0.2 �M probesfor 18SrRNA, 0.3 and 0.2 �M of primers andprobes, respectively, for PBP C3 and 0.7 and 0.3�M of primers and probes, respectively, forTRPM-2. For sequences of the primers andprobes, see Table 1 and Table 2. The PCR pro-gram followed the manufacturer’s instructions(LightCycler-RNA Amplification Kit Hybridiza-tion Probes or LightCycler-DNA Master Hy-bridization Probes, Roche) except when Taq startantibody (Clontech) was employed where the de-naturation step prior to the amplification wasincreased to 2 min. The Taq start antibody (Clon-
Table 1Sequences of the sense and antisense primers for TRPM-2, PBP C3, �-actin, and 18S rRNA, respectively
tech) (0.16 �l per 20 �l reaction) was incubated withthe RT-PCR enzyme mix or the DNA MasterHybridization Probes mix at room temperature for5 min prior to addition of the rest of the compo-nents (water, MgCl2, RT-PCR Reaction Mix Hy-bridization Probes (only for one-step RT-PCR),primers and hybridization probes). A standardcurve was achieved by diluting one sample of RNAor cDNA followed by amplification with respect to18SrRNA or �-actin. The signal of the samples ofinterest were then quantified from the standardcurve and the expression of TRPM-2 or PBP C3determined relative to the housekeeping genes,18SrRNA or �-actin, for each RNA or cDNAsample. Gene expressions were analysed at leastthree times for each animal. When dual colourdetection was employed, the LC hybridizationprobe for the housekeeping gene was labelled withLC705 and a colour compensation file was used.
2.4. Hormone analysis
Rat luteinizing hormone (LH) was measuredusing the time-resolved fluorimetric assay (IFMA,Delfia, Wallac OY, Turku, Finland) as described(Haavisto et al., 1993). Biotinylated and Eu-la-belled monoclonal antibodies, streptavidine coatedplates, assay buffer, wash solution and enhance-ment solution were obtained from Wallac. Thestandard rLH RP-3 was kindly provided byNIDDK, NIH (Baltimore, MD).
2.5. Statistical analyses
The statistical analysis was performed in the SAS6.12 (SAS Institute Inc, Cary, NC, USA). Aone-way analysis of variance (ANOVA) was per-formed. All data were log-transformed to normal-ize the distributions and eliminate heterogeneity of
variance between treatment groups. When signifi-cant treatment effects were detected in theANOVA, two-tailed t-tests comparisons (leastsquare means) were used to compare the testos-terone group with the castrated, the flutamidetreated or the vinclozolin treated groups. Thegeneral level of significance was set to P�0.05.Therefore, no individual P-values are given.
3. Results
3.1. Optimization of the PCR amplifications
To quantify the expression of androgen-sensitivegenes relative to the expression of an endogenoushousekeeping gene such as �-actin or 18S rRNA,we employed both external and internal standardsand investigated the efficiency of the PCR reac-tions. The efficiency of PCR amplification is deter-mined by e.g. fragment length, the amount of GC�sin the sequence, as well as the quantity and qualityof the RNA or cDNA sample. To be able tocompare the starting amount of two differenttranscripts, the efficiency and thereby the slope ofthe standard curve for each PCR run have to besimilar. An example of a standard curve can beobserved in Fig. 1.
The amplification efficiency (E) of a PCR run inthe logarithmic phase is described as:
E= 10−1/slope
(Technical note no. LC10/2000, Roche)
The maximal theoretical E is 2.0 with a slope ofthe standard curve of −3.3.
It was determined whether the expression of twogenes could be investigated in the same capillary,
C. Nellemann et al. / Toxicology 163 (2001) 29–38 33
called dual-colour detection. Dual-colour detectionwould exclude differences in the PCR amplificationbetween two capillaries, but the presence of eightDNA-oligos per capillary could inhibit the amplifi-cation. As example, the optimal efficiency for�-actin was obtained by use of one-colour detectionper capillary, antibody and 0.1�M of both primersand probes resulting in E=1.9 for a two-stepRT-PCR. The results for one-step RT-PCR for�-actin and for the other genes can be seen in Table
3. It was found that the dynamic range of measur-ing two transcripts in the same capillary was muchless than when external standards were used. Dif-ferences in E between two gene expressions willintroduce variances in the relative gene expressionand should not exceed � 0.05 (Technical note no.LC10/2000, Roche). The efficiencies observed forone-step RT-PCR were not as reproducible as fortwo-step RT-PCR and the mean efficiencies ob-tained normally lower for one-step RT-PCR. Based
Fig. 1. Example of a standard curve for �-actin with r= −1.00, error=0.0615 and slope= −3.406 is seen on the lower of thefigure with cycle number against the logarithm of the known concentrations of cDNA in pg: 1000, 2500, 5000, 7500 and 10 000 pgcDNA. The calculated efficiency, E, equals 1.97. The top of the figure shows fluorescence against cycle number for the five differentconcentrations of cDNA, with the highest concentration resulting in the first signal.
C. Nellemann et al. / Toxicology 163 (2001) 29–3834
Table 3The efficiency coefficients (E) of the PCR for one- or two-step RT-PCR, for one-colour or dual-colour detection in each capillary,and for the genes TRPM-2, PBP C3, �-actin, and 18S rRNA, respectivelya
on these data, we chose to continue usingtwo-step RT-PCR, external standards andone-colour detection.
3.2. Quantitati�e determination of the relati�eexpression of TRPM-2 and PBP C3
We detected a statistically significant increase(P�0.05) in the expression of the TRPM-2 genein the animals treated with testosterone in additionto flutamide or to a low dose of vinclozolincompared to the testosterone-treated group (Fig. 2and Table 4). Additionally, we detected a signifi-cant decline in the expression of PBP C3 in the sametwo groups (Fig. 3 and Table 4).
3.3. Weights of reproducti�e organs
Weights of the chosen reproductive organs wereinvestigated in the four animal groups after thetreatment period of 7 days. A clear effect oftestosterone treatment was observed when com-pared to the control rats (Table 5). The bodyweights of the control rats were significantly lowerthan for the testosterone-treated animals, but thebody weights of the flutamide- and vinclozolin-treated groups were not different from that of thetestosterone-treated animals. The effect on abso-lute organ weights was significantly diminished byflutamide compared to the testosterone group forthe ventral prostate, the �esicula seminalis and thele�ator ani plus bulboca�ernosus muscles. In thevinclozolin-treated animals this was also the casefor the �esicula seminalis, but not for ventralprostate or musculus le�ator ani/bulboca�ernosus.
The weights of liver, kidneys and adrenal glandswere not changed for the vinclozolin group com-pared to the testosterone group. The same wasfound for the flutamide animals except that theweights of the adrenal glands were significantlyincreased (data not shown). The liver and kidneyweights were significantly increased in the testos-terone group compared to the control group (datanot shown). The statistically significant results of allabsolute organ weights for the four groups werereproduced when the weights were taken relative tothe body weights (data not shown). An exceptionwas that the relative weight of the kidneys in thetestosterone group was not significantly increasedcompared to the control group as seen for theabsolute organ weights.
3.4. Hormone analysis
The potential effect of the test compounds on theLH level was evaluated using a very sensitiveimmunofluorometric assay (Haavisto et al., 1993).Testosterone treatment of the animals resulted, asexpected, in a decreased LH level compared to thecontrol group and this effect was reversed by thetreatment of the animals with flutamide (P�0.05).The dose of vinclozolin (50 mg/kg/day) adminis-tered subcutaneously did not show any effect on thelevel of LH (Fig. 4).
4. Discussion
We have described a real-time RT-PCR methodsuitable for quantitative determination of expres-
C. Nellemann et al. / Toxicology 163 (2001) 29–38 35
Fig. 2. Expression of TRPM-2 relative to the expression of �-actin (mean � SEM) for the four animal groups. The TRPM-2/�-actinvalue is significantly decreased for the testosterone animals compared to the control animals. The TRPM-2/�-actin value issignificantly increased for the flutamide- and vinclozolin-treated animals compared to the testosterone group. * Denotes astatistically significant difference from the testosterone treated animals (P�0.05).
sion of the androgen-sensitive genes, PBP C3 andTRPM-2. This kind of gene expression analysispromises to be a very sensitive and rapid methodto detect suspected antiandrogenic compounds andmay be a valuable supplement to the traditionalendpoints, as weights of reproductive organs. Thesensitivity of the method is of great importance andoffers the ability to study weak antiandrogens, tostudy RNA from very small amounts of tissue aswell as to study the effect of antiandrogens at lowand more relevant levels.
Many genes contain an androgen-responsive-ele-ment in the promotor region and their expressionis directly influenced by the amount of androgen orantiandrogen available. Genes without the andro-gen-responsive-element in the promoter area canalso be influenced by the presence of androgens viaactivation of transcription factors, circulating hor-mones etc. (Verhoeven & Swinnen, 1999). It is wellknown that lack of androgens results in increase inTRPM-2 expression and decrease in PBP C3 ex-pression in the rat ventral prostate which has oftenbeen described in castrated animals (Bossyns et al.,1986; Bettuzzi et al., 1989), but also followingexposure to antiandrogens (Kelce et al., 1997).Besides androgens, other compounds such asgrowth hormone or prolactin have been shown to
have influence on the expression of PBP C3 in therat ventral prostate, (Reiter et al., 1995a,b). How-ever, the changes observed in these experiments arebelieved to be a result of androgen regulation only.
As expected, flutamide showed antiandrogeniceffect based on both organ weight and LH measure-ments, and both TRPM-2 levels and PBP C3 levelswere significantly different from the ones of thetestosterone animals. The dose of vinclozolin (50mg/kg/day) administered subcutaneouly for 7 days
Table 4Expression of TRPM-2 and PBP C3 relative to �-actin in theventral prostatesa
a * denotes a statistically significant difference from thetestosterone treated animals (P�0.05). Data shows themean�SEM for the indicated number of animals.
C. Nellemann et al. / Toxicology 163 (2001) 29–3836
Fig. 3. Expression of PBP C3 relative to the expression of �-actin (mean � SEM). For all three groups compared to the testosteronegroup, a significant decrease is observed. * Denotes statistically significant difference from the testosterone treated animals(P�0.05).
Fig. 4. The level of rat LH in �g/L (mean � SEM) in the four groups. The level of LH showed a statistically significant differencefrom the testosterone treated animals (P�0.05) in the control as well as the flutamide treated groups but not in the vinclozolingroup. * Denotes statistically significant difference from the testosterone treated group.
Table 5The body weight (in g) and the absolute organ weights (mean�SEM) of the reproductive organs investigated (in mg)a
a * Denotes statistically significant difference from testosterone treated animals (P�0.05).
C. Nellemann et al. / Toxicology 163 (2001) 29–38 37
was chosen on purpose to be at a level that wasexpected only to give rise to borderline activity onreproductive organ weights and hormone levels.The normally used dose of vinclozolin is 100mg/kg/day administered perorally often for alonger treatment period (Kelce et al., 1997; Ashby& Lefevre, 2000) and as the metabolites of vinclo-zolin, M1 and M2, are the active components thisway of administration may result in a greaterresponse. We observed a significantly decrease inthe weight of �esicula seminalis in the vinclozolingroup compared to the testosterone group, but nodifference for the weights of the two other repro-ductive organs investigated or for the rest of theorgans. No change was observed in the level of LH,but expressions of both the TRPM-2 and PBP C3genes were significantly different from the expres-sions in the testosterone-treated group in spite ofthe use of a suboptimal dose of vinclozolin. Thetypical coefficient of variation (cv) in the LightCy-cler for 5 000 – 25 000 copy numbers of a purifiedplasmid varies from 3.4% to 9.1% depending onPCR fragment length (Betzl et al., 2000).
The results described in this paper show that geneexpression analysis can be a supplement as well aspossibly a more sensitive parameter than bothorgan weights and hormone analysis, and thatLightCycler technology is suitable for quantitativeevaluation of antiandrogen function.
Acknowledgements
The NIDDK’s National Hormone and PituitaryProgram and Dr. A.F. Parlow are greatly acknowl-edged for having provided rLH. In addition, wewant to thank Lonnie Sand, Birgitte Plesning,Karen Roswall and Lillian Sztuk for excellenttechnical assistance. The Danish Medical ResearchCouncil is greatly acknowledged for financial sup-port (no. 9801270).
References
Ashby, J., Lefevre, P., 2000. Preliminary evaluation of themajor protocol variables for the Hershberger castratedmale rat assay for the detection of androgens, antiandro-gens and metabolic modulators. Regul. Toxicol. Pharma-col. 31, 92–105.
Bettuzzi, S., Hiipakka, R.A., Gilna, P., Liao, S.T., 1989.Identification of an androgen-repressed mRNA in rat ven-tral prostate as coding for sulphated glycoprotein 2 bycDNA cloning and sequence analysis. Biochem. J. 257,293–296.
Betzl, G., Seller, M., Eichner, C., Cornelia Eichner, AndreasKalbe, Maria Geyer, Wilhelm Kleider, Rob vanMiltenburg (2000) Reproducibility of PCR on the Light-Cycler System. In: Biochemica, 1, 22–26, Roche MolecularBiochemicals.
Bossyns, D., Delaey, B., Rombauts, W., Heyns, W., 1986.Influence of castration and androgen treatment on thesynthesis of prostatic binding protein and the concentra-tion of its mRNA in the rat ventral prostate. J. Steroid.Biochem. 24, 661–668.
Buttyan, R., 1991. Genetic Response of Prostate Cells toAndrogen Deprivation: Insights into the Cellular Mecha-nism of Apoptosis in Apoptosis. In: The Molecular Basisof Cell Death. Cold Spring Harbor Laboratory Press.
Carlsen, E., Giwercman, A., Keiding, N., Skakkebaek, N.E.,1992. Evidence for decreasing quality of semen during past50 years. Br. Med. J. 305, 609–613.
Chang, C., Saltzman, A., Yeh, S., Young, W., Keller, E., Lee,H.J., Wang, C., Mizokami, A., 1995. Androgen receptor:an overview. Crit. Rev. Eukaryot. Gene. Expr. 5, 97–125.
Chia, S.E., 2000. Endocrine disruptors and male reproductivefunction-a short review. Int. J. Androl. 23, 45–46.
Emig, M., Saussele, S., Wittor, H., Weisser, A., Reiter, A.,Willer, A., Berger, U., Hehlmann, R., Cross, N.C., Hoch-haus, A., 1999. Accurate and rapid analysis of residualdisease in patients with CML using specific fluorescenthybridization probes for real time quantitative RT-PCR.Leukemia 13, 1825–1832.
Garcia-Rodriguez, J., Garcia-Martin, M., Nogueras-Ocana,M., de Dios Luna-del-Castillo, J., Espigares Garcia, M.,Olea, N., Lardelli-Claret, P., 1996. Exposure to Pesticidesand Cryptorchidism: Geographical Evidence of a PossibleAssociation. Environ. Health. Perspec. 104, 1090–1095.
Gray, L.E. Jr., Ostby, J.S., Kelce, W.R., 1994. Developmentaleffects of an environmental antiandrogen: the fungicidevinclozolin alters sex differentiation of the male rat. Toxi-col. Appl. Pharmacol. 129, 46–52.
Haavisto, A.M., Pettersson, K., Bergendahl, M., Perheentupa,A., Roser, J.F., Huhtaniemi, I., 1993. A supersensitiveimmunofluorometric assay for rat luteinizing hormone.Endocrinology 132, 1687–1691.
Hershberger, L.G., Shipley, E., Meyer, R., 1953. MyotrophicActivity of 19-Nortestosterone and other steriods deter-mined by modified levator ani muscle method. MyotrophicActivity of Androgens, p. 1953.
Heyns, W., Van Damme, B., De Moor, P., 1978. Secretion ofprostatic binding protein by rat ventral prostate: influenceof age and androgen. Endocrinology 103, 1090–1095.
Ho, S.M., Leav, I., Ghatak, S., Merk, F., Jagannathan, V.S.,Mallery, K., 1998. Lack of association between enhancedTRPM-2/clusterin expression and increased apoptotic ac-tivity in sex-hormone-induced prostatic dysplasia of theNoble rat. Am. J. Pathol. 153, 131–139.
C. Nellemann et al. / Toxicology 163 (2001) 29–3838
Jensen, T.K.J., Toppari, J., Keiding, N., Skakkebaek, N.E.,1995. Do environmental estrogens contribute to the declinein male reproductive health? Clin. Chem. 41, 1896–1901.
Kelce, W.R., Monosson, E., Gamcsik, M.P., Laws, S.C.,Gray, L.E. Jr., 1994. Environmental hormone disruptors:evidence that vinclozolin developmental toxicity is medi-ated by antiandrogenic metabolites. Toxicol. Appl. Phar-macol. 126, 276–285.
Montpetit, M.L., Lawless, K.R., Tenniswood, M., 1986. An-drogen-repressed messages in the rat ventral prostate.Prostate 8, 25–36.
Møller, H., 1993. Clues to the Aetiology of Testicular GermCell Tumours from Descriptive Epidemiology. Eur. Urol.23, 8–15.
Ostby, J., Kelce, W.R., Lambright, C., Wolf, C.J., Mann, P.,Gray, L.E. Jr., 1999. The fungicide procymidone alterssexual differentiation in the male rat by acting as anandrogen-receptor antagonist in �i�o and in �itro. Toxicol.Ind. Health. 15, 80–93.
Page, M.J., Parker, M.G., 1982. Effect of androgen on thetranscription of rat prostatic binding protein genes. Mol.Cell. Endocrinol. 27, 343–355.
Paulozzi, L.J., 1999. International trends in rates of hypospa-
dias and cryptorchidism. Environ. Health. Perspect. 107,297–302.
Reiter, E., Lardinois, S., Klug, M., Sente, B., Hennuy, B.,Bruyninx, M., Closset, J., Hennen, G., 1995a. Androgen-independent effects of prolactin on the different lobes ofthe immature rat prostate. Mol. Cell. Endocrinol. 112,113–122.
Reiter, E., Kecha, O., Hennuy, B., Lardinois, S., Klug, M.,Bruyninx, M., Closset, J., Hennen, G., 1995b. Growthhormone directly affects the function of the different lobesof the rat prostate. Endocrinology 136, 3338–3345.
Roche Molecular Biochemicals, Overview of LightCyclerQuantification Methods, Technical noteLC 10/2000.
Vercaeren, I., Vanaken, H., Devos, A., Peeters, B., Verhoeven,G., Heyns, W., 1996. Androgens transcriptionally regulatethe expression of cystatin-related protein and the C3 com-ponent of prostatic binding protein in rat ventral prostateand lacrimal gland. Endocrinology 137, 4713–4720.
Weidner, I.S., Moller, H., Jensen, T.K., Skakkebaek, N.E.,1998. Cryptorchidism and Hypospadias in Sons of Gar-deners. Environ. Health. Perspec. 106, 793–796.
