Altered Gene Profiles in Fetal Rat Testes after in Utero Exposure to Di(n-butyl) Phthalate

Abstract

Di(n-butyl) phthalate (DBP) has antiandrogenic-like effects on the developing reproductive tract in the male rat and produces regions of interstitial cell hyperplasia and gonocyte degeneration in the developing fetal testes at maternal doses of 100–500 mg/kg/day. Neither DBP nor its primary metabolites interact with the androgen receptor in vitro. The present study was performed to examine gene expression in the fetal rat testes following in utero DBP exposure. Pregnant Sprague-Dawley rats received corn oil, DBP (500 mg/kg/day), or flutamide (reference antiandrogen, 50 mg/kg/day) by gavage daily from gestation day (GD) 12 to 21. Dose levels were selected to maximize fetal response with minimal maternal toxicity. Testes were isolated on GD 16, 19, and 21. Global changes in gene expression were determined by microarray analysis. Selected genes were further examined by quantitative RT-PCR. DBP, but not flutamide, reduced expression of the steroidogenic enzymes cytochrome P450 side chain cleavage, cytochrome P450c17, and steroidogenic acute regulatory protein. Testicular testosterone and androstenedione were decreased on GD 19 and 21, while progesterone was increased on GD 19 in DBP-exposed testes. Testosterone-repressed prostate message-2 (TRPM-2) was upregulated, while c-kit (stem cell factor receptor) mRNA was downregulated following DBP exposure. TRPM-2 and bcl-2 protein staining was elevated in GD 21 DBP-exposed Leydig and Sertoli cells. Results of this study have led to the identification of several possible mechanisms by which DBP can induce its antiandrogenic effects on the developing male reproductive tract without direct interaction with the androgen receptor. Our results suggest that the antiandrogenic effects of DBP are due to decreased testosterone synthesis. In addition, enhanced expression of cell survival proteins such as TRPM-2 and bcl-2 may be involved in DBP-induced Leydig cell hyperplasia, whereas, downregulation of c-kit may play a role in gonocyte degeneration. Future studies will explore the link between these identified gene expression alterations and ultimate adverse responses.

Di(n-butyl) phthalate (DBP) is a plasticizer used to impart flexibility to products containing nitrocellulose, polyvinyl acetate, and polyvinyl chloride such as food wraps and blood bags (Brandt, 1985). DBP is also used in cosmetics as a solvent and fixative for perfumes, a suspension agent for solids, a lubricant for aerosol valves, an antifoamer, a skin emollient, and a plasticizer in nail polish, fingernail elongators, and hair spray (Brandt, 1985). DBP is rapidly metabolized and eliminated, and levels of DBP metabolites are detectable in human urine (Saillenfait et al., 1998; Tanaka et al., 1978). Current estimates for human exposure to DBP range from 0.84 to 113 μg/kg/day as determined from urine samples from a reference population of 289 adults (Blount et al., 2000; Kohn et al., 2000). Eight of the 10 highest individual exposures from that study were women in the 20–40 year age group, raising the concern that women of childbearing age may have higher exposure to DBP than the rest of the general population (Blount et al., 2000).

DBP is a male reproductive toxicant in the rat, and the developing testis is a primary target (Mylchreest et al., 1998, 1999, 2000). In a dose-response study (0.5, 50, 100, 500 mg/kg/day), testes of male rat offspring from dams treated with DBP at 100 and 500 mg/kg/day from gestation day (GD) 12 to 21 exhibit histological abnormalities characterized by enlarged seminiferous tubules containing multinucleated degenerating germ cells and regions of Leydig cell hyperplasia (Mylchreest et al., 1998). A small percentage of the Leydig cell hyperplastic lesions develop into adenomas as early as 3 months of age (Mylchreest et al., 1999). No adverse effects were observed in the offspring of dams treated with 50 mg/kg/day (Mylchreest et al., 1998). The physiological effects of DBP on the developing male reproductive tract are similar to the antiandrogen flutamide. Unlike flutamide, however, neither DBP nor its primary metabolite monobutyl phthalate (MBP) interacts with the androgen receptor (Foster et al., 2001), and DBP differs from flutamide in its ability to induce regions of Leydig cell hyperplasia and gonocyte degeneration in the fetal testis.

We hypothesize that DBP induces its antiandrogenic effects by altering both androgen-dependent and independent signaling pathways during male reproductive development. Consequently, the studies reported here examined alterations in gene expression induced by DBP in the developing fetal testes and compared them with alterations elicited by flutamide. We demonstrate that DBP exposure results in altered expression of numerous genes within fetal testes, including genes associated with steroidogenesis and cell survival. The pattern of DBP-mediated gene expression alterations was distinct and yet overlapped with flutamide-induced alterations in gene expression.

MATERIALS AND METHODS

Animal model.

Female Sprague-Dawley outbred CD rats approximately 8-weeks old were timed-mated at Charles River Laboratories, Inc. (Raleigh, NC). The day sperm was found in the vagina of the mated female was considered GD 0. Dams were delivered to the CIIT animal facility on GD 1. Allocation of animals to dose groups was done by body weight randomization to ensure equal weight distribution among groups.

Animals were housed in the CIIT animal facility, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC), and maintained in a HEPA-filtered, mass air displacement room on a 12-h light-dark cycle at approximately 18–26°C with a relative humidity of 30–70%. This study followed federal guidelines for the care and use of laboratory animals (National Academy of Sciences, 1996) and was approved by the Institutional Animal Care and Use Committee at CIIT.

Dams were treated by gavage daily from GD 12 to 21 with corn oil (vehicle control, Sigma Chemical Co., St. Louis, MO), DBP (500 mg/kg/day, 99.8%, Aldrich Chemical Co., Milwaukee, WI), and flutamide (50 mg/kg/day, 99%, Sigma Chemical Co.) as previously described (Mylchreest et al., 1999). Dose levels were selected to maximize fetal response with minimal maternal toxicity (Mylchreest et al., 1998, 1999, 2000). Body weights were recorded daily before dosing. Food consumption was monitored biweekly throughout the dosing period.

Six dams for each dose group were sacrificed at GD 16, 19, and 21 by carbon dioxide asphyxiation. Upon sacrifice, fetuses were immediately removed and weighed. Fetuses from GD 19 and 21 were sexed by measuring anogenital distance using a dissecting microscope with a micrometer lens (accuracy 0.05 mm). All fetuses from GD 16, and flutamide-exposed fetuses from GD 19 and 21, were sexed by internal examination. Testes and epididymides were isolated from male fetuses and either immediately snap-frozen for RNA isolation or fixed in buffered formalin for immunocytochemical analysis.

Microarray hybridization and analysis.

Total RNA was isolated from both testes of 1 male fetus per dam and 3 dams per treatment. Reverse transcription reactions were performed using 5 μg of total RNA, [³²P]-dATP, and superscript II MMLV-RT (Gibco BRL, Gaithersburg, MD) for 50 min at 50°C. Following purification, probes were counted, and equal numbers of cpm (minimum of 2 × 10⁶ cpm/ml) were added to each rat 588 gene cDNA expression array (Clontech, Palo Alto, CA). Arrays were hybridized with cDNA from 1 fetus per dam, and 3 arrays were hybridized per treatment group at GD 19 and 21. RNA was pooled from the testes from 5 fetuses per array, and 3 arrays were hybridized per treatment group at GD 16 (5 fetuses from a single dam per array, 3 arrays per treatment). Hybridization and washing were performed according to manufacturer's instructions. Digital images were collected on a BioRad phosphorimager (Hercules, CA) and analyzed using Clontech's Atlas Image software.

Within treatment groups, gene expression values for each array were logarithmically transformed (log2) and comparison between arrays presented as xy-scatterplots. All 3 arrays in each treatment group were analyzed in a pairwise fashion. Each dot on the xy-scatterplot represents 1 of the genes on the array with the x-value representing the expression value for that gene from 1 blot and the y-value representing the expression value for that gene from a second blot. The least-squares regression line y = ax + b is the line with the smallest sum of squared vertical distances between the points of the xy-scatterplot and the line. R-squared value is the Pearson coefficient of the least-squared regression line. This analysis will generate a straight line if the expression values of each individual gene are equal between 2 arrays. Variation in gene expression creates a spread of values along the linear regression line. The lower the correlation coefficient the greater the difference between the 2 arrays. For comparison between treatment groups, expression data from 3 array blots (3 control or treated) were first averaged to generate a single value for scatterplot analysis. Scatterplots were created and analyzed using the graphical and statistical program Prism (GraphPad, San Diego, CA).

Real-time PCR analysis of gene expression.

Total RNA (1 μg) was treated with DNase I (Amersham Pharmacia Biotech Inc., Newark, NJ) at 37°C for 30 min in the presence of RNasin (PE Applied Biosystems, Foster City, CA). Dnase I was heat inactivation at 75°C for 5 min. cDNA was synthesized using random hexamers and TaqMan reverse transcription reagents (PE Applied Biosystems) according to the manufacturer's suggested protocol. Real-time PCR (TaqMan) was performed on a Perkin-Elmer/Applied Biosystems 7700 Prism using SYBR Green according to the manufacturer's instructions for quantification of relative gene expression (User Bulletin # 2; P/N 4303859). Rat-specific primers (Table 1) were generated using Primer Express software, production of a single PCR product was confirmed using gel electrophoresis, and primer efficiency was determined according to manufacturer's recommended protocol (Applied Biosystems, Foster City, CA).

Immunohistochemical analysis of TRPM-2 and bcl-2 expression.

The right testis from randomly selected animals from different litters (4 to 7 per dose group) was immersed in 10% neutral-buffered formalin for 24 h, transferred to 70% ethanol, and embedded in paraffin; 5-μm sections were cut and mounted on slides. Immunohistochemistry of antigen-retrieved paraffin sections was performed as previously described (You and Sar, 1998). Deparaffinized sections were heated in a microwave oven for 3 or 6 min using citrate buffer (pH 5.5–5.7, BioGenex, San Ramona, CA). The sections were washed in PBS (pH 7.4) and then treated with 10% powdered nonfat milk in PBS followed by 2% normal rabbit serum for goat polyclonal antibody TRPM-2 or 2% normal goat serum for rabbit polyclonal antibody bcl-2 for 15 min each to reduce nonspecific staining. The sections were incubated with antigoat TRPM-2 antibody (5 μg/ml, Santa Cruz Biotechnology, CA) or antirabbit bcl-2 (1:1000 dilution, BD Pharmingen, San Diego, CA) at 4°C overnight. Sections were then incubated with the biotinylated secondary antibody, rabbit antigoat IgG, or goat antirabbit IgG and then avidin-biotin peroxidase (1:200, Vector Labs, Burlington, CA) for 30 min at room temperature. After a 5-min wash, the sections were treated with liquid diaminobenzidine (BioGenex) and counterstained with hematoxylin. The specificity of immunostaining was established by incubating the sections with preadsorbed antibody at a similar concentration to that of the primary antibody. For the preparation of the adsorbed antibody, primary antibody was incubated with the peptide antigen in 4-fold excess 24 h prior to use.

Radioimmunoassay of fetal testicular steroid concentration.

Fetal testicular testosterone, androstenedione, and progesterone steroid hormone concentrations were determined from individual fetuses using the method of vom Saal et al. (1990) with the following modifications. The procedure was the same for testosterone, androstenedione, and progesterone, with the appropriate steroid hormone standard, radiolabeled steroid hormone, and steroid hormone antibodies used for each assay. After homogenization of testes in 100 μl of PBS-gel buffer, the homogenate was extracted 3 times with a total of 1 ml of a fresh mixture of ethylacetate and chloroform (4:1). Extracts were dried under nitrogen and resuspended in 1 ml methanol. An aliquot (5 or 10 μl) was taken for analysis. An equal volume of extraction solvent was added to standards (0.25–128 pg of steroid hormone per tube) and recovery tubes (25 μl 3[H]steroid hormone; 5000 dpm) and dried under nitrogen. Dextran-coated charcoal (DCC) stripped serum (25 μl) was added to recovery tubes. Rabbit antisteroid hormone antibody (ICN) was diluted (1:800,000) with phosphate-buffered saline containing 0.01% gamma-globulin and 0.1% gelatin (PBS-Gel); 100 μl was added to each tube, gently mixed, and incubated overnight at 4°C. ¹²⁵I-steroid hormone (100 μl; 15,000 cpm) was added, and tubes were incubated for 4 h at room temperature. The second antibody (100 μl; goat antirabbit IgG diluted 1:9–1:11, ICN) was added, and tubes incubated for 1 h in a water bath at 38°C. Following addition of PBS-Gel (3 ml), tubes were centrifuged for 1 h at 1500 × g. The supernatant was decanted, the tubes blotted on absorbent paper, and the pellet counted for 2 min per tube in a Cobra gamma counter (D5005, Packard Instrument Co., Downers Grove, IL).

Statistics.

Each data point is an average of 3 animals per group, with each analysis performed in triplicate. Error bars represent the SE, with all values represented as -fold change compared to control treatment group average of 1.0. Significance was determined by one-tailed, nonpaired student's t-test with p < 0.05.

RESULTS

Microarray Analysis of Gene Expression

Microarray analysis was performed using radiolabeled cDNA isolated from control and DBP- and flutamide-exposed fetal testes. Testes from a single male fetus were used for each microarray, and 3 microarrays were hybridized per treatment group at GD 19 and 21. RNA was pooled from the testes of 5 male fetuses per array for each treatment group at GD 16. Individual microarrays were first compared within groups. Representative xy-scatterplots using logarithmic-transformed gene expression values between 2 arrays within each of the different treatment groups from GD 21 are presented in Figures 1A–1C. Correlation coefficients varied from 0.86 to 0.96, indicating a high degree of array reproducibility within treatment groups. Similar results were obtained with GD 19 arrays (data not shown). Scatterplot analysis within groups on GD 16 revealed considerable variability (data not shown); as a result, GD 16 blots were not considered for further analysis.

Average microarray values between treatment groups were then determined for GD 19 and 21. Logarithmic-transformed gene expression values from triplicate blots within groups were averaged, and the average gene expression values between groups were compared by xy-scatterplot. Results for GD 21 are presented in Figures 1D–1F. Correlation coefficients varied from 0.93 to 0.96, indicating that most of the 588 different genes examined by microarray were similarly expressed between the different treatment groups. Similar results were obtained with GD 19 array analysis (data not shown). Genes that had at least a 2-fold change in average expression values relative to average control values were identified and are listed in Table 2.

Quantitative RT-PCR

The exact role in the developing testes of the genes listed in Table 2 remains to be determined. However, many of the genes listed can be directly or indirectly associated with testicular steroidogenesis, cell proliferation, and cell survival. Expression of selected genes from Table 2, as well as some associated genes, was further analyzed by quantitative RT-PCR.

Genes examined by quantitative RT-PCR associated with steroidogenesis include cytochrome P450 side chain cleavage (P450scc), scavenger receptor class B1 (SR-B1), and myristoylated alanine-rich C-kinase substrate (MARCKS). Additional steroidogenic genes not on the array but also examined include cytochrome P450c17 (P450c17) and steroidogenic acute regulatory protein (StAR). P450scc mRNA declined almost 2-fold relative to control values at GD 19 in DBP-exposed fetal testes (Fig. 2A). In contrast, P450scc mRNA doubled at GD 19 in flutamide-treated fetal testes. P450scc mRNA values returned to control levels by GD 21 in both DBP and flutamide-treated fetal testes. P450c17 was reduced approximately 2-fold in GD 19 DBP-exposed fetal testes (Fig. 2B). P450c17 mRNA values in DBP-exposed fetal testes returned to control levels by GD 21. P450c17 mRNA values increased in flutamide-exposed fetal testes greater than 2-fold above control by GD 21. StAR mRNA was significantly downregulated in DBP-exposed testes on GD 16, 19, and 21 (Fig. 2C). In flutamide-exposed testes, StAR mRNA was significantly downregulated on GD 16 and 21 and upregulated on GD 19. SR-B1 mRNA was significantly reduced in DBP-exposed testes on GD 16, 19, and 21 (Fig. 2D). SR-B1 mRNA in flutamide-exposed testes was significantly downregulated on GD 21. MARCKS mRNA was induced 4-fold and 2-fold by flutamide on GD 19 and 21, respectively (Fig. 2E). MARCKS mRNA expression was not significantly altered with DBP-treatment.

Genes from Table 2 associated with cell proliferation and cell survival that were examined by quantitative RT-PCR include testosterone-repressed prostate message-2 (TRPM-2), proliferating cell nuclear antigen (PCNA), and stem cell factor receptor (c-kit). TRPM-2, also known as clusterin, was increased over 2-fold above control values in DBP-treated fetal testes on GD 21 (Fig. 2F). In contrast, TRPM-2 expression values from flutamide-exposed fetal testes were not significantly different from control values. PCNA mRNA expression levels were increased 2-fold in flutamide treated fetal testes on GD 16 and 19 and decreased almost 2-fold on GD 21 (Fig. 2G). PCNA values were not significantly altered in DBP-exposed fetal testes. C-kit was reduced approximately 7-fold (14% of control value) on GD 19 in DBP-exposed fetal testes (Fig. 2H). C-kit mRNA values returned to control levels by GD 21 in DBP-exposed fetal testes. C-kit gene expression values in flutamide-exposed fetal testes were not significantly different from control values.

Additional genes examined by either semiquantitative or quantitative RT-PCR not significantly altered in expression in response to DBP or flutamide include androgen receptor (AR), steroidogenic factor-1 (SF-1), epidermal growth factor, epidermal growth factor receptor, glutathione peroxidase, and superoxide dismutase (data not shown).

Radioimmunoassay

Radioimmunoassay analysis of fetal testes revealed a reduction in both testosterone and androstenedione in DBP-treated fetal testes on GD 19 and 21 (Figs. 3A and 3B). Progesterone, a precursor to androstenedione in the steroidogenic pathway, was increased in DBP-exposed testes on GD 19 and not different from control or flutamide values on GD 21 (Fig. 3C).

TRPM-2 and Bcl-2 Immunocytochemistry

An increase in the intensity and number of cells expressing TRPM-2 protein was observed in GD 21 DBP-exposed fetal testes (Figs. 4A–4C). The immunoreactivity localized primarily to Sertoli cells and Leydig cells. A similar staining pattern was observed with GD 19 testes (data not shown). A similar pattern of immunostaining was also observed with bcl-2 antibody in DBP-exposed fetal testes (Figs. 4D–4F). PCNA immunostaining was also performed and showed no change (data not shown). The TRPM-2 and bcl-2 immunostaining in testes was specific since no positive immunoreactivity was detected when preadsorbed antibody was used instead of the primary antibody.

DISCUSSION

The antiandrogenic effects of DBP on the developing male fetus include reduced anogenital distance, nipple retention, hypospadias, and testicular toxicity (Mylchreest et al., 1998, 1999). These effects resemble those caused by AR antagonists such as flutamide (Mylchreest et al., 1998, 1999). However, neither DBP nor its primary metabolite MBP interact with the AR in vitro (Foster et al., 2001). In addition, DBP is distinct from flutamide in its ability to induce fetal Leydig cell hyperplasia and primordial germ cell (gonocyte) degeneration (Mylchreest et al., 1999). Together, these results suggest that DBP acts without direct interaction with the AR to induce antiandrogenic effects as well as its own unique toxic responses. In this study, we demonstrate that DBP alters expression of a number of genes in the developing fetal testes. The pattern of DBP-mediated gene expression alterations is distinct and yet overlaps with those produced by flutamide. Thus DBP and flutamide alter both common and distinct molecular pathways within the developing fetal testes.

DBP-exposed fetal testes have a significant reduction in testosterone concentration, and this may be due in part to down-regulation in the expression of genes linked directly or indirectly to steroidogenesis. Steroidogenic genes downregulated on GD 19 in DBP-exposed fetal testes include P450scc, P450c17, StAR, and SR-B1 (Fig. 5). P450scc cleaves the side chain of cholesterol to form pregnenolone, which is both the initial step as well as the rate-limiting enzymatic step in steroid biosynthesis (Omura and Morohashi, 1995). P450c17 has both 17α-hydroxylase and 17,20-lyase activity and converts progesterone to 17-OH progesterone and then to androstenedione (Omura and Morohashi, 1995). Under normal conditions, steroidogenic enzymes are expressed in excess, and delivery of cholesterol to P450scc is generally the rate-controlling process in testosterone synthesis (Crivello and Jefcoate, 1980; Privalle et al., 1983; Temel et al., 1997). Cholesterol uptake into the cell is mediated by SR-B1, also known as high-density lipoprotein receptor (Acton et al., 1996). StAR performs a similar function to mediate the transfer of cholesterol from the outer to the inner mitochondrial membrane. P450scc is located on the matrix side of the inner mitochondrial membrane (Stocco, 1999). Downregulation of either SR-B1 or StAR results in a reduction in steroidogenesis (Leers-Sucheta et al., 1999; Temel et al., 1997; Walsh et al., 2000; Walsh and Stocco, 2000).

Luteinizing hormone secreted by the pituitary is the primary regulator of testosterone synthesis in the Leydig cells of the testes. However, initiation of testosterone synthesis by the fetal Leydig cell from GD 16 to 19 occurs in the absence of luteinizing hormone and is due to undefined paracrine factors produced within the fetal testes (El-Gehani et al., 1998). Luteinizing hormone does not reach levels within the testes capable of regulating testosterone synthesis until after GD 19.5 (El-Gehani et al., 1998). Luteinizing hormone receptor (LHR) mRNA, as well as the mRNA of an associated regulatory protein, β-arrestin 2, are downregulated in GD 21 DBP-exposed fetal testes by microarray analysis (Table 2). β-arrestin 2 is a scaffold protein that facilitates LHR signal transduction by bringing LHR together with components of the mitogen-activated protein kinase pathway upon LHR activation (McDonald et al., 2000; Mukherjee et al., 2000). These results indicate a general reduction in gene expression of proteins involved in testosterone synthesis through both LH-dependent and independent mechanisms following DBP exposure.

Both DBP and flutamide downregulated expression of genes involved in lipogenesis (fatty acid and cholesterol synthesis), including long chain specific acyl-CoA, acetyl-CoA carboxylase, steryl sulfatase, and low-density lipoprotein receptor as determined by microarray analysis (Table 2). Androgens upregulate fatty acid and cholesterol synthesis through a cascade of events involving androgen-dependent activation of sterol regulatory element binding proteins (SREBP) (Brown and Goldstein, 1998; Swinnen et al., 1998, 1997). Downregulation of genes involved in lipogenesis is likely the result of reduced AR signaling, due to decreased testicular testosterone levels in DBP-exposed testes and AR blockade in flutamide-exposed testes (Fig. 5).

Despite a similar downregulation in gene expression of proteins involved in intracellular cholesterol synthesis, testosterone levels were not reduced in flutamide-exposed fetal testes. Induction of steroidogenic enzyme gene expression in flutamide-exposed fetal testes may compensate for the reduction in intracellular cholesterol synthesis. P450scc gene expression was induced over 2-fold on GD 19 by flutamide and P450c17 was induced 6-fold on GD 21. Flutamide-exposed testes had enhanced expression of MARCKS (Fig. 2E), a protein kinase C (PKC) substrate that mediates PKC signaling through phosphorylation and subsequent association with filamentous actin and calmodulin (Ramsden, 2000). Actin cytoskeletal rearrangement by MARCKS enhances cholesterol transport to the mitochondria (Betancourt-Calle et al., 1999). StAR mRNA expression is also enhanced on GD 19, although StAR mRNA is significantly reduced on GD 21 compared with control. Enhanced cholesterol transport together with increased levels of P450scc and P450c17 may allow the flutamide-exposed testes to compensate for reduced de novo cholesterol synthesis and maintain testosterone synthesis at normal levels. In contrast, DBP-exposed fetal testes had a combined reduction in gene expression of cholesterol synthesizing enzymes and steroidogenic enzymes as well as a reduction in gene expression of SR-B1 and StAR, which are involved in transport of cholesterol into the steroidogenic pathway. This combined effect on cholesterol synthesis, transport, and steroidogenesis in DBP-exposed fetal testes may explain the reduction in fetal testicular testosterone in DBP-exposed, but not flutamide-exposed, testes (Fig. 5).

Leydig cell proliferation and differentiation are permanently altered by developmental exposure to DBP (Mylchreest et al., 1999). Regions of Leydig cell hyperplasia are present in 25% of the fetuses at GD 19 and 21 exposed to 500 mg/kg/day DBP (Mylchreest et al., 1999). A small percentage of exposed fetuses develop adenomas by 3 months of age, an extremely rare event in young adult Sprague-Dawley rats (Mylchreest et al., 2000). PCNA, a general marker of proliferation, is not elevated in DBP-exposed testes at either the level of mRNA (Fig. 2G) or protein (data not shown). However, DBP exposure did result in the increased expression of 2 proteins involved in cell survival, TRPM-2 and bcl-2 in Sertoli cells and Leydig cells (Fig. 4). TRPM-2, also known as clusterin, was originally believed to be associated with apoptosis because expression is upregulated in the regressing prostate following androgen ablation (Montpetit et al., 1986). More recent studies show that TRPM-2 expression is reduced in cells undergoing apoptosis and enhanced in surviving cells (French et al., 1994; Koch-Brandt and Morgans, 1996). In addition, TRPM-2 has been shown to inhibit apoptosis and enhance survival of cells in culture (Miyake et al., 2000a,b,c; Zwain and Amato, 2000). Enhanced bcl-2 expression is also associated with cell survival in the testes (Beumer et al., 2000; Fujisawa et al., 2000; Suzuki et al., 1996). Together, these results suggest that regions of fetal Leydig cell hyperplasia may be due, in part, to inhibition of apoptosis and enhanced cell survival.

Primordial germ cells (gonocytes) in DBP-exposed fetal testes undergo degeneration as indicated by a reduction in cell number together with the presence of large multinucleated cells (Mylchreest et al., 1999). DBP exposure resulted in the specific downregulation of stem cell factor receptor (c-kit) mRNA (Fig. 2H), which is expressed exclusively in gonocytes in the developing testes (Kissel et al., 2000; Mauduit et al., 1999; Ohta et al., 2000). Stem cell factor is produced by the Sertoli cells, and regulates gonocyte migration, proliferation, and survival (Mauduit et al., 1999; Ohta et al., 2000) by interacting with its receptor (c-kit) on the gonocyte cell surface and initiating a kinase cascade involving rac-beta serine/threonine kinase (Blume-Jensen et al., 1998, et al., 2000; Feng et al., 2000; McCubrey et al., 2000). Neurofibroma factor-1 (NF-1) modulates c-kit signaling (Blume-Jensen et al., 1998, 2000; Feng et al., 2000). Both rac-beta serine/threonine kinase and NF-1 were downregulated in DBP-exposed fetal testes on GD 21 (Table 2). The coordinate downregulation of c-kit together with NF-1 and rac-1 following DBP exposure suggest that reduced stem cell factor signaling may be a factor in germ cell degeneration following DBP exposure.

Microarray data and quantitative RT-PCR data do not exactly match for all genes compared under both methods. For example, MARCKS was upregulated greater than 2-fold by microarray analysis in both DBP and flutamide-exposed testes. However, MARCKS was upregulated only in flutamide-exposed testes by quantitative RT-PCR. The reasons for these differences in results between these 2 methods have not been determined. Microarrays provide the advantage of being able to investigate expression of hundreds of genes at a time. Combined with cluster analysis, the effect of a chemical on multiple molecular pathways can be determined. However, microarrays are less sensitive and less quantitative than RT-PCR. RT-PCR allows for the accurate quantitation of individual genes but is a labor-intensive procedure and requires more mRNA per gene analysis than microarrays. Under limiting conditions, and given the state of the art of the different technologies, microarrays currently provide the best overall approach for identifying multiple biological pathways targeted by a chemical. Our results underscore the qualitative aspect of microarrays and the need for confirmation of microarray results through additional methods including cluster analysis of genes known to be coordinately regulated, quantitative RNA and protein analysis, and through linking gene expression alterations with ultimate biological response in target tissues.

In summary, we demonstrate that DBP exposure in utero results in altered expression of a number of genes, including those involved in testosterone synthesis and cell survival. The role of many of these genes in fetal testis development and DBP toxicity remains to be determined. Additional studies are necessary to examine dose-response relationships for gene expression alterations and DBP toxicity. We confirmed by quantitative PCR the expression of a select group of genes known to play a significant role in testicular development. Alteration in expression of these genes can be linked to the actions of DBP in the developing male reproductive tract, including reduced testicular testosterone, Leydig cell hyperplasia, and gonocyte degeneration. The pattern of gene expression is distinct from and yet overlaps with gene expression alterations in flutamide-exposed testes. The mechanism by which DBP induces these changes in gene expression remains to be determined. The diverse cell populations within the fetal testes as well as the dramatic changes in gene expression and cell function that occur normally during fetal testicular development complicates efforts to understand the mechanism of DBP-induced male reproductive toxicity. Future efforts will examine alterations in gene expression and protein expression in specific cell populations within the fetal testes.

The authors thank Kim Lehmann, Duncan Wallace, and Drs. Katrina Waters and Scott Boley for their technical assistance and helpful discussions and Dr. Barbara Kuyper for editorial assistance.

REFERENCES