Activation of Penile Proadipogenic Peroxisome Proliferator-Activated Receptor γ with an Estrogen: Interaction with Estrogen Receptor Alpha during Postnatal Development

Exposure to the estrogen receptor alpha (ERα) ligand diethylstilbesterol (DES) between neonatal days 2 to 12 induces penile adipogenesis and adult infertility in rats. The objective of this study was to investigate the in vivo interaction between DES-activated ERα and the proadipogenic transcription factor peroxisome proliferator-activated receptor gamma (PPARγ). Transcripts for PPARs α, β, and γ and γ1a splice variant were detected in Sprague-Dawley normal rat penis with PPARγ predominating. In addition, PPARγ1b and PPARγ2 were newly induced by DES. The PPARγ transcripts were significantly upregulated with DES and reduced by antiestrogen ICI 182, 780. At the cellular level, PPARγ protein was detected in urethral transitional epithelium and stromal, endothelial, neuronal, and smooth muscular cells. Treatment with DES activated ERα and induced adipocyte differentiation in corpus cavernosum penis. Those adipocytes exhibited strong nuclear PPARγ expression. These results suggest a biological overlap between PPARγ and ERα and highlight a mechanism for endocrine disruption.


INTRODUCTION
Endocrine disruption, originally limited to steroid receptor signaling, now extends to include other members of the 48 reported nuclear receptor superfamily [1]. Both peroxisome proliferator-activated receptor gamma (PPARγ) and estrogen receptor alpha (ERα) are targets for endocrine disrupting chemicals [2][3][4]. Recently, Goyal et al. showed that neonatal exposure of rats to the estrogenic endocrine disruptor diethylstilbestrol (DES) induced adipogenesis in penile corpus cavernosum by activation of ERα [5][6][7][8]. In this model of DES-ERα activation, DES exposure at a dose of 0.1 to 0.12 mg/kg bw/day, on alternate days, from postnatal days 2 to12, resulted in infertility in 100% of the treated male rats. Loss of fertility was associated with abnormal accumulation of fat cells in the corpus cavernosum penis, and the associated loss of cavernous spaces apparent as early as postnatal day 18 (reviewed in [9]). It remains unknown, however, whether this penile ERα-induced adipogenesis is mediated by activation of a constitutively expressed or DES-induced PPARγ.
Both ERα and PPARγ pathways are implicated in fat regulation. First, recent findings suggest that PPARγ and ERα pathways involve shared coactivators that promote differentiation of preadipocytes into mature fat cells. For example, constitutive coactivator of PPARγ (CCPG) is described as a bona fide coactivator that cross reacts with ERα independent of its ligand and contains four LXXLL motifs that are characteristic of nuclear receptor coactivators [10]. Second, studies have shown that forced expression of PPARγ2 or PPARγ1 can trigger the differentiation of fibroblasts to adipocytes resulting in the activation of adipocyte-specific genes and lipid accumulation [11].

PPAR Research
The PPAR family consists of three isotypes that include PPARα (NR1C1), PPARβ (also known as PPARδ, NR1C2, FAAR, or NUC-1), and PPARγ (NR1C3) [12][13][14]. A nuclear receptor, PPARγ, is known to play a central role in fat metabolism and adipocyte differentiation [15,16]. The PPARγ is present in two key isoforms, PPARγ1 and PPARγ2. The two isoforms stem from alternate promoters [17]. Compared to PPARγ1, PPARγ2 has an additional 30 amino acids at the N-terminal end and is distinctively expressed in adipose tissue, where it plays a key role in adipogenesis [18]. These nonsteroidal receptors (i.e., do not mediate effects of steroids) form part of a class I nuclear hormone receptor superfamily [19] and function as ligand-activated transcription factors [20][21][22].
Each of the three PPAR isotypes is constitutively expressed in certain reproductive and nonreproductive rat tissues [23,24], but their temporal and cell-specific expression in penile tissue, with the exception of a limited demonstration of PPARγ in penile corporal smooth muscle cells [25], has not been shown. Further, no specific link is known between neonatal activation of ERα and penile PPARγ. This is important given the expanded definition of the term endocrine disruptors to include activation of metabolic sensors such as PPARs. A number of findings suggest involvement of PPARs in endocrine disruption either through direct receptor activation or indirectly through crosstalk with other nuclear receptors. First, in vitro studies demonstrated that PPARγ and ERα (the iconic receptor involved in endocrine disruption) are implicated in cross-talk [26][27][28]. Second, some endocrine disruptor chemicals, such as monethylhexyl phthalate (MEHP), a primary metabolite of diethylhexyl phthalate (DEHP), mediate their toxic effect by PPARγ activation [29,30]. Third, several nonbiological xenobiotics compounds can activate PPARγ. For example, activation of PPARγ with synthetic PPARγ activators, such as antidiabetic drugs thiazolidinediones (TZDs), improve insulin sensitivity but they undesirably increase preadipocyte differentiation and white adipose tissue mass [31][32][33]. Consistent with this adipogenic effect, reduced PPARγ level, as in mice with heterozygous (PPARγ +/− ) deficiency, is associated with reduced white adipose tissue mass [34].
Findings related to interaction between ERα and PPARγ in the aforementioned DES-penile rat model will illuminate a potential molecular mechanism by which estrogen exposure at critical period of development perturbs reproductive tissues. Therefore, we hypothesize that DES-induced penile adipogenesis is associated with ERα-mediated activation of PPARγ. Objectives of this study were to (1) determine the basal expression of PPARs (α, β, and γ) in rat penis and (2) evaluate the neonatal modulatory effect of ERαactivator DES on penile PPARγ as a marker of undesirable adipogenesis.

Animals and treatments
This DES study was performed in collaboration with Dr. Hari Goyal at Tuskegee University using male pups from pregnant female Sprague-Dawley (SD) rats (Harlan Sprague-Dawley, Indianapolis, Ind, USA). All animal procedures were approved by Institutional Animal Care and Use Committee at Tuskegee University. In all experiments, rats were maintained using standard housing conditions including constant temperature of 22 • C, ad libitum water and feeding, and 12:12 hours light dark cycle. Two experiments were conducted. In experiment 1, three groups of male pups (n = 5 per group, all were littermates) received subcutaneous injections of 25 μL of olive oil (control), oil containing DES (0.1 mg/kg, Sigma-Aldrich, St. Louis, Miss, USA), or DES plus ICI 182, 780 (16.6 mg/kg, ICI; Tocris Bioscience, Ellisville, Miss, USA) daily on postnatal day 2 to 6. Rats in experiment 1 were sacrificed at 28 day of age. ICI 182, 780 is a high-affinity estrogen receptor antagonist (IC 50 = 0.29 nM) and is also considered a high-affinity ligand for the membrane estrogen receptor GPR30 (Tocris Bioscience). In experiment 2, two groups of male pups (n = 4 per group) received DES (1 mg/kg) or olive oil (control) every other day for 6 days starting at postnatal day 2. Penile tissues were collected from rats sacrificed at 120 days of age (adulthood). Small sections of the penile shaft tissue from each rat in experiment 1 and 2 were fixed overnight in 4% paraformaldehyde for IHC or fat staining, and the remainder of the shaft tissue was frozen in liquid nitrogen and stored at −80 • C for RNA extraction and PCR analysis. The doses used for end-point evaluation at 28 and 120 days post-treatment were based on previous publications from our group that showed DES prenatal exposure (between postnatal days 2 to12) at a dose range of 0.1 to 0.12 mg/kg/day, or higher (1 mg/kg/day) result in similar abnormal penile development and adipogenesis [5,8].

Total RNA isolation
Total RNA was isolated from the body of the penis using TRIZOL reagent (Invitrogen-Life Technologies Inc., Carlsbad, Calif, USA), according to the manufacturer's protocol. RNA concentrations were estimated at 260 nm and the ratio of 260/280 was determined using UV spectrophotometry (DU640, Beckman Coulter Fullerton, Calif, USA). The integrity of each RNA sample, indicated by the presence of intact 28S and 18S ribosomal RNA, was verified by denaturing agarose gel electrophoresis. RNA samples were treated with DNase (Ambion Inc.) to remove possible genomic DNA contamination. Samples with 260/280 ratio of ≥1.8 were used.

Conventional end-point and real-time PCR
Expression of mRNA for PPAR (α, β, and γ) isotypes was initially determined by conventional end-point RT-PCR with primers designed using primer quest software and synthesized by Integrated DNA Technology (IDT Inc, Coralville, Iowa, USA) from previously published rat sequences (see Table 1). Subsequently, semiquantitative RT-PCR for coamplification of PPARs and S-15 (known as Rig; small subunit ribosomal protein used as a house keeping gene) was performed to determine the relative expression levels of Table 1: PCR primer sets, sequence, product size (bp), nucleotide (nt) location, and GenBank accession numbers for rat PPARs used in this study. Note that a common antisense oligoprimer (sequence in bold) was used for PPARγ1a, PPARγ1b, and PPARγ2. PPAR isotypes. Verification of accurate PCR products was confirmed by determination of the expected size of PCR bands and by sequence analysis of generated amplicons at Auburn University sequencing facility. The resulting sequences for the three PPAR isotypes were matched with previously published rat sequences in GenBank (accession number NM013196, U40064, and NM013124 for PPARα, PPARβ, and PPARγ, resp.) using Chromas 2.31 software (Technelysium Pty ltd, Tewantin Qld 4565, Australia). PPARγ splice variants or subtypes were identified using specific primers designed for rat PPARγ1a and PPARγ1b synthesized by IDT Inc. (Table 1). Liver and white adipose tissues from adult Sprague-Dawley rats in experiment 2 were used as positive controls for PPARγ1 [35] and PPARγ2 [18], respectively. The amplification protocol was as follows: initial cycle for 3 minutes at 95 • C, and 30 cycles each at (95 • C for 30 seconds, 55 • C for 30 seconds, and 72 • C for 30 seconds) followed by a final extension cycle at 72 • C for 7 minutes. PCR reactions were performed on a Robocycler (Stratagene Inc, La Jolla, Calif, USA) and products were analyzed electrophoretically on 2% (w/v) agarose gels. The intensity of the PCR bands was determined using Fluor-S multi-imaging analysis system (Bio-Rad, Hercules, Calif, USA). Level of mRNA for PPARs was normalized to the levels of S-15 housekeeping gene. Quantitative real-time PCR (Bio-Rad, MyiQ TM ) for determination of expression levels of PPARγ and ERα mRNA was performed in 25-μL reaction mixture containing 12.5 μL RT 2 real-time SYBR/Fluorescein Green PCR master mix, 1 μL first strand cDNA, 1 μL RT 2 validated PCR primer set for PPARγ or ERα (Super Array Bioscience Corporation, Frederic, Md, USA), and 10.5 μL PCR-grade water (Ambion Inc). Samples were run in 96-well PCR plates (Bio-Rad, Hercules, Calif, USA) in duplicates, and the results were normalized to GAPDH (see primer set in Table 1) expression. The amplification protocol was set at 95 • C for 15 minutes for one cycle, and 40 cycles each at (95 • C for 30 seconds, 55 • C for 30 seconds, and 72 • C for 30 seconds) followed by melting curve determination between 55 • C and 95 • C to ensure detection of a single PCR product. Template RNA from rat white adipose tissue and penis were used for determination of amplification efficiencies for (ERα/PPARγ) targets and GAPDH by generating standard curves. Curves were generated by using serial 10-fold dilutions total RNA 4 PPAR Research and plotting the log dilution against C T (threshold cycle) value obtained for each dilution. The Pearson's correlation coefficient (r) value for each generated standard curve was ≥0.98, and the calculated amplification efficiency was between 98.5 to 99%.

Immunohistochemistry (IHC)
Immunolocalization of PPARγ in penile tissue was performed using mouse anti-PPARγ IgG1 monoclonal antibody (sc7273, Santa Cruz Biotechnology Inc, Santa Cruz, Calif, USA) raised against a C-terminus sequence of human and mouse PPARγ (similar to the corresponding rat sequence). The antibody detects PPARγ1, PPARγ2 and, to a lesser extent, PPARα and PPARβ of rat, mouse, and human by IHC using paraplast-embedded tissues. Approximately 5-mmlong penis sections from the middle of the body of the penis were fixed in 4% paraformaldehyde for 48 hours, embedded in Paraplast (Sigma-Aldrich), and cut at 5-μm thickness [7]. Mounted penis sections were deparaffinized in Hemo-D (Scientific Safety Solvents, Keller, Tex, USA) and hydrated to distilled water (dH 2 O). The slides were transferred to a rack and placed in 1 L of 10 mM sodium citrate (pH 6.0). The beaker was placed on a hot plate, allowed to come to a boil and tissues were boiled for 20 minutes. When the citrate solution cooled to near room temperature (RT), the slides were transferred to a glass staining dish and equilibrated in phosphate buffered saline (PBS) (Sigma-Aldrich, ST Louis, Miss, USA). After 20 minutes incubation in blocker (5% normal goat serum, Sigma-Aldrich) and 2.5% BSA (Sigma) in PBS, slides were washed briefly in PBS. Anti-PPARγ, diluted 1:20 in blocker, was applied and the sections were left to incubate overnight at RT. Next day, slides were washed 3x in PBS, 3 minutes each, and tissues were incubated with Alexa 488-conjugated goat antimouse IgG (Molecular Probes, Eugene, Ore, USA) for 1 hour at RT. After washing two times in PBS, 3 minutes each, slides were mounted with VectaShield (Vector Laboratories, Burlingame, Calif, USA), and the coverslips were sealed. The sections were examined using a Nikon TE2000E microscope and digital images were generated using an attached Retiga EX CCD digital camera (Q Imaging, Burnaby, BC, Canada). Penile tissue sections from all 28-day treated rats were examined. Representative micrographs from different penile histological structures were shown for untreated control rats, and for rats treated with DES or DES + ICI.

Fat staining
Histochemical demonstration of fat was performed as previously described [7]. Briefly, tissue sections from penile body, approximately 5 mm-long, were fixed for 24 hours in 4% formaldehyde, followed by en bloc staining of fat for 8 hours with 1% osmium tetroxide dissolved in 2.5% potassium dichromate solution. Specimens were then processed for paraplast embedding and cut at 5-μm thickness. Deparaffinized sections were examined for black staining indicative of fat cells using light microscopy.

Statistical analyses
Analysis of real-time PCR data for relative gene expression level (fold change of target relative to control) was performed using a modification of the delta delta Ct method (ΔΔ CT) as described previously [36]. Statistical differences between treatment groups were performed using Sigma Stat statistical software (Jandel Scientific, Chicago, Ill, USA). Δ CT for real-time PCR data [37], and intensity values (for semiquantitative RT-PCR data) were subjected to analyses of variance. Experimental groups with means significantly different (P < .05) from controls were identified using Holm-Sidak and Tukey tests. When data were not distributed normally, or heterogeneity of variance was identified, analyses were performed on transformed or ranked data.

Detection and sequence analysis of PPAR and ERα transcripts in the body of the penis
Primer sets used in this study are shown in Table 1.
Transcripts for three PPAR isoforms (α, β, and γ) were detected, albeit at different levels, in penile tissue from normal control adult (120 days) rats (Figure 1, parts A1 and A2). Semiquantitative RT-PCR analysis of PPARs indicated predominant expression of PPARγ mRNA when compared with PPAR (α and β) isoforms (Figure 1(B)). Sequence analysis and alignment with published sequence data confirmed the identity of all three PPAR isoforms. Treatment with DES induced over three-fold-increase (3.38) in ERα transcripts in 28-day-old rats compared to over two-foldincrease (2.5) in 120-day-old adult rats when each age group was compared with its respective untreated controls ( Figure 2). Similarly, DES induced slightly over seven-foldincrease (7.1) in PPARγ transcription level in 28-day-old rats compared with over six-fold-increase (6.8) in 120day-old adult rats ( Figure 3). The upregulation of PPARγ expression by DES in 28-day-old rats was abrogated when rats were cotreated with DES and ICI 182, 780 ( Figure 4). The differences in the transcriptional level of penile ERα and PPARγ between the DES-treated rat groups (28 versus 120day-old rats) were not significantly different. Because of the relatively high expression of penile PPARγ in the 28-day-old rats subsequent studies for determination of splice variants and PPARγ protein expression were performed in the 28day-old rats.

Detection of PPARγ splice variants and real-time PCR data
In order to determine which PPARγ splice variant is expressed in the body of normal and DES-treated rats, primers (Table 1) were designed to amplify the two known rat PPARγ1a and PPARγ1b splice variants using conventional end-point RT-PCR. Splice variant analyses revealed expression of PPARγ1a in normal 28-day-old rat penis. However, in addition to PPARγ1a, PPARγ1b and PPARγ2 were newly induced by DES treatment (Figure 5).

Immunohistochemistry and fat staining
Immunohistochemistry results revealed PPARγ protein localization in transitional epithelium of the urethra, and the surrounding corpus spongiosum penis. It is also expressed in stromal, endothelial, neuronal, and smooth muscular cells of the cavernous sinuses located in the corpus cavernousm region of normal 28-day-old rat penis (Figures 6(a) and 6(b)). Treatment with DES induced a strong staining intensity for PPARγ protein in the peripherally located nuclei of newly induced adipocytes (Figure 6(a), Panel (c) with a magnified inset-box view in C2). PPARγ immunostaining was markedly reduced by ICI 182,780 treatment ( Figure 6(b)). In unstained penile sections from 28-day-old and adult DES-   treated rats, the new adipocytes were seen as empty spaces similar to fat cells and were specifically localized in the corpus cavernosum region of the penis (Figure 7, panels (b) and (d)). In addition, staining with 1% osmium tetroxide confirmed that the empty spaces were cluster of fat cells (stained as black granules in Figure 7, panels (c) and (e)). No fat cells were seen in penile sections from rats treated with DES + ICI (Figure 7, panels (f) and (g)).

DISCUSSION
This study demonstrated that three PPAR transcripts (α, β, and γ) are constitutively coexpressed in normal rat penis   with PPARγ as the predominant isotype. In addition, it established that some ERα synthetic ligands, such as DES, can activate PPARγ subtypes when administered at early perinatal days. Further, upregulation of ERα by DES was associated with a corresponding increase in PPARγ suggesting a synergistic interaction between the two receptors. Previous studies that used in situ hybridization to determine the distribution of PPARs in rat tissues, including reproductive organs, showed expression of PPARα and PPARβ in somatic (Sertoli and Leydig) and in germ cells of the testis, but did not address expression of these two receptors in penile tissue [23,24]. The role of PPARα and PPARβ in the testis, however, remains unknown. Detailed study addressing expression of PPARγ isotypes in penile tissue is also lacking, with the exception of a study that showed limited penile PPARγ expression in corporal smooth muscle cells [25].
In this study, PPARγ and PPARγ1a were detected in normal rat penis. However, DES as ERα activator distinctively induced expression of PPARγ1b and PPARγ2 splice variants that were not present in control untreated penile tissue. The induction of splice variant PPARγ1b is in agreement with previous in vitro studies that demonstrated activation of PPARγ1 by the endocrine disruptor monoethyl-hexylphthalate in C2C12 mouse skeletal muscle cell line [2], and with MCF-7 breast cancer cells stimulated with E2, the natural ERα ligand [38]. Further, the induction of PPARγ2 concurs with increased adipogenesis observed in the corpus cavernousm penis as PPARγ2 is considered a unique marker for mature adipocytes, and its forced induction is associated with terminal differentiation of preadipocytes or fibroblast cells to functional mature adipocytes [11,22]. The upregulation of PPARγ was abrogated by coadministration of the type-II antiestrogen ICI 182,780, indicating that DES effects were mediated, at least in part, via the estrogen receptor system. It is possible, however, that ICI may have directly repressed activation of PPARγ as ICI was previously shown to inhibit the action of the selective PPARγ agonist BRL 48, 482 in MDA-MB 231 breast cancer cell culture in the absence of ER [38].
One important difference between this study and previous in vitro studies that addressed signal cross-talk between PPARγ and ERα using MCF-7 cells [38][39][40] is that the activation of ERα by DES in our study is associated with selective induction of PPARγ1b and PPARγ2. This unique effect resulted in generation of de novo adipocytes that provide direct functional proof for PPARγ2 induction. In contrast to our study, activated ERα by E2 lowers both basal and ligand-stimulated PPARγ-mediated gene reporter activity in MCF-7 cancer cell culture [38]. Likewise, activation of PPARγ in MCF-7 cell culture with the natural PPARγ ligand cyclopentenone 15-deoxy-Δ 12,14 prostaglandin J2 (15d-PGJ2) inhibited estrogen-responsive elements [40]. Consequently, the MCF-7 cell culture studies suggest that ERα and PPARγ negatively regulate each other. The reason for the difference between our study and the aforementioned in vitro data could be related to differences between in vitro and in vivo milieu or to the deletional mutants used in the in vitro studies compared with the in vivo wild type receptors in our study. Another reason for the disagreement could be due to differences in coactivators and corepressors present in MCF-7 and penile tissue cells or more importantly to differences in the ligands used. One plausible hypothesis, however, for the increased transcriptional activation of PPARγ1b and PPARγ2 by DES-activated ERα is that exposure of rats to DES at a critical neonatal period of days 1 to 12 is uniquely associated with reprogramming of penile stromal or preadipocytes to mature adipocytes [5][6][7][8].
In support of this concept, it is known that postnatal days 1 to 5 in rodents coincide with a period for reproductive tract and adipocyte differentiation [41]. Further, data from other laboratories indicated that neonatal exposure of rodents to DES is associated with increased whole body fat at adulthood 8 PPAR Research [42]. This novel adipogenic effect of DES was proposed as a model for the study of what is called developmental obesity mediated by early exposure to endocrine disruptors [43].
The molecular mechanism involved in DES-ERα-PPARγ transactivation could be related to two factors. First, activated-ERα could directly bind to PPAR response elements (PPREs) because the two receptors share the capacity to bind to the AGGTCA half-sites consensus sequences contained as palindrome or direct repeat in estrogen response elements (EREs) and PPRE sequences, respectively [44]. This mechanism could result in bidirectional activation of shared target sequences between ERα and PPARγ depending on activated receptor involved. Second, it is known that estrogen could induce enzymatic conversion of prostaglandin D2 (PGD2) and the endogenous metabolites of the latter can directly activate PPARγ [45]. The latter effect, however, was not associated with induced PPARγ mRNA [46] suggesting that the first mechanism could be in play in our study.
The strong PPARγ protein expression in normal transitional epithelium of the urethra and the dorsal artery and vein of the penis indicates possible physiological role for PPARγ in the penis vasculature and the urothelium of the urinary tract. Although this study did not address functionality of PPARγ in the penis, current evidence suggests that its constitutive expression in some tissues is linked to eicosanoids and prostaglandins (PGs) actions [47,48]. In this regard, the terminal metabolite of the J series of PG, 15d-PGJ2, is considered the natural activator of PPARγ [48]. Sources of penile PGs could include synthesis by local penile cells and/or cells of the renal medulla where PGs can be transported via the ureter and pelvic urethra to the penis [49]. Among other functions, PGs are important mediators of inflammation, vascular homoeostasis, and pain all of which may be relevant to the pathophysiology of the penis.
Staining with osmium confirmed the presence of new lipid-laden adipocytes in penile tissues of DES-treated rats. Previously, our group showed that Sprague-Dawley rats treated neonatally with DES accumulated fat in the corpus cavernous penis [5][6][7][8] just as observed for the rats in the present study. The histological demonstration of DESinduced lipid buildup in the corpus cavernosum penis concurs with the newly induced adipocyte marker PPARγ2 detected with RT-PCR.
In penile tissue direct pharmacological activation of PPARγ by the antidiabetic TZD pioglitazone reportedly blocked corporal veno-occlusive dysfunction in rat model of type 2 diabetes mellitus [25]. However, this effect was associated with fat buildup suggesting that direct activation of penile PPARγ by TZDs or indirectly by ERα ligands, as in this study, could be a potential pathway for development of undesirable adipogenesis. In conclusion, PPARs are currently considered potential drug targets for diverse conditions including, vascular homoeostasis, diabetes mellitus, hyperlipidemia, inflammation, cancer, and infertility [50][51][52][53][54]. This study furthers our knowledge of mechanisms of endocrine disruption mediated by PPARγ in male subjects. The ERα-PPARγ signal pathway activation by DES is analogous in some way to mechanisms postulated for endocrine disruptor MEHP and other phthalates esters and organotins which directly activates PPARγ and promotes adipogenesis in cell culture models [2,3,55].