| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Departments of Pharmacology and Therapeutics (S.S., B.R.) and Obstetrics and Gynecology (B.R.) , McGill University, Montréal, Québec, Canada H3G 1Y6
Address all correspondence and requests for reprints to: Bernard Robaire, Department of Obstetrics and Gynecology, 3655 Promenade Sir William Osler, McGill University, Montréal, Québec, Canada H3G 1Y6. E-mail: bernard.robaire{at}mcgill.ca.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-reduced metabolite of testosterone (12, 13, 14, 15, 16, 17).
Various approaches have been used to study the role of androgens in the epididymis. Androgen deprivation has been achieved through bilateral orchidectomy (14, 18, 19), treatment with androgen antagonists (20, 21), or treatment with GnRH antagonists (22, 23). Orchidectomy is the most commonly used method to study the changes in epididymal morphology and gene expression associated with androgen deprivation; it is often followed by androgen replacement to ascertain the ability of testosterone to reverse these changes (14, 24). Recently rats were treated with inhibitors of 5-reductases, the enzymes responsible for the reduction of testosterone to DHT, to differentiate the effects of DHT from those of testosterone on gene expression in the different epididymal regions (25). The major drawback with all these in vivo approaches is the difficulties in interpreting the results because of the different cell types present in the tissue; consequently, it has been impossible to distinguish the direct effects of androgens on a particular cell type.
Several pure epididymal rodent cell lines have been developed recently (26, 27, 28, 29, 30). The mouse PC-1 cell line is the first immortalized epithelial cell line of the epididymis that has been shown to display tissue- and caput-specific gene expression (26). It was derived from primary cultures of epididymal cells from transgenic mice harboring a temperature-sensitive simian virus 40 large T antigen (26). It is a pure population of caput principal cells, the most numerous cell type of the epididymis, and the cell of central importance in the development of the tissue. Although insertion of the large T antigen is random and may compromise gene expression, data obtained to date indicate that the characteristics of principal cells are maintained in vitro (26, 31). The use of PC-1 cells in promoter studies and in characterizing the regulation of epididymal genes has proved invaluable (26, 31).
Androgen dependence has not been explored previously in an epididymal cell line. Because the caput segment is the region most active in terms of secretory activities of the mammalian epididymis and principal cells are the most sensitive epididymal cell type to the removal of circulating androgens (2, 32, 33), the PC-1 cell line is particularly relevant in studying the effects of androgen withdrawal and supplementation on gene expression. Principal cells are very active in protein synthesis and the absorption and secretion of fluid and small molecules (2); these functions, as well as cell morphology, are compromised in an androgen-deprived state (32). Understanding how androgens regulate gene expression in these cells is therefore crucial to elucidate the underlying causes of its impaired structure and functions in the absence of androgen. DNA microarrays were used to gain a comprehensive insight into gene expression in PC-1 cells after androgen withdrawal and supplementation.
This study uncovered differential responses of genes to androgen withdrawal in principal cells; whereas the expression levels of a number of genes decrease and increase, the majority of genes show an early or late transient increase in expression. Interestingly, many of these genes lose the ability to respond to androgens after prolonged deprivation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-DHT. All cell culture products were purchased from Invitrogen-Life Technologies (Carlsbad, CA). PC-1 cells were cultured at 33 C with 5% CO2. Two passages before the experimental procedure, FBS was replaced by charcoal-filtered FBS (Cocalico Biologicals, Inc., Reamstown, PA), and the latter was subsequently used for the duration of the experiment.
The cells were plated (2 x 105 cells/well) onto 6-well plates in the media described above and subjected to different androgenic conditions. The control group was grown in 1 nM 5-DHT for 2 d, the androgen withdrawal groups were grown without androgen for 2, 4, or 6 d, and the androgen supplementation groups were grown without androgen for 2 or 4 d, after which 1 nM 5-DHT was added to the media for 2 d. Media were changed every 24 h.
Cell viability assay
The effects of treatment on cell viability were assessed using the CellTiter-Glo luminescent cell viability assay (Promega, Madison, WI.) according to the manufacturer’s instructions. Briefly, cells from each experimental group as well as from cells grown in 1 nM 5-DHT for 4 or 6 d (three flasks/group) were collected and counted, and an aliquot of each was seeded onto 96-well plates. The CellTiter-Glo reagent was added to the aliquots, and the wells were read twice and compared with a standard curve using a microplate luminometer (LUMIstar Galaxy; BMG Lab Technologies, Durham, NC).
RNA extraction and DNase treatment
Each experimental group was comprised of five replicates with 2 wells/replicate to ensure sufficient RNA. The cells were washed twice with PBS (pH 7.4) buffer and lysed directly in the wells as described for the RNeasy minikit (QIAGEN Inc., Mississauga, Ontario, Canada). Total RNA was then extracted from each sample and subjected to DNase1 treatment using the kit. RNA concentrations were determined by absorbance at 260 nm (DU7 spectrophotometer; Beckman, Montreal, Québec, Canada), and RNA quality was verified by conventional gel electrophoresis.
cDNA arrays and hybridization
RNA samples were used to probe cDNA arrays (Atlas Mouse 1.2K array; BD Biosciences, San Jose, CA) consisting of 1176 genes, according to the manufacturer’s instructions. Five arrays per experimental group (control; 2, 4, or 6 d of androgen withdrawal; and 2 d of DHT supplementation after 2 or 4 d of androgen withdrawal) were probed and are referred to as replicates. The arrays were exposed to PhosphorImager plates (Molecular Dynamics, Inc., Sunnyvale, CA) overnight at room temperature before scanning with a PhosphorImager (Storm; Molecular Dynamics). Analysis of array images with Atlas Image (version 2.0; BD Biosciences) was done to quantify the intensity of each cDNA spot, which reflects the relative abundance of that RNA in the sample. The raw data for each gene (intensity of each spot on the array minus the background) were imported into GeneSpring 7.2 (Silicon Genetics, Redwood, CA) for further analysis. For each replicate array in a given experimental group, a gene was considered detected if its intensity was above threshold, with threshold defined as 2 times the average background of that individual array. A gene was considered expressed if it was detected in at least three replicates in that group.
To minimize experimental variation and allow for comparison of different experimental groups, data were normalized with the standard experiment-to-experiment normalization. More precisely, the median level of expression on each array was defined as 1, and the expression of each gene on that array was then normalized relative to 1. This value was averaged for all replicates in a group to generate what is referred to as the relative intensity for a given gene. We focused on changes in gene expression of at least 1.5-fold (i.e. 33% decrease or 50% increase). Statistical analyses were done using Wilcoxon-Mann-Whitney test (significance level set at P < 0.05) with the GeneSpring software. Genes changing by 1.5-fold that were not statistically significant were excluded from the analysis. For k-means cluster analysis, gene-to-gene normalization was done in addition to experiment-to-experiment normalization as described previously (19). This method normalizes the signal strength of each gene relative to the median of all measurements taken for that gene in each experiment, defined as 1. It was used to visualize the expression profile of all genes on the same vertical axis. Clustering was done on genes statistically changing by 1.5-fold at any time after androgen withdrawal.
PathwayStudio 4.0 (Ariadne Genomics, Rockville, MD) and ResNet-3.0 database were used to visualize direct relationships between genes differentially affected by androgen supplementation after 2 or 4 d of androgen deprivation. Objects were limited to proteins and pathways. Controls were limited to binding, expression, protein modification, and regulation. Four starting groups were used: genes rescued at 4 d, genes not rescued at 4 d, genes rescued at 6 d, and genes not rescued at 6 d (see supplemental Table 3, A and B, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org). Mediators of androgen action [androgen receptor (AR) and the two steroid 5-reductases] were also included in these groups. Rescued genes are defined as genes for which the expression levels change significantly by at least 1.5-fold from control after androgen deprivation but return to control levels after DHT supplementation.
All gene names and symbols were retrieved in April 2006 from the Mouse Genome Database, Mouse Genome Informatics Web site (The Jackson Laboratory, Bar Harbor, ME; http://www.informatics.jax.org). Genes were classified by their primary function (or one of their primary functions) obtained from Mouse Genome Informatics and PathwayStudio 4.0.
Quantitative real-time RT-PCR
The expression of selected genes (Table 1
) was quantified by real-time RT-PCR using the LightCycler system (Roche Diagnostics, Laval, Québec, Canada). The genes were normalized against peptidylprolyl isomerase A (Ppia, cyclophilin A), a housekeeping gene with mRNA levels that do not change with androgen manipulation (34). Gene-specific primer sequences (Table 1) were obtained from the literature (35, 36, 37) or designed using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3.cgi/) with the exception of steroid 5-reductase 2 (Srd5a2) (catalog no. QT00128464, QuantiTect Primer Assays; QIAGEN). The primers were synthesized by the Sheldon Biotechnology Centre, McGill University. Total RNA isolated for the microarray analysis was reversed transcribed and amplified with the QuantiTect SRBR green RT-PCR kit (QIAGEN) according to the manufacturer’s protocol. Briefly, each quantitative RT-PCR contained 5 ng of RNA (50 ng RNA for Srd5a2), 0.5 µM of gene-specific primers, 10 µl of the 2x QuantiTect SYBR green RT-PCR master mix, and 0.2 µl of the QuantiTect reverse transcriptase mix. The cycling conditions were as follows: reverse transcriptase for 20 min at 50 C, followed by initial denaturation/enzyme activation for 15 min at 95 C and 50 cycles of denaturation at 94 C for 15 sec, annealing at 55–60 C for 30 sec, and elongation at 72 C for 30 sec. The production of a single product was confirmed by melting curve analysis (temperature elevation from 65 to 95 C at 0.2 C/sec) and conventional gel electrophoresis. For each analysis, a standard curve was prepared from six serial dilutions of a standard containing equal amounts of RNA from the different experimental groups. All standards and samples were assayed in duplicate. One-way ANOVA followed by Dunnett post hoc test was used to detect significant effects of androgen withdrawal on gene expression. The significant effects of androgen supplementation were determined using paired t tests. The level of significance was taken as P < 0.05.
|
The samples, 20 µg protein per lane, were boiled with loading buffer for 5 min and fractionated by SDS-PAGE using 12.5% acrylamide gels (38). Prestained precision standards (Bio-Rad Laboratories) were used as molecular weight markers. The fractionated proteins were transferred to a Hybond-P membrane (Amersham Biosciences UK, Buckinghamshire, UK). The blots were blocked with 5% nonfat dried milk in 137 mM NaC1, 20 mM Tris (pH 7.4), and 0.1% Tween 20 at room temperature for 1 h and then incubated overnight at 4 C with a primary goat polyclonal antibody to secreted phosphoprotein 1 (SPP1) (1:500, sc-10593; Santa Cruz Biotechnology, Santa Cruz, CA). Antibody binding was detected by incubating with donkey antigoat IgG conjugated to horseradish peroxidase (1:10,000, sc-2056: Santa Cruz Biotechnology). Cyclophilin A (1:2500, catalog no. 07-313; Upstate, Lake Placid, NY) was used as a loading control according to the manufacturer’s instructions and was detected using a secondary, donkey antirabbit IgG-horseradish peroxidase antibody (1:5000, NA934V; Amersham Biosciences UK).
Western blots were visualized with the Enhanced Chemiluminescence Plus kit (Amersham Biosciences UK) and Hyperfilm enhanced chemiluminescence (Amersham Biosciences UK). Quantification of Western blot data were done by line densitometry using a ChemiImager 4000 imaging system (Alpha Innotech, San Leandro, CA) with AlphaEase (version 5.5 software; Alpha Innotech). SPP1 levels for each experimental group were expressed relative to the corresponding cyclophilin A value.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). There was no change in the doubling time or number of viable cells when they were deprived of androgen for 2, 4 or 6 d when compared with cells grown in the presence of DHT for the same duration (Fig. 1). Similarly, supplementing the media with androgen after 2 or 4 d of androgen deprivation had no effect on cell viability (Fig. 1). Significant cell death was observed in over 40% of cells 8 d after androgen withdrawal (data not shown).
|
). Some of these genes are associated with housekeeping functions [ubiquitin B (Ubb), ß-actin (Actb), glyceraldehyde-3-phosphate dehydrogenase (Gapdh), and ribosomal proteins S29 (Rps29) and SA (Rpsa)], whereas others are involved in DNA replication and cell cycle regulation [minichromosome maintenance deficient 2 mitotin (Mcm2), minichromosome maintenance deficient 5 (Mcm5), recombination activating gene 1 activating protein 1 (Rag1ap1), proliferation-associated 2G4 (Pa2g4), cell division cycle 20 homolog (Cdc20), CDC28 protein kinase regulatory subunit 2 (Cks2), and the cyclins A2 (Ccna2), B1 (Ccnb1), and B2(Ccnb2)]. Genes involved in the oxidative stress response [glutathione S-transferase µ2 (Gstm2), glutathione S-transferase 
1 (Gstp1), and glutathione reductase 1 (Gsr)] as well as genes associated with calcium signaling [S100 calcium binding protein A10 (S100a10), RAD21 homolog (Rad21), calmodulin 1 (Calm1), stromal cell derived factor 4 (Sdf4), Spp1, and calcitonin gene-related peptide-receptor component protein (Crcp)] were also highly expressed.
|
|
) had been shown previously to be expressed in the epididymis (19, 40, 41). These include a number of transcriptional activators and repressors [Y box protein 1 (Ybx1), myelocytomatosis oncogene (Myc), paired box gene 8 (Pax8), v-rel reticuloendotheliosis viral oncogene homolog A (Rela, Rel-A), and 3-phosphoglycerate dehydrogenase (Phgdh)], some cytoskeletal proteins [vimentin (Vim), cofilin 1 (Cfl1), and villin 2 (Vil2)], and genes involved in metabolism (glucose phosphate isomerase 1 (Gpi), tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein 
-polypeptide (Ywhah), and ornithine decarboxylase (Odc1)]. Heat shock protein (HSP) 90-kDa -class B member 1 (Hsp90ab1, HSP-90ß), a chaperone associated with steroid receptor function, was also one of three HSPs that were very highly expressed in the cell line. Whereas most of these transcripts were identified by Northern blot and/or array analysis in the epididymis, only a few have been immunolocalized specifically to the principal cells of the caput epididymidis in various species. Protective protein for ß-galactosidase (Ppgb, cathepsin A) (42), solute carrier family 9 member 1 (Slc9a1, sodium/hydrogen exchanger 1) (43), and vimentin (44) are examples of such genes. Transcripts for the kinase expressed in nonmetastatic cells 2 (Nme2, nucleoside diphosphate kinase B) were found at very high levels in the cell line. This enzyme catalyzes the last step in the production of nucleoside triphosphates, and its high levels could most certainly provide principal cells with sufficient energy to carry out the extensive protein synthesis and secretion observed in the caput epididymidis. Transcript levels for the immunoglobulin protein basigin (Bsg) were also abundant in PC-1 cells, suggesting a role for this protein in the proximal region of the epididymis. This glycoprotein is essential in the later stages of spermatogenesis and undergoes molecular processing and deglycosylation at the surface of spermatozoa during its transit in the epididymis (45), but its expression in the epididymal epithelium has not been elucidated previously.
Consequences of androgen withdrawal on overall gene expression
Androgen withdrawal led to important changes in gene expression. The lack of effect on cell viability for the 6-d study window indicates that the changes seen are the consequences of androgen deprivation and are not the result of cell death. Of the 167 genes detected in the PC-1 cells, 50 genes changed by at least 1.5-fold (increase or decrease) and 20 genes increased from undetected levels. There were 38 genes decreasing at any one time after androgen withdrawal and 33 genes increasing. HSP-90ß was the only transcript both increasing (at 2 d) and decreasing (at 6 d) by 1.5-fold during the course of the experiment. A list of genes significantly decreasing or increasing in expression by at least 1.5-fold from control is found in the supplemental Table 1, A and B, respectively.
The number of genes increasing in expression 4 d after androgen deprivation was far greater than at the other two time points combined (Fig. 2), suggesting that removal of the inhibitory effects of androgens on gene expression occurred most dramatically at 4 d. The number of genes decreasing in expression, however, was most prominent at 6 d. There were 20 more genes decreasing at 6 d than at 4 d (Fig. 2), and all the genes decreasing at 4 d also decreased at 6 d with the exception of tubulin-ß 4 (Tubb4). Thus, these data suggest that deprivation of the stimulatory effects of androgens on gene expression had the most consequences at 6 d. Whereas there is a maximal effect at 4 d on the number of genes increasing, the number of genes decreasing displayed a gradual increase with time (Fig. 2). The few genes changing by at least 1.5-fold 2 d after the removal of androgens were not the same genes changing at 4 and 6 d. These genes may be early responders of androgen withdrawal.
|
). These patterns were grouped into four distinct gene expression profiles: decreasing genes, early and late transiently increasing genes, and increasing genes. Nineteen percent of genes progressively decreased in gene expression during androgen withdrawal (profile 1). They were essentially regulators of apoptosis, cytoskeletal proteins, cell adhesion proteins, and a few cell cycle proteins. The removal of androgens led to the transient increase in expression levels of the majority of genes. This is consistent with a previous study that found transiently increasing genes to be most prominent in the caput region of the epididymis after castration (19). These genes were further divided into two profiles based on time of increase. Thirty-three percent of the genes partitioned into profile 2, in which a transient increase occurred early at 2 d. Most of these genes are involved in cell cycle regulation and transcription. Proteins involved in cell adhesion and cell proliferation/growth also belonged to this profile. Profile 3 consisted of 31% of genes changing by at least 1.5-fold; their transient increase occurred late, between 2 and 4 d. They regroup many intracellular enzymes, transcription factors, and cell signaling molecules. The early transient increase may be indicative of genes directly regulated by androgens, whereas the genes displaying a late transient increase may be indirectly regulated by androgens through the intermediate of another gene or transcription factor. The remaining 17% of genes increased in gene expression levels (profile 4); they displayed a sharp increase between 2 and 4 d, and their expression levels were maintained at 6 d. They are essentially proteins involved in cell adhesion, cell signaling, and transcription. Whereas genes involved in transcription and cell adhesion were found in all profiles, albeit at different proportions, genes with functions in cell structure, apoptosis, and protein folding were restricted to unique profiles. A list of the genes associated with each of these profiles is found in supplemental Table 2.
|
depicts the number of genes expressed at 4 and 6 d after the start of the experiment. Gene lists associated with each section of the pie chart for 4 and 6 d are found in supplemental Table 3, A and B, respectively. The larger number of total genes expressed at 4 d was attributed to the higher number of genes increasing from undetectable levels at this time point. At both 4 and 6 d, about a quarter of the genes expressed were affected by androgen withdrawal. The expression of the majority of these genes increased at 4 d but decreased at 6 d. A significant effect of androgen supplementation was found in almost 60% of the genes increasing and decreasing in expression at 4 d, in contrast to the effects of DHT at 6 d. At 6 d, only 33% of the genes increasing and less than 12% of the genes decreasing in expression had retained their ability to respond to DHT. Androgen supplementation was, in most instances, successful at returning gene expression levels to that of control (no significant difference between the two groups) with the exception of two genes. A partial rescue was observed with gap junction membrane channel protein-ß3 (Gjb3) and periostin (Postn) after giving back DHT 2 or 4 d after androgen withdrawal, respectively.
|
Effects of androgen supplementation on related genes at 4 and 6 d
Known regulatory relationships between genes that were rescued after androgen supplementation at 4 and 6 d and those that were not significantly affected by supplementation at the two time points were examined; mediators of androgen action were also included in the analysis (Fig. 5). A large number of genes rescued at 4 d were functionally linked using this analysis; they were found to directly influence the expression and/or the regulation of one another (Fig. 5A). Some of the genes were also linked by protein binding. For example, RAD21 homolog and RAD50 homolog form a DNA repair complex that can possibly reinitiate collapsed replication forks during S phase (46). The transcription factors early growth response (EGR)-1 and POU domain class 2 transcription factor 1 (POU2F1, also known as OCT-1) can form complexes with each other and with the AR to regulate genes. Whereas EGR1 modulates cell growth, apoptosis, and differentiation, POU2F1 is involved in the regulation of a wide variety of genes and can interact positively or negatively with the AR, in a promoter-specific manner to enhance AR coactivator recruitment (47). Presumably, the rescue of Pou2f1 transcript levels at 4 d also allows the transcript levels of some androgen-regulated genes to return to control values at this time point. EGR1, POU2F1, and AR are all linked to IGF-I. In fact, most of these genes rescued at d 4 are regulated by IGF-I. These include genes involved in cell adhesion [breast cancer antiestrogen resistance 1 (BCAR1), amyloid-ß A4 precursor protein (APP), SPP1, and integrin ß 1 (ITGB1)], apoptosis [pleckstrin homology-like domain, family A, member 1 (PHLDA1), APP, SPP1, and EGR1)], and cell growth [EGR1 and IGF binding protein (IGFBP) 2]. These results suggest that androgens may directly regulate one or more genes in this pathway and through various direct interactions at the protein level, other genes are indirectly regulated by androgens. Only one direct association was revealed among genes not rescued at d 4; proviral integration site 1 (PIM1) was found to influence the expression of BCL2L1 (not shown).
|
); a large number of them are involved in cell cycle regulation [cyclin (CCN) A2, CCNB1, CCND1, CCNE1, CDC20, and budding uninhibited by benzimidazoles 1 homolog-ß (BUB1B)]. This functional association map includes four genes that were rescued at d 4 [Egr1, Spp1, Rad21, fibroblast growth factor receptor 1 (Fgfr1)], of which FGFR1 directly influences the regulation of SPP1. It also highlights three AR coregulators [amino-terminal enhancer of split (AES), CCND1, and CCNE1] whose expression levels were not rescued at d 6, suggesting there is inadequate AR signaling at the promoter region of certain genes and consequently the ability of these androgen-regulated genes to respond to androgen is compromised.
Mediators of androgen action and their responses to androgens
Quantitative real-time RT-PCR was done to determine the effects of androgen withdrawal and supplementation on the transcript levels of Ar, Srd5a1, and Srd5a2. The Ar transcript was undetected in the arrays, whereas the 5-reductases, the enzymes responsible for the reduction of testosterone to the more potent DHT, were not on the array. These genes displayed distinct patterns of expression (Fig. 6). There was a small, progressive increase in Ar mRNA levels after androgen withdrawal. When DHT was added back to the media after 2 d of deprivation, Ar transcript levels decreased to that of control (P = 0.01) (Fig. 6A). This rescue in expression levels observed after the addition of DHT was absent if the cells were deprived of androgens for an additional 2 d (Fig. 5A). The 5-reductases responded differently to androgens. Srd5a1 mRNA levels decreased slightly after androgen removal, and there was a partial return to control levels when DHT was added after 2 d (P = 0.14) or 4 d (P = 0.05) of androgen deprivation (Fig. 6B). In contrast, Srd5a2 mRNA levels showed a transient increase 2 d after androgen removal and decreased to undetectable levels for the duration of the treatment (Fig. 6C). Srd5a2 transcript levels were unexpectedly very low under control conditions, and the low levels were not found to be a consequence of the charcoal-filtered FBS (data not shown).
|
) to confirm the expression data from the microarray study (Fig. 7A). The expression profiles of the four genes were found to be practically identical with the two methods. Rela transcript levels were essentially unaffected by androgens. DHT supplementation after 2 d of androgen deprivation led to a small increase in transcript levels, closer to the levels of control. The removal of androgens resulted in a decrease in Igfbp2 transcripts between 2 and 4 d and adding DHT had no effect on gene expression at 4 or 6 d. Whereas the role of androgens in regulating Rela and Igfbp2 expression was found to be minor, androgen withdrawal and supplementation had considerable consequences on Postn and Spp1 expression levels. Postn expression increased sharply from undetectable levels 4 d after androgen withdrawal followed by a smaller increase 2 d later. DHT supplementation after 2 d of deprivation successfully suppressed the increase in Postn transcripts but after 4 d of deprivation led only to a slight decrease in transcript levels. Spp1 was a highly expressed gene under control conditions. Androgen withdrawal resulted in a steep decline in Spp1 transcripts at 4 d, which leveled off at 6 d. DHT supplementation was able to prevent this decrease when added back after 2 d of withdrawal but had no significant effect if it was added 4 d after the removal of androgens. Western blots were done to confirm whether the changes observed at the mRNA level after androgen manipulations were also seen at the protein level. Similar to its mRNA levels, SPP1 protein levels decreased after androgen removal from the cell culture media (Fig. 8). The decrease at 2 d after androgen deprivation was more pronounced than that observed at the mRNA level and may be due to a higher proportion of protein that is secreted into the extracellular matrix. The decrease was also not as significant as that observed at 4 and 6 d after the removal of androgens. The stability of the protein may account for this discrepancy. The decrease in expression levels was prevented when androgen was added back to the media after 2 d but not after 4 d, of androgen deprivation (Fig. 8).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The regulation of genes by androgens has been investigated in many studies, but their differential response to time of supplementation after androgen withdrawal has not. The loss in the ability of the majority of genes to restore gene expression levels to that of control after 4 d of androgen deprivation may be the consequence of inadequate AR function. AR coregulators are important components of AR signaling; for instance, type 1 coregulators facilitate DNA occupancy, chromatin remodeling, and the recruitment of general transcription factors associated with the RNA polymerase II holocomplex (49). At 4 d, we identified direct interactions between POU2F1 and the AR among the genes rescued with DHT supplementation, but at 6 d, the transcript levels of three AR coregulators were not rescued. CCNE1 is a strong coactivator of the AR and has been shown to increase the transactivation activity of the AR in the presence of DHT by binding to the COOH terminus portion of the receptor’s AB domain (50). CCNB1, on the other hand, was found to inhibit ligand-dependent AR transactivation in the prostate; this function of CCND1 was independent of its role in cell cycle progression and is likely elicited through its ability to form a specific complex with AR (51). AES is another selective repressor of ligand-dependent AR-mediated transcription but it specifically acts by targeting the basal transcription machinery (52). The inability of these AR coregulators to respond to DHT after 4 d of DHT deprivation presumably would have profound consequences at the promoter of many androgen-regulated genes.
The implications may even be broader; after physiological conditions of temporary androgen suppression, such as hormonal contraception or the therapeutic administration of testosterone in aging men, the inability of AR coregulators to respond to androgens will affect the expression of many androgen-regulated genes, thereby affecting many physiological processes. In the epididymis, three members of the 160-kDa protein family were recently described (53); they recruit protein complexes containing histone acetyltransferases and histone methyltransferases to the target gene promoter. One of the members, p/CIP/RAC3/ACTR/AIB1/TRAM-1, was immunolocalized mainly to the cytoplasm of epithelial cells in the caput epididymidis (53), but the consequences of prolonged androgen deprivation on its transcript levels or localization are unknown.
Many of the genes rescued by DHT supplementation at 4 d were functionally linked by direct interactions and all but two of these genes (Rad21 and Rad50) converged on IGF-I. The role of this growth factor appears to be central as it can directly influence the regulation and expression of many genes rescued at this time point. IGF-I is involved in a variety of biological processes, some of which are illustrated by the genes it regulates, and it is an important element controlling tissue growth and differentiation (54). IGF-I transcripts increase from undetected levels in the cell line with removal of androgen from the media, presumably leading to the induction of AR activation. IGF-I was shown to regulate AR-mediated transcription in the absence of, or at low levels of, ligand in the prostate by increasing cellular levels of ß-catenin, an activator of the receptor (55). When androgen is added back to the media, IGF-I is presumably no longer needed, and its levels decrease to below threshold, becoming undetected. Whereas its specific role and localization within adult epithelial epididymal cells is unknown, IGF-I induced the maturation of epididymal principal cells in vitro (56) and was immunolocalized to the subapical and apical cytoplasmic compartments of principal cells in prepubescent rat caput epididymidis (57). Similarly, IGF-I plays a crucial role in Leydig cell maturation in the testis and absence of this growth factor results in undifferentiated cells, altered expression of testosterone biosynthetic and metabolizing enzymes leading to decrease levels of androgens, and infertility (58, 59). The distal portions of the epididymis are also severely affected in IGF-I null mice (58). We hypothesize that during development and conditions of androgen deprivation, IGF-I expression serves to amplify AR signaling.
Whereas many genes rescued at 4 d seem to converge to IGF-I, it is particularly interesting that most of the directly related genes not rescued at 6 d belong to the cell cycle pathway. All but cyclin B1 display an early transient increase in expression levels after the removal of androgens from the media. This increase in expression occurs at 2 d, but by 6 d transcript levels of these genes have decreased to more than 1.5-fold below control levels, indicating that androgens inhibit cell cycle genes in principal cells under physiological conditions. In fact, the lack of effect of androgens on mitotic rate in the adult epididymis distinguishes it from other androgen-dependent tissues such as the prostate and seminal vesicles (60). The increase in gene expression is transient, and presumably, transcription factors and cell signaling molecules capable of inhibiting cell proliferation are up-regulated before an increase in cell viability is observed when compared with control conditions.
Many studies, including those using DNA microarrays, have explored androgen regulation in the epididymis. Androgen withdrawal and supplementation were examined in only one such study; this study compared the gene expression profiles of mice 9 d after orchidectomy with those supplemented with oil or DHT for the last 2 d and identified androgen-regulated genes in the caput, corpus, and cauda epididymidis (61). Our study focused on the early responses of a specific cell type of the epididymis to androgen deprivation and examined the time dependence of rescue with DHT supplementation. Another gene expression study investigated changes in the different regions of the rat epididymis occurring at several time points within the first week after orchidectomy (19); this study profiled 474 stress-related genes and described four patterns of gene expression that were similar to the ones we identified by cluster analysis. The transient increase in gene expression levels observed in the different regions within the first week after orchidectomy was particularly prevalent in our study, with the majority of genes displaying an early or late transient increase in expression levels. Whereas the aim of this study was to examine the effects of testicular factors and androgens on gene expression, our focus was directed to androgens and their ability to rescue specific genes in the most androgen sensitive epididymal cell type (32).
Androgens regulate the expression of many genes and gene families in the epididymis, most likely through interactions with the AR present in the tissue. Previous studies established that the AR itself is regulated by androgens. Both androgen-mediated up-regulation and down-regulation of Ar mRNA have been described in many cell lines and tissues (62, 63, 64, 65). In the rat epididymis, Ar mRNA levels were increased almost 1.5-fold with androgen withdrawal and decreased below control levels after androgen stimulation (63). Similarly, we observed a small progressive increase in mRNA levels after the removal of androgens, but androgen supplementation was successful only at bringing the levels down to those of control after 2 d and not after 4 d of androgen deprivation.
Other important mediators of androgen action are the steroid 5-reductases. Although both the type 1 and type 2 enzymes catalyze the same reaction, they differ with respect to their biochemical properties, pharmacological characterization, and tissue distribution patterns (66, 67, 68). Their mRNA levels are also differentially regulated in the rat epididymis (69). Srd5a1 mRNA levels decreased in PC-1 cells after androgen removal. This decrease was not as significant as the one described for the rat epididymis after orchidectomy (70), but the time points examined were also earlier. Unlike for the majority of genes described in this study, androgen supplementation after either 2 or 4 d of androgen deprivation led to an increase in Srd5a1 transcript levels. The ability to rescue gene expression levels to those of control was not compromised with time and was also observed in the rat caput epididymidis after testosterone-filled capsules were implanted subdermally immediately after orchidectomy (70). Steady-state Srd5a2 mRNA expression in the PC-1 cells was remarkably low; it was almost 10-fold lower than that of Srd5a1. This was in sharp contrast to the high levels of Srd5a2 mRNA expressed in the proximal caput region of both the mouse and rat epididymides, in which its expression levels were more than 10-fold higher than the type 1 enzyme (69, 71). The low levels of Srd5a2 in the PC-1 cell line may be due to its sensitivity to in vitro culture conditions or to the absence of one or many regulatory factors crucial at maintaining its high expression levels. Its expression levels in other caput epididymal cell lines would need to be investigated to ascertain whether its low levels were specific to PC-1 cells or all epididymal cell lines. Srd5a2 mRNA decreased to undetectable levels 4 d after androgen deprivation and remained undetected for the duration of the treatment, suggesting that it is particularly sensitive to the androgen levels in culture media.
Androgen manipulations led to small, nonsignificant changes in Rela and Igfbp2 transcript levels in the PC-1 cell line. Rel-A is a ubiquitous member of the nuclear factor-
B family of transcription factors and is often found as a heterodimer with the p50 subunit (reviewed in Ref. 72). Among its numerous functions, it has been described as both an activator and repressor of the AR gene. Androgens, however, have not been reported to regulate the transcript levels of any of the nuclear factor-B subunits, nor have they resulted in the activation and translocation of the complex to the nucleus. Although our study supports the lack of effect on the transcript levels, the possibility does remain that Rel-A is regulated posttranslationally by androgens.
Igfbp2 belongs to the family of IGFBPs that modulate cellular growth and differentiation directly (73, 74) or through their interactions with IGFs (75). It is the second most abundant IGFBP in the circulation and is found in a variety of mammalian fluids and tissues (reviewed in Refs. 76, 77). The regulation of Igfbp2 expression and abundance is highly complex and influenced by multiple growth factors and hormones, including androgens (reviewed in Ref. 78). Whereas castration had no significant effect on Igfbp2 expression in a model of androgen-dependent neoplasia (79), it stimulated Igfbp2 production in both LNCaP cells (80) and the rat (81). Androgens have also been shown to stimulate Igfbp2 production in the LNCaP prostate cancer cell line (82) and increase Igfbp2 mRNA and protein in human fetal osteoblastic cells (83). In the PC-1 cells, androgen withdrawal led to a nonsignificant decrease in Igfbp2 expression levels and androgen supplementation after 2 or 4 d of deprivation had no effect on expression levels. Androgens appear to play a lesser role in regulating Igfbp2 expression in PC-1 cells, compared with other tissues and cell lines; whether this also holds true for the epididymis still remains to be determined.
Periostin and Spp1 are examples of two genes differentially regulated by androgen manipulations. Periostin, formerly known as osteoblast-specific factor 2, is a cell-adhesion molecule that was first described in bone in the context of bone formation (84, 85). It has since been found to be expressed in a wide range of adult normal tissues, with negligible expression in the prostate and testis (86). Recently mouse prostate stromal cells were found to express a number of osteogenic molecules, including periostin (87). This secreted protein binds members of the integrin family of receptors and is consequently involved in cytosolic signaling cascades mediating cell proliferation, cell survival, and cell migration in addition to its role in cell adhesion (86, 88, 89, 90). It has also been demonstrated that stress responses can stimulate periostin transcription (91, 92, 93). Similarly, periostin mRNA levels increase from undetected levels after androgen withdrawal, and alleviation of this stressor after 2 d returns the expression levels to control. Its up-regulation may prevent stress-induced apoptosis in the cell line by activating the Akt/protein kinase B signaling pathway, as described for colon cancer cells (88). Although the role of periostin in the epididymis remains to be elucidated, our studies establish, for the first time, the androgen dependence of this protein.
Spp1, also known as osteopontin, is a cell surface protein with many isoforms and proposed functions (reviewed in Refs. 94 and 95), which can be regulated by many factors. Pou2f1, IGF-I, Igfbp2, Gpi, and Fgfr1 have all been identified through direct functional association studies as factors that can bind or influence the expression or regulation of Spp1. The role of these factors in cell proliferation/growth, apoptosis, metabolism, development, and organogenesis provide testimony to the numerous functions of Spp1. In the rat epididymis, Spp1 expression was found to be region and cell specific, with principal cells of the distal caput epididymidis exhibiting the highest expression levels (96). This glycoprotein is proposed to be an important means by which the epididymis regulates calcium ions in the lumen to prevent mineral accumulation and eventual decrease in sperm fertilizing ability (96). Although no androgen-responsive element has been identified (94), Spp1 was found to be regulated by androgens. This regulation was restricted to the principal cells of the epididymis (96). Consistent with castration studies, Spp1 expression in PC-1 cells decreased after the removal of androgens from the media. The effects on its protein levels have also been previously investigated in the rat proximal caput epididymidis after orchidectomy and testosterone supplementation (96); rescue was seen at all the time points examined when testosterone was administered immediately after orchidectomy.
We have characterized gene expression in a proximal caput epididymidis cell line under control conditions and after androgen manipulations. We have observed four distinct responses of genes to the removal of androgens and, for the first time, identified a differential rescue of gene expression levels in principal cells deprived of DHT for different lengths of time. This differential rescue is associated with particular gene families and is mediated in part by changes in AR coregulator expression after androgen manipulations. Our study presents a novel approach for understanding mechanisms of androgen regulation in the epididymis and provides an in-depth look into the regulation of gene expression in its most abundant and androgen-sensitive cell type.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Disclosure Statement: The authors have nothing to disclose.
First Published Online October 5, 2006
Abbreviations: AR, Androgen receptor; CCN, cyclin; DHT, dihydrotestosterone; EGR, early growth response; FBS, fetal bovine serum; HSP, heat shock protein; IGFBP, IGF binding protein; POU2F1, POU domain class 2 transcription factor 1; SPP1, secreted phosphoprotein 1.
Received July 26, 2006.
Accepted for publication September 27, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-glutamyl transpeptidase catalytic activity and protein level in the initial segment of the rat epididymis by testicular factors: role of basic fibroblast growth factor. Biol Reprod 58:197–206-androstan-3-one (dihydrotestosterone) by prostate and epididymis. J Clin Endocrinol Metab 29:970–977[Medline]
4–5-reductase and 3-hydroxysteroid dehydrogenase in the rat epididymis. Endocrinology 101:1379–1390[Medline] reductase in the rat. Steroids 30:41–51[CrossRef][Medline]-steroid reductase inhibitors. Experientia 37:1031–1032[CrossRef][Medline]-reductase inhibitor, on gene expression in the rat epididymis. J Endocrinol 181:245–261[Abstract]-reductase type 2 promoter in the rat epididymis. Biol Reprod 72:851–861-glutamyl transpeptidase messenger ribonucleic acid transcripts in the adult rat epididymis is differentially regulated by androgens and testicular factors in a region-specific manner. Endocrinology 135:1146–1156[Abstract]
) and those decreasing (genes 
) after androgen withdrawal. The genes with increasing and decreasing expression were subdivided into those in which androgen supplementation had a significant effect and rescued gene expression (diagonal stripes) and those where it did not (solid black). The number of genes corresponding to each piece of the chart is indicated adjacent to the piece and the total number of genes expressed at each time point is indicated at the bottom of the chart. The difference in the total number of genes expressed at 4 and 6 d is attributed to the genes increasing from undetected levels.
).
