The Role of Androgens in Sertoli Cell Proliferation and Functional Maturation: Studies in Mice with Total or Sertoli Cell-Selective Ablation
http://www.100md.com
内分泌学杂志 2005年第6期
0Medical Research Council Human Reproductive Sciences Unit (K.A.L.T., M.W., R.M.S., P.T.K.S.), Centre for Reproductive Biology, University of Edinburgh, Edinburgh EH16 4SB, Scotland, United Kingdom; Department of Developmental Biology (K.D.G., E.D., G.V.), Laboratory for Experimental Medicine and Endocrinology, Catholic University of Leuven, B-3000 Leuven, Belgium; and Institute of Experimental Morphology and Anthropology (N.A.), Bulgarian Academy of Science, 1113 Sofia, Bulgaria
Address all correspondence and requests for reprints to: Professor G. Verhoeven; Laboratory for Experimental Medicine and Endocrinology, Onderwijs en Navorsing, Gasthuisberg, Herestraat 49 bus 902, B-3000 Leuven, Belgium. E-mail: guido.verhoeven@med.kuleuven.ac.be.
Abstract
The role of androgens in the proliferation and maturation of Sertoli cells (SC) and the development of their capacity to support spermatogenesis remains poorly understood. We evaluated these functions in complete androgen receptor knockout (ARKO) and SC-selective androgen receptor knockout (SCARKO) mice. Compared with controls, ARKO mice exhibited a progressive reduction in SC number/testis, whereas SCARKOs showed minor changes, suggesting that androgen effects on SC number are not mediated via direct action on SCs. Immunoexpression of anti-Müllerian hormone (AMH), p27kip1, GATA-1, and sulfated glycoprotein-2, which changes according to SC maturational status, occurred normally in ARKOs and SCARKOs. Functional capacity of SCs to support spermatogonia was similar in SCARKOs and controls, whereas ARKOs showed reduced capacity with age. SC capacity to support total germ cells revealed major deficits in ARKO and SCARKO adults, particularly with respect to postmeiotic germ cells. Using quantitative RT-PCR, the expression of SC markers was compared in d 50 testes. In ARKOs, expression of Pem, fatty acid binding protein, platelet-derived growth factor-A, and transferrin were all significantly reduced, whereas FSH receptor and AMH were increased. In SCARKOs, there were modest reductions in expression of cystatin-related gene highly expressed in testis and epididymis (cystatin-TE) and claudin-11, whereas expression of Pem, fatty acid binding protein, and platelet-derived growth factor-A was markedly reduced, highlighting these as potentially androgen-regulated SC genes that merit further study. In conclusion, androgen action is not required for maturation-dependent changes in immunoexpression of the SC markers AMH, p27kip1, GATA-1, and sulfated glycoprotein-2 but is essential for expression of other SC genes, the attainment of normal SC number, and the support of meiotic and postmeiotic germ cell development.
Introduction
IT IS WELL ESTABLISHED that androgens play the central role in masculinization of the reproductive tract, genitalia, and many other organ systems during the sexual differentiation process in the male (1, 2). It is also established that, in adulthood, androgen action on seminiferous tubules is essential for full, quantitatively normal spermatogenesis and fertility, with the majority of evidence pointing to this effect being mediated by effects on Sertoli cells (SCs), although the mechanisms remain unclear (3, 4, 5). Between the periods of testicular/sexual differentiation in fetal life and the establishment of complete spermatogenesis in adulthood, two other key changes occur to SCs that are essential for normal fertility in adulthood. First, SCs undergo intense proliferation during fetal, neonatal, and peripubertal life (6) to ensure that high levels of sperm production can be achieved in adulthood; SC number is the main determinant of the sperm production rate per testis in adulthood (6, 7, 8). Second, SCs have to undergo a maturation process (terminal differentiation), whereby they lose the ability to proliferate and switch on functions that are essential for the support of germ cells during spermatogenesis (6).
There are various lines of evidence that point to a role for androgens in the SC maturation process, although this is largely circumstantial (4, 6). Such a role for androgens seems intrinsically logical when considering that maturation and acquisition of full androgen responsiveness are prerequisites for SCs to support full spermatogenesis in adulthood, and notable in this regard is that the onset of androgen receptor (AR) expression in SCs precedes onset of SC maturation (4, 6). In contrast, until recently, it was generally believed that (in rodents) androgens played little if any role in SC proliferation, primarily because the AR is not normally expressed in SCs for the majority of the periods during which they proliferate (4, 6, 9). However, recent data from mice lacking a functional AR (10), microarray analysis of androgen-regulated transcripts in neonatal mice (11), and studies involving manipulation of androgen production/action in neonatal rats (12) demonstrate that androgens probably play a physiological role in promoting SC proliferation in fetal and early postnatal life in rodents. There is similar evidence for a role for androgens in postnatal SC proliferation in nonhuman primates (13, 14). These new findings raise a conundrum, namely how do androgens exert such a positive effect on SC proliferation in the absence of AR immunoexpression in fetal and early neonatal life? One obvious explanation is that androgens act via peritubular myoid cells to modify SC proliferation because there is clear in vitro evidence that the function of SCs can be modified by such interactions (15, 16). Alternatively, the proliferative effects of androgens on SCs might be mediated via membrane ARs or other nongenomic pathways because there is growing evidence from cell lines for a role for such systems, including effects on cell proliferation (17, 18). Indeed, it is apparent from a number of studies that such systems are present in SCs (19), and a recent study has identified one pathway via which nongenomic actions of androgens on SCs could play a role in spermatogenesis (20). Such developments might explain the inability of researchers to identify clear pathways via which genomic actions of androgens on SCs support spermatogenesis (4).
One approach to gaining insight into the above issues is to assess SC development and adult function in the absence of a functional (genomic) AR in AR0/Y mice (10, 21). However, this suffers from several potential problems. First, it is impossible in such animals to establish whether any observed effects result specifically from lack of androgen action on SCs, as opposed to effects on other testicular cells such as peritubular myoid cells. Second, because testes in such animals do not descend into the scrotum (which fails to form), there is the complication in adulthood of the confounding effects of elevated testicular temperature (10). The recent generation by ourselves (22) and others (23, 24) of mice with SC-selective knockout of the AR (SCARKO mice) has opened up new possibilities for elucidating the role that androgens play in regulating SC proliferation, maturation, and function. Moreover, by comparing these processes in SCARKO and complete AR knockout (AR0/Y; ARKO) mice, it is possible to gain key insights into the relative importance of direct (SC) and indirect (non-SC-mediated) effects of androgens on SCs. We used such an approach in the present studies and show that neither the effects of androgens on SC number nor the maturation-dependent change in expression of four SC markers require AR expression in SCs. In contrast, the ability of the SCs to support complete spermatogenesis, as well as expression of a number of SC genes, does require AR expression in SCs.
Materials and Methods
Generation of SCARKO and ARKO mice
The ARKOs were generated using Cre/loxP technology. ARflox/+ female animals (129/Swiss) with exon 2 of the AR floxed were either crossed with anti-Müllerian hormone (AMH)-Cre+/+ male mice (C57BL/6) expressing Cre recombinase (under the AMH gene promoter) selectively in SCs to generate the SCARKO line or to phosphoglycerate kinase-1 (PGK)-Cre+/+ male animals (C57BL/6) expressing Cre ubiquitously to produce the ARKO line. Full details are provided elsewhere (22). All animals were treated according to the National Institutes of Health guide for the care and use of laboratory animals and all experiments were approved by the local ethical committee.
Antibodies for immunohistochemistry and their dilutions
Immunolocalization of AR used a rabbit polyclonal antibody (N20) raised against a peptide within the N-terminal domain of the human AR (sc-816; Santa Cruz Biotechnology, Santa Cruz, CA), which was diluted at 1:200. Mouse monoclonal antibody raised against p27kip1 (NCL-p27; Novocastra, Newcastle Upon Tyne, UK) was used at 1:40. GATA-1 (sc-0266; Santa Cruz Biotechnology) rat monoclonal IgG2a antibody was diluted 1:500 for use and goat polyclonal antibody (sc-6886; Santa Cruz Biotechnology) specific for AMH was used at 1:1000. Rabbit anti-sulfated glycoprotein (SGP)-2 polyclonal antibody, used at 1:100, was a gift from Dr. S. R. Sylvester (Department of Biochemistry and Biophysics, Washington State University, Pullman, WA). Pem antibody, used at 1:500, was kindly donated by Professor M. F. Wilkinson (Department of Immunology, University of Texas M. D. Anderson Cancer Center, Houston, TX).
Immunohistochemistry
Urogenital systems from wild-type (WT), ARflox, AMH-Cre, ARKO, and SCARKO males at ages d 2, 8, 12, 20, 50, and/or 140 were removed, weighed, and fixed in Bouins fluid for 4–6 h before transferring into 70% (vol/vol) ethanol. Testes and epididymides were dissected and processed into paraffin wax using standard methods. Immunohistochemistry was performed on dewaxed 5-μm sections in conjunction with heat-induced antigen retrieval for 5 min in 0.01 M citrate buffer (pH 6.0) (Sigma-Aldrich, Dorset, UK) using a pressure cooker (for AR, GATA-1, p27kip1, and Pem) or no retrieval (for AMH and SGP-2). This was followed by endogenous peroxidase blocking in 3% (vol/vol) H2O2 in methanol for 30 min at room temperature. All washes between antibody or reagent incubations comprised 2 x 5 min at room temperature in Tris-buffered saline [TBS, 0.05 M Tris (pH 7.4), 0.85% (vol/vol) saline]. Tissue sections were first blocked in the appropriate serum elaborated as follows for the different primary antibodies. Rabbit polyclonal primary antibodies were diluted in TBS containing normal swine serum (1:4; Diagnostics Scotland, Carluke, Lanarkshire, UK) and 5% (wt/vol) BSA (Sigma-Aldrich), whereas goat polyclonal and mouse and rat monoclonal antibodies were diluted in TBS containing normal rabbit serum (1:4; Diagnostics Scotland) and 5% (wt/vol) BSA. A swine antirabbit biotinylated second antibody (E0353; Dako, Cambridge, UK) was used in conjunction with rabbit polyclonal antibodies, whereas a rabbit antimouse biotinylated second antibody (E0464; Dako) was used to detect p27kip1 antibody. Biotinylated rabbit antirat secondary antibody (BA-4000; Vector Laboratories, Peterborough, UK) and rabbit antigoat antibody (BA-5000; Vector Laboratories) were used in combination with GATA-1 and AMH antibodies, respectively. All biotinylated secondary antibodies were diluted 1:500 in the appropriate blocking serum in TBS/BSA and incubated at room temperature for 30 min.
Bound antibodies were visualized by incubating the sections with avidin biotin complex-horseradish peroxidase reagent (K0355; Dako) for 30 min followed by color development using 3,3'-diaminobenzidine tetra-hydrochloride chromogenic substrate (K3468, liquid DAB+ kit; Dako), monitored microscopically. Sections were counterstained with hematoxylin, dehydrated, and mounted with Pertex. Images were captured using a Provis microscope (Olympus Optical Co., London, UK) equipped with a DCS330 camera (Eastman Kodak Co., Rochester, NY). Captured images were stored on a Macintosh computer and compiled using Photoshop 7.0 (Adobe, Mountain View, CA).
To enable comparative evaluation of the immunostaining, sections of tissues from control and knockout animals were processed in parallel on at least three occasions to ensure reproducibility of results; on each occasion tissue sections from four to six animals in each group were run. To ensure direct comparability of staining intensities, one section each from control, ARKO, and SCARKO mice were mounted on the same slide.
Determination of SC nuclear volume/number per testis and germ cell nuclear volume per unit SC
Standard stereological methods were used to determine cell nuclear volumes and/or number per testis in groups of five WT, ARflox, AMH-Cre, ARKO, and SCARKO mice (d 2, 50, and 140), using methods detailed elsewhere (12, 25, 26). In brief, cross-sections of testes were examined under oil immersion using a Leitz 63 x plan apo objective fitted to a Leitz laborlux microscope and a 121-point eyepiece graticule. Using a systematic clock-face sampling pattern from a random starting point, 16 fields were counted. Points falling over the nuclei of SC or germ cell nuclei were scored and expressed as a percentage of the total points counted. For each animal, the values for percentage nuclear volume were converted to absolute nuclear volumes per testis by reference to testis volume (weight) because shrinkage was minimal. SC nuclear volumes per testis equate to SC numbers per testis, assuming no change in average nuclear size. However, because the individual nuclear volume of SCs might be influenced by reduced germ cell complement in adult ARKO and SCARKO males, average SC nuclear volume was also determined in four to five animals per group at d 140, using methods similar to those described previously (26, 27). Briefly, images were captured from an Olympus BH2 microscope using a video camera (HV-C20; Hitachi, Tokyo, Japan) and were analyzed with Image Pro Plus software with a Stereology 5.0 plug-in (Media Cybernetics, Berkshire, UK). An area of interest was created by drawing around the SC nucleus, within which the computer program then determined the average length of several diameters measured at two-degree intervals, which passed through the center of the nucleus. This was measured for a minimum of 100 SC nuclei per testis and mean nuclear volume then determined. Data for SC nuclear volume per testis was then converted to absolute numbers of SC per testis by dividing by the average SC nuclear volume.
RNA analysis
Tissue samples were removed and snap frozen in liquid nitrogen. RNA of WT, AMH-Cre, ARflox, and SCARKO testes was prepared using the RNeasy midi kit (QIAGEN, Chatsworth, CA). Due to the small size, RNA from ARKO testes and its appropriate control, PGK-Cre testes, was extracted using the RNeasy mini kit (QIAGEN). Synthesis of cDNA from DNase I-treated total RNA (RNase-free DNase I set; QIAGEN) used Superscript II RNaseH– reverse transcriptase and random hexamer primers (Invitrogen, Carlsbad, CA). To allow specific mRNA levels to be expressed per testis and control for the efficiency of RNA extraction, RNA degradation and the reverse transcription step, an external standard was used (10). The external standard was luciferase mRNA (Promega, Madison, WI), and 10 ng were added to each testis at the start of the RNA extraction procedure. For quantification of gene expression, the ABI Prism 7700 PCR detection system (Applied Biosystems, Foster City, CA) was used with a two-step RT-quantitative-PCR protocol. Components for real-time PCR were obtained from Applied Biosystems, apart from primers and probes (Eurogentec, Sar-Tilman, Belgium) and Sybrgreen (Sigma, St. Louis, MO). Each 25-μl real-time PCR contained 1x buffer A, 5 mM MgCl2, 400 μM deoxynucleotide triphosphates, 200 nM of each primer, 0.4x Sybrgreen, and 0.025 U/μl Amplitaq Gold enzyme. Amplified samples were electrophoresed on polyacrylamide gels to ensure that only a single band was amplified in each PCR. Primer sequences were for androgen binding protein (ABP): forward, GGAACAAATTCTTCTTGGCTTAACA, reverse, CTTAGGTTCTGCCTCGGAAGAC; transferrin: forward, GGACGCCATGACTTTGGATG, reverse, GCCATGACAGGCACTAGACCA; or as described elsewhere [AMH, claudin-11, luciferase, FSH receptor, and platelet-derived growth factor (PDGF)-A (10); cystatin-related gene highly expressed in testis and epididymis (cystatin-TE) and fatty acid binding protein (FABP) (28)]. For Pem, a probe was used to ensure specificity [primers and probe (22)], and 100 nM of each primer and 400 nM probe were added to the reaction. The quantity of measured mRNA was expressed relative to the luciferase standard in the same sample. All samples and standard curves were run in triplicate.
Statistical analysis
With the exception of the real-time PCR data whereby a two-sample t test was employed, statistical analysis was performed by using one-way ANOVA, supplemented with a Fisher multiple comparison test by using NCSS 2000 software (NCSS Statistical Analysis and Data Analysis Software, Kaysville, UT).
Results
To establish that any changes in phenotype observed in SCARKO males did not stem from aberrant changes arising in the founder transgenic lines (ARflox, AMH-Cre, or PGK-Cre lines), we demonstrated that testis weight, SC nuclear volume per testis, AR immunoexpression, and functional maturation of SCs (using the markers described below) were comparable in the founder lines and WT mice at d 2 (to reflect fetal development) and at d 50 and 140 to encompass early- and midadulthood, respectively (data not shown). Therefore, for the main studies reported below, either WT or AMH-Cre mice were used as controls, unless otherwise stated.
SC nuclear volume/number per testis in ARKO and SCARKO males
At d 2, SC nuclear volume per testis was already reduced significantly by 32% in ARKO males, compared with controls. This decrement increased to 57% on d 50 and to more than 70% on d 140 (Fig. 1). In contrast, in SCARKO males SC nuclear volume per testis was comparable with controls at d 2 and 50, whereas a small (16%) decrement was evident at d 140 (Fig. 1). Determination of SC number at d 140 in the same animals revealed changes in ARKO and SCARKO animals almost identical with that shown for SC nuclear volume in Fig. 1 (control AMH-Cre 3.14 ± 0.12 x 106 SC per testis; ARKO 0.56 ± 0.10 x 106; SCARKO 2.66 ± 0.18 x 106; means ± SEM, n = 4–5 per group). The SC number in ARKOs was significantly different (P < 0.001) from the controls, but the difference in SCARKOs only approached statistical significance (0.05 < P < 0.1).
FIG. 1. SC nuclear volume per testis at d 2, 50, and 140 in control, ARKO, and SCARKO animals. Numerical figures at d 2 reflect primarily SC proliferation in fetal life, whereas values at d 50 and 140 reflect final SC nuclear volume (number) per testis (note difference in scale at d 2). Values are means ± SEM for four to nine (d 50 and 140) or five to 11 (d 2) animals. *, P < 0.05, ***, P < 0.001, in comparison with respective control group.
Maturation-dependent changes in immunoexpression of SC markers in ARKO and SCARKO males
SC maturational development was assessed at d 2, 12, 20, and 50 of age in four to seven animals per group using four specific protein markers. We selected a protein that should be switched off during maturation (AMH), two that should be switched on acutely at around the time of SC maturation (p27kip1, GATA-1), and one that is switched on progressively during SC maturation (SGP-2) (6). Additionally, because AR expression is normally switched on in SCs just before their maturation (5, 9), it was checked whether expression occurred at the appropriate time in all control animals (WT and founder lines) and did not occur at any age in SCARKO and ARKO males.
In controls, AMH expression in SC cytoplasm was intense at d 2, had declined but was still evident at d 12, and was nondetectable at d 20 and 50. SCARKO and ARKO males showed an identical age-related change in AMH immunoexpression (Fig. 2). The absence of any difference in AMH expression between male controls and SCARKOs was confirmed by quantitative PCR (d 50) (see below) and Western blotting (d 2–50) (data not shown).
FIG. 2. Immunoexpression profiles for AMH and p27kip1 in SCs from control, ARKO, and SCARKO testes at d 2, 12, 20, and 50. Note that AMH is expressed in SC cytoplasm predominantly at d 2 with a major decrease by d 12 and thereafter, whereas the opposite pattern is seen with nuclear expression of p27kip1. Bar, 100 μm.
The cyclin-dependent kinase inhibitor, p27kip1, inhibits D-cyclin mediated G1 to S progression in the cell cycle and its nuclear expression in the SCs is first evident at around the time of terminal differentiation (maturation) and cessation of proliferation (6, 29). Consistent with this, SC in control mice were immunonegative for p27kip1 at d 2, expression was weakly apparent in some SC nuclei at d 12, and was intense in all SC nuclei at d 20 and 50. Comparable age-dependent patterns of p27kip1 immunoexpression were found in both SCARKO and ARKO males (Fig. 2).
SGP-2 is one of the major SC-secreted proteins (30), although its function remains unclear. In controls, SGP-2 immunoexpression was weak in SC cytoplasm at d 2 and increased progressively in intensity between d 12 and 50, during which period the intensity of its immunoexpression became increasingly stage dependent; a similar age-dependent increase in SGP-2 immunoexpression was evident in SCARKO and ARKO animals (see supplemental data, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org).
GATA-1 is a zinc finger transcription factor that is first expressed in SCs during their maturation and development in the first wave of spermatogenesis (31). GATA-1 immunoexpression in controls was nondetectable at d 2 but was prominent in SC nuclei at d 12 and thereafter, and a similar age-dependent increase in GATA-1 immunoexpression was evident in SCARKO and ARKO animals (supplemental data).
AR immunoexpression showed an identical age-dependent pattern of expression in SCs in all controls (AMH-Cre, PGK-Cre, ARflox, and WT). Nuclear immunostaining was not detectable in SCs at d 2 but was observed from d 8 onward and was notably stage dependent in adulthood (Fig. 3). In all controls, there was uniform nuclear AR immunoexpression in Leydig and peritubular myoid cells at all ages (Fig. 3). In ARKO and SCARKO males, no AR immunoexpression in SCs was evident at any age, but in SCARKO animals normal AR immunoexpression was observed in Leydig and peritubular myoid cells, comparable with that in controls (Fig. 3).
FIG. 3. Immunoexpression profile for the AR in control, ARKO, and SCARKO testes at d 2–50 of age to illustrate ontogeny of AR expression in SCs. Note the AR immunoexpression in d 12–50 testes of control animals (B–D) with stage-dependent SC (arrowheads) nuclear staining at d 50 (D) and that SCARKO testes exhibit no SC AR immunoexpression at any age (J–M), whereas nuclei of Leydig cells (L) and PTMCs (arrows) are immunopositive for AR expression at all ages in control and SCARKO mice (A–D and J–M, respectively). Note the complete absence of AR immunoexpression in ARKO testes (E–H). Bar, 100 μm.
Having established that SCs in ARKO and SCARKO males had exhibited a normal age-dependent pattern of change in the expression of four recognized SC nuclear and cytoplasmic markers, we next investigated their functional competence in two ways, first by quantifying their capacity to support germ cell development at various ages and second by quantifying mRNA expression for nine genes shown previously to be expressed by adult SCs and/or to be androgen regulated.
Functional capacity of SCs to support germ cells in ARKO and SCARKO males and testis weight
SCs in SCARKO males supported a normal volume of spermatogonia at d 2, 50, and 140, although there was a downward trend at d 140. In contrast, total germ cell volume supported per unit SC volume in SCARKOs was decreased by 69 and 76% at d 50 and d 140, respectively (Fig. 4). At d 2, ARKO SCs supported significantly more (126%) spermatogonia than in controls, but at d 50 and 140, they supported a significantly lower volume of spermatogonia, with approximately a 60% reduction relative to controls or SCARKOs (Fig. 4). This impaired function of ARKO SCs was manifested more dramatically with total germ cell volume per unit SCs, which was reduced by approximately 95% at d 50 and 140 relative to controls (Fig. 4). The age-dependent changes in germ cell volume per SC in ARKO and SCARKO males was matched by changes in overall testis weight (Table 1).
FIG. 4. Functional ability of SCs to support germ cells at d 2, 50, and 140 in control, ARKO, and SCARKO animals. Data are expressed as germ cell nuclear volume per unit SC nuclear volume to take account of differences in SC nuclear volume shown in Fig. 1 and are shown separately for spermatogonia (top row) and total germ cells (spermatogonia + spermatocytes + spermatids) (bottom row). *, P < 0.05, ***, P < 0.001, in comparison with respective control group.
TABLE 1. Testis weights (mean ± SEM) at age d 2, 50, and 140 in WT, ARKO, and SCARKO males and in their respective controls (PGK-Cre for ARKOs; AMH-Cre for SCARKOs)
Expression of SC genes
Nine genes were selected for study based on one or more of the following criteria: expression restricted to SCs in the testis, evidence of androgen dependence, role in spermatogenesis, expression in SCs modifiable by peritubular myoid cell factors, and previous study in ARKO mice (10, 21, 23, 32, 33, 34, 35, 36, 37, 38). Results are presented in Fig. 5. The amount of mRNA in d 50 SCARKO and ARKO testes were each compared with their respective controls (AMH-Cre and PGK-Cre littermates, respectively). The expression levels of these nine genes normalized to absolute levels per testis (using the known spiked concentrations of luciferase added per testis before RNA extraction) are thus not affected by testicular germ cell composition and are shown relative to their controls (arbitrarily set equal to 1). Because there were also significant changes in SC nuclear volume/number per testis, especially in ARKOs, mRNA expression has also been shown in Fig. 5 after correction for SC nuclear volume, using the mean values per group for d 50 animals illustrated in Fig. 1.
FIG. 5. Changes in expression of SC genes in 50-d-old ARKO and SCARKO mice, compared with the respective controls (PGK-Cre for ARKO and AMH-Cre for SCARKO). Testicular RNA was extracted and cDNA was prepared as described in Materials and Methods. Real-time PCR was used to measure cDNA levels relative to an external standard (luciferase) added during RNA extraction. Upper panel, To compare expression levels per testis, results were calculated as the number of cDNA copies per 106 copies of luciferase and were expressed as a fraction of the respective control, arbitrarily set equal to 1. Lower panel, To compare expression levels per SC nuclear volume, the number of cDNA copies per 106 copies of luciferase was divided by the appropriate SC nuclear volume for each genotype (derived from the data in Fig. 1). Results are expressed as a fraction of the respective control, arbitrarily set equal to 1. Values shown are means ± SEM for six to eight animals per group. Data indicated by an asterisk are significantly different (P < 0.05) from the respective control as assessed by a two-sample t test.
In both AR knockout models, the most dramatic change in expression was observed for Pem. Pem transcript levels in the testis were reduced to less than 0.4% of control values in ARKOs as well as SCARKOs (Fig. 5), a change matched by Pem protein immunohistochemistry (Fig. 6).
FIG. 6. Immunoexpression of Pem protein in SC in d 50 control, ARKO, and SCARKO testes. Bar, 100 μm.
In the AR Knockout (ARKO) testis, significant reductions in gene expression for ABP, cystatin-TE, claudin-11, and the FSH receptor were found (Fig. 5), although this reduction was no longer evident after correction for SC nuclear volume per testis, which even resulted in a small but significant increase in FSH receptor mRNA expression (Fig. 5). A markedly more severe reduction in gene expression was observed in ARKO testes for FABP, transferrin, and PDGF-A, and correction for SC nuclear volume per testis only slightly attenuated these decreases (Fig. 5). In contrast, no reduction in AMH gene expression was observed in ARKOs, and once the reduction in SC number was taken into account, AMH expression per SC was more than doubled (Fig. 5). These findings are in accordance with the observations of Johnston et al. (10).
In SCARKO testes no overexpression of AMH mRNA was observed (Fig. 5), in contrast with a previous report (23). Similarly, ABP and the FSH receptor mRNA were expressed at levels comparable with the control, whereas transferrin expression was slightly reduced (29%), and reductions of 36–60% were noted for FABP, PDGF-A, cystatin-TE, and claudin-11 (Fig. 5); the latter changes were still evident after correction for SC nuclear volume per testis (Fig. 5).
Discussion
The primary aim of the present study was to evaluate the role that androgen action in the testis plays in determining the numbers and functional development of SCs. For this purpose we compared ARKO mice, which lack AR in all testicular cells, with SCARKO mice in which AR expression was deleted selectively in SCs.
A role for androgens in regulating SC proliferation/final number in rodents had not been seriously considered in the past, perhaps because SCs do not normally express the AR for much of the period in which proliferation takes place (5, 9). This perspective has changed recently due to a study (10) that has shown that testicular feminized mice (tfm) exhibit an approximately 50% deficit in SC numbers at birth and a 75% deficit in adulthood and another study (12) showing that, in the neonatal rat, endogenous androgens play a physiological role in increasing SC number. Our present findings in ARKO males, which are analogous phenotypically to tfm mice, confirm the observations of Johnston et al. (10) by showing reductions in SC nuclear volume/number per testis of 31, 58, and 71–82% at d 2, 50, and 140, respectively, in ARKO, compared with control males. However, what the present findings in SCARKOs demonstrate is that AR expression in SCs is not required for attainment of normal SC number, indicating that AR expression in another testicular cell type must be important for increasing SC proliferation perinatally in mice. Although the present findings do not allow unambiguous identification of this cell type(s), it seems highly likely that it is the peritubular myoid cell (PTMC). These cells express AR intensely throughout fetal and postnatal life, and there is abundant evidence in the literature, based mainly on coculture of isolated PTMCs and SCs, to show that secretions from the PTMCs can modify various aspects of SC function (15, 16), including the production of transferrin (16, 39, 40). It is also well established that PTMCs and SCs cooperate to lay down the basal lamina of the seminiferous cords (41, 42) to which the SCs adhere, and this could be a factor that limits SC proliferation via processes such as contact inhibition (43). Moreover, there is evidence that activin A produced by PTMCs may stimulate SC proliferation (44). Whether the production of activin A or other growth factors by PTMCs is androgen controlled remains to be shown. If proliferation of SCs is regulated indirectly via androgen action on neighboring PTMCs, it would provide an intriguing parallel with the rest of the developing male reproductive tract in which proliferation and differentiation of epithelial cells is regulated indirectly via androgen action on neighboring mesenchymal/stromal cells (45).
SC maturation, or terminal differentiation, is the process via which SCs lose their proliferative ability and acquire functions that will enable them to support spermatogenesis (6). Among other changes this involves marked down-regulation of AMH expression (6) and the switching on of p27kip1 (29) and GATA-1 (31) as well as the progressive up-regulation of SGP-2 (6, 46). We selected these markers on the basis of their SC specificity in the testis and because of their differing patterns and locations of expression (two nuclear vs. two cytoplasmic). To our surprise, the maturation-dependent changes in expression of all four proteins occurred normally in both ARKO and SCARKO males, demonstrating that at least some key aspects of SC functional maturation do not require either direct or indirect androgen action via the AR. In this regard, our finding that AMH expression is down-regulated normally in SCARKO testes at the same age as in controls, contrasts with the report of another group that also generated SC-selective AR knockouts (23). These authors did not study AMH expression throughout development as we have done, nor did they examine expression of other SC markers modulated during SC maturation. Our finding of normal maturation-dependent reduction in AMH expression in ARKOs (despite increased mRNA levels when expressed per SC in adulthood) is consistent with earlier data reported for tfm mice (10, 47).
The present findings confirm and extend previous findings by ourselves (22) and others (10, 23, 47) in AR-deficient mice, in showing that SCs in adulthood are incapable of supporting meiotic and postmeiotic germ cell development in the absence of androgen action via the AR. In this regard, ARKO males were considerably more deficient than were SCARKOs in adulthood because not only did their SC support fewer germ cells per SC, but also even SC support of spermatogonia was severely compromised. In contrast, SCs in SCARKOs supported normal numbers of spermatogonia and showed deficits only in postmeiotic germ cell support. The extent to which this difference reflects the abdominal vs. scrotal position of testes in ARKO and SCARKO adults, respectively, is unclear, but it is notable that before the time of normal testis descent (d 2), SCs in ARKO clearly supported normal/supranormal numbers of spermatogonia.
Data on gene expression in adult ARKO and SCARKO mice cover seven genes representative for diverse SC functions that have been analyzed previously in normal and tfm mice (10). Two genes encoding lipid-binding proteins (ABP, FABP), a tissue remodeling factor (cystatin-TE), a component of the tight junction complex (claudin-11), and three proteins related to endocrine or paracrine functions (PDGF-A, AMH, and FSH receptor). In addition, we included Pem as a control for direct androgen action on SCs and transferrin as a SC product, the secretion of which is known to be modulated by PTMCs. Gene expression data should be interpreted with caution because aberrant expression may be related to not only the absence of AR (ubiquitous or SC specific) but also indirect effects caused by absence of postmeiotic germ cells, changes in FSH secretion, and (in ARKOs) abnormal location of the testis. Data on AMH gene expression have been discussed above.
In confirmation of our previous studies, Pem expression was virtually absent in both ARKOs and SCARKOs. Immunostaining studies confirmed these differences at the protein level and indicated SC localization compatible with the proposed direct regulation by androgens (36). The expression of PDGF-A and FABP was also clearly reduced in SCARKOs and even more so in ARKOs (even after correction for the reduction in SC nuclear volume per testis). Whereas the data of Johnston et al. (10) in tfm mice suggest that the decrease in these parameters may be due to cryptorchidism, our data in SCARKOs, which unlike ARKOs have normally descended testes, indicate that this may not be the correct interpretation. One explanation could be that impaired expression of these genes is due to the decreased germ cell complement. In the adult testis, PDGF-A is expressed primarily in SCs, although weak expression in occasional interstitial cells may also occur (48); expression in the SCs occurs in a stage-dependent manner (48), a phenomenon frequently associated with germ cell modulation of SC function (49). Because PDGF-A has been shown to play an essential role in adult Leydig cell development (48), the marked reduction in expression of PDGF-A in testes of both ARKOs and SCARKOs could suggest that Leydig cell development is altered in these animals, and this is currently under investigation.
The limited reduction in the expression of the transferrin gene in SCARKOs and the severe reduction in ARKOs are of interest because earlier studies have indicated that PTMCs and their secretion products positively modulate transferrin production by SCs with some evidence of this being androgen regulated (16, 39, 40). The present findings are compatible with this hypothesis, but the slight reduction in transferrin mRNA in SCARKOs suggests that there may be additional control mechanisms involving either direct effects of androgens on SCs or indirect effects caused by the reduced germ cell complement (50, 51).
Surprisingly, cystatin-TE and claudin-11 expression levels were approximately normal in ARKOs after correction for SC number but were significantly reduced in SCARKOs. The normal expression levels of cystatin-TE and claudin-11 in ARKOs confirms a previous study in tfm animals (10), but the partial suppression of mRNA for these genes in SCARKOs is intriguing. It seems unlikely that this decrease could be explained by absence of particular germ cell types because ARKO animals lacked all of the same germ cell types as did SCARKOs but did not exhibit significant down-regulation of cystatin-TE and claudin-11 after correction for SC nuclear volume per testis. An alternative explanation might be that the higher FSH levels observed in ARKOs, compared with SCARKOs (22), could have led to increased expression of the mRNA for cystatin-TE and claudin-11 in the former. Claudin-11 has recently been demonstrated to be regulated at least in part by FSH (38).
Finally, when corrected for SC nuclear volume per testis, mRNA levels for ABP were normal in SCARKOs and ARKOs. A similar absence of androgen regulation of ABP expression has previously been observed in irradiated rats treated with GnRH analogs and antiandrogens (52). In SCARKOs and ARKOs, expression of the FSH receptor gene was slightly elevated after correction for SC nuclear volume per testis, and in the case of ARKOs, this was statistically significant. Whether this change is also related to altered FSH levels in these two lines (22) or is due to lack of androgen action on the SCs is a matter for speculation. Taken together, the described gene expression data show that, despite a normal pattern of change in expression of four proteins associated with SC maturation, significant changes occur in the expression of several SC genes in ARKOs and SCARKOs. Comparison of the gene expression pattern in these two lines, after correction for change in SC nuclear volume/number, provides a starting point for dissecting which genes might be regulated directly by androgen action on SCs. In this regard, the change in expression of Pem was dramatic (but expected), but the reduced expression of FABP and PDGF-A was also intriguing. Although it remains possible that expression of either or both of these genes by SCs could be secondary to altered signaling to the SCs from the reduced germ cell complement in ARKOs and SCARKOs, the alternative explanation that they might to some extent be directly androgen regulated merits further study.
In conclusion, the present studies show that complete ablation of androgen action in ARKO males results in major deficits in SC numbers, although at least some aspects of the SC maturation process appear to occur normally in these animals. In the selective absence of androgen action on SCs (SCARKO), these cells develop normally with respect to both their numbers and the expression of four selected maturational markers, but they are clearly unable to support late meiotic and postmeiotic germ cells; this is associated with changes in expression of a number of SC-expressed genes, in particular Pem, FABP, and PDGF-A. Given their normal testicular descent and the normality of expression of at least some indicators of SC functional maturation, SCARKO males will hopefully provide the means via which the SC functions necessary for survival of late meiotic and postmeiotic germ cells can be identified.
Acknowledgments
We thank Arantza Esnal, Hilde Geeraerts, and Ludo Deboel for their excellent technical assistance.
References
Hughes IA 2001 Minireview: sex differentiation. Endocrinology 142:3281–3287
Sultan C, Paris F, Terouanne B, Balaguer P, Georget V, Poujol N, Jeandel C, Lumbroso S, Nicolas JC 2001 Disorders linked to insufficient androgen action in male children. Hum Reprod Update 7:314–322
Sharpe RM 1994 Regulation of spermatogenesis. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven Press; 1363–1433
Hill CM, Anway MD, Zirkin BR, Brown TR 2004 Intratesticular androgen levels, androgen receptor localization, and androgen receptor expression in adult rat Sertoli cells. Biol Reprod 71:1348–1358
Sharpe RM 2005 Sertoli cell endocrinology and signal transduction: androgen regulation. In: Griswold M, Skinner M, eds. Sertoli cell biology. San Diego: Academic Press; 199–216
Sharpe RM, McKinnell C, Kivlin C, Fisher JS 2003 Proliferation and functional maturation of Sertoli cells, and their relevance to disorders of testis function in adulthood. Reproduction 125:769–784
Johnson L, Zane RS, Petty CS, Neaves WB 1984 Quantification of the human Sertoli cell population: its distribution, relation to germ cell numbers, and age-related decline. Biol Reprod 31:785–795
Orth JM, Gunsalus GL, Lamperti AA 1988 Evidence from Sertoli cell-depleted rats indicates that spermatid number in adults depends on numbers of Sertoli cells produced during perinatal development. Endocrinology 122:787–794
Bremner WJ, Millar MR, Sharpe RM, Saunders PT 1994 Immunohistochemical localization of androgen receptors in the rat testis: evidence for stage-dependent expression and regulation by androgens. Endocrinology 135:1227–1234
Johnston H, Baker PJ, Abel M, Charlton HM, Jackson G, Fleming L, Kumar TR, O’Shaughnessy PJ 2004 Regulation of Sertoli cell number and activity by follicle-stimulating hormone and androgen during postnatal development in the mouse. Endocrinology 145:318–329
Zhou Q, Shima JE, Nie R, Friel PJ, Griswold MD, 2005 Androgen-regulated transcripts in the neonatal mouse testis as determined through microarray analysis. Biol Reprod 72:1010–1019
Atanassova NN, Walker M, McKinnell C, Fisher JS, Sharpe RM 2005 Evidence that androgens and oestrogens, as well as follicle-stimulating hormone, can alter Sertoli cell number in the neonatal rat. J Endocrinol 184:107–117
Schlatt S, Arslan M, Weinbauer GF, Behre HM, Nieschlag E 1995 Endocrine control of testicular somatic and premeiotic germ cell development in the immature testis of the primate Macaca mulatta. Eur J Endocrinol 133:235–247
Ramaswamy S, Plant TM, Marshall GR 2000 Pulsatile stimulation with recombinant single chain human luteinizing hormone elicits precocious Sertoli cell proliferation in the juvenile male rhesus monkey (Macaca mulatta). Biol Reprod 63:82–88
Skinner MK, Fritz IB 1985 Testicular peritubular cells secrete a protein under androgen control that modulates Sertoli cell functions. Proc Natl Acad Sci USA 82:114–118
Swinnen K, Cailleau J, Heyns W, Verhoeven G 1990 Prostatic stromal cells and testicular peritubular cells produce similar paracrine mediators of androgen action. Endocrinology 126:142–150
Migliaccio A, Castoria G, Di Domenico M, de Falco A, Bilancio A, Lombardi M, Barone MV, Ametrano D, Zannini MS, Abbondanza C, Auricchio F 2000 Steroid-induced androgen receptor-oestradiol receptor ?-Src complex triggers prostate cancer cell proliferation. EMBO J 19:5406–5417
Losel RM, Falkenstein E, Feuring M, Schultz A, Tillmann HC, Rossol-Haseroth K, Wehling M 2003 Nongenomic steroid action: controversies, questions, and answers. Physiol Rev 83:965–1016
Silva FR, Leite LD, Wassermann GF 2002 Rapid signal transduction in Sertoli cells. Eur J Endocrinol 147:425–433
Fix C, Jordan C, Cano P, Walker WH 2004 Testosterone activates mitogen-activated protein kinase and the cAMP response element binding protein transcription factor in Sertoli cells. Proc Natl Acad Sci USA 101:10919–10924
Yeh S, Tsai MY, Xu Q, Mu XM, Lardy H, Huang KE, Lin H, Yeh SD, Altuwaijri S, Zhou X, Xing L, Boyce BF, Hung MC, Zhang S, Gan L, Chang C, Hung MC 2002 Generation and characterization of androgen receptor knockout (ARKO) mice: an in vivo model for the study of androgen functions in selective tissues. Proc Natl Acad Sci USA 99:13498–13503
De Gendt K, Swinnen JV, Saunders PTK, Schoonjans L, Dewerchin M, Devos A, Tan K, Atanassova N, Claessens F, Lécureuil C, Heyns W, Carmeliet P, Guillou F, Sharpe RM, Verhoeven G 2004 A Sertoli cell-selective knockout of the androgen receptor causes spermatogenic arrest in meiosis. Proc Natl Acad Sci USA 101:1327–1332
Chang C, Chen YT, Yeh SD, Xu Q, Wang RS, Guillou F, Lardy H, Yeh S 2004 Infertility with defective spermatogenesis and hypotestosteronemia in male mice lacking the androgen receptor in Sertoli cells. Proc Natl Acad Sci USA 101:6876–6881
Holdcraft RW, Braun RE 2004 Androgen receptor function is required in Sertoli cells for the terminal differentiation of haploid spermatids. Development 131:459–467
Atanassova N, McKinnell C, Walker M, Turner KJ, Fisher JS, Morley M, Millar MR, Groome NP, Sharpe RM 1999 Permanent effects of neonatal estrogen exposure in rats on reproductive hormone levels, Sertoli cell number, and the efficiency of spermatogenesis in adulthood. Endocrinology 140:5364–5373
Sharpe RM, Walker M, Millar MR, Atanassova NN, Morris K, McKinnell C, Saunders PTK, Fraser HM 2000 Effect of neonatal gonadotropin-releasing hormone antagonist administration on Sertoli cell number and testicular development in the marmoset: comparison with the rat. Biol Reprod 62:1685–1693
Mahood IK, McKinnell C, Fisher JS, Walker M, Hallmark N, Sharpe RM 2005 Abnormal Leydig cell aggregation in the fetal testis of rats exposed to Di(n-butyl) phthalate and its possible role in testicular dysgenesis. Endocrinology 146:613–623
O’Shaughnessy PJ, Fleming L, Baker PJ, Jackson G, Johnston H 2003 Identification of developmentally regulated genes in the somatic cells of the mouse testis using serial analysis of gene expression. Biol Reprod 69:797–808
Beumer TL, Kiyokawa H, Roepers-Gajadien HL, van den Bos LA, Lock TM, Gademan IS, Rutgers DH, Koff A, de Rooij DG 1999 Regulatory role of p27kip1 in the mouse and human testis. Endocrinology 140:1834–1840
Sharpe RM, Millar M, McKinnell C 1993 Relative roles of testosterone and the germ cell complement in determining stage-dependent changes in protein secretion by isolated rat seminiferous tubules. Int J Androl 16:71–81
Yomogida K, Ohtani H, Harigae H, Ito E, Nishimune Y, Engel JD, Yamamoto M 1994 Developmental stage- and spermatogenic cycle-specific expression of transcription factor GATA-1 in mouse Sertoli cells. Development 120:1759–1766
Lindsey JS, Wilkinson MF 1996 Pem: a testosterone- and LH-regulated homeobox gene expressed in mouse Sertoli cells and epididymis. Dev Biol 179:471–484
Danzo BJ, Pavlou SN, Anthony HL 1990 Hormonal regulation of androgen-binding protein in the rat. Endocrinology 127:2829–2838
Skinner MK, Griswold MD 1982 Secretion of testicular transferrin by cultured Sertoli cells is regulated by hormones and retinoids. Biol Reprod 27:211–221
Loveland KL, Zlatic K, Stein-Oakley A, Risbridger G, deKretser DM 1995 Platelet-derived growth factor ligand and receptor subunit mRNA in the Sertoli and Leydig cells of the rat testis. Mol Cell Endocrinol 108:155–159
Li Y, Friel PJ, Robinson MO, McLean DJ, Griswold MD 2002 Identification and characterization of testis- and epididymis-specific genes: cystatin SC and cystatin TE-1. Biol Reprod 67:1872–1880
Byers S, Graham R, Dai HN, Hoxter B 1991 Development of Sertoli cell junctional specializations and the distribution of the tight-junction-associated protein ZO-1 in the mouse testis. Am J Anat 191:35–47
Hellani A, Ji J, Mauduit C, Deschildre C, Tabone E, Benahmed M 2000 Developmental and hormonal regulation of the expression of oligodendrocyte-specific protein/claudin-11 in mouse testis. Endocrinology 141:3012–3019
Janecki A, Steinberger A 1987 Vectorial secretion of transferrin and androgen binding protein in Sertoli cell cultures: effect of extracellular matrix, peritubular myoid cells and medium composition. Mol Cell Endocrinol 52:125–135
Anthony CT, Rosselli M, Skinner MK 1991 Actions of the testicular paracrine factor (P-Mod-S) on Sertoli cell transferrin secretion throughout pubertal development. Endocrinology 129:353–360
Tung PS, Fritz IB 1987 Morphogenetic restructuring and formation of basement membranes by Sertoli cells and testis peritubular cells in co-culture: inhibition of the morphogenetic cascade by cyclic AMP derivatives and by blocking direct cell contact. Dev Biol 120:139–153
Skinner MK, Tung PS, Fritz IB 1985 Cooperativity between Sertoli cells and testicular peritubular cells in the production and deposition of extracellular matrix components. J Cell Biol 100:1941–1947
Schlatt S, de Kretser DM, Loveland KL 1996 Discriminative analysis of rat Sertoli and peritubular cells and their proliferation in vitro: evidence for follicle-stimulating hormone-mediated contact inhibition of Sertoli cell mitosis. Biol Reprod 55:227–235
Buzzard JJ, Farnworth PG, De Kretser DM, O’Connor AE, Wreford NG, Morrison JR 2003 Proliferative phase Sertoli cells display a developmentally regulated response to activin in vitro. Endocrinology 144:474–483
Cunha GR, Foster B, Thomson A, Sugimura Y, Tanji N, Tsugi M, Terada N, Finch PW, Donjacour AA 1995 Growth factors as mediators of androgen action during development of the male urogenital tract. World J Urol 13:264–276
Zakeri Z, Curto M, Hoover D, Wightman K, Engelhardt J, Smith FF, Kierszenbaum AL, Gleeson T, Tenniswood M 1992 Developmental expression of the S35–S45/SGP-2/TRPM-2 gene in rat testis and epididymis. Mol Reprod Dev 33:373–384
Al-Attar L, No?l K, Dutertre M, Belville C, Forest MG, Burgoyne PS, Josso N, Rey R 1997 Hormonal and cellular regulation of Sertoli cell anti-Müllerian hormone production in the postnatal mouse. J Clin Invest 100:1335–1343
Gnessi L, Basciani S, Mariani S, Arizzi M, Spera G, Wang C, Bondjers C, Karlsson L, Betsholtz C 2000 Leydig cell loss and spermatogenic arrest in platelet-derived growth factor (PDGF)-A-deficient mice. J Cell Biol 149:1019–1026
Jegou B 1993 The Sertoli-germ cell communication network in mammals. Int Rev Cytol 147:25–96
Stallard BJ, Griswold MD 1990 Germ cell regulation of Sertoli cell transferrin mRNA levels. Mol Endocrinol 4:393–401
Maguire SM, Millar MR, Sharpe RM, Gaughan J, Saunders PT 1997 Investigation of the potential role of the germ cell complement in control of the expression of transferrin mRNA in the prepubertal and adult rat testis. J Mol Endocrinol 19:67–77
Verhoeven G, Deboel L, Swinnen J, Rombauts L, Vanderhoydonc F, Rosseels J, Hoeben E, Heyns W 1995 Effects of androgens on the germ cell-depleted testes of prenatally irradiated rats. Int J Androl 16:22–34(Karen A. L. Tan1, Karel D)
Address all correspondence and requests for reprints to: Professor G. Verhoeven; Laboratory for Experimental Medicine and Endocrinology, Onderwijs en Navorsing, Gasthuisberg, Herestraat 49 bus 902, B-3000 Leuven, Belgium. E-mail: guido.verhoeven@med.kuleuven.ac.be.
Abstract
The role of androgens in the proliferation and maturation of Sertoli cells (SC) and the development of their capacity to support spermatogenesis remains poorly understood. We evaluated these functions in complete androgen receptor knockout (ARKO) and SC-selective androgen receptor knockout (SCARKO) mice. Compared with controls, ARKO mice exhibited a progressive reduction in SC number/testis, whereas SCARKOs showed minor changes, suggesting that androgen effects on SC number are not mediated via direct action on SCs. Immunoexpression of anti-Müllerian hormone (AMH), p27kip1, GATA-1, and sulfated glycoprotein-2, which changes according to SC maturational status, occurred normally in ARKOs and SCARKOs. Functional capacity of SCs to support spermatogonia was similar in SCARKOs and controls, whereas ARKOs showed reduced capacity with age. SC capacity to support total germ cells revealed major deficits in ARKO and SCARKO adults, particularly with respect to postmeiotic germ cells. Using quantitative RT-PCR, the expression of SC markers was compared in d 50 testes. In ARKOs, expression of Pem, fatty acid binding protein, platelet-derived growth factor-A, and transferrin were all significantly reduced, whereas FSH receptor and AMH were increased. In SCARKOs, there were modest reductions in expression of cystatin-related gene highly expressed in testis and epididymis (cystatin-TE) and claudin-11, whereas expression of Pem, fatty acid binding protein, and platelet-derived growth factor-A was markedly reduced, highlighting these as potentially androgen-regulated SC genes that merit further study. In conclusion, androgen action is not required for maturation-dependent changes in immunoexpression of the SC markers AMH, p27kip1, GATA-1, and sulfated glycoprotein-2 but is essential for expression of other SC genes, the attainment of normal SC number, and the support of meiotic and postmeiotic germ cell development.
Introduction
IT IS WELL ESTABLISHED that androgens play the central role in masculinization of the reproductive tract, genitalia, and many other organ systems during the sexual differentiation process in the male (1, 2). It is also established that, in adulthood, androgen action on seminiferous tubules is essential for full, quantitatively normal spermatogenesis and fertility, with the majority of evidence pointing to this effect being mediated by effects on Sertoli cells (SCs), although the mechanisms remain unclear (3, 4, 5). Between the periods of testicular/sexual differentiation in fetal life and the establishment of complete spermatogenesis in adulthood, two other key changes occur to SCs that are essential for normal fertility in adulthood. First, SCs undergo intense proliferation during fetal, neonatal, and peripubertal life (6) to ensure that high levels of sperm production can be achieved in adulthood; SC number is the main determinant of the sperm production rate per testis in adulthood (6, 7, 8). Second, SCs have to undergo a maturation process (terminal differentiation), whereby they lose the ability to proliferate and switch on functions that are essential for the support of germ cells during spermatogenesis (6).
There are various lines of evidence that point to a role for androgens in the SC maturation process, although this is largely circumstantial (4, 6). Such a role for androgens seems intrinsically logical when considering that maturation and acquisition of full androgen responsiveness are prerequisites for SCs to support full spermatogenesis in adulthood, and notable in this regard is that the onset of androgen receptor (AR) expression in SCs precedes onset of SC maturation (4, 6). In contrast, until recently, it was generally believed that (in rodents) androgens played little if any role in SC proliferation, primarily because the AR is not normally expressed in SCs for the majority of the periods during which they proliferate (4, 6, 9). However, recent data from mice lacking a functional AR (10), microarray analysis of androgen-regulated transcripts in neonatal mice (11), and studies involving manipulation of androgen production/action in neonatal rats (12) demonstrate that androgens probably play a physiological role in promoting SC proliferation in fetal and early postnatal life in rodents. There is similar evidence for a role for androgens in postnatal SC proliferation in nonhuman primates (13, 14). These new findings raise a conundrum, namely how do androgens exert such a positive effect on SC proliferation in the absence of AR immunoexpression in fetal and early neonatal life? One obvious explanation is that androgens act via peritubular myoid cells to modify SC proliferation because there is clear in vitro evidence that the function of SCs can be modified by such interactions (15, 16). Alternatively, the proliferative effects of androgens on SCs might be mediated via membrane ARs or other nongenomic pathways because there is growing evidence from cell lines for a role for such systems, including effects on cell proliferation (17, 18). Indeed, it is apparent from a number of studies that such systems are present in SCs (19), and a recent study has identified one pathway via which nongenomic actions of androgens on SCs could play a role in spermatogenesis (20). Such developments might explain the inability of researchers to identify clear pathways via which genomic actions of androgens on SCs support spermatogenesis (4).
One approach to gaining insight into the above issues is to assess SC development and adult function in the absence of a functional (genomic) AR in AR0/Y mice (10, 21). However, this suffers from several potential problems. First, it is impossible in such animals to establish whether any observed effects result specifically from lack of androgen action on SCs, as opposed to effects on other testicular cells such as peritubular myoid cells. Second, because testes in such animals do not descend into the scrotum (which fails to form), there is the complication in adulthood of the confounding effects of elevated testicular temperature (10). The recent generation by ourselves (22) and others (23, 24) of mice with SC-selective knockout of the AR (SCARKO mice) has opened up new possibilities for elucidating the role that androgens play in regulating SC proliferation, maturation, and function. Moreover, by comparing these processes in SCARKO and complete AR knockout (AR0/Y; ARKO) mice, it is possible to gain key insights into the relative importance of direct (SC) and indirect (non-SC-mediated) effects of androgens on SCs. We used such an approach in the present studies and show that neither the effects of androgens on SC number nor the maturation-dependent change in expression of four SC markers require AR expression in SCs. In contrast, the ability of the SCs to support complete spermatogenesis, as well as expression of a number of SC genes, does require AR expression in SCs.
Materials and Methods
Generation of SCARKO and ARKO mice
The ARKOs were generated using Cre/loxP technology. ARflox/+ female animals (129/Swiss) with exon 2 of the AR floxed were either crossed with anti-Müllerian hormone (AMH)-Cre+/+ male mice (C57BL/6) expressing Cre recombinase (under the AMH gene promoter) selectively in SCs to generate the SCARKO line or to phosphoglycerate kinase-1 (PGK)-Cre+/+ male animals (C57BL/6) expressing Cre ubiquitously to produce the ARKO line. Full details are provided elsewhere (22). All animals were treated according to the National Institutes of Health guide for the care and use of laboratory animals and all experiments were approved by the local ethical committee.
Antibodies for immunohistochemistry and their dilutions
Immunolocalization of AR used a rabbit polyclonal antibody (N20) raised against a peptide within the N-terminal domain of the human AR (sc-816; Santa Cruz Biotechnology, Santa Cruz, CA), which was diluted at 1:200. Mouse monoclonal antibody raised against p27kip1 (NCL-p27; Novocastra, Newcastle Upon Tyne, UK) was used at 1:40. GATA-1 (sc-0266; Santa Cruz Biotechnology) rat monoclonal IgG2a antibody was diluted 1:500 for use and goat polyclonal antibody (sc-6886; Santa Cruz Biotechnology) specific for AMH was used at 1:1000. Rabbit anti-sulfated glycoprotein (SGP)-2 polyclonal antibody, used at 1:100, was a gift from Dr. S. R. Sylvester (Department of Biochemistry and Biophysics, Washington State University, Pullman, WA). Pem antibody, used at 1:500, was kindly donated by Professor M. F. Wilkinson (Department of Immunology, University of Texas M. D. Anderson Cancer Center, Houston, TX).
Immunohistochemistry
Urogenital systems from wild-type (WT), ARflox, AMH-Cre, ARKO, and SCARKO males at ages d 2, 8, 12, 20, 50, and/or 140 were removed, weighed, and fixed in Bouins fluid for 4–6 h before transferring into 70% (vol/vol) ethanol. Testes and epididymides were dissected and processed into paraffin wax using standard methods. Immunohistochemistry was performed on dewaxed 5-μm sections in conjunction with heat-induced antigen retrieval for 5 min in 0.01 M citrate buffer (pH 6.0) (Sigma-Aldrich, Dorset, UK) using a pressure cooker (for AR, GATA-1, p27kip1, and Pem) or no retrieval (for AMH and SGP-2). This was followed by endogenous peroxidase blocking in 3% (vol/vol) H2O2 in methanol for 30 min at room temperature. All washes between antibody or reagent incubations comprised 2 x 5 min at room temperature in Tris-buffered saline [TBS, 0.05 M Tris (pH 7.4), 0.85% (vol/vol) saline]. Tissue sections were first blocked in the appropriate serum elaborated as follows for the different primary antibodies. Rabbit polyclonal primary antibodies were diluted in TBS containing normal swine serum (1:4; Diagnostics Scotland, Carluke, Lanarkshire, UK) and 5% (wt/vol) BSA (Sigma-Aldrich), whereas goat polyclonal and mouse and rat monoclonal antibodies were diluted in TBS containing normal rabbit serum (1:4; Diagnostics Scotland) and 5% (wt/vol) BSA. A swine antirabbit biotinylated second antibody (E0353; Dako, Cambridge, UK) was used in conjunction with rabbit polyclonal antibodies, whereas a rabbit antimouse biotinylated second antibody (E0464; Dako) was used to detect p27kip1 antibody. Biotinylated rabbit antirat secondary antibody (BA-4000; Vector Laboratories, Peterborough, UK) and rabbit antigoat antibody (BA-5000; Vector Laboratories) were used in combination with GATA-1 and AMH antibodies, respectively. All biotinylated secondary antibodies were diluted 1:500 in the appropriate blocking serum in TBS/BSA and incubated at room temperature for 30 min.
Bound antibodies were visualized by incubating the sections with avidin biotin complex-horseradish peroxidase reagent (K0355; Dako) for 30 min followed by color development using 3,3'-diaminobenzidine tetra-hydrochloride chromogenic substrate (K3468, liquid DAB+ kit; Dako), monitored microscopically. Sections were counterstained with hematoxylin, dehydrated, and mounted with Pertex. Images were captured using a Provis microscope (Olympus Optical Co., London, UK) equipped with a DCS330 camera (Eastman Kodak Co., Rochester, NY). Captured images were stored on a Macintosh computer and compiled using Photoshop 7.0 (Adobe, Mountain View, CA).
To enable comparative evaluation of the immunostaining, sections of tissues from control and knockout animals were processed in parallel on at least three occasions to ensure reproducibility of results; on each occasion tissue sections from four to six animals in each group were run. To ensure direct comparability of staining intensities, one section each from control, ARKO, and SCARKO mice were mounted on the same slide.
Determination of SC nuclear volume/number per testis and germ cell nuclear volume per unit SC
Standard stereological methods were used to determine cell nuclear volumes and/or number per testis in groups of five WT, ARflox, AMH-Cre, ARKO, and SCARKO mice (d 2, 50, and 140), using methods detailed elsewhere (12, 25, 26). In brief, cross-sections of testes were examined under oil immersion using a Leitz 63 x plan apo objective fitted to a Leitz laborlux microscope and a 121-point eyepiece graticule. Using a systematic clock-face sampling pattern from a random starting point, 16 fields were counted. Points falling over the nuclei of SC or germ cell nuclei were scored and expressed as a percentage of the total points counted. For each animal, the values for percentage nuclear volume were converted to absolute nuclear volumes per testis by reference to testis volume (weight) because shrinkage was minimal. SC nuclear volumes per testis equate to SC numbers per testis, assuming no change in average nuclear size. However, because the individual nuclear volume of SCs might be influenced by reduced germ cell complement in adult ARKO and SCARKO males, average SC nuclear volume was also determined in four to five animals per group at d 140, using methods similar to those described previously (26, 27). Briefly, images were captured from an Olympus BH2 microscope using a video camera (HV-C20; Hitachi, Tokyo, Japan) and were analyzed with Image Pro Plus software with a Stereology 5.0 plug-in (Media Cybernetics, Berkshire, UK). An area of interest was created by drawing around the SC nucleus, within which the computer program then determined the average length of several diameters measured at two-degree intervals, which passed through the center of the nucleus. This was measured for a minimum of 100 SC nuclei per testis and mean nuclear volume then determined. Data for SC nuclear volume per testis was then converted to absolute numbers of SC per testis by dividing by the average SC nuclear volume.
RNA analysis
Tissue samples were removed and snap frozen in liquid nitrogen. RNA of WT, AMH-Cre, ARflox, and SCARKO testes was prepared using the RNeasy midi kit (QIAGEN, Chatsworth, CA). Due to the small size, RNA from ARKO testes and its appropriate control, PGK-Cre testes, was extracted using the RNeasy mini kit (QIAGEN). Synthesis of cDNA from DNase I-treated total RNA (RNase-free DNase I set; QIAGEN) used Superscript II RNaseH– reverse transcriptase and random hexamer primers (Invitrogen, Carlsbad, CA). To allow specific mRNA levels to be expressed per testis and control for the efficiency of RNA extraction, RNA degradation and the reverse transcription step, an external standard was used (10). The external standard was luciferase mRNA (Promega, Madison, WI), and 10 ng were added to each testis at the start of the RNA extraction procedure. For quantification of gene expression, the ABI Prism 7700 PCR detection system (Applied Biosystems, Foster City, CA) was used with a two-step RT-quantitative-PCR protocol. Components for real-time PCR were obtained from Applied Biosystems, apart from primers and probes (Eurogentec, Sar-Tilman, Belgium) and Sybrgreen (Sigma, St. Louis, MO). Each 25-μl real-time PCR contained 1x buffer A, 5 mM MgCl2, 400 μM deoxynucleotide triphosphates, 200 nM of each primer, 0.4x Sybrgreen, and 0.025 U/μl Amplitaq Gold enzyme. Amplified samples were electrophoresed on polyacrylamide gels to ensure that only a single band was amplified in each PCR. Primer sequences were for androgen binding protein (ABP): forward, GGAACAAATTCTTCTTGGCTTAACA, reverse, CTTAGGTTCTGCCTCGGAAGAC; transferrin: forward, GGACGCCATGACTTTGGATG, reverse, GCCATGACAGGCACTAGACCA; or as described elsewhere [AMH, claudin-11, luciferase, FSH receptor, and platelet-derived growth factor (PDGF)-A (10); cystatin-related gene highly expressed in testis and epididymis (cystatin-TE) and fatty acid binding protein (FABP) (28)]. For Pem, a probe was used to ensure specificity [primers and probe (22)], and 100 nM of each primer and 400 nM probe were added to the reaction. The quantity of measured mRNA was expressed relative to the luciferase standard in the same sample. All samples and standard curves were run in triplicate.
Statistical analysis
With the exception of the real-time PCR data whereby a two-sample t test was employed, statistical analysis was performed by using one-way ANOVA, supplemented with a Fisher multiple comparison test by using NCSS 2000 software (NCSS Statistical Analysis and Data Analysis Software, Kaysville, UT).
Results
To establish that any changes in phenotype observed in SCARKO males did not stem from aberrant changes arising in the founder transgenic lines (ARflox, AMH-Cre, or PGK-Cre lines), we demonstrated that testis weight, SC nuclear volume per testis, AR immunoexpression, and functional maturation of SCs (using the markers described below) were comparable in the founder lines and WT mice at d 2 (to reflect fetal development) and at d 50 and 140 to encompass early- and midadulthood, respectively (data not shown). Therefore, for the main studies reported below, either WT or AMH-Cre mice were used as controls, unless otherwise stated.
SC nuclear volume/number per testis in ARKO and SCARKO males
At d 2, SC nuclear volume per testis was already reduced significantly by 32% in ARKO males, compared with controls. This decrement increased to 57% on d 50 and to more than 70% on d 140 (Fig. 1). In contrast, in SCARKO males SC nuclear volume per testis was comparable with controls at d 2 and 50, whereas a small (16%) decrement was evident at d 140 (Fig. 1). Determination of SC number at d 140 in the same animals revealed changes in ARKO and SCARKO animals almost identical with that shown for SC nuclear volume in Fig. 1 (control AMH-Cre 3.14 ± 0.12 x 106 SC per testis; ARKO 0.56 ± 0.10 x 106; SCARKO 2.66 ± 0.18 x 106; means ± SEM, n = 4–5 per group). The SC number in ARKOs was significantly different (P < 0.001) from the controls, but the difference in SCARKOs only approached statistical significance (0.05 < P < 0.1).
FIG. 1. SC nuclear volume per testis at d 2, 50, and 140 in control, ARKO, and SCARKO animals. Numerical figures at d 2 reflect primarily SC proliferation in fetal life, whereas values at d 50 and 140 reflect final SC nuclear volume (number) per testis (note difference in scale at d 2). Values are means ± SEM for four to nine (d 50 and 140) or five to 11 (d 2) animals. *, P < 0.05, ***, P < 0.001, in comparison with respective control group.
Maturation-dependent changes in immunoexpression of SC markers in ARKO and SCARKO males
SC maturational development was assessed at d 2, 12, 20, and 50 of age in four to seven animals per group using four specific protein markers. We selected a protein that should be switched off during maturation (AMH), two that should be switched on acutely at around the time of SC maturation (p27kip1, GATA-1), and one that is switched on progressively during SC maturation (SGP-2) (6). Additionally, because AR expression is normally switched on in SCs just before their maturation (5, 9), it was checked whether expression occurred at the appropriate time in all control animals (WT and founder lines) and did not occur at any age in SCARKO and ARKO males.
In controls, AMH expression in SC cytoplasm was intense at d 2, had declined but was still evident at d 12, and was nondetectable at d 20 and 50. SCARKO and ARKO males showed an identical age-related change in AMH immunoexpression (Fig. 2). The absence of any difference in AMH expression between male controls and SCARKOs was confirmed by quantitative PCR (d 50) (see below) and Western blotting (d 2–50) (data not shown).
FIG. 2. Immunoexpression profiles for AMH and p27kip1 in SCs from control, ARKO, and SCARKO testes at d 2, 12, 20, and 50. Note that AMH is expressed in SC cytoplasm predominantly at d 2 with a major decrease by d 12 and thereafter, whereas the opposite pattern is seen with nuclear expression of p27kip1. Bar, 100 μm.
The cyclin-dependent kinase inhibitor, p27kip1, inhibits D-cyclin mediated G1 to S progression in the cell cycle and its nuclear expression in the SCs is first evident at around the time of terminal differentiation (maturation) and cessation of proliferation (6, 29). Consistent with this, SC in control mice were immunonegative for p27kip1 at d 2, expression was weakly apparent in some SC nuclei at d 12, and was intense in all SC nuclei at d 20 and 50. Comparable age-dependent patterns of p27kip1 immunoexpression were found in both SCARKO and ARKO males (Fig. 2).
SGP-2 is one of the major SC-secreted proteins (30), although its function remains unclear. In controls, SGP-2 immunoexpression was weak in SC cytoplasm at d 2 and increased progressively in intensity between d 12 and 50, during which period the intensity of its immunoexpression became increasingly stage dependent; a similar age-dependent increase in SGP-2 immunoexpression was evident in SCARKO and ARKO animals (see supplemental data, published on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org).
GATA-1 is a zinc finger transcription factor that is first expressed in SCs during their maturation and development in the first wave of spermatogenesis (31). GATA-1 immunoexpression in controls was nondetectable at d 2 but was prominent in SC nuclei at d 12 and thereafter, and a similar age-dependent increase in GATA-1 immunoexpression was evident in SCARKO and ARKO animals (supplemental data).
AR immunoexpression showed an identical age-dependent pattern of expression in SCs in all controls (AMH-Cre, PGK-Cre, ARflox, and WT). Nuclear immunostaining was not detectable in SCs at d 2 but was observed from d 8 onward and was notably stage dependent in adulthood (Fig. 3). In all controls, there was uniform nuclear AR immunoexpression in Leydig and peritubular myoid cells at all ages (Fig. 3). In ARKO and SCARKO males, no AR immunoexpression in SCs was evident at any age, but in SCARKO animals normal AR immunoexpression was observed in Leydig and peritubular myoid cells, comparable with that in controls (Fig. 3).
FIG. 3. Immunoexpression profile for the AR in control, ARKO, and SCARKO testes at d 2–50 of age to illustrate ontogeny of AR expression in SCs. Note the AR immunoexpression in d 12–50 testes of control animals (B–D) with stage-dependent SC (arrowheads) nuclear staining at d 50 (D) and that SCARKO testes exhibit no SC AR immunoexpression at any age (J–M), whereas nuclei of Leydig cells (L) and PTMCs (arrows) are immunopositive for AR expression at all ages in control and SCARKO mice (A–D and J–M, respectively). Note the complete absence of AR immunoexpression in ARKO testes (E–H). Bar, 100 μm.
Having established that SCs in ARKO and SCARKO males had exhibited a normal age-dependent pattern of change in the expression of four recognized SC nuclear and cytoplasmic markers, we next investigated their functional competence in two ways, first by quantifying their capacity to support germ cell development at various ages and second by quantifying mRNA expression for nine genes shown previously to be expressed by adult SCs and/or to be androgen regulated.
Functional capacity of SCs to support germ cells in ARKO and SCARKO males and testis weight
SCs in SCARKO males supported a normal volume of spermatogonia at d 2, 50, and 140, although there was a downward trend at d 140. In contrast, total germ cell volume supported per unit SC volume in SCARKOs was decreased by 69 and 76% at d 50 and d 140, respectively (Fig. 4). At d 2, ARKO SCs supported significantly more (126%) spermatogonia than in controls, but at d 50 and 140, they supported a significantly lower volume of spermatogonia, with approximately a 60% reduction relative to controls or SCARKOs (Fig. 4). This impaired function of ARKO SCs was manifested more dramatically with total germ cell volume per unit SCs, which was reduced by approximately 95% at d 50 and 140 relative to controls (Fig. 4). The age-dependent changes in germ cell volume per SC in ARKO and SCARKO males was matched by changes in overall testis weight (Table 1).
FIG. 4. Functional ability of SCs to support germ cells at d 2, 50, and 140 in control, ARKO, and SCARKO animals. Data are expressed as germ cell nuclear volume per unit SC nuclear volume to take account of differences in SC nuclear volume shown in Fig. 1 and are shown separately for spermatogonia (top row) and total germ cells (spermatogonia + spermatocytes + spermatids) (bottom row). *, P < 0.05, ***, P < 0.001, in comparison with respective control group.
TABLE 1. Testis weights (mean ± SEM) at age d 2, 50, and 140 in WT, ARKO, and SCARKO males and in their respective controls (PGK-Cre for ARKOs; AMH-Cre for SCARKOs)
Expression of SC genes
Nine genes were selected for study based on one or more of the following criteria: expression restricted to SCs in the testis, evidence of androgen dependence, role in spermatogenesis, expression in SCs modifiable by peritubular myoid cell factors, and previous study in ARKO mice (10, 21, 23, 32, 33, 34, 35, 36, 37, 38). Results are presented in Fig. 5. The amount of mRNA in d 50 SCARKO and ARKO testes were each compared with their respective controls (AMH-Cre and PGK-Cre littermates, respectively). The expression levels of these nine genes normalized to absolute levels per testis (using the known spiked concentrations of luciferase added per testis before RNA extraction) are thus not affected by testicular germ cell composition and are shown relative to their controls (arbitrarily set equal to 1). Because there were also significant changes in SC nuclear volume/number per testis, especially in ARKOs, mRNA expression has also been shown in Fig. 5 after correction for SC nuclear volume, using the mean values per group for d 50 animals illustrated in Fig. 1.
FIG. 5. Changes in expression of SC genes in 50-d-old ARKO and SCARKO mice, compared with the respective controls (PGK-Cre for ARKO and AMH-Cre for SCARKO). Testicular RNA was extracted and cDNA was prepared as described in Materials and Methods. Real-time PCR was used to measure cDNA levels relative to an external standard (luciferase) added during RNA extraction. Upper panel, To compare expression levels per testis, results were calculated as the number of cDNA copies per 106 copies of luciferase and were expressed as a fraction of the respective control, arbitrarily set equal to 1. Lower panel, To compare expression levels per SC nuclear volume, the number of cDNA copies per 106 copies of luciferase was divided by the appropriate SC nuclear volume for each genotype (derived from the data in Fig. 1). Results are expressed as a fraction of the respective control, arbitrarily set equal to 1. Values shown are means ± SEM for six to eight animals per group. Data indicated by an asterisk are significantly different (P < 0.05) from the respective control as assessed by a two-sample t test.
In both AR knockout models, the most dramatic change in expression was observed for Pem. Pem transcript levels in the testis were reduced to less than 0.4% of control values in ARKOs as well as SCARKOs (Fig. 5), a change matched by Pem protein immunohistochemistry (Fig. 6).
FIG. 6. Immunoexpression of Pem protein in SC in d 50 control, ARKO, and SCARKO testes. Bar, 100 μm.
In the AR Knockout (ARKO) testis, significant reductions in gene expression for ABP, cystatin-TE, claudin-11, and the FSH receptor were found (Fig. 5), although this reduction was no longer evident after correction for SC nuclear volume per testis, which even resulted in a small but significant increase in FSH receptor mRNA expression (Fig. 5). A markedly more severe reduction in gene expression was observed in ARKO testes for FABP, transferrin, and PDGF-A, and correction for SC nuclear volume per testis only slightly attenuated these decreases (Fig. 5). In contrast, no reduction in AMH gene expression was observed in ARKOs, and once the reduction in SC number was taken into account, AMH expression per SC was more than doubled (Fig. 5). These findings are in accordance with the observations of Johnston et al. (10).
In SCARKO testes no overexpression of AMH mRNA was observed (Fig. 5), in contrast with a previous report (23). Similarly, ABP and the FSH receptor mRNA were expressed at levels comparable with the control, whereas transferrin expression was slightly reduced (29%), and reductions of 36–60% were noted for FABP, PDGF-A, cystatin-TE, and claudin-11 (Fig. 5); the latter changes were still evident after correction for SC nuclear volume per testis (Fig. 5).
Discussion
The primary aim of the present study was to evaluate the role that androgen action in the testis plays in determining the numbers and functional development of SCs. For this purpose we compared ARKO mice, which lack AR in all testicular cells, with SCARKO mice in which AR expression was deleted selectively in SCs.
A role for androgens in regulating SC proliferation/final number in rodents had not been seriously considered in the past, perhaps because SCs do not normally express the AR for much of the period in which proliferation takes place (5, 9). This perspective has changed recently due to a study (10) that has shown that testicular feminized mice (tfm) exhibit an approximately 50% deficit in SC numbers at birth and a 75% deficit in adulthood and another study (12) showing that, in the neonatal rat, endogenous androgens play a physiological role in increasing SC number. Our present findings in ARKO males, which are analogous phenotypically to tfm mice, confirm the observations of Johnston et al. (10) by showing reductions in SC nuclear volume/number per testis of 31, 58, and 71–82% at d 2, 50, and 140, respectively, in ARKO, compared with control males. However, what the present findings in SCARKOs demonstrate is that AR expression in SCs is not required for attainment of normal SC number, indicating that AR expression in another testicular cell type must be important for increasing SC proliferation perinatally in mice. Although the present findings do not allow unambiguous identification of this cell type(s), it seems highly likely that it is the peritubular myoid cell (PTMC). These cells express AR intensely throughout fetal and postnatal life, and there is abundant evidence in the literature, based mainly on coculture of isolated PTMCs and SCs, to show that secretions from the PTMCs can modify various aspects of SC function (15, 16), including the production of transferrin (16, 39, 40). It is also well established that PTMCs and SCs cooperate to lay down the basal lamina of the seminiferous cords (41, 42) to which the SCs adhere, and this could be a factor that limits SC proliferation via processes such as contact inhibition (43). Moreover, there is evidence that activin A produced by PTMCs may stimulate SC proliferation (44). Whether the production of activin A or other growth factors by PTMCs is androgen controlled remains to be shown. If proliferation of SCs is regulated indirectly via androgen action on neighboring PTMCs, it would provide an intriguing parallel with the rest of the developing male reproductive tract in which proliferation and differentiation of epithelial cells is regulated indirectly via androgen action on neighboring mesenchymal/stromal cells (45).
SC maturation, or terminal differentiation, is the process via which SCs lose their proliferative ability and acquire functions that will enable them to support spermatogenesis (6). Among other changes this involves marked down-regulation of AMH expression (6) and the switching on of p27kip1 (29) and GATA-1 (31) as well as the progressive up-regulation of SGP-2 (6, 46). We selected these markers on the basis of their SC specificity in the testis and because of their differing patterns and locations of expression (two nuclear vs. two cytoplasmic). To our surprise, the maturation-dependent changes in expression of all four proteins occurred normally in both ARKO and SCARKO males, demonstrating that at least some key aspects of SC functional maturation do not require either direct or indirect androgen action via the AR. In this regard, our finding that AMH expression is down-regulated normally in SCARKO testes at the same age as in controls, contrasts with the report of another group that also generated SC-selective AR knockouts (23). These authors did not study AMH expression throughout development as we have done, nor did they examine expression of other SC markers modulated during SC maturation. Our finding of normal maturation-dependent reduction in AMH expression in ARKOs (despite increased mRNA levels when expressed per SC in adulthood) is consistent with earlier data reported for tfm mice (10, 47).
The present findings confirm and extend previous findings by ourselves (22) and others (10, 23, 47) in AR-deficient mice, in showing that SCs in adulthood are incapable of supporting meiotic and postmeiotic germ cell development in the absence of androgen action via the AR. In this regard, ARKO males were considerably more deficient than were SCARKOs in adulthood because not only did their SC support fewer germ cells per SC, but also even SC support of spermatogonia was severely compromised. In contrast, SCs in SCARKOs supported normal numbers of spermatogonia and showed deficits only in postmeiotic germ cell support. The extent to which this difference reflects the abdominal vs. scrotal position of testes in ARKO and SCARKO adults, respectively, is unclear, but it is notable that before the time of normal testis descent (d 2), SCs in ARKO clearly supported normal/supranormal numbers of spermatogonia.
Data on gene expression in adult ARKO and SCARKO mice cover seven genes representative for diverse SC functions that have been analyzed previously in normal and tfm mice (10). Two genes encoding lipid-binding proteins (ABP, FABP), a tissue remodeling factor (cystatin-TE), a component of the tight junction complex (claudin-11), and three proteins related to endocrine or paracrine functions (PDGF-A, AMH, and FSH receptor). In addition, we included Pem as a control for direct androgen action on SCs and transferrin as a SC product, the secretion of which is known to be modulated by PTMCs. Gene expression data should be interpreted with caution because aberrant expression may be related to not only the absence of AR (ubiquitous or SC specific) but also indirect effects caused by absence of postmeiotic germ cells, changes in FSH secretion, and (in ARKOs) abnormal location of the testis. Data on AMH gene expression have been discussed above.
In confirmation of our previous studies, Pem expression was virtually absent in both ARKOs and SCARKOs. Immunostaining studies confirmed these differences at the protein level and indicated SC localization compatible with the proposed direct regulation by androgens (36). The expression of PDGF-A and FABP was also clearly reduced in SCARKOs and even more so in ARKOs (even after correction for the reduction in SC nuclear volume per testis). Whereas the data of Johnston et al. (10) in tfm mice suggest that the decrease in these parameters may be due to cryptorchidism, our data in SCARKOs, which unlike ARKOs have normally descended testes, indicate that this may not be the correct interpretation. One explanation could be that impaired expression of these genes is due to the decreased germ cell complement. In the adult testis, PDGF-A is expressed primarily in SCs, although weak expression in occasional interstitial cells may also occur (48); expression in the SCs occurs in a stage-dependent manner (48), a phenomenon frequently associated with germ cell modulation of SC function (49). Because PDGF-A has been shown to play an essential role in adult Leydig cell development (48), the marked reduction in expression of PDGF-A in testes of both ARKOs and SCARKOs could suggest that Leydig cell development is altered in these animals, and this is currently under investigation.
The limited reduction in the expression of the transferrin gene in SCARKOs and the severe reduction in ARKOs are of interest because earlier studies have indicated that PTMCs and their secretion products positively modulate transferrin production by SCs with some evidence of this being androgen regulated (16, 39, 40). The present findings are compatible with this hypothesis, but the slight reduction in transferrin mRNA in SCARKOs suggests that there may be additional control mechanisms involving either direct effects of androgens on SCs or indirect effects caused by the reduced germ cell complement (50, 51).
Surprisingly, cystatin-TE and claudin-11 expression levels were approximately normal in ARKOs after correction for SC number but were significantly reduced in SCARKOs. The normal expression levels of cystatin-TE and claudin-11 in ARKOs confirms a previous study in tfm animals (10), but the partial suppression of mRNA for these genes in SCARKOs is intriguing. It seems unlikely that this decrease could be explained by absence of particular germ cell types because ARKO animals lacked all of the same germ cell types as did SCARKOs but did not exhibit significant down-regulation of cystatin-TE and claudin-11 after correction for SC nuclear volume per testis. An alternative explanation might be that the higher FSH levels observed in ARKOs, compared with SCARKOs (22), could have led to increased expression of the mRNA for cystatin-TE and claudin-11 in the former. Claudin-11 has recently been demonstrated to be regulated at least in part by FSH (38).
Finally, when corrected for SC nuclear volume per testis, mRNA levels for ABP were normal in SCARKOs and ARKOs. A similar absence of androgen regulation of ABP expression has previously been observed in irradiated rats treated with GnRH analogs and antiandrogens (52). In SCARKOs and ARKOs, expression of the FSH receptor gene was slightly elevated after correction for SC nuclear volume per testis, and in the case of ARKOs, this was statistically significant. Whether this change is also related to altered FSH levels in these two lines (22) or is due to lack of androgen action on the SCs is a matter for speculation. Taken together, the described gene expression data show that, despite a normal pattern of change in expression of four proteins associated with SC maturation, significant changes occur in the expression of several SC genes in ARKOs and SCARKOs. Comparison of the gene expression pattern in these two lines, after correction for change in SC nuclear volume/number, provides a starting point for dissecting which genes might be regulated directly by androgen action on SCs. In this regard, the change in expression of Pem was dramatic (but expected), but the reduced expression of FABP and PDGF-A was also intriguing. Although it remains possible that expression of either or both of these genes by SCs could be secondary to altered signaling to the SCs from the reduced germ cell complement in ARKOs and SCARKOs, the alternative explanation that they might to some extent be directly androgen regulated merits further study.
In conclusion, the present studies show that complete ablation of androgen action in ARKO males results in major deficits in SC numbers, although at least some aspects of the SC maturation process appear to occur normally in these animals. In the selective absence of androgen action on SCs (SCARKO), these cells develop normally with respect to both their numbers and the expression of four selected maturational markers, but they are clearly unable to support late meiotic and postmeiotic germ cells; this is associated with changes in expression of a number of SC-expressed genes, in particular Pem, FABP, and PDGF-A. Given their normal testicular descent and the normality of expression of at least some indicators of SC functional maturation, SCARKO males will hopefully provide the means via which the SC functions necessary for survival of late meiotic and postmeiotic germ cells can be identified.
Acknowledgments
We thank Arantza Esnal, Hilde Geeraerts, and Ludo Deboel for their excellent technical assistance.
References
Hughes IA 2001 Minireview: sex differentiation. Endocrinology 142:3281–3287
Sultan C, Paris F, Terouanne B, Balaguer P, Georget V, Poujol N, Jeandel C, Lumbroso S, Nicolas JC 2001 Disorders linked to insufficient androgen action in male children. Hum Reprod Update 7:314–322
Sharpe RM 1994 Regulation of spermatogenesis. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven Press; 1363–1433
Hill CM, Anway MD, Zirkin BR, Brown TR 2004 Intratesticular androgen levels, androgen receptor localization, and androgen receptor expression in adult rat Sertoli cells. Biol Reprod 71:1348–1358
Sharpe RM 2005 Sertoli cell endocrinology and signal transduction: androgen regulation. In: Griswold M, Skinner M, eds. Sertoli cell biology. San Diego: Academic Press; 199–216
Sharpe RM, McKinnell C, Kivlin C, Fisher JS 2003 Proliferation and functional maturation of Sertoli cells, and their relevance to disorders of testis function in adulthood. Reproduction 125:769–784
Johnson L, Zane RS, Petty CS, Neaves WB 1984 Quantification of the human Sertoli cell population: its distribution, relation to germ cell numbers, and age-related decline. Biol Reprod 31:785–795
Orth JM, Gunsalus GL, Lamperti AA 1988 Evidence from Sertoli cell-depleted rats indicates that spermatid number in adults depends on numbers of Sertoli cells produced during perinatal development. Endocrinology 122:787–794
Bremner WJ, Millar MR, Sharpe RM, Saunders PT 1994 Immunohistochemical localization of androgen receptors in the rat testis: evidence for stage-dependent expression and regulation by androgens. Endocrinology 135:1227–1234
Johnston H, Baker PJ, Abel M, Charlton HM, Jackson G, Fleming L, Kumar TR, O’Shaughnessy PJ 2004 Regulation of Sertoli cell number and activity by follicle-stimulating hormone and androgen during postnatal development in the mouse. Endocrinology 145:318–329
Zhou Q, Shima JE, Nie R, Friel PJ, Griswold MD, 2005 Androgen-regulated transcripts in the neonatal mouse testis as determined through microarray analysis. Biol Reprod 72:1010–1019
Atanassova NN, Walker M, McKinnell C, Fisher JS, Sharpe RM 2005 Evidence that androgens and oestrogens, as well as follicle-stimulating hormone, can alter Sertoli cell number in the neonatal rat. J Endocrinol 184:107–117
Schlatt S, Arslan M, Weinbauer GF, Behre HM, Nieschlag E 1995 Endocrine control of testicular somatic and premeiotic germ cell development in the immature testis of the primate Macaca mulatta. Eur J Endocrinol 133:235–247
Ramaswamy S, Plant TM, Marshall GR 2000 Pulsatile stimulation with recombinant single chain human luteinizing hormone elicits precocious Sertoli cell proliferation in the juvenile male rhesus monkey (Macaca mulatta). Biol Reprod 63:82–88
Skinner MK, Fritz IB 1985 Testicular peritubular cells secrete a protein under androgen control that modulates Sertoli cell functions. Proc Natl Acad Sci USA 82:114–118
Swinnen K, Cailleau J, Heyns W, Verhoeven G 1990 Prostatic stromal cells and testicular peritubular cells produce similar paracrine mediators of androgen action. Endocrinology 126:142–150
Migliaccio A, Castoria G, Di Domenico M, de Falco A, Bilancio A, Lombardi M, Barone MV, Ametrano D, Zannini MS, Abbondanza C, Auricchio F 2000 Steroid-induced androgen receptor-oestradiol receptor ?-Src complex triggers prostate cancer cell proliferation. EMBO J 19:5406–5417
Losel RM, Falkenstein E, Feuring M, Schultz A, Tillmann HC, Rossol-Haseroth K, Wehling M 2003 Nongenomic steroid action: controversies, questions, and answers. Physiol Rev 83:965–1016
Silva FR, Leite LD, Wassermann GF 2002 Rapid signal transduction in Sertoli cells. Eur J Endocrinol 147:425–433
Fix C, Jordan C, Cano P, Walker WH 2004 Testosterone activates mitogen-activated protein kinase and the cAMP response element binding protein transcription factor in Sertoli cells. Proc Natl Acad Sci USA 101:10919–10924
Yeh S, Tsai MY, Xu Q, Mu XM, Lardy H, Huang KE, Lin H, Yeh SD, Altuwaijri S, Zhou X, Xing L, Boyce BF, Hung MC, Zhang S, Gan L, Chang C, Hung MC 2002 Generation and characterization of androgen receptor knockout (ARKO) mice: an in vivo model for the study of androgen functions in selective tissues. Proc Natl Acad Sci USA 99:13498–13503
De Gendt K, Swinnen JV, Saunders PTK, Schoonjans L, Dewerchin M, Devos A, Tan K, Atanassova N, Claessens F, Lécureuil C, Heyns W, Carmeliet P, Guillou F, Sharpe RM, Verhoeven G 2004 A Sertoli cell-selective knockout of the androgen receptor causes spermatogenic arrest in meiosis. Proc Natl Acad Sci USA 101:1327–1332
Chang C, Chen YT, Yeh SD, Xu Q, Wang RS, Guillou F, Lardy H, Yeh S 2004 Infertility with defective spermatogenesis and hypotestosteronemia in male mice lacking the androgen receptor in Sertoli cells. Proc Natl Acad Sci USA 101:6876–6881
Holdcraft RW, Braun RE 2004 Androgen receptor function is required in Sertoli cells for the terminal differentiation of haploid spermatids. Development 131:459–467
Atanassova N, McKinnell C, Walker M, Turner KJ, Fisher JS, Morley M, Millar MR, Groome NP, Sharpe RM 1999 Permanent effects of neonatal estrogen exposure in rats on reproductive hormone levels, Sertoli cell number, and the efficiency of spermatogenesis in adulthood. Endocrinology 140:5364–5373
Sharpe RM, Walker M, Millar MR, Atanassova NN, Morris K, McKinnell C, Saunders PTK, Fraser HM 2000 Effect of neonatal gonadotropin-releasing hormone antagonist administration on Sertoli cell number and testicular development in the marmoset: comparison with the rat. Biol Reprod 62:1685–1693
Mahood IK, McKinnell C, Fisher JS, Walker M, Hallmark N, Sharpe RM 2005 Abnormal Leydig cell aggregation in the fetal testis of rats exposed to Di(n-butyl) phthalate and its possible role in testicular dysgenesis. Endocrinology 146:613–623
O’Shaughnessy PJ, Fleming L, Baker PJ, Jackson G, Johnston H 2003 Identification of developmentally regulated genes in the somatic cells of the mouse testis using serial analysis of gene expression. Biol Reprod 69:797–808
Beumer TL, Kiyokawa H, Roepers-Gajadien HL, van den Bos LA, Lock TM, Gademan IS, Rutgers DH, Koff A, de Rooij DG 1999 Regulatory role of p27kip1 in the mouse and human testis. Endocrinology 140:1834–1840
Sharpe RM, Millar M, McKinnell C 1993 Relative roles of testosterone and the germ cell complement in determining stage-dependent changes in protein secretion by isolated rat seminiferous tubules. Int J Androl 16:71–81
Yomogida K, Ohtani H, Harigae H, Ito E, Nishimune Y, Engel JD, Yamamoto M 1994 Developmental stage- and spermatogenic cycle-specific expression of transcription factor GATA-1 in mouse Sertoli cells. Development 120:1759–1766
Lindsey JS, Wilkinson MF 1996 Pem: a testosterone- and LH-regulated homeobox gene expressed in mouse Sertoli cells and epididymis. Dev Biol 179:471–484
Danzo BJ, Pavlou SN, Anthony HL 1990 Hormonal regulation of androgen-binding protein in the rat. Endocrinology 127:2829–2838
Skinner MK, Griswold MD 1982 Secretion of testicular transferrin by cultured Sertoli cells is regulated by hormones and retinoids. Biol Reprod 27:211–221
Loveland KL, Zlatic K, Stein-Oakley A, Risbridger G, deKretser DM 1995 Platelet-derived growth factor ligand and receptor subunit mRNA in the Sertoli and Leydig cells of the rat testis. Mol Cell Endocrinol 108:155–159
Li Y, Friel PJ, Robinson MO, McLean DJ, Griswold MD 2002 Identification and characterization of testis- and epididymis-specific genes: cystatin SC and cystatin TE-1. Biol Reprod 67:1872–1880
Byers S, Graham R, Dai HN, Hoxter B 1991 Development of Sertoli cell junctional specializations and the distribution of the tight-junction-associated protein ZO-1 in the mouse testis. Am J Anat 191:35–47
Hellani A, Ji J, Mauduit C, Deschildre C, Tabone E, Benahmed M 2000 Developmental and hormonal regulation of the expression of oligodendrocyte-specific protein/claudin-11 in mouse testis. Endocrinology 141:3012–3019
Janecki A, Steinberger A 1987 Vectorial secretion of transferrin and androgen binding protein in Sertoli cell cultures: effect of extracellular matrix, peritubular myoid cells and medium composition. Mol Cell Endocrinol 52:125–135
Anthony CT, Rosselli M, Skinner MK 1991 Actions of the testicular paracrine factor (P-Mod-S) on Sertoli cell transferrin secretion throughout pubertal development. Endocrinology 129:353–360
Tung PS, Fritz IB 1987 Morphogenetic restructuring and formation of basement membranes by Sertoli cells and testis peritubular cells in co-culture: inhibition of the morphogenetic cascade by cyclic AMP derivatives and by blocking direct cell contact. Dev Biol 120:139–153
Skinner MK, Tung PS, Fritz IB 1985 Cooperativity between Sertoli cells and testicular peritubular cells in the production and deposition of extracellular matrix components. J Cell Biol 100:1941–1947
Schlatt S, de Kretser DM, Loveland KL 1996 Discriminative analysis of rat Sertoli and peritubular cells and their proliferation in vitro: evidence for follicle-stimulating hormone-mediated contact inhibition of Sertoli cell mitosis. Biol Reprod 55:227–235
Buzzard JJ, Farnworth PG, De Kretser DM, O’Connor AE, Wreford NG, Morrison JR 2003 Proliferative phase Sertoli cells display a developmentally regulated response to activin in vitro. Endocrinology 144:474–483
Cunha GR, Foster B, Thomson A, Sugimura Y, Tanji N, Tsugi M, Terada N, Finch PW, Donjacour AA 1995 Growth factors as mediators of androgen action during development of the male urogenital tract. World J Urol 13:264–276
Zakeri Z, Curto M, Hoover D, Wightman K, Engelhardt J, Smith FF, Kierszenbaum AL, Gleeson T, Tenniswood M 1992 Developmental expression of the S35–S45/SGP-2/TRPM-2 gene in rat testis and epididymis. Mol Reprod Dev 33:373–384
Al-Attar L, No?l K, Dutertre M, Belville C, Forest MG, Burgoyne PS, Josso N, Rey R 1997 Hormonal and cellular regulation of Sertoli cell anti-Müllerian hormone production in the postnatal mouse. J Clin Invest 100:1335–1343
Gnessi L, Basciani S, Mariani S, Arizzi M, Spera G, Wang C, Bondjers C, Karlsson L, Betsholtz C 2000 Leydig cell loss and spermatogenic arrest in platelet-derived growth factor (PDGF)-A-deficient mice. J Cell Biol 149:1019–1026
Jegou B 1993 The Sertoli-germ cell communication network in mammals. Int Rev Cytol 147:25–96
Stallard BJ, Griswold MD 1990 Germ cell regulation of Sertoli cell transferrin mRNA levels. Mol Endocrinol 4:393–401
Maguire SM, Millar MR, Sharpe RM, Gaughan J, Saunders PT 1997 Investigation of the potential role of the germ cell complement in control of the expression of transferrin mRNA in the prepubertal and adult rat testis. J Mol Endocrinol 19:67–77
Verhoeven G, Deboel L, Swinnen J, Rombauts L, Vanderhoydonc F, Rosseels J, Hoeben E, Heyns W 1995 Effects of androgens on the germ cell-depleted testes of prenatally irradiated rats. Int J Androl 16:22–34(Karen A. L. Tan1, Karel D)