Development and Function of the Adult Generation of Leydig Cells in Mice with Sertoli Cell-Selective or Total Ablation of the Androgen Recep
http://www.100md.com
《内分泌学杂志》
Laboratory for Experimental Medicine and Endocrinology, Department of Developmental Biology, Catholic University of Leuven (K.D.G., E.D., G.V.), B-3000 Leuven, Belgium
Medical Research Council Human Reproductive Sciences Unit, Center for Reproductive Biology (N.A., K.A.L.T., C.M., R.M.S., P.T.K.S.), and Division of Reproductive and Developmental Science, University of Edinburgh (J.I.M.), Edinburgh, Scotland, EH16 4SB United Kingdom
Institute of Experimental Morphology and Anthropology, Bulgarian Academy of Science (N.A.), 1113 Sofia, Bulgaria
Laboratory of Cellular Biology, Department of Morphology, Instituto de Ciências Biológicas/Federal University of Minas Gerais (L.R.d.F., G.G.P.), Belo Horizonte-MG, Brazil
Institute for Hormone and Fertility Research, University of Hamburg (S.H.), D-20251 Hamburg, Germany
School of Molecular and Biomedical Sciences, University of Adelaide (R.I.), Adelaide 5005, Australia
Abstract
It is established that androgens and unidentified Sertoli cell (SC)-derived factors can influence the development of adult Leydig cells (LC) in rodents, but the mechanisms are unclear. We evaluated adult LC development and function in SC-selective androgen receptor (AR) knockout (SCARKO) and complete AR knockout (ARKO) mice. In controls, LC number increased 26-fold and LC size increased by approximately 2-fold between 12 and 140 d of age. LC number in SCARKOs was normal on d 12, but was reduced by more than 40% at later ages, although LC were larger and contained more lipid droplets and mitochondria than control LC by adulthood. ARKO LC number was reduced by up to 83% at all ages compared with controls, and LC size did not increase beyond d 12. Serum LH and testosterone levels and seminal vesicle weights were comparable in adult SCARKOs and controls, whereas LH levels were elevated 8-fold in ARKOs, although testosterone levels appeared normal. Immunohistochemistry and quantitative PCR for LC-specific markers indicated steroidogenic function per LC was probably increased in SCARKOs and reduced in ARKOs. In SCARKOs, insulin-like factor-3 and estrogen sulfotransferase (EST) mRNA expression were unchanged and increased 3-fold, respectively, compared with controls, whereas the expression of both was reduced more than 90% in ARKOs. Changes in EST expression, coupled with reduced platelet-derived growth factor-A expression, are potential causes of altered LC number and function in SCARKOs. These results show that loss of androgen action on SC has major consequences for LC development, and this could be mediated indirectly via platelet-derived growth factor-A and/or estrogens/EST.
Introduction
IN ADULTHOOD, TESTOSTERONE action on seminiferous tubules is essential for full, quantitatively normal spermatogenesis and fertility, an effect mediated largely by action on Sertoli cells (SC) via the androgen receptor (AR), although the downstream mechanisms remain unclear (1, 2). Testosterone is produced by the adult population of Leydig cells (LC), which develop during puberty from presumptive progenitor cells that are present before this time (3, 4, 5, 6). Development of the LC involves at least three steps, namely the proliferation of LC precursors, their differentiation into immature LC and their final differentiation into adult LC (3, 5, 7). The mechanisms that regulate the numbers of both adult and precursor LC and the differentiation of the latter into the former are incompletely understood, but hormones such as LH (8), FSH (9, 10), thyroid hormone (11, 12, 13), androgens (3, 14), and estrogens (15, 16) probably play a role. Additionally, numerous studies involving a variety of experimental approaches have demonstrated that SC play a role in regulating the development and function of the adult LC (9, 17). For example, FSH action on SC has been shown to advance the development of LC, and other studies have shown that the status of spermatogenesis (i.e. the germ cell complement) can alter the development of neighboring LC, presumably via the action of SC-derived (unknown) factors (9). Indeed, some (11, 18), but not all (19), studies suggest that the number of SC per testis may determine the number of adult LC.
Differentiation of LC precursors into immature LC is a key step in development of the adult LC population (3, 5), and androgens appear to play a positive role in this process (14). Perhaps the most dramatic demonstration of the potential importance of androgens in LC development is the observation that in mice lacking a functional AR, there is a major reduction in adult LC number and function (20). Whether this deficit arises because of the absence of effect of androgens directly on LC or their precursors or is mediated partly or wholly via the absence of effects on the SC (or even on the peritubular myoid cells) is unclear. Another complication in this model is that the testes are cryptorchid, and the consequent elevation of testicular temperature and/or the associated depletion of germ cells may also play a role. The recent generation by ourselves (21) and others (22, 23) 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 the development and function of testicular cells. Moreover, by comparing these processes in SCARKO and complete AR knockout [AR0/Y; ARKO] mice, it is possible to gain insights into the relative importance of SC-mediated and non-SC-mediated effects of androgens on developing LC. We used such an approach in the present studies and show that SC-mediated effects of androgens may be important for the development of normal numbers of adult LC.
Materials and Methods
Generation of SCARKO and ARKO mice
The AR knockout animals were generated using cyclization recombination (Cre) recombinase /loxP technology. ARflox/+ female animals (129/Swiss) with exon 2 of the AR floxed were either crossed with anti-Mullerian hormone (AMH)-Cre+/+ male mice (C57BL/6) expressing Cre recombinase (under the AMH gene promoter) selectively in SC 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 were provided previously (21). 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 committees.
Tissue and blood sampling
Urogenital systems from control, ARKO and SCARKO males (aged 12, 20, 50, or 140 d) were removed, fixed in Bouin’s fluid for 4–6 h, then transferred to 70% ethanol. Testes and seminal vesicles (if present) were dissected out and weighed. Blood was taken by cardiac puncture under ether anesthesia and allowed to clot overnight at 4 C. Serum was isolated by two centrifugal steps of 10 min at 13,000 rpm and stored at –20 C until assayed.
Antibodies for immunohistochemistry and their dilutions
Rabbit polyclonal antibody (AB1244, Chemicon International, Temecula, CA) recognizing cytochrome P450 side-chain cleavage (P450scc) was used at a 1:200 dilution. Anti-17-hydroxylase/C17–20lyase (anti-P450c17) and 3-hydroxysteroid dehydrogenase (3-HSD) rabbit polyclonal antibodies, used at 1:300 and 1:3000 dilutions, respectively, were generated in-house, as was the rabbit polyclonal antibody to insulin-like factor-3 (Insl3), which was used at a 1:200 dilution. 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, Inc., Santa Cruz, CA) and was diluted 1:500. Mouse monoclonal estrogen receptor (ER) antibody (NCL-ER-6F11/2, Novocastra Laboratories Ltd., Newcastle upon Tyne, UK) was diluted 1:40.
Immunohistochemistry
Immunohistochemistry was performed on dewaxed sections without antigen retrieval, except for sections used for ER immunostaining, which were subjected to heat-induced antigen retrieval for 5 min in 0.01 M citrate buffer, pH 6.0 (Sigma-Aldrich Corp., St. Louis, MO) using a pressure cooker. This was followed by endogenous peroxidase blocking [3% (vol/vol) H2O2 (BDH, Poole, UK)] in methanol (BDH) for 30 min at room temperature. All washes between antibody or reagent incubations comprised two washes, 5 min each time, at room temperature in Tris-(Tris-hydroxymethyl methylamine)-buffered saline [TBS; 0.05 M Tris, pH 7.4, 0.85% (vol/vol) saline] unless otherwise stated. Tissue sections were first blocked in TBS containing normal swine serum or normal rabbit serum for ER- (1:4 dilution; Diagnostics Scotland, Carluke, UK) and 5% (wt/vol) BSA [Sigma-Aldrich Corp.; normal swine serum (NSS) or normal rabbit serum (NRS)] before incubation with primary antibodies diluted in NSS or NRS. A swine antirabbit biotinylated secondary antibody (E0353, DakoCytomation, Carpinteria, CA) or rabbit antimouse biotinylated secondary antibody (E0464, DakoCytomation), both applied at 1:500, were diluted in NSS or NRS, respectively, and incubated at room temperature for 30 min. Bound antibodies were visualized by incubating the sections with avidin-biotin-horseradish peroxidase agent (K0355, DakoCytomation) for 30 min, followed by color development with 3,3'-diaminobenzidine tetrahydrochloride chromogenic substrate (K3468, Liquid DAB+ kit, DakoCytomation), monitored microscopically. Sections were counterstained with hematoxylin, dehydrated, and mounted with Pertex (Histolab, Gteborg, Sweden). Images were captured using a Provis microscope (Olympus Optical Co., London, UK) equipped with a Kodak DCS330 camera (Eastman Kodak Co., Rochester, NY). Captured images were stored on a Macintosh G4 computer (Apple Computer, Cupertino, CA) and compiled using Photoshop 7.0 (Adobe Systems, Mountain View, CA).
For double immunostaining for AR and 3-HSD, dewaxed sections were subjected to heat-induced antigen retrieval, as described above, before endogenous peroxidase blocking. Testis sections were blocked in normal goat serum [goat serum (Diagnostics Scotland) diluted 1:4 in TBS with 5% (wt/vol) BSA (Sigma-Aldrich Corp.); NGS] and incubated with anti-AR (Santa Cruz Biotechnology, Inc.) before the addition of a goat antirabbit peroxidase secondary antibody (P0448, DakoCytomation) diluted 1:200 and subsequent color development with DAB+ (DakoCytomation). Slides were then boiled in glycine/EDTA [0.05 M glycine and 0.01% (wt/vol) EDTA, pH 3.5] for 2 min before blocking again in NGS and incubation with 3-HSD antibody. A biotinylated goat antirabbit antibody (E0432, DakoCytomation), diluted 1:500, was used in combination with Strept ABComplex/alkaline phosphatase (K0391, DakoCytomation) and Fast Blue [1 mg/ml Fast Blue BB salt (Sigma-Aldrich Corp.) in 0.1 M Tris-(hydroxymethyl)methylamine (pH 8.2), 200 μg/ml Naphthol AS-MX phosphate (Sigma-Aldrich Corp.), and 2% (vol/vol) dimethylformamide buffer] to enable visualization of 3-HSD protein expression. Sections were counterstained with hematoxylin and aqueous mounted in Hydromount (National Diagnostics, Highland Park, NJ). All washes for double immunostaining comprise of one wash in TBS and 0.05% (vol/vol) Tween 20, followed by a second wash in TBS, and all antibodies were diluted in NGS.
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 was mounted on the same slide.
Measurement of Leydig (3-HSD-positive) cytoplasmic volume and number per testis
Testicular sections were immunostained for 3-HSD as described previously (19, 24) and counterstained with hematoxylin. The volume of 3-HSD-positive cells per testis was then determined using point-counting methods detailed previously (19, 25). In brief, testicular cross-sections from each of three to 12 animals per group and age were examined, and points falling over 3-HSD-positive cytoplasm or over the nuclei of cells with 3-HSD-positive cytoplasm were scored separately. Both were then independently expressed as relative volumes per testis. This data were converted to absolute volumes per testis by multiplying by testis weight (=volume), because shrinkage was minimal. Separately, the diameter of 100 LC nuclei in three to five animals per group was measured, and the mean value obtained was used to convert data for LC nuclear volume per testis to LC numbers per testis using standard procedures, as described previously (25).
LC morphology and ultrastructure in adult control and SCARKO mice
The testes of three control and three 50-d-old SCARKO mice were perfusion-fixed with 4% (vol/vol) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) preceded by a brief saline wash. The testes were then diced into small pieces, placed into the same fixative for 1 h, washed in cacodylate buffer overnight, postfixed with 1% (wt/vol) osmium/1.25% (wt/vol) potassium ferrocyanide mixture, dehydrated in ethanol, and embedded in Araldite (CY 212). Thin sections were then prepared, mounted on 200-mesh grids, stained with uranyl acetate and lead citrate, and examined on an EM-10 electron microscope (Zeiss, Oberkochen, Germany). Photomicrographs were taken at x6,500 and, with a final magnification of approximately x15,000, the volume density of mitochondria, endoplasmic reticulum, and lipid droplets was determined by point counting using a multipurpose grid (line length, 1 cm). A minimum of 10 LC were analyzed in each animal. The volume density of organelles per cell was determined as detailed previously (26).
Measurement of plasma LH and testosterone levels
Serum LH was measured via a double-antibody RIA using reagents supplied by Dr. A. F. Parlow (Harbor-University of California-Los Angeles, Torrance, CA) and the National Institute of Diabetes and Digestive and Kidney Diseases, National Hormone and Peptide Program. The standard preparation used was mLH-RP (lot AFP5306A), the tracer was prepared from rLH-I-10, and the antiserum was anti rLH-S-11. Serum testosterone levels were measured using the Testo-RIA-CT kit (BioSource International, Camarillo, CA; detection limit, 0.05 ng/ml), according to the instructions of the manufacturer. The within-assay coefficients of variation for testosterone and LH were 4.7% and 7.8%; the interassay coefficients of variation were 6.2% and 8.1% for testosterone and LH, respectively.
RNA analysis
Tissue samples were removed and snap-frozen in liquid nitrogen. RNA from control (AMH-Cre) and SCARKO testes was prepared using the RNeasy midi kit (Qiagen, Chatsworth, CA). Due to their small testicular size, RNA from ARKO testes and their appropriate control, PGK-Cre testes, were extracted using the RNeasy mini kit (Qiagen). Synthesis of cDNA from deoxyribonuclease I-treated total RNA (RNase-Free DNase I Set, Qiagen) used SuperScript II ribonuclease H– reverse transcriptase and random hexamer primers (Invitrogen Life Technologies, Carlsbad, CA). To allow specific mRNA levels to be expressed per testis and to control for the efficiency of RNA extraction, RNA degradation, and the RT step, 10 ng luciferase mRNA (Promega Corp., Madison, WI) were added to each testis at the start of the RNA extraction procedure and used as an internal standard. Gene expression was quantified using the ABI PRISM 7700 PCR detection system (Applied Biosystems, Foster City, CA) with a quantitative two-step RT-PCR protocol. Components for real-time PCR were obtained from Applied Biosystems, except for primers and probes (Eurogentec, Sar-Tilman, Belgium) and SYBR Green (Sigma-Aldrich Corp.). Each 25-μl real-time PCR contained 1x buffer A, 5 mM MgCl2, 400 μM deoxy-NTPs, 200 nM of each primer, 0.4x SYBR Green, 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 for platelet-derived growth factor (PDGF) receptor- were: forward, 5'-CCTTACGACTCCAGATGGGAAT-3'; reverse, 5'-ATGCACCGGATCCCAAAA-3'; those used for PDGF-A, Insl3, P450scc, 3-HSDI, P450c17, and estrogen sulfotransferase (EST) were described previously (20). 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, for which a two-sample t test was employed, statistical analysis was performed using one-way ANOVA supplemented with Fisher’s multiple comparison test using NCSS 2000 software (NCSS Statistical Analysis and Data Analysis Software, Kaysville, UT).
Results
Testis weight in SCARKO and ARKO mice
In SCARKO mice, testis weight on d 12 was comparable to that in controls, but was reduced to 53% of the control value on d 20 and to 25–30% of the control value in adulthood (Table 1). In ARKO mice, testis weight was already reduced to 43% of the control value on d 12, and the magnitude of this decrease increased progressively thereafter, such that by d 140, testis weight was less than 6% of that in controls (Table 1).
Development of the adult LC population in SCARKOs and ARKOs
In controls, LC number increased 26-fold from d 12 to d 140, with most of this increase occurring during the period between 12 and 50 d of age (Fig. 1A). In SCARKO animals on d 12, LC number (mean ± SEM, 0.36 ± 0.05 x 106; n = 5) was comparable to the control value (0.40 ± 0.06 x 106; n = 10), but on d 20, 50, and 140, LC number was reduced by 42–54% compared with controls (Fig. 1A), although the temporal pattern of change in LC number was broadly similar in SCARKO mice and controls. In ARKO mice, LC number was already reduced (by 37%, although nonsignificantly) on d 12 (0.25 ± 0.03 x 106; n = 5) compared with controls, and a major decrease in LC number was evident at all later ages, such that the LC number achieved on d 140 was only 22% of that in controls (Fig. 1A). On d 50 and 140, LC number was significantly higher in SCARKO than in ARKO animals (Fig. 1A).
LC cytoplasmic volume in controls was relatively constant up to d 20, but more than doubled between d 20 and 50, a change equally evident in SCARKO animals (Fig. 1B). Between d 50 and 140, LC cytoplasmic volume in controls remained constant, but there was additionally a 32% increase (P < 0.05) in cytoplasmic volume of LC in SCARKO animals between d 50 and 140, such that at the latter age their LC cytoplasmic volume was 23% greater than that in controls (P < 0.05; Fig. 1B). In contrast, LC cytoplasmic volume in ARKO animals remained virtually unchanged from d 12 to 140, and at the latter age it was only 60% of that in controls and 48% of that in SCARKO mice (Fig. 1B).
LC function in adulthood in SCARKO and ARKO mice
To evaluate the function of LC in adulthood (d 50–140), the mRNA expression and immunoexpression of four proteins connected to LC hormone secretory function (P450ssc, 3-HSD, P450c17, and Insl3) were studied (Fig. 2) as well as the expression of the mRNA for EST; data for mRNA expression were expressed relative to LC number (Fig. 3).
Despite the reduction in LC number in SCARKO and ARKO testes, the substantial reduction in testis size and seminiferous tubule volume means that more LC are apparent in testicular cross-sections (Fig. 2). This has to be taken into account when evaluating whether the immunoexpression of steroidogenic enzymes in LC is altered in such testes. The intensity of immunoexpression of P450ssc and P450c17 in SCARKOs displayed little change from controls, whereas immunoexpression of P450ssc was increased and that of P450c17 was reduced in ARKO mice (Fig. 2). Immunoexpression of 3-HSD revealed no major difference in the intensity of expression among the three groups (Fig. 2); it should also be noted that the antibody used was not 3-HSD type specific. Immunoexpression of Insl3 was comparable to that in controls in SCARKO mice, but was reduced markedly in intensity in ARKO testes (Fig. 2).
Expression per LC of the five mRNAs studied revealed major differences between SCARKOs and ARKOs. Thus, mRNA expression for each of the three steroidogenic enzymes investigated (P450ssc, 3-HSD type 1, and P450c17) was approximately doubled in SCARKO mice compared with controls, whereas ARKO mice showed a different pattern of change for each enzyme; namely, expression of 3-HSD type 1 mRNA was increased 5-fold, and that of P450ssc was increased by 2-fold, whereas mRNA for P450c17 was reduced by approximately 70% (Fig. 3). In SCARKO mice, the expression per LC of Insl3 was normal, and that of EST increased more than 3-fold compared with controls, whereas in ARKO mice, the expression of Insl3 and EST was reduced by more than 90% compared with controls (Fig. 3).
Serum testosterone levels were only determined on d 50 and 140 and showed very wide variation between animals within genotype groups. Although mean values in both SCARKO and ARKO animals were lower than those in controls, there was no significant difference from controls (Fig. 4). In contrast, serum LH levels were elevated more than 8-fold in ARKO mice compared with controls, presumably due to the lack of androgen negative feedback via ARs, but LH levels were normal in SCARKO mice, consistent with there being a normal level of androgen feedback (Fig. 4). Additional evidence of normal serum testosterone levels overall in SCARKO mice was that seminal vesicle weight on d 50 and 140 was comparable to that in controls (Table 2); the seminal vesicles were absent in ARKOs due to failure of Wolffian duct development (21).
LC morphology and ultrastructure
The data presented above suggested that LC in SCARKO mice, although reduced in number compared with controls, were larger and expressed higher levels of the mRNAs for three key steroidogenic enzymes. Therefore, LC morphology was investigated at the electron microscopic level on d 50 (Fig. 5), and the volume of cytoplasmic organelles per LC was evaluated. This demonstrated that the volumes of mitochondria and lipid droplets were approximately doubled in SCARKO animals compared with controls (Fig. 5D), although only the increase in lipid droplets was statistically significant (P < 0.05). In contrast, the volume of endoplasmic reticulum per LC was nonsignificantly decreased in SCARKO mice compared with controls (Fig. 5D).
Evaluation of possible explanations for altered LC development/function in SCARKO mice
Three approaches were used to provide additional insight into the possible origins of the altered LC number/function in SCARKO animals. First, colocalization of AR and 3-HSD by immunohistochemistry revealed that on d 12, in both control and SCARKO mice, only a minority of cells identified as LC by 3-HSD immunostaining of their cytoplasm also expressed AR in their nuclei (Fig. 6); other interstitial cells, which were immunonegative for 3-HSD, were AR immunopositive. In contrast, by d 20 and through to d 50, most 3-HSD-immunopositive cells were AR immunopositive in both control and SCARKO mice (Fig. 6); there was no AR immunoexpression at any age in ARKO testes (Fig. 6). Second, immunostaining for ER revealed that the intensity of ER immunoexpression was reduced in SCARKO LC and increased in ARKO LC compared with controls (supplementary figure), perhaps indicating parallel alterations in estrogen action within LC.
Because no difference in interstitial cell AR immunoexpression or in LH and testosterone levels was found in control and SCARKO mice, our conclusion was that the altered LC development in SCARKOs most likely stemmed from altered androgen action on SC to modify the secretion of LC growth regulatory factors. We therefore evaluated the expression of mRNAs for PDGF-A and its receptor, because our earlier study of SC development had indicated reduced expression of PDGF-A mRNA in adulthood in both SCARKOs and ARKOs (24), and there is evidence for an important role of this growth factor in LC development (27). Quantitative PCR confirmed that PDGF-A gene expression per testis was reduced markedly in SCARKO mice on both d 20 and 50, whereas the expression of PDGF receptor-, when expressed per LC, was unaltered in SCARKO mice, except for a small increase on d 50 (Fig. 7). It was notable that the expression of mRNA for PDGF receptor- declined more than 10-fold from d 12 to 50 in both control and SCARKO mice (Fig. 7).
Discussion
The primary aim of the present study was to evaluate the effect of SC-selective knockout of the AR on the development and function of the adult LC population. To provide perspective for this aim, we also compared LC development in ARKO animals, because previous studies have shown that the development of the adult LC population is severely compromised in such animals (20). Our findings confirmed this, but show that in SCARKO males, LC numbers are not reduced as severely as in ARKOs, and there are some major differences in LC development and function between these two knockout models. Some of these differences may be explained by the fact that all testes are cryptorchid in ARKO/testicular feminized mice (tfm) mice (20, 21), whereas testes are scrotal in SCARKO mice (21). Our analysis of LC gene and protein expressions provide some potentially important insights into the pathways via which LC number and function may have been altered in SCARKO animals.
The major observed differences in LC development between ARKO and SCARKO mice were 1) LC number was reduced only by about 50% in SCARKO mice compared with about 75% in ARKO mice; and 2) LC size was normal or increased in SCARKO mice compared with controls, whereas LC size was consistently reduced in ARKO mice. The reduced LC numbers in ARKO animals are not explained by the concurrent cryptorchidism (20). At face value, the comparison of LC number in the two knockout models could indicate that the development of approximately 50% of adult LC number is dependent on androgen action via the AR in SC, whereas the development of normal adult LC size is not dependent on such action. In contrast, androgen action on testicular cell types other than SC is essential for the development of normal adult LC size and also accounts for approximately 25% of the final adult LC number. The latter finding would be consistent with evidence that androgen action on LC precursors plays a role in their differentiation into immature LC (14), and the present observation that the AR was mainly expressed in 3-HSD-immunonegative cells (i.e. not in immature and adult LC, which are both 3-HSD immunopositive) in control and SCARKO animals at 12 d of age is consistent with this. Although these straightforward interpretations may be partly true, there are numerous publications that attest to potentially important roles for locally produced hormones and growth factors in regulating LC development and function (3, 5, 7, 10), and the expression of at least two of these were altered dramatically in SCARKO and ARKO mice, namely, EST and PDGF-A, as discussed below.
EST, which is expressed in LC, plays a key role in the inactivation of biologically active estrogens. In EST knockout mice, there is reduced testicular expression of P450c17 (28), reduced testosterone production (28), and progressive LC hyperplasia/hypertrophy with aging (29). These changes can be recapitulated by administering estrogens to mice and these exert their effect via ER (30, 31). Conversely, knockout of aromatase, leading to ablation of estrogen action in the testis, leads to LC hyperplasia and/or hypertrophy and elevated testosterone levels (32). The dramatic down-regulation of EST expression in ARKO mice in the present studies and also reported in tfm mice in other studies (33) may therefore account for their reduced expression of P450c17, although the categorical absence of LC hyperplasia and hypertrophy in ARKO mice means that altered EST expression does not provide a complete explanation of their LC phenotype. In any case, the reduced EST expression in ARKO mice is probably explained by the abdominal location of the testes (20). In contrast to ARKO animals, the expression of EST in LC of adult SCARKO mice was increased 3-fold, a change that would be expected to increase the rate of estrogen catabolism within LC and thus reduce local estrogen action within the LC. ER knockout mice, in which estrogen action on LC via ER is prevented, exhibit increased testosterone production per LC in fetal (31) as well as in adult (30) life together with increased expression of P450scc and P450c17, changes also observed in adult LC of SCARKO mice in the present studies. Estrogen action in mice can also lead to LC hyperplasia and hypertrophy (33, 34), but whether underactivity of estrogens can lead to reduced LC number in mice, as found presently in SCARKO animals, is unknown. In view of the evidence just discussed, it seems reasonable to propose that some of the LC changes observed in both SCARKO and ARKO mice are attributable to altered EST expression and consequent alteration in local estrogen action within LC. Our observations of increased intensity of ER immunoexpression in LC nuclei of ARKO mice and reduced ER immunoexpression in LC nuclei of SCARKO mice are consistent with this interpretation, because estrogens are recognized to positively regulate ER immunoexpression (35). Although EST expression in the epididymis (36) and LC (33) is reported to be androgen dependent, the mechanism by which androgen action on SC is able to alter EST expression in LC in SCARKO mice remains to be defined, but one potential candidate for mediating SC-LC communication is PDGF-A, which is expressed predominantly in SC in the testis (27, 37).
In a recent study of SC development in SCARKO and ARKO mice, we reported that both models exhibited a major decrease in the expression of PDGF-A (24), a change confirmed in the present study for SCARKO animals. Because the reduction in PDGF-A expression occurred in both models, it is unlikely that the cryptorchidism in ARKOs (20, 21) is a major factor in this decrease. The reduction in PDGF-A could account for the reduced LC number in SCARKO and ARKO mice described in this study, because interference with PDGF-A action via knockout of its receptor results in impairment of development of fetal LC (38), and knockout of PDGF-A itself leads to gross impairment of adult LC development (27). In this regard, the relative degree of reduction in PDGF-A gene expression in SCARKO mice (present study and Ref. 24) and ARKO mice (24) parallels the magnitude of reduction in LC number.
Another SC-derived factor that has been shown to affect LC development in rodents is AMH. LC development is slightly compromised in AMH-deficient mice (39), whereas aberrantly high/continued expression of AMH beyond early puberty inhibits the development of adult LC generation and steroidogenesis (40, 41). However, our previous study indicated that the expression of AMH by SC showed a normal age-related decline in both SCARKO and ARKO mice (24), so AMH is unlikely to be involved in the LC changes in these animals.
In view of the key role that SC appear to play in regulating LC development and function, it has been suggested that an alteration of SC number, such as occurs in animals in which thyroid hormones are manipulated perinatally could lead to a parallel change in LC numbers (11, 18). In ARKO animals, the approximately 75% reduction in LC number shown here and previously (20) parallels a similar reduction in SC number (24). However, SC number remains largely unchanged from control values in SCARKO mice (24), yet these same animals exhibit an approximately 50% reduction in adult LC number. The present data together with other findings (19) therefore do not support the concept of a simple relationship between SC and LC numbers in the adult testis.
One of the most puzzling observations of the present studies was the apparent disparity between the major reduction in LC number in SCARKO and ARKO mice and the lack of change in blood levels of testosterone. Although the latter measurements showed huge between-animal variation, which may have obscured a reduction in testosterone levels in ARKO mice, three independent lines of evidence support the normality of testosterone levels in SCARKO animals; thus, serum testosterone levels were normal, serum LH levels were normal, and seminal vesicle weight was normal. The ability of LC in SCARKO mice to maintain normal testosterone levels despite an approximately 50% reduction in LC number and no increase in LH levels can only indicate that testosterone production per LC must be increased, perhaps doubled, in young adult SCARKO animals. Our observations of increased LC cytoplasmic volume in adulthood, increased volume of lipid droplets and mitochondria, and increased expression per LC of several of the key steroidogenic enzymes (P450ssc, 3-HSD, and P450c17) in adult SCARKO animals are consistent with this interpretation. The finding that testosterone levels are normal in SCARKO mice means that, physiologically, androgen-dependent processes outside the SC may be entirely normal in these animals. However, the fact that major adaptive changes have to occur in LC function/gene expression in SCARKO mice to achieve normal testosterone levels despite the halving of LC number, emphasizes the importance of a thorough analysis of testicular cell development in transgenic models, because there may be secondary consequences of these adaptive changes (e.g. altered estrogen action within the testis).
Although direct measurement of serum testosterone levels did not show any significant reduction in ARKO mice, our suspicion is that a reduction was masked by the high variability in control values, because earlier studies of tfm and ARKO male mice have reported up to 80% reductions in testosterone levels (42, 43). Because there were no seminal vesicles in ARKO mice, and serum LH levels were grossly elevated due to lack of viable androgen negative feedback, there was no independent measure of the normality of testosterone levels in ARKO animals. In fact, the marked reduction in LC cytoplasmic volume and in the expression of P450c17 per LC in ARKO mice would fit more readily with reduced testosterone production; in this regard, it is possible that the reduction in P450c17 expression per LC may result from the almost complete suppression of EST expression in ARKO animals, because a similar reduction in P450c17 expression was observed in EST null mice (28); the reduced EST expression in ARKO mice is probably a consequence of the cryptorchid position of the testes (20).
The present data also show that Insl3 expression in adult LC is almost obliterated in ARKO mice, but is normal in SCARKO animals; the massive reduction in ARKO mice is not explained by cryptorchid position of the testes (20), but is consistent with the view that fully adult LC do not differentiate in ARKO/tfm testes (20), although a role for androgens in regulating Insl3 expression is also possible. In this regard, the normal LC expression of Insl3 in SCARKO mice shows that androgen action on SC is not required for the expression of this protein in LC.
In conclusion, the present findings demonstrate that the absence of AR-mediated androgen action selectively on SC has important consequences for LC development, in particular for the development of normal LC numbers. In terms of LC function and LC-specific gene expression, there are major differences between SCARKO and ARKO animals, which may indicate that androgen action via cell types other than SC, probably including LC and their precursors, is important for the development of normal LC hormone (testosterone and Insl3) secretory function; however, this conclusion is tempered by the occurrence of cryptorchidism in ARKO mice, which itself is undoubtedly an important cause of changes in LC gene expression (20). Nevertheless, it is likely that estrogens (via altered EST expression) and PDGF-A may play some part in the LC changes, in SCARKO animals at least, although it is likely that factors other than these are also involved. The altered LC function in SCARKO mice, in particular the likelihood of reduced estrogen levels intratesticularly due to elevation of EST expression, must be taken into account when using this model for investigation of the regulation of spermatogenesis.
Acknowledgments
We thank Arantza Esnal, Hilde Geeraerts, Ludo Deboel, and Adriano Moreira for their excellent technical assistance.
Footnotes
This work was supported by Concerted Research Action (Research Fund, Katholieke Universiteit Leuven), Fund for Scientific Research Flanders (Belgium), Medical Research Council (United Kingdom), Contract QLK4-CT-2002–00603 (EDEN) from the European Union, and a Wellcome International Research Training Fellowship (to N.A.).
1 K.D.G. and N.A. contributed equally to this study.
Abbreviations: AMH, Anti-Mullerian hormone; AR, androgen receptor; ARKO, complete androgen receptor knockout; Cre recombinase, cyclization recombination recombinase; EST, estrogen sulfotransferase; 3-HSD, 3-hydroxysteroid dehydrogenase; Insl3, insulin-like factor-3; LC, Leydig cell; NGS, normal goat serum; NRS, normal rabbit serum; NSS, normal swine serum; PDGF-A, platelet-derived growth factor-A; P450scc, P450 cholesterol side chain cleavage enzyme; P450c17 17-hydroxylase/C17–20lyase; PGK, phosphoglycerate kinase-1; SC, Sertoli cell; SCARKO, Sertoli cell-selective androgen receptor knockout; TBS, Tris-buffered saline; tfm, testicular feminized mice; WT, wild type.
References
Sharpe RM 1994 Regulation of spermatogenesis. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven Press; 1363–1434
Sharpe RM 2005 Sertoli cell endocrinology and signal transduction: androgen regulation. In: Griswold M, Skinner M, eds. Sertoli cell biology. 1st ed. San Diego: Academic Press; 199–216
Ge R-S, Shan L-X, Hardy MP 1996 Pubertal development of Leydig cells. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 159–172
Ge RS, Hardy MP 1998 Variation in the end products of androgen biosynthesis and metabolism during postnatal differentiation of rat Leydig cells. Endocrinology 139:3787–3795
Mendis-Handagma SM, Ariyartne HB 2001 Differentiation of the adult Leydig cell population in the postnatal testis. Biol Reprod 65:660–671
Davidoff MS, Middendorf R, Enikolopov G, Riethmacher D, Holsetin AF, Muller D 2004 Progenitor cells of the testosterone-producing Leydig cells revealed. J Cell Biol 167:935–944
Haider SG 2004 Cell biology of Leydig cells in the testis. Int Rev Cytol 233:181–241
Teerds KJ 1996 Regeneration of Leydig cells after depletion by EDS: a model for postnatal Leydig cell renewal. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 203–219
Sharpe RM 1993 Experimental evidence for Sertoli cell-germ cell and Sertoli cell-Leydig cell interactions. In: Russell LD, Griswold MD, eds. The Sertoli cell. 1st ed. Clearwater: Cache River Press; 391–419
Saez JM, Lejeune H 1996 Regulation of Leydig cell functions by hormones and growth factors other than LH and IGF-1. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 383–406
Hardy MP, Sharma RS, Arambepola NK, Sottas CM, Russell LD, Bunick D, Hess RA, Cooke PS 1996 Increased proliferation of Leydig cells induced by neonatal hypothyroidism in the rat. J Androl 17:231–238
Teerds KJ, de Rooij DG, de Jong FH, van Haaster LH 1998 Development of the adult type Leydig cell population in the rat is affected by neonatal thyroid hormone levels. Biol Reprod 59:344–350
Ariyaratne HB, Mills N, Mason JI, Mendis-Handagama SM 2000 Effects of thyroid hormone on Leydig cell regeneration in the adult rat following ethane dimethane sulfonate treatment. Biol Reprod 63:1115–1123
Hardy MP, Kelce WR, Klinefelter GR, Ewing LL 1990 Differentiation of Leydig cell precursors in vitro: a role for androgen. Endocrinology 127:488–490
Abney TO, Myers RB 1991 17-Estradiol inhibition of Leydig cell regeneration in the ethane dimethylsulfonate-treated mature rat. J Androl 12:295–304
Abney TO 1999 The potential roles of estrogens in regulating Leydig cell development and function: a review. Steroids 64:610–617
Verhoeven G, Cailleau J 1990 Influence of co-culture with Sertoli cells on steroidogenesis in immature rat Leydig cells. Mol Cell Endocrinol 71:239–251
Hardy MP, Kirby JD, Hess RA, Cooke PS 1993 Leydig cells increase their numbers but decline in steroidogenic function in the adult rat after neonatal hypothyroidism. Endocrinology 132:2417–2426
Sharpe RM, Rivas A, Walker M, McKinnell C, Fisher JS 2003 Effect of neonatal treatment of rats with potent or weak (environmental) oestrogens, or with a GnRH antagonist, on Leydig cell development and function through puberty into adulthood. Int J Androl 26:26–36
O’Shaughnessy PJ, Johnston H, Willerton L, Baker PJ 2002 failure of normal adult Leydig cell development in androgen-receptor-deficient mice. J Cell Sci 115:3491–3496
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
Tan K, De Gendt K, Atanassova N, Sharpe RM, Saunders PTK, Denolet E, Verhoeven G 2005 The role of androgens in Sertoli cell proliferation and functional maturation: studies in mice with total (ARKO) or Sertoli cell-selective (SCARKO) ablation of the androgen receptor. Endocrinology 146:2674–2683
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
Frana LR, Ye S-J, Ying L, Sandberg M, Russell LD 1995 Morphometry of rat germ cells during spermatogenesis. Anat Rec 241:181–204
Gnessi L, Bascianai 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 149:1019–1025
Tong MH, Christenson LK, Song WC 2004 Aberrant cholesterol transport and impaired steroidogenesis in Leydig cells lacking estrogen sulfotransferase. Endocrinology 145:2487–2497
Quian YM, Sun XJ, Tong MH, Li XP, Richa J, Song WC 2001 Targeted disruption of the mouse estrogen sulfotransferase gene reveals a role of estrogen metabolism in intracrine and paracrine estrogen regulation. Endocrinology 142:5342–5350
Akingbemi BT, ge R, Rosenfeld CS, Newton LG, Hardy DO, Catterall JF, Lubahn DB, Korach KS, Hardy MP 2003 Estrogen receptor- gene deficiency enhances androgen biosynthesis in the mouse Leydig cell. Endocrinology 144:84–93
Delbes G, Levacher C, Duquenne C, Racine C, Pakarinen P, Habert R 2005 Endogenous estrogens inhibit mouse fetal Leydig cell development via estrogen receptor . Endocrinology 146:2454–2461
Robertson KM, O’Donnell L, Jones ME, Meachme SJ, Boon WC, Fisher CR, Graves KH, McLachlan RI, Simpson ER 1999 Impairment of spermatogenesis in mice lacking a functional aromatase (cyp 19) gene. Proc Natl Acad Sci USA 96:7986–7991
Quian YM, Song WC 1999 Regulation of estrogen sulfotransferase expression in Leydig cells by adenosine 3',5'-monophsophate and androgen. Endocrinology 140:1048–1053
Cook J, Klinefelter GR, Hardisty JF, Sharpe RM, Foster PMD 1999 Rodent Leydig cell tumorigenesis: a review of the physiology, pathology, mechanisms and relevance to humans. Crit Rev Toxicol 29:169–261
Kos M, Reid G, Denger S, Gannon F 2001 Minireview: genomic organization of the human ER gene promoter region. Mol Endocrinol 15:2057–2063
Tong MH, Song WC 2002 Estrogen sulfotransferase: discrete and androgen-dependent expression in the male reproductive tract and demonstration of an in vivo function in the mouse epididymis. Endocrinology 143:3144–3151
Loveland KL, Zlatic K, Stein-Oakley A, Risbridger G, de Kretser 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
Brennan J, Tilmann C, Capel B 2003 PDGFR- mediates testis cord organization and fetal Leydig cell development in the XY gonad. Genes Dev 17:800–810
Wu X, Arumugam R, Baker SP, Lee MM 2005 Pubertal and adult Leydig cell function in Mullerian inhibiting substance-deficient mice. Endocrinology 146:589–595
Fynn-Thompson E, Cheng H, Teixeira J 2003 Inhibition of steroidogenesis in Leydig cells by Mullerian-inhibiting substance. Mol Cell Endocrinol 211:99–104
Salva A, Hardy MP, Wu XF, Sottas CM, MacLaughlin DT, Donahoe PK, Lee MM 2004 Mullerian-inhibiting substance inhibits rat Leydig cell regeneration after ethylene dimethanesulphonate ablation. Biol Reprod 70:600–607
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
Jones RD, Pugh PJ, Hall J, Channer KS, Jones TH 2003 Altered circulating hormone levels, endothelial function and vascular reactivity in the testicular feminized mouse. Eur J Endocrinol 148:111–120(Karel De Gendt1, Nina Ata)
Medical Research Council Human Reproductive Sciences Unit, Center for Reproductive Biology (N.A., K.A.L.T., C.M., R.M.S., P.T.K.S.), and Division of Reproductive and Developmental Science, University of Edinburgh (J.I.M.), Edinburgh, Scotland, EH16 4SB United Kingdom
Institute of Experimental Morphology and Anthropology, Bulgarian Academy of Science (N.A.), 1113 Sofia, Bulgaria
Laboratory of Cellular Biology, Department of Morphology, Instituto de Ciências Biológicas/Federal University of Minas Gerais (L.R.d.F., G.G.P.), Belo Horizonte-MG, Brazil
Institute for Hormone and Fertility Research, University of Hamburg (S.H.), D-20251 Hamburg, Germany
School of Molecular and Biomedical Sciences, University of Adelaide (R.I.), Adelaide 5005, Australia
Abstract
It is established that androgens and unidentified Sertoli cell (SC)-derived factors can influence the development of adult Leydig cells (LC) in rodents, but the mechanisms are unclear. We evaluated adult LC development and function in SC-selective androgen receptor (AR) knockout (SCARKO) and complete AR knockout (ARKO) mice. In controls, LC number increased 26-fold and LC size increased by approximately 2-fold between 12 and 140 d of age. LC number in SCARKOs was normal on d 12, but was reduced by more than 40% at later ages, although LC were larger and contained more lipid droplets and mitochondria than control LC by adulthood. ARKO LC number was reduced by up to 83% at all ages compared with controls, and LC size did not increase beyond d 12. Serum LH and testosterone levels and seminal vesicle weights were comparable in adult SCARKOs and controls, whereas LH levels were elevated 8-fold in ARKOs, although testosterone levels appeared normal. Immunohistochemistry and quantitative PCR for LC-specific markers indicated steroidogenic function per LC was probably increased in SCARKOs and reduced in ARKOs. In SCARKOs, insulin-like factor-3 and estrogen sulfotransferase (EST) mRNA expression were unchanged and increased 3-fold, respectively, compared with controls, whereas the expression of both was reduced more than 90% in ARKOs. Changes in EST expression, coupled with reduced platelet-derived growth factor-A expression, are potential causes of altered LC number and function in SCARKOs. These results show that loss of androgen action on SC has major consequences for LC development, and this could be mediated indirectly via platelet-derived growth factor-A and/or estrogens/EST.
Introduction
IN ADULTHOOD, TESTOSTERONE action on seminiferous tubules is essential for full, quantitatively normal spermatogenesis and fertility, an effect mediated largely by action on Sertoli cells (SC) via the androgen receptor (AR), although the downstream mechanisms remain unclear (1, 2). Testosterone is produced by the adult population of Leydig cells (LC), which develop during puberty from presumptive progenitor cells that are present before this time (3, 4, 5, 6). Development of the LC involves at least three steps, namely the proliferation of LC precursors, their differentiation into immature LC and their final differentiation into adult LC (3, 5, 7). The mechanisms that regulate the numbers of both adult and precursor LC and the differentiation of the latter into the former are incompletely understood, but hormones such as LH (8), FSH (9, 10), thyroid hormone (11, 12, 13), androgens (3, 14), and estrogens (15, 16) probably play a role. Additionally, numerous studies involving a variety of experimental approaches have demonstrated that SC play a role in regulating the development and function of the adult LC (9, 17). For example, FSH action on SC has been shown to advance the development of LC, and other studies have shown that the status of spermatogenesis (i.e. the germ cell complement) can alter the development of neighboring LC, presumably via the action of SC-derived (unknown) factors (9). Indeed, some (11, 18), but not all (19), studies suggest that the number of SC per testis may determine the number of adult LC.
Differentiation of LC precursors into immature LC is a key step in development of the adult LC population (3, 5), and androgens appear to play a positive role in this process (14). Perhaps the most dramatic demonstration of the potential importance of androgens in LC development is the observation that in mice lacking a functional AR, there is a major reduction in adult LC number and function (20). Whether this deficit arises because of the absence of effect of androgens directly on LC or their precursors or is mediated partly or wholly via the absence of effects on the SC (or even on the peritubular myoid cells) is unclear. Another complication in this model is that the testes are cryptorchid, and the consequent elevation of testicular temperature and/or the associated depletion of germ cells may also play a role. The recent generation by ourselves (21) and others (22, 23) 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 the development and function of testicular cells. Moreover, by comparing these processes in SCARKO and complete AR knockout [AR0/Y; ARKO] mice, it is possible to gain insights into the relative importance of SC-mediated and non-SC-mediated effects of androgens on developing LC. We used such an approach in the present studies and show that SC-mediated effects of androgens may be important for the development of normal numbers of adult LC.
Materials and Methods
Generation of SCARKO and ARKO mice
The AR knockout animals were generated using cyclization recombination (Cre) recombinase /loxP technology. ARflox/+ female animals (129/Swiss) with exon 2 of the AR floxed were either crossed with anti-Mullerian hormone (AMH)-Cre+/+ male mice (C57BL/6) expressing Cre recombinase (under the AMH gene promoter) selectively in SC 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 were provided previously (21). 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 committees.
Tissue and blood sampling
Urogenital systems from control, ARKO and SCARKO males (aged 12, 20, 50, or 140 d) were removed, fixed in Bouin’s fluid for 4–6 h, then transferred to 70% ethanol. Testes and seminal vesicles (if present) were dissected out and weighed. Blood was taken by cardiac puncture under ether anesthesia and allowed to clot overnight at 4 C. Serum was isolated by two centrifugal steps of 10 min at 13,000 rpm and stored at –20 C until assayed.
Antibodies for immunohistochemistry and their dilutions
Rabbit polyclonal antibody (AB1244, Chemicon International, Temecula, CA) recognizing cytochrome P450 side-chain cleavage (P450scc) was used at a 1:200 dilution. Anti-17-hydroxylase/C17–20lyase (anti-P450c17) and 3-hydroxysteroid dehydrogenase (3-HSD) rabbit polyclonal antibodies, used at 1:300 and 1:3000 dilutions, respectively, were generated in-house, as was the rabbit polyclonal antibody to insulin-like factor-3 (Insl3), which was used at a 1:200 dilution. 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, Inc., Santa Cruz, CA) and was diluted 1:500. Mouse monoclonal estrogen receptor (ER) antibody (NCL-ER-6F11/2, Novocastra Laboratories Ltd., Newcastle upon Tyne, UK) was diluted 1:40.
Immunohistochemistry
Immunohistochemistry was performed on dewaxed sections without antigen retrieval, except for sections used for ER immunostaining, which were subjected to heat-induced antigen retrieval for 5 min in 0.01 M citrate buffer, pH 6.0 (Sigma-Aldrich Corp., St. Louis, MO) using a pressure cooker. This was followed by endogenous peroxidase blocking [3% (vol/vol) H2O2 (BDH, Poole, UK)] in methanol (BDH) for 30 min at room temperature. All washes between antibody or reagent incubations comprised two washes, 5 min each time, at room temperature in Tris-(Tris-hydroxymethyl methylamine)-buffered saline [TBS; 0.05 M Tris, pH 7.4, 0.85% (vol/vol) saline] unless otherwise stated. Tissue sections were first blocked in TBS containing normal swine serum or normal rabbit serum for ER- (1:4 dilution; Diagnostics Scotland, Carluke, UK) and 5% (wt/vol) BSA [Sigma-Aldrich Corp.; normal swine serum (NSS) or normal rabbit serum (NRS)] before incubation with primary antibodies diluted in NSS or NRS. A swine antirabbit biotinylated secondary antibody (E0353, DakoCytomation, Carpinteria, CA) or rabbit antimouse biotinylated secondary antibody (E0464, DakoCytomation), both applied at 1:500, were diluted in NSS or NRS, respectively, and incubated at room temperature for 30 min. Bound antibodies were visualized by incubating the sections with avidin-biotin-horseradish peroxidase agent (K0355, DakoCytomation) for 30 min, followed by color development with 3,3'-diaminobenzidine tetrahydrochloride chromogenic substrate (K3468, Liquid DAB+ kit, DakoCytomation), monitored microscopically. Sections were counterstained with hematoxylin, dehydrated, and mounted with Pertex (Histolab, Gteborg, Sweden). Images were captured using a Provis microscope (Olympus Optical Co., London, UK) equipped with a Kodak DCS330 camera (Eastman Kodak Co., Rochester, NY). Captured images were stored on a Macintosh G4 computer (Apple Computer, Cupertino, CA) and compiled using Photoshop 7.0 (Adobe Systems, Mountain View, CA).
For double immunostaining for AR and 3-HSD, dewaxed sections were subjected to heat-induced antigen retrieval, as described above, before endogenous peroxidase blocking. Testis sections were blocked in normal goat serum [goat serum (Diagnostics Scotland) diluted 1:4 in TBS with 5% (wt/vol) BSA (Sigma-Aldrich Corp.); NGS] and incubated with anti-AR (Santa Cruz Biotechnology, Inc.) before the addition of a goat antirabbit peroxidase secondary antibody (P0448, DakoCytomation) diluted 1:200 and subsequent color development with DAB+ (DakoCytomation). Slides were then boiled in glycine/EDTA [0.05 M glycine and 0.01% (wt/vol) EDTA, pH 3.5] for 2 min before blocking again in NGS and incubation with 3-HSD antibody. A biotinylated goat antirabbit antibody (E0432, DakoCytomation), diluted 1:500, was used in combination with Strept ABComplex/alkaline phosphatase (K0391, DakoCytomation) and Fast Blue [1 mg/ml Fast Blue BB salt (Sigma-Aldrich Corp.) in 0.1 M Tris-(hydroxymethyl)methylamine (pH 8.2), 200 μg/ml Naphthol AS-MX phosphate (Sigma-Aldrich Corp.), and 2% (vol/vol) dimethylformamide buffer] to enable visualization of 3-HSD protein expression. Sections were counterstained with hematoxylin and aqueous mounted in Hydromount (National Diagnostics, Highland Park, NJ). All washes for double immunostaining comprise of one wash in TBS and 0.05% (vol/vol) Tween 20, followed by a second wash in TBS, and all antibodies were diluted in NGS.
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 was mounted on the same slide.
Measurement of Leydig (3-HSD-positive) cytoplasmic volume and number per testis
Testicular sections were immunostained for 3-HSD as described previously (19, 24) and counterstained with hematoxylin. The volume of 3-HSD-positive cells per testis was then determined using point-counting methods detailed previously (19, 25). In brief, testicular cross-sections from each of three to 12 animals per group and age were examined, and points falling over 3-HSD-positive cytoplasm or over the nuclei of cells with 3-HSD-positive cytoplasm were scored separately. Both were then independently expressed as relative volumes per testis. This data were converted to absolute volumes per testis by multiplying by testis weight (=volume), because shrinkage was minimal. Separately, the diameter of 100 LC nuclei in three to five animals per group was measured, and the mean value obtained was used to convert data for LC nuclear volume per testis to LC numbers per testis using standard procedures, as described previously (25).
LC morphology and ultrastructure in adult control and SCARKO mice
The testes of three control and three 50-d-old SCARKO mice were perfusion-fixed with 4% (vol/vol) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.3) preceded by a brief saline wash. The testes were then diced into small pieces, placed into the same fixative for 1 h, washed in cacodylate buffer overnight, postfixed with 1% (wt/vol) osmium/1.25% (wt/vol) potassium ferrocyanide mixture, dehydrated in ethanol, and embedded in Araldite (CY 212). Thin sections were then prepared, mounted on 200-mesh grids, stained with uranyl acetate and lead citrate, and examined on an EM-10 electron microscope (Zeiss, Oberkochen, Germany). Photomicrographs were taken at x6,500 and, with a final magnification of approximately x15,000, the volume density of mitochondria, endoplasmic reticulum, and lipid droplets was determined by point counting using a multipurpose grid (line length, 1 cm). A minimum of 10 LC were analyzed in each animal. The volume density of organelles per cell was determined as detailed previously (26).
Measurement of plasma LH and testosterone levels
Serum LH was measured via a double-antibody RIA using reagents supplied by Dr. A. F. Parlow (Harbor-University of California-Los Angeles, Torrance, CA) and the National Institute of Diabetes and Digestive and Kidney Diseases, National Hormone and Peptide Program. The standard preparation used was mLH-RP (lot AFP5306A), the tracer was prepared from rLH-I-10, and the antiserum was anti rLH-S-11. Serum testosterone levels were measured using the Testo-RIA-CT kit (BioSource International, Camarillo, CA; detection limit, 0.05 ng/ml), according to the instructions of the manufacturer. The within-assay coefficients of variation for testosterone and LH were 4.7% and 7.8%; the interassay coefficients of variation were 6.2% and 8.1% for testosterone and LH, respectively.
RNA analysis
Tissue samples were removed and snap-frozen in liquid nitrogen. RNA from control (AMH-Cre) and SCARKO testes was prepared using the RNeasy midi kit (Qiagen, Chatsworth, CA). Due to their small testicular size, RNA from ARKO testes and their appropriate control, PGK-Cre testes, were extracted using the RNeasy mini kit (Qiagen). Synthesis of cDNA from deoxyribonuclease I-treated total RNA (RNase-Free DNase I Set, Qiagen) used SuperScript II ribonuclease H– reverse transcriptase and random hexamer primers (Invitrogen Life Technologies, Carlsbad, CA). To allow specific mRNA levels to be expressed per testis and to control for the efficiency of RNA extraction, RNA degradation, and the RT step, 10 ng luciferase mRNA (Promega Corp., Madison, WI) were added to each testis at the start of the RNA extraction procedure and used as an internal standard. Gene expression was quantified using the ABI PRISM 7700 PCR detection system (Applied Biosystems, Foster City, CA) with a quantitative two-step RT-PCR protocol. Components for real-time PCR were obtained from Applied Biosystems, except for primers and probes (Eurogentec, Sar-Tilman, Belgium) and SYBR Green (Sigma-Aldrich Corp.). Each 25-μl real-time PCR contained 1x buffer A, 5 mM MgCl2, 400 μM deoxy-NTPs, 200 nM of each primer, 0.4x SYBR Green, 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 for platelet-derived growth factor (PDGF) receptor- were: forward, 5'-CCTTACGACTCCAGATGGGAAT-3'; reverse, 5'-ATGCACCGGATCCCAAAA-3'; those used for PDGF-A, Insl3, P450scc, 3-HSDI, P450c17, and estrogen sulfotransferase (EST) were described previously (20). 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, for which a two-sample t test was employed, statistical analysis was performed using one-way ANOVA supplemented with Fisher’s multiple comparison test using NCSS 2000 software (NCSS Statistical Analysis and Data Analysis Software, Kaysville, UT).
Results
Testis weight in SCARKO and ARKO mice
In SCARKO mice, testis weight on d 12 was comparable to that in controls, but was reduced to 53% of the control value on d 20 and to 25–30% of the control value in adulthood (Table 1). In ARKO mice, testis weight was already reduced to 43% of the control value on d 12, and the magnitude of this decrease increased progressively thereafter, such that by d 140, testis weight was less than 6% of that in controls (Table 1).
Development of the adult LC population in SCARKOs and ARKOs
In controls, LC number increased 26-fold from d 12 to d 140, with most of this increase occurring during the period between 12 and 50 d of age (Fig. 1A). In SCARKO animals on d 12, LC number (mean ± SEM, 0.36 ± 0.05 x 106; n = 5) was comparable to the control value (0.40 ± 0.06 x 106; n = 10), but on d 20, 50, and 140, LC number was reduced by 42–54% compared with controls (Fig. 1A), although the temporal pattern of change in LC number was broadly similar in SCARKO mice and controls. In ARKO mice, LC number was already reduced (by 37%, although nonsignificantly) on d 12 (0.25 ± 0.03 x 106; n = 5) compared with controls, and a major decrease in LC number was evident at all later ages, such that the LC number achieved on d 140 was only 22% of that in controls (Fig. 1A). On d 50 and 140, LC number was significantly higher in SCARKO than in ARKO animals (Fig. 1A).
LC cytoplasmic volume in controls was relatively constant up to d 20, but more than doubled between d 20 and 50, a change equally evident in SCARKO animals (Fig. 1B). Between d 50 and 140, LC cytoplasmic volume in controls remained constant, but there was additionally a 32% increase (P < 0.05) in cytoplasmic volume of LC in SCARKO animals between d 50 and 140, such that at the latter age their LC cytoplasmic volume was 23% greater than that in controls (P < 0.05; Fig. 1B). In contrast, LC cytoplasmic volume in ARKO animals remained virtually unchanged from d 12 to 140, and at the latter age it was only 60% of that in controls and 48% of that in SCARKO mice (Fig. 1B).
LC function in adulthood in SCARKO and ARKO mice
To evaluate the function of LC in adulthood (d 50–140), the mRNA expression and immunoexpression of four proteins connected to LC hormone secretory function (P450ssc, 3-HSD, P450c17, and Insl3) were studied (Fig. 2) as well as the expression of the mRNA for EST; data for mRNA expression were expressed relative to LC number (Fig. 3).
Despite the reduction in LC number in SCARKO and ARKO testes, the substantial reduction in testis size and seminiferous tubule volume means that more LC are apparent in testicular cross-sections (Fig. 2). This has to be taken into account when evaluating whether the immunoexpression of steroidogenic enzymes in LC is altered in such testes. The intensity of immunoexpression of P450ssc and P450c17 in SCARKOs displayed little change from controls, whereas immunoexpression of P450ssc was increased and that of P450c17 was reduced in ARKO mice (Fig. 2). Immunoexpression of 3-HSD revealed no major difference in the intensity of expression among the three groups (Fig. 2); it should also be noted that the antibody used was not 3-HSD type specific. Immunoexpression of Insl3 was comparable to that in controls in SCARKO mice, but was reduced markedly in intensity in ARKO testes (Fig. 2).
Expression per LC of the five mRNAs studied revealed major differences between SCARKOs and ARKOs. Thus, mRNA expression for each of the three steroidogenic enzymes investigated (P450ssc, 3-HSD type 1, and P450c17) was approximately doubled in SCARKO mice compared with controls, whereas ARKO mice showed a different pattern of change for each enzyme; namely, expression of 3-HSD type 1 mRNA was increased 5-fold, and that of P450ssc was increased by 2-fold, whereas mRNA for P450c17 was reduced by approximately 70% (Fig. 3). In SCARKO mice, the expression per LC of Insl3 was normal, and that of EST increased more than 3-fold compared with controls, whereas in ARKO mice, the expression of Insl3 and EST was reduced by more than 90% compared with controls (Fig. 3).
Serum testosterone levels were only determined on d 50 and 140 and showed very wide variation between animals within genotype groups. Although mean values in both SCARKO and ARKO animals were lower than those in controls, there was no significant difference from controls (Fig. 4). In contrast, serum LH levels were elevated more than 8-fold in ARKO mice compared with controls, presumably due to the lack of androgen negative feedback via ARs, but LH levels were normal in SCARKO mice, consistent with there being a normal level of androgen feedback (Fig. 4). Additional evidence of normal serum testosterone levels overall in SCARKO mice was that seminal vesicle weight on d 50 and 140 was comparable to that in controls (Table 2); the seminal vesicles were absent in ARKOs due to failure of Wolffian duct development (21).
LC morphology and ultrastructure
The data presented above suggested that LC in SCARKO mice, although reduced in number compared with controls, were larger and expressed higher levels of the mRNAs for three key steroidogenic enzymes. Therefore, LC morphology was investigated at the electron microscopic level on d 50 (Fig. 5), and the volume of cytoplasmic organelles per LC was evaluated. This demonstrated that the volumes of mitochondria and lipid droplets were approximately doubled in SCARKO animals compared with controls (Fig. 5D), although only the increase in lipid droplets was statistically significant (P < 0.05). In contrast, the volume of endoplasmic reticulum per LC was nonsignificantly decreased in SCARKO mice compared with controls (Fig. 5D).
Evaluation of possible explanations for altered LC development/function in SCARKO mice
Three approaches were used to provide additional insight into the possible origins of the altered LC number/function in SCARKO animals. First, colocalization of AR and 3-HSD by immunohistochemistry revealed that on d 12, in both control and SCARKO mice, only a minority of cells identified as LC by 3-HSD immunostaining of their cytoplasm also expressed AR in their nuclei (Fig. 6); other interstitial cells, which were immunonegative for 3-HSD, were AR immunopositive. In contrast, by d 20 and through to d 50, most 3-HSD-immunopositive cells were AR immunopositive in both control and SCARKO mice (Fig. 6); there was no AR immunoexpression at any age in ARKO testes (Fig. 6). Second, immunostaining for ER revealed that the intensity of ER immunoexpression was reduced in SCARKO LC and increased in ARKO LC compared with controls (supplementary figure), perhaps indicating parallel alterations in estrogen action within LC.
Because no difference in interstitial cell AR immunoexpression or in LH and testosterone levels was found in control and SCARKO mice, our conclusion was that the altered LC development in SCARKOs most likely stemmed from altered androgen action on SC to modify the secretion of LC growth regulatory factors. We therefore evaluated the expression of mRNAs for PDGF-A and its receptor, because our earlier study of SC development had indicated reduced expression of PDGF-A mRNA in adulthood in both SCARKOs and ARKOs (24), and there is evidence for an important role of this growth factor in LC development (27). Quantitative PCR confirmed that PDGF-A gene expression per testis was reduced markedly in SCARKO mice on both d 20 and 50, whereas the expression of PDGF receptor-, when expressed per LC, was unaltered in SCARKO mice, except for a small increase on d 50 (Fig. 7). It was notable that the expression of mRNA for PDGF receptor- declined more than 10-fold from d 12 to 50 in both control and SCARKO mice (Fig. 7).
Discussion
The primary aim of the present study was to evaluate the effect of SC-selective knockout of the AR on the development and function of the adult LC population. To provide perspective for this aim, we also compared LC development in ARKO animals, because previous studies have shown that the development of the adult LC population is severely compromised in such animals (20). Our findings confirmed this, but show that in SCARKO males, LC numbers are not reduced as severely as in ARKOs, and there are some major differences in LC development and function between these two knockout models. Some of these differences may be explained by the fact that all testes are cryptorchid in ARKO/testicular feminized mice (tfm) mice (20, 21), whereas testes are scrotal in SCARKO mice (21). Our analysis of LC gene and protein expressions provide some potentially important insights into the pathways via which LC number and function may have been altered in SCARKO animals.
The major observed differences in LC development between ARKO and SCARKO mice were 1) LC number was reduced only by about 50% in SCARKO mice compared with about 75% in ARKO mice; and 2) LC size was normal or increased in SCARKO mice compared with controls, whereas LC size was consistently reduced in ARKO mice. The reduced LC numbers in ARKO animals are not explained by the concurrent cryptorchidism (20). At face value, the comparison of LC number in the two knockout models could indicate that the development of approximately 50% of adult LC number is dependent on androgen action via the AR in SC, whereas the development of normal adult LC size is not dependent on such action. In contrast, androgen action on testicular cell types other than SC is essential for the development of normal adult LC size and also accounts for approximately 25% of the final adult LC number. The latter finding would be consistent with evidence that androgen action on LC precursors plays a role in their differentiation into immature LC (14), and the present observation that the AR was mainly expressed in 3-HSD-immunonegative cells (i.e. not in immature and adult LC, which are both 3-HSD immunopositive) in control and SCARKO animals at 12 d of age is consistent with this. Although these straightforward interpretations may be partly true, there are numerous publications that attest to potentially important roles for locally produced hormones and growth factors in regulating LC development and function (3, 5, 7, 10), and the expression of at least two of these were altered dramatically in SCARKO and ARKO mice, namely, EST and PDGF-A, as discussed below.
EST, which is expressed in LC, plays a key role in the inactivation of biologically active estrogens. In EST knockout mice, there is reduced testicular expression of P450c17 (28), reduced testosterone production (28), and progressive LC hyperplasia/hypertrophy with aging (29). These changes can be recapitulated by administering estrogens to mice and these exert their effect via ER (30, 31). Conversely, knockout of aromatase, leading to ablation of estrogen action in the testis, leads to LC hyperplasia and/or hypertrophy and elevated testosterone levels (32). The dramatic down-regulation of EST expression in ARKO mice in the present studies and also reported in tfm mice in other studies (33) may therefore account for their reduced expression of P450c17, although the categorical absence of LC hyperplasia and hypertrophy in ARKO mice means that altered EST expression does not provide a complete explanation of their LC phenotype. In any case, the reduced EST expression in ARKO mice is probably explained by the abdominal location of the testes (20). In contrast to ARKO animals, the expression of EST in LC of adult SCARKO mice was increased 3-fold, a change that would be expected to increase the rate of estrogen catabolism within LC and thus reduce local estrogen action within the LC. ER knockout mice, in which estrogen action on LC via ER is prevented, exhibit increased testosterone production per LC in fetal (31) as well as in adult (30) life together with increased expression of P450scc and P450c17, changes also observed in adult LC of SCARKO mice in the present studies. Estrogen action in mice can also lead to LC hyperplasia and hypertrophy (33, 34), but whether underactivity of estrogens can lead to reduced LC number in mice, as found presently in SCARKO animals, is unknown. In view of the evidence just discussed, it seems reasonable to propose that some of the LC changes observed in both SCARKO and ARKO mice are attributable to altered EST expression and consequent alteration in local estrogen action within LC. Our observations of increased intensity of ER immunoexpression in LC nuclei of ARKO mice and reduced ER immunoexpression in LC nuclei of SCARKO mice are consistent with this interpretation, because estrogens are recognized to positively regulate ER immunoexpression (35). Although EST expression in the epididymis (36) and LC (33) is reported to be androgen dependent, the mechanism by which androgen action on SC is able to alter EST expression in LC in SCARKO mice remains to be defined, but one potential candidate for mediating SC-LC communication is PDGF-A, which is expressed predominantly in SC in the testis (27, 37).
In a recent study of SC development in SCARKO and ARKO mice, we reported that both models exhibited a major decrease in the expression of PDGF-A (24), a change confirmed in the present study for SCARKO animals. Because the reduction in PDGF-A expression occurred in both models, it is unlikely that the cryptorchidism in ARKOs (20, 21) is a major factor in this decrease. The reduction in PDGF-A could account for the reduced LC number in SCARKO and ARKO mice described in this study, because interference with PDGF-A action via knockout of its receptor results in impairment of development of fetal LC (38), and knockout of PDGF-A itself leads to gross impairment of adult LC development (27). In this regard, the relative degree of reduction in PDGF-A gene expression in SCARKO mice (present study and Ref. 24) and ARKO mice (24) parallels the magnitude of reduction in LC number.
Another SC-derived factor that has been shown to affect LC development in rodents is AMH. LC development is slightly compromised in AMH-deficient mice (39), whereas aberrantly high/continued expression of AMH beyond early puberty inhibits the development of adult LC generation and steroidogenesis (40, 41). However, our previous study indicated that the expression of AMH by SC showed a normal age-related decline in both SCARKO and ARKO mice (24), so AMH is unlikely to be involved in the LC changes in these animals.
In view of the key role that SC appear to play in regulating LC development and function, it has been suggested that an alteration of SC number, such as occurs in animals in which thyroid hormones are manipulated perinatally could lead to a parallel change in LC numbers (11, 18). In ARKO animals, the approximately 75% reduction in LC number shown here and previously (20) parallels a similar reduction in SC number (24). However, SC number remains largely unchanged from control values in SCARKO mice (24), yet these same animals exhibit an approximately 50% reduction in adult LC number. The present data together with other findings (19) therefore do not support the concept of a simple relationship between SC and LC numbers in the adult testis.
One of the most puzzling observations of the present studies was the apparent disparity between the major reduction in LC number in SCARKO and ARKO mice and the lack of change in blood levels of testosterone. Although the latter measurements showed huge between-animal variation, which may have obscured a reduction in testosterone levels in ARKO mice, three independent lines of evidence support the normality of testosterone levels in SCARKO animals; thus, serum testosterone levels were normal, serum LH levels were normal, and seminal vesicle weight was normal. The ability of LC in SCARKO mice to maintain normal testosterone levels despite an approximately 50% reduction in LC number and no increase in LH levels can only indicate that testosterone production per LC must be increased, perhaps doubled, in young adult SCARKO animals. Our observations of increased LC cytoplasmic volume in adulthood, increased volume of lipid droplets and mitochondria, and increased expression per LC of several of the key steroidogenic enzymes (P450ssc, 3-HSD, and P450c17) in adult SCARKO animals are consistent with this interpretation. The finding that testosterone levels are normal in SCARKO mice means that, physiologically, androgen-dependent processes outside the SC may be entirely normal in these animals. However, the fact that major adaptive changes have to occur in LC function/gene expression in SCARKO mice to achieve normal testosterone levels despite the halving of LC number, emphasizes the importance of a thorough analysis of testicular cell development in transgenic models, because there may be secondary consequences of these adaptive changes (e.g. altered estrogen action within the testis).
Although direct measurement of serum testosterone levels did not show any significant reduction in ARKO mice, our suspicion is that a reduction was masked by the high variability in control values, because earlier studies of tfm and ARKO male mice have reported up to 80% reductions in testosterone levels (42, 43). Because there were no seminal vesicles in ARKO mice, and serum LH levels were grossly elevated due to lack of viable androgen negative feedback, there was no independent measure of the normality of testosterone levels in ARKO animals. In fact, the marked reduction in LC cytoplasmic volume and in the expression of P450c17 per LC in ARKO mice would fit more readily with reduced testosterone production; in this regard, it is possible that the reduction in P450c17 expression per LC may result from the almost complete suppression of EST expression in ARKO animals, because a similar reduction in P450c17 expression was observed in EST null mice (28); the reduced EST expression in ARKO mice is probably a consequence of the cryptorchid position of the testes (20).
The present data also show that Insl3 expression in adult LC is almost obliterated in ARKO mice, but is normal in SCARKO animals; the massive reduction in ARKO mice is not explained by cryptorchid position of the testes (20), but is consistent with the view that fully adult LC do not differentiate in ARKO/tfm testes (20), although a role for androgens in regulating Insl3 expression is also possible. In this regard, the normal LC expression of Insl3 in SCARKO mice shows that androgen action on SC is not required for the expression of this protein in LC.
In conclusion, the present findings demonstrate that the absence of AR-mediated androgen action selectively on SC has important consequences for LC development, in particular for the development of normal LC numbers. In terms of LC function and LC-specific gene expression, there are major differences between SCARKO and ARKO animals, which may indicate that androgen action via cell types other than SC, probably including LC and their precursors, is important for the development of normal LC hormone (testosterone and Insl3) secretory function; however, this conclusion is tempered by the occurrence of cryptorchidism in ARKO mice, which itself is undoubtedly an important cause of changes in LC gene expression (20). Nevertheless, it is likely that estrogens (via altered EST expression) and PDGF-A may play some part in the LC changes, in SCARKO animals at least, although it is likely that factors other than these are also involved. The altered LC function in SCARKO mice, in particular the likelihood of reduced estrogen levels intratesticularly due to elevation of EST expression, must be taken into account when using this model for investigation of the regulation of spermatogenesis.
Acknowledgments
We thank Arantza Esnal, Hilde Geeraerts, Ludo Deboel, and Adriano Moreira for their excellent technical assistance.
Footnotes
This work was supported by Concerted Research Action (Research Fund, Katholieke Universiteit Leuven), Fund for Scientific Research Flanders (Belgium), Medical Research Council (United Kingdom), Contract QLK4-CT-2002–00603 (EDEN) from the European Union, and a Wellcome International Research Training Fellowship (to N.A.).
1 K.D.G. and N.A. contributed equally to this study.
Abbreviations: AMH, Anti-Mullerian hormone; AR, androgen receptor; ARKO, complete androgen receptor knockout; Cre recombinase, cyclization recombination recombinase; EST, estrogen sulfotransferase; 3-HSD, 3-hydroxysteroid dehydrogenase; Insl3, insulin-like factor-3; LC, Leydig cell; NGS, normal goat serum; NRS, normal rabbit serum; NSS, normal swine serum; PDGF-A, platelet-derived growth factor-A; P450scc, P450 cholesterol side chain cleavage enzyme; P450c17 17-hydroxylase/C17–20lyase; PGK, phosphoglycerate kinase-1; SC, Sertoli cell; SCARKO, Sertoli cell-selective androgen receptor knockout; TBS, Tris-buffered saline; tfm, testicular feminized mice; WT, wild type.
References
Sharpe RM 1994 Regulation of spermatogenesis. In: Knobil E, Neill JD, eds. The physiology of reproduction. 2nd ed. New York: Raven Press; 1363–1434
Sharpe RM 2005 Sertoli cell endocrinology and signal transduction: androgen regulation. In: Griswold M, Skinner M, eds. Sertoli cell biology. 1st ed. San Diego: Academic Press; 199–216
Ge R-S, Shan L-X, Hardy MP 1996 Pubertal development of Leydig cells. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 159–172
Ge RS, Hardy MP 1998 Variation in the end products of androgen biosynthesis and metabolism during postnatal differentiation of rat Leydig cells. Endocrinology 139:3787–3795
Mendis-Handagma SM, Ariyartne HB 2001 Differentiation of the adult Leydig cell population in the postnatal testis. Biol Reprod 65:660–671
Davidoff MS, Middendorf R, Enikolopov G, Riethmacher D, Holsetin AF, Muller D 2004 Progenitor cells of the testosterone-producing Leydig cells revealed. J Cell Biol 167:935–944
Haider SG 2004 Cell biology of Leydig cells in the testis. Int Rev Cytol 233:181–241
Teerds KJ 1996 Regeneration of Leydig cells after depletion by EDS: a model for postnatal Leydig cell renewal. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 203–219
Sharpe RM 1993 Experimental evidence for Sertoli cell-germ cell and Sertoli cell-Leydig cell interactions. In: Russell LD, Griswold MD, eds. The Sertoli cell. 1st ed. Clearwater: Cache River Press; 391–419
Saez JM, Lejeune H 1996 Regulation of Leydig cell functions by hormones and growth factors other than LH and IGF-1. In: Payne AH, Hardy MP, Russell LD, eds. The Leydig cell. 1st ed. Clearwater: Cache River Press; 383–406
Hardy MP, Sharma RS, Arambepola NK, Sottas CM, Russell LD, Bunick D, Hess RA, Cooke PS 1996 Increased proliferation of Leydig cells induced by neonatal hypothyroidism in the rat. J Androl 17:231–238
Teerds KJ, de Rooij DG, de Jong FH, van Haaster LH 1998 Development of the adult type Leydig cell population in the rat is affected by neonatal thyroid hormone levels. Biol Reprod 59:344–350
Ariyaratne HB, Mills N, Mason JI, Mendis-Handagama SM 2000 Effects of thyroid hormone on Leydig cell regeneration in the adult rat following ethane dimethane sulfonate treatment. Biol Reprod 63:1115–1123
Hardy MP, Kelce WR, Klinefelter GR, Ewing LL 1990 Differentiation of Leydig cell precursors in vitro: a role for androgen. Endocrinology 127:488–490
Abney TO, Myers RB 1991 17-Estradiol inhibition of Leydig cell regeneration in the ethane dimethylsulfonate-treated mature rat. J Androl 12:295–304
Abney TO 1999 The potential roles of estrogens in regulating Leydig cell development and function: a review. Steroids 64:610–617
Verhoeven G, Cailleau J 1990 Influence of co-culture with Sertoli cells on steroidogenesis in immature rat Leydig cells. Mol Cell Endocrinol 71:239–251
Hardy MP, Kirby JD, Hess RA, Cooke PS 1993 Leydig cells increase their numbers but decline in steroidogenic function in the adult rat after neonatal hypothyroidism. Endocrinology 132:2417–2426
Sharpe RM, Rivas A, Walker M, McKinnell C, Fisher JS 2003 Effect of neonatal treatment of rats with potent or weak (environmental) oestrogens, or with a GnRH antagonist, on Leydig cell development and function through puberty into adulthood. Int J Androl 26:26–36
O’Shaughnessy PJ, Johnston H, Willerton L, Baker PJ 2002 failure of normal adult Leydig cell development in androgen-receptor-deficient mice. J Cell Sci 115:3491–3496
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
Tan K, De Gendt K, Atanassova N, Sharpe RM, Saunders PTK, Denolet E, Verhoeven G 2005 The role of androgens in Sertoli cell proliferation and functional maturation: studies in mice with total (ARKO) or Sertoli cell-selective (SCARKO) ablation of the androgen receptor. Endocrinology 146:2674–2683
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
Frana LR, Ye S-J, Ying L, Sandberg M, Russell LD 1995 Morphometry of rat germ cells during spermatogenesis. Anat Rec 241:181–204
Gnessi L, Bascianai 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 149:1019–1025
Tong MH, Christenson LK, Song WC 2004 Aberrant cholesterol transport and impaired steroidogenesis in Leydig cells lacking estrogen sulfotransferase. Endocrinology 145:2487–2497
Quian YM, Sun XJ, Tong MH, Li XP, Richa J, Song WC 2001 Targeted disruption of the mouse estrogen sulfotransferase gene reveals a role of estrogen metabolism in intracrine and paracrine estrogen regulation. Endocrinology 142:5342–5350
Akingbemi BT, ge R, Rosenfeld CS, Newton LG, Hardy DO, Catterall JF, Lubahn DB, Korach KS, Hardy MP 2003 Estrogen receptor- gene deficiency enhances androgen biosynthesis in the mouse Leydig cell. Endocrinology 144:84–93
Delbes G, Levacher C, Duquenne C, Racine C, Pakarinen P, Habert R 2005 Endogenous estrogens inhibit mouse fetal Leydig cell development via estrogen receptor . Endocrinology 146:2454–2461
Robertson KM, O’Donnell L, Jones ME, Meachme SJ, Boon WC, Fisher CR, Graves KH, McLachlan RI, Simpson ER 1999 Impairment of spermatogenesis in mice lacking a functional aromatase (cyp 19) gene. Proc Natl Acad Sci USA 96:7986–7991
Quian YM, Song WC 1999 Regulation of estrogen sulfotransferase expression in Leydig cells by adenosine 3',5'-monophsophate and androgen. Endocrinology 140:1048–1053
Cook J, Klinefelter GR, Hardisty JF, Sharpe RM, Foster PMD 1999 Rodent Leydig cell tumorigenesis: a review of the physiology, pathology, mechanisms and relevance to humans. Crit Rev Toxicol 29:169–261
Kos M, Reid G, Denger S, Gannon F 2001 Minireview: genomic organization of the human ER gene promoter region. Mol Endocrinol 15:2057–2063
Tong MH, Song WC 2002 Estrogen sulfotransferase: discrete and androgen-dependent expression in the male reproductive tract and demonstration of an in vivo function in the mouse epididymis. Endocrinology 143:3144–3151
Loveland KL, Zlatic K, Stein-Oakley A, Risbridger G, de Kretser 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
Brennan J, Tilmann C, Capel B 2003 PDGFR- mediates testis cord organization and fetal Leydig cell development in the XY gonad. Genes Dev 17:800–810
Wu X, Arumugam R, Baker SP, Lee MM 2005 Pubertal and adult Leydig cell function in Mullerian inhibiting substance-deficient mice. Endocrinology 146:589–595
Fynn-Thompson E, Cheng H, Teixeira J 2003 Inhibition of steroidogenesis in Leydig cells by Mullerian-inhibiting substance. Mol Cell Endocrinol 211:99–104
Salva A, Hardy MP, Wu XF, Sottas CM, MacLaughlin DT, Donahoe PK, Lee MM 2004 Mullerian-inhibiting substance inhibits rat Leydig cell regeneration after ethylene dimethanesulphonate ablation. Biol Reprod 70:600–607
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
Jones RD, Pugh PJ, Hall J, Channer KS, Jones TH 2003 Altered circulating hormone levels, endothelial function and vascular reactivity in the testicular feminized mouse. Eur J Endocrinol 148:111–120(Karel De Gendt1, Nina Ata)