|Home | About | Journals | Submit | Contact Us | Français|
Proliferating cells express cyclins, cell cycle regulatory proteins that regulate the activity of cyclin-dependent kinases (CDKs). The actions of CDKs are regulated by specific inhibitors, the CDK inhibitors (CDKIs), which are comprised of the Cip/Kip and INK4 families. Expression of the Cip/Kip CDKI 1B (Cdkn1b, encoding protein CDKN1B, also called p27kip1) in developing Leydig cells (LCs) has been reported, but the function of CDKN1B in LCs is unclear. The goal of the present study was to determine the effects of CDKN1B on LC proliferation and steroidogenesis by examining these parameters in Cdkn1b knockout (Cdkn1b−/−) mice. LC proliferation was measured by bromodeoxyuridine incorporation. Testicular testosterone levels, mRNA levels, and enzyme activities of steroidogenic enzymes were compared in Cdkn1b−/− and Cdkn1b+/+ mice. The labeling index of LCs in Cdkn1b−/− mice was 1.5% ± 0.2%, almost 7-fold higher than 0.2% ± 0.08% (P < 0.001) in the Cdkn1b+/+ control mice. LC number per testis in Cdkn1b−/− mice was 2-fold that seen in the Cdkn1b+/+ control mice. However, testicular testosterone levels, mRNA levels of steroidogenic acute regulatory protein (Star), cholesterol side-chain cleavage enzyme (Cyp11a1), and 3beta-hydroxtsteroid dehydrogenase 6 (Hsd3b6), and their respective proteins, were significantly lower in Cdkn1b−/− mice. We conclude that deficiency of CDKN1B increased LC proliferation, but decreased steroidogenesis. Thus, CDKN1B is an important regulator of LC development and function.
Cyclin-dependent kinase inhibitor (CDKI) 1B (CDKN1B, also designated as p27kip1) is a member of the Cip/Kip family of CDKIs that regulates the activity of cyclins and cyclin-dependent kinase complexes, and thus serves to prevent progression of the cell cycle from the G1 phase to the S phase; as a result of CDKN1B gene (Cdkn1b) expression, cell division slows or stops . Experimentally, increased expression of Cdkn1b in oligodendrocyte precursor cells results in low cell proliferation , and decreased expression of Cdkn1b using anti-sense technology leads to an increased number of BALB/c-3T3 fibroblasts in the cell cycle .
In the seminiferous tubule of the testis, CDKN1B is found in Sertoli cells, but not in germ cells. Adult levels of CDKN1B are high in the testis and inversely correlated with rates of Sertoli cell proliferation during development [4, 5]. Levels of CDKN1B are low in rapidly proliferating neonatal rat Sertoli cells, and high in postmitotic adult Sertoli cells . Levels of CDKN1B are also low in rapidly proliferating Sertoli cell tumors , but high in adjacent normal Sertoli cells that are not proliferating. Hypothyroid mice have low levels of Sertoli cell CDKN1B, and this results in increased proliferation and delayed maturation in these cells . Overall, these findings show that CDKN1B is an inhibitor of Sertoli cell proliferation, and may be involved in overall Sertoli cell maturation .
In the testis, CDKN1B is also present in Leydig cells (LCs) , the endocrine cells that produce androgen. When neonatal rats are made transiently hypothyroid, these rats have higher numbers of LCs at adulthood than littermate controls , suggesting that thyroid hormone is also an important regulator of LC development and the adult LC population. Therefore, like Sertoli cells in hypothyroid mice, the increased LC numbers could be involved in the reduced level of CDKN1B, although the functional role of CDKN1B in LC development and the establishment of the adult LC population have not been previously addressed. In the present study, we have used Cdkn1b knockout (Cdkn1b−/−) mice to test the hypothesis that CDKN1B may play a critical role in LC development. A lack of CDKN1B resulted in increased LC proliferation and large increases in the adult LC population. Conversely, LC steroidogenic function decreased. Our results indicate that CDKN1B is an important regulator of LC mitogenesis and maturation.
Cdkn1b−/− and Cdkn1b+/+ mice (background strain C57BL/6) were a gift from Dr. Hiro Kiyokawa (Feinberg School of Medicine, Northwestern University, Chicago, IL). All mice were housed and bred in the College of Veterinary Medicine at the University of Illinois. Mice were housed five per cage under controlled environmental conditions (temperature 22 ± 2°C; 12L:12D, with lights on from 0600 to 1800 h) and given water and a standard rodent diet ad libitum until being killed at 90 to 120 days of age. All animals were handled to become adapted for at least 3 days prior to the beginning of the experiment. All animal experiments were approved by the Institutional Animal Care and Use Committee of the University of Illinois and conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
At 2 h prior to being killed, all animals received an i.p. injection of bromodeoxyuridine (BrdU; 40 μg/g body weight [BW]; Boehringer Mannheim, GmbH, Mannheim, Germany) in order to label dividing cells. The animals were killed under deep anesthesia by i.p. injection of sodium pentobarbital (25 mg/100 g BW; Abbott Laboratories, North Chicago, IL), and one testis from each animal was removed and frozen in liquid nitrogen for subsequent study of testicular T levels and enzyme activity, as well as Western blot analysis. The contralateral testis from each animal was fixed by whole-body perfusion through the left ventricle of the heart with Bouin solution . After dehydration in ethanol and xylene, the testes were embedded in paraffin. Transverse sections were cut at a thickness of 5 μm and mounted on glass slides (Fisher Scientific Co., Fairlawn, NJ) for immunohistochemical analysis.
The LC was identified by staining for 3β-hydroxysteroid dehydrogenase (3β-HSD, HSD3B6). In brief, the avidin-biotin immunostaining was performed using kits from Vector Laboratories (Burlingame, CA) according to the manufacturer's instructions. Antigen retrieval was carried out by microwave irradiation for 10 min in 10 mM (pH 6.0) citrate buffer, and endogenous peroxidase was blocked with 0.5% H2O2 in methanol for 30 min. The sections were then incubated with a monoclonal anti-BrdU antibody (RPN 202; Amersham Biosciences, Little Chalfont, UK) for 30 min at room temperature. The antibody bound to the nuclei was visualized with diaminobenzidine (catalog no. SK-4100; Vector Laboratories) and the labeled nuclei were stained black by adding a nickel solution to the chromogen. After washing, the sections were double labeled by incubation with an HSD3B6 polyclonal antibody diluted 1:300 (provided by Dr. Van Luu-The, Laval University, Laval, Quebec, Canada) for 1 h at room temperature. The antibody-antigen complexes were visualized with diaminobenzidine alone, resulting in brown cytoplasmic staining in positively labeled LCs. The sections were counterstained with Mayer hematoxylin, dehydrated in graded concentrations of alcohol, and cover slipped with Permount resin (Fisher Scientific). In control experiments, sections were incubated with non-immune rabbit IgG (HSD3B6) or mouse IgG (BrdU) using the same working dilution as the primary antibody.
LC number was determined by stereology using the fractionator method . In brief, 15–20 randomly selected fields in each of three nonadjacent sections per testis were captured using a Nikon Eclipse E800 microscope. The total number of cells was calculated by multiplying the number of cells counted in a known fraction of the testis by the inverse of the sampling probability.
Image analysis was performed as previously described . Dividing cells, as marked by BrdU incorporation, had black nuclei. Labeling index (LI) was calculated as BrdU-labeled cells/total cells for LCs. At least 200 cells were counted.
Total RNAs were extracted from Cdkn1b−/− and Cdkn1b+/+ mouse whole testes in Trizol (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The 12 genes analyzed and the primers are listed in Table 1. The RNA was reverse transcribed using random hexamers and MMLV reverse transcriptase (Promega Biosciences, Inc., San Luis Obispo, CA). The cDNA synthesis step and real-time PCR were performed as previously described . The RNA level for the housekeeping gene, ribosomal protein S16 (Rps16), was also assayed in all samples as an internal control. Messenger RNA measurements were subjected to a robust global normalization algorithm. All control crossing threshold values were corrected by the median difference in all samples from Rps16. Messenger RNA levels were calculated relative to Rps16 mRNA. Real-time PCR was carried out in a 25-μl volume with SYBER green. Reactions were carried out and fluorescence was detected on a GeneAmp 5700 system (PE Applied Biosystems).
Testicular testosterone (T) concentration was measured using the method of Knorr et al. . In brief, testes were homogenized in 5 ml of 70% methanol using a glass homogenizer. The homogenates were transferred to 15-ml screw cap test tubes. Tracer steroid (1000 counts per minute [cpm] of tritiated T) was added to the homogenate to correct for recovery. The homogenates were left standing overnight at room temperature. The tubes were centrifuged at 3000 × g, and the supernatants were aspirated and dried under nitrogen to remove the methanol. The water layer was extracted twice with HPLC-grade diethylether. The ether extracts were then resuspended in 400 μl of TBSG buffer (50 mM Tris, 1 mM MgCl2, 150 mM NaCl, 0.1% [w/v] gelatin, pH 7.8), and 100 μl was removed for measurement of recovery (the cpm value in 100 μl × 4 ÷ 1000). The remaining 300 μl was used to measure testicular T levels with a tritium-based radioimmunoassay, as previously described . Values for interassay variation were 5%; the sensitivity of the assay was 10 pg/ml.
The testes were homogenized in 1 ml ice-cold 0.1 M PBS (pH 7.2) containing 0.25 M sucrose. Supernatants were collected by centrifugation at 700 × g for 30 min. Supernatants were used to measure enzyme activities of steroidogenic enzymes, HSD3B6 (a T biosynthetic enzyme that catalyzes pregnenolone to progesterone), P450-dependent 17α-hydroxylase/20-lyase (CYP17A1, a T biosynthetic enzyme that catalyzes progesterone into 17α-hydroxyprogesterone and further into androstenedione), and 17β-hydroxysteroid dehydrogenase (HSD17B3, the last-step T biosynthetic enzyme that catalyzes androstenedione into T). For the assay of protein levels of acute regulatory protein (STAR, a rate-limiting protein to transfer cholesterol from cytosol into mitochondria) and mitochondrial enzyme cholesterol side-chain cleavage enzyme (CYP11A1, a T biosynthetic enzyme to catalyze the first step from cholesterol into pregnenolone), the resultant supernatant was centrifuged at 9000 × g for 30 min to isolate mitochondria. The protein concentrations were determined using a kit (no. 500-0006; Bio-Rad Laboratories, Inc., Hercules, CA) with bovine serum albumin (BSA) as a standard.
Mitochondrial protein (25 μg) was boiled in equal volumes of sample loading buffer, a Tris-Cl buffer (pH 6.8) containing 20% glycerol, 5% SDS, 3.1% dithiothreitol, and 0.001% bromophenol blue. Protein samples were electrophoresed on 10% polyacrylamide gels containing SDS. Proteins were electrophoretically transferred onto nitrocellulose membranes, and, after 30-min exposure to 10% nonfat milk to block nonspecific binding, the membranes were incubated with a STAR polyclonal rabbit antibody (1:1000; kindly donated by Dr. D.M. Stocco, Texas Tech Health Sciences Center, Lubbock, TX) or a CYP11A1 polyclonal rabbit antibody (catalog no. RDI-p450sccabr; RDI Research Diagnostics, Inc., Flanders, NJ). The membranes were then washed and incubated with a 1:5000 dilution of second antibody that was conjugated to horseradish peroxidase (HRP) (anti-rabbit Ig, HRP-linked whole antibody produced in donkey; Amersham Biosciences Corp., Piscataway, NJ). The washing step was repeated, and immunoreactive bands were visualized with an enhanced chemiluminescence kit (ECL; Amersham, Arlington Heights, IL). The antibody in the membrane was then stripped away and probed again by an antibody against β-actin (ACTB; 1:1000, no. A2228; Sigma). The second anti-mouse Ig, HRP-linked whole antibody produced in sheep (Amersham Biosciences, Piscataway, NJ) was used and the actin band was visualized by chemiluminescence. Protein levels were measured by densitometry of the films and normalized to ACTB.
Activities of the androgen biosynthetic enzymes, HSD3B6, CYP17A1, and HSD17B3 were determined by thin layer chromatography. In brief, the testes were homogenized in 1 ml 0.1 M PBS (pH 7.2) containing 2.5 M sucrose, and the testicular proteins were extracted. The protein concentrations were determined using a kit (Bio-Rad Laboratories, Inc.) with BSA as a standard. The reaction mixtures (total volume of 250 μl) containing 25 μg protein, 400 μM cofactors, and 5 μM radiolabeled steroid substrates (pregnenolone for HSD3B6, progesterone for CYP17A1, and androstenedione for HSD17B3) were incubated in a shaking water bath at 37°C for 20 min. The reaction rates were linear throughout the incubation period in the presence of 5 μM substrate. The steroids were extracted from the reaction mixture with 2 ml of ice-cold ether, and the organic layer was evaporated under nitrogen gas. The extract of steroids was resuspended in 70 μl ether and then spotted on thin layer plates (Baker-flex, Phillipsburg, NJ). The steroids were separated chromatographically in chloroform and methanol (97/3 [v/v]). The radioactivity was measured with a scanning radiometer (System 200/AC3000; Bioscan, Inc., Washington, DC). The conversion of substrates to products (progesterone for HSD3B6, androstenedione for CYP17A1, and T for HSD17B3) was calculated as a percentage of the total radioactivity found in the product.
Data are expressed as mean ± SEM. Significant differences between Cdkn1b−/− and Cdkn1b+/+ groups were identified using statistical analysis software (GraphPad 4.0; GraphPad, San Diego, CA). When a significant difference was observed between groups, a Student t-test was employed to separate the means. Differences were regarded as significant at a P value less than 0.05.
The LC was identified by immunohistochemical staining for HSD3B6 (brown cytosolic staining, Fig. 1A). Proliferating cells were identified by BrdU incorporation (black nuclear staining, Fig. 1B). LI was calculated to compare rates of proliferation. The Cdkn1b−/− group had a significantly higher LI (Fig. 1C) and in LC number per testis was doubled in Cdkn1b−/− compared to Cdkn1b+/+ mice (Fig. 1D). Although testis weight was increased in Cdkn1b−/− mice (by 35%, Fig. 1E), the adjusted LC density per 100 mg testis (Fig. 1F) was also significantly higher (P < 0.01) in Cdkn1b−/− compared to Cdkn1b+/+ mice. However, LC diameters in Cdkn1b−/− and Cdkn1b+/+ mice were similar (Fig. 1G).
Testicular T concentrations were significantly lower in Cdkn1b−/− mice when expressed either as ng T per mg testis (Fig. 2A) or as μg T per testis (Fig. 2B). These results suggest that the T production of LCs in Cdkn1b−/− testis was lower.
The primary function of LCs is to synthesize and secrete androgenic steroids under the stimulation of luteinizing hormone (LH). As serum LH levels were normal in Cdkn1b−/− mice , we measured mRNA levels of LC-specific genes. The mRNA levels of Lhcgr (encoding LH receptor), Star (encoding STAR), Cyp11a1 (encoding CYP11A1) and Hsd3b6 (encoding 3β-HSD) were significantly lower in Cdkn1b−/− mice (Fig. 3). However, mRNA levels of other steroidogenic enzymes including Cyp17a1 (encoding CYP17A1) and Hsd17b3 (encoding HSD17B3) were similar to controls. Insl3, which is a LC specific gene, was not affected, and thus we infer that the effects were confined to the steroidogenic pathway . We found that mRNA levels of Sertoli cell specific genes (such as Fshr and Kitl), spermatogonia specific gene Kit and the cell proliferation marker Pcna were similar in Cdkn1b−/− and Cdkn1b+/+ mice. Thus, although overall Sertoli cell number is increased in Cdkn1b−/− mice , relative concentrations of mRNA for Sertoli cell proteins appear normal.
Steroidogenic capacities of LCs were evaluated by measuring the amounts of STAR and various steroidogenic enzymes activities in the testis. Significantly reduced amounts of STAR and CYP11A1 protein levels were detected by Western blot analysis (Fig. 4). HSD3B6 activity was significantly lower in Cdkn1b−/− mice, while two other enzymes involved in T biosynthesis (CYP17A1 and HSD17B3) were not different in Cdkn1b−/− and Cdkn1b+/+ mice (Fig. 5).
In the present study we found increased numbers of LCs and testis weight, but lower testicular T in Cdkn1b−/− mice. These findings indicate that CDKN1B is a major regulator of both LC development and adult LC function. LCs in the adult testis rarely undergo mitosis . Consistent with these previous observations, in the present study 0.2% of LCs in Cdkn1b+/+ mice were labeled by BrdU. The LI of LCs was 7-fold higher in Cdkn1b−/− mice, indicating that the proliferation rate of LCs was significantly increased in Cdkn1b−/− mice compared to Cdkn1b+/+ controls. Our results support the hypothesis that CDKN1B is a major regulator of adult LC proliferation; CDKN1B normally serves to inhibit LC proliferation and in the absence of CDKN1B, mitogenesis of adult LCs is increased several fold. A lack of CDKN1B may inhibit one or more steps in the maturational sequence that converts a proliferative mitogenic LC into a steroidogenic cell that is not actively proliferating, and this allows a larger fraction of LCs to remain in the proliferative pool even into adulthood.
LCs develop from a precursor stem cell that has been recently characterized [19, 20]. These stem cells give rise to undifferentiated LC precursor cells that then pass through several developmental steps on their path to obtaining the adult LC phenotype and steroidogenic capacity. All of these early stages of LC differentiation, such as the stem cells, precursor cells and immature LCs, also divide during their developmental progression to the adult LCs . The present results showing a critical role of CDKN1B in adult LC proliferation suggest that CDKN1B could also be involved in the proliferation of some or all of these earlier stages of LC development, although exactly how the absence of Cdkn1b affects proliferation in the various stages of LC development was not directly evaluated in the present study.
Our morphometric data indicate that there is a two-fold higher LC population in Cdkn1b−/− mice compared to Cdkn1b+/+ mice. Previous work has indicated that adult testis weight is increased in Cdkn1b−/− mice , and that these increases in testis weight are accompanied by increases in adult Sertoli cell numbers and sperm production . The 30% increase in testis weight observed in the present study is consistent with similar increases reported previously in the Cdkn1b−/− mice, and the present results further indicate that LC hyperplasia is also a contributor to the overall testis weight increase in these animals. Therefore, the increased testicular size in Cdkn1b−/− mice may reflect more global increases in the population of its constituent cells, rather than a selective increase in only the Sertoli and spermatogenic lineages. The doubling of the LC population exceeds the overall increase of 30% in testis weight, indicating that LCs may be selectively more affected than other testicular cell types (Fig. 1).
It is interesting to note that despite the 7-fold increase in LC proliferation and the 1.5% LI during adulthood, the adult LC population in Cdkn1b−/− mice is only doubled, while an even larger increase would be predicted if the 1.5% LI seen during adulthood had been maintained during the entire lifespan of the mice. This phenomenon has been seen previously in the prostate, where epithelial proliferation in the Cdkn1b−/− prostate is increased 2- to almost 4-fold in the various prostatic lobes, but these animals do not have an increase in prostatic weight, due to a concomitant increase in epithelial cell apoptosis . A similar mechanism may exist in the Cdkn1b−/− LCs and account for the less than expected increase in adult population despite a high LI, but this remains to be directly addressed experimentally.
The importance of CDKN1B in cell cycle control has been shown by the observations of multi-organ hyperplasia, gigantism and tumor tendencies in Cdkn1b−/− mice [4, 21, 22]. Strikingly, Cdkn1b heterozygous mice are haplo-insufficient for tumor suppression, thus even a 50% reduction of Cdkn1b expression has serious consequences for cancer susceptibility in these mice . An increased rate of LC tumors was not found in Cdkn1b−/− mouse testis although there was significant LC hyperplasia.
In the Cdkn1b−/− mouse testis, testicular T concentration was reduced approximately 60% when expressed as T production per mg of testis. As indicated above, there is an increase in testicular size in these animals, but even when testicular T is expressed per testis, the amount of T present is still reduced by almost half, and this is significant considering the 2-fold increase in LCs. This decrease in T concentration appeared to result from impairments in multiple steps of the steroidogenic pathway. For example, we observed significant decreases in Lhcgr mRNA levels in Cdkn1b−/− testes whereas Fshr and Kitl (both found in Sertoli cells) mRNA levels were unchanged. We observed specific deficiencies in the LC steroidogenic pathway: Star, Cyp11a1, and Hsd3b6 mRNA levels and their respective protein levels or enzyme activities were significantly reduced. Most importantly, the protein level of STAR, which is the rate limiting step for ferrying cholesterol across the mitochondrial membrane and providing the starting material for the steroidogenic pathway, was significantly lower in the Cdkn1b−/− mouse testis. Therefore, the lower testicular T levels may largely be attributed to the decreased expression of Star. The low levels of Lhcgr mRNA might be one of causes of lower expression levels of Star and steroidogenic enzyme genes. The decrease of CYP11A1 protein level follows the change of its mRNA levels. Although we did not measure its activity, its protein level usually corresponds to its enzyme activity as shown in our previous studies [24, 25]. Interestingly, this decrease in steroidogenesis did not result from a global decrease in all steroidogenic enzymes, as the mRNA levels and enzyme activities of two T biosynthetic enzymes CYP17A1 and HSD17B3 did not change. Thus, the steroidogenic lesions in Cdkn1b−/− mice are selective, rather than universal. The decease of HSD3B6 enzyme activity is somewhat unexpected because this enzyme is normally largely constitutive in the LCs.
Pituitary gland or hypothalamic activities were probably not involved as the serum LH level was unchanged , indicating that the impaired steroidogenesis was the result of changes at a point downstream of LH. In further support of this, Insl3 mRNA levels, which are normally upregulated by serum LH levels , did not change in Cdkn1b−/− mice. Therefore, it is reasonable to assign a role for CDKN1B in LC differentiation, as it specifically affects several key points in LC cholesterol transporting and steroidogenesis. Thus, the role of Cdkn1b−/− in LC development appears to be consistent with that suggested by Bryja et al. , who postulated a dual role for CDKN1B as a classical regulator of cell cycle via inhibition of cyclin-dependent kinases, and as a participant, in conjunction with D-type cyclins, in the establishment and/or maintenance of differentiated status. T levels are low in Cdkn1b−/− mice (Fig. 2), but an elevation in LH levels was not seen in the previous study , indicating that there was a defect in hypothalamic-pituitary level in these mice. The developmental changes in the LCs are worthy of further investigation.
Sertoli cells have been reported to regulate the proliferation and differentiation of LCs (reviewed in ). For example, injections of follicle-stimulating hormone to animals with low or absent serum LH stimulates LC differentiation and steroidogenic activity, and Sertoli cell Ar knockout mice had reduced steroidogenic capacity (reviewed in ). Adult levels of CDKN1B are high in the testis and inversely correlated with rates of Sertoli cell proliferation during development [4, 5], and hypothyroid mice with low levels of Sertoli cell CDKN1B have increased proliferation and delayed maturation in these cells [7, 20]. Therefore, Cdkn1b knockout in Sertoli cells might also have an impact on LC proliferation and differentiation. However, determining which Sertoli cell factors will be affected requires further investigation. Apparently, the possible Sertoli cell factor, Kitl, did not change in Cdkn1b knockout mice compared with wild-type mice (Fig. 3).
In conclusion, the role of CDKN1B in the control of the cell cycle appears to be critical for LC proliferation, the establishment of the adult LC population, and the functional differentiation of LCs, as indicated by their expression of steroidogenic enzymes and subsequent T production. This suggests that the absence of CDKN1B may hold LCs in the cell cycle longer, and result in the LCs undergoing more extensive proliferation than in the controls. LCs normally express their full complement of steroidogenic enzymes and reach maximal T production as they come out of the cell cycle and attain full differentiation. As such, the present results indicate that the absence of CDKN1B may impair LC functional differentiation simply as a consequence of the retention of a greater proportion of LCs in the proliferative pool.
We thank Ming Pan for excellent technical support and Dianne O. Hardy for her critical comments. This paper is in memory of Dr. Matthew P. Hardy, who untimely passed away on November 4, 2007.
1Supported in part by National Institutes of Health grants RO1 HD050570, AG030598 to R.-S.G., and T32 HD 07028 to D.R.H., and by the Billie A. Field Endowment, University of Illinois, to P.S.C.