Leydig cell genes change their expression and association with polysomes in a stage-specific manner in the adult mouse testis

Abstract

Spermatogenesis in mammals occurs in a very highly organized manner within the seminiferous epithelium regulated by different cell types in the testis. Testosterone produced by Leydig cells regulates blood–testis barrier formation, meiosis, spermiogenesis, and spermiation. However, it is unknown whether Leydig cell function changes with the different stages of the seminiferous epithelium. This study utilized the WIN 18,446 and retinoic acid (RA) treatment regime combined with the RiboTag mouse methodology to synchronize male germ cell development and allow for the in vivo mapping of the Leydig cell translatome across the different stages of one cycle of the seminiferous epithelium. Using microarrays analysis, we identified 11 Leydig cell-enriched genes that were expressed in stage-specific manner such as the glucocorticoid synthesis and transport genes, Cyp21a1 and Serpina6. In addition, there were nine Leydig cell transcripts that change their association with polysomes in correlation with the different stages of the spermatogenic cycle including Egr1. Interestingly, the signal intensity of EGR1 and CYP21 varied among Leydig cells in the adult asynchronous testis. However, testosterone levels across the different stages of germ cell development did not cycle. These data show, for the first time, that Leydig cell gene expression changes in a stage-specific manner during the cycle of the seminiferous epithelium and indicate that a heterogeneous Leydig cell population exists in the adult mouse testis.

Summary Sentence

Synchronizing spermatogenesis utilizing WIN18,446/RA treatment combined with the RiboTag methodology revealed that Leydig cell gene expression events change across one spermatogenic cycle but does not alter testosterone levels

Introduction

Mammalian spermatogenesis is regulated via interactions between somatic and germ cells. Germ cell development occurs within the seminiferous tubules, composed of the germ cells intertwined between Sertoli cells and surrounded by the peritubular myoid (PTM) cells. The Leydig cells, located in the interstitium between tubules, secrete testosterone that diffuses into the seminiferous epithelium and signals through the androgen receptor (AR) in Sertoli cells, regulating processes such as the maintenance of the blood–testis barrier, meiosis, Sertoli-spermatid adhesion, and spermiation [1–6]. All of these events align to occur at a specific stage during spermatogenesis. However, it is unknown whether Leydig cell function or testosterone levels correlate with the different stages of the seminiferous epithelium, thereby contributing to the regulation of cyclic germ cell development.

In mice and rats, germ cell development has been classified into 12 and 14 different stages, respectively, based on the repetitive cellular associations that are seen along seminiferous tubules. Germ cells progress through the different stages along the tubules in a sequential and cyclic manner [7]. This complex organization of asynchronous spermatogenesis in the murine testis results in continuous sperm production. During asynchronous spermatogenesis many cellular functions and gene expression events occur in a stage-specific manner [8–12]. In the murine testis, AR expression and androgen action in Sertoli cells are stage-specific, playing a more active role during Stages VII and VIII. Ar mRNA is maximally expressed in Stage VII, while the AR protein is highly expressed in Stages VII and VIII [10, 13, 14]. Knockout of Ar in Sertoli cells or ablation of testosterone causes a blockage at the end of meiosis, resulting in decreased numbers of round spermatids and missing elongated spermatids [3, 4, 15]. Similarly, destruction of Leydig cells with ethane dimethanesulfonate impairs germ cell development by blocking meiosis specifically in tubules in Stages VII-XI [16]. In the murine testis, different levels of testosterone are required for different spermatogenesis processes. If testosterone levels fall below the threshold of 70 mM, spermatogenesis is disrupted starting in Stages VII and VIII [17, 18]. In rats, formation of elongated spermatids requires 12% of normal testosterone levels, while only 3% of normal testosterone levels is needed for the completion of meiosis [19]. However, it is unknown whether impairment of spermatogenesis in specific stages by ablation of androgen action is a direct consequence of alterations in cyclic AR expression, levels of testosterone, and/or Leydig cell function.

Several published studies have investigated testosterone levels across the different stages of the seminiferous epithelium and reached various conclusions. Parvinen et al. mechanically separated rat seminiferous tubules and classified them into different stages to measure testosterone levels across one cycle. This study reported that testosterone levels were highest at Stages VII-VIII [20]. While this method allowed for the study of each spermatogenic stage independently, the method used takes time and disrupts the Leydig cells, thereby likely altering measurement accuracy. To circumvent this, other studies have performed testosterone measurements following the induction of synchronized spermatogenesis by generating vitamin A deficient (VAD) animals, usually through diet, followed by injection of retinoids [21, 22]. Synchronized testes contain tubules with only one to three different stages of the seminiferous epithelium, making it easier to study specific events during the different stages. Contradictory to the results obtained by Parvinen et al., no difference in testosterone levels were detected across the different stages of the seminiferous epithelium following VAD/retinoid-induced synchronous spermatogenesis [21, 22]. However, while this method allows for stage enrichment, there are also drawbacks, including compromised animal health and incomplete restoration of spermatogenesis.

To overcome asynchronous spermatogenesis and the drawbacks associated with the VAD/retinoids protocol, our laboratory previously developed a protocol using a bisdichloroacetyldiamine (WIN 18,446) and retinoic acid (RA), the active metabolite of vitamin A, to synchronize spermatogenesis faster and without compromising animal health, allowing for the study of specific spermatogenic stages in healthy adult mouse testis tissue [12, 23, 24]. Using this method, our laboratory has successfully investigated cyclic changes of retinoid metabolizing enzymes and RA levels across one cycle of the seminiferous epithelium in a synchronized testis [12, 25, 26]. However, this method has yet to be employed as a means to investigate testosterone levels across the cycle of the seminiferous epithelium and whether events in the seminiferous epithelium are regulated by stage-specific functions of Leydig cells.

In this study, we utilized our novel WIN 18,446/RA treatment regime to synchronize male germ cell development [12, 23, 24] and RiboTag-positive/Cyp17iCre-positive male mice to investigate changes in Leydig cell gene expression across the cycle of the seminiferous epithelium [27, 28]. These two methodologies together have allowed us, for the first time, to map the Leydig cell “translatome” across the different stages of the spermatogenic cycle in vivo and investigate the Leydig cell-enriched cyclic genes and Leydig cell genes that change association with polysomes in a cyclic manner. In addition, we measured levels of testosterone across the different stages of cycle of the seminiferous epithelium in the adult mouse testis. Importantly, this study has identified Leydig cell genes that might be involved in paracrine regulation between Leydig cells and germ cell development within the seminiferous epithelium.

Materials and methods

Animals and tissues

All animal procedures were conducted according to the guidelines stated in the United States Public Health Service's Guide for the Care and Use of Laboratory Animals and approved by the WSU Institutional Animal Care and Use committee. Testosterone measurements were conducted using C57BL/6-129 mice. All other experiments were performed using double heterozygous RiboTag-positive/Cyp17icre-positive experimental male mice containing Leydig cells expressing a human influenza hemagglutinin (HA) tag under the ribosomal protein l22 (Rpl22) gene [27, 28]. The experimental animals were generated by crossing the RiboTag mouse line [27] to a steroidogenic cell-specific cre, Cyp17iCre [28]. The RiboTag mouse line was provided by Dr. Paul Ameiux from the University of Washington in Seattle and the Cyp17iCre mouse line by Dr. CheMyong Ko from the University of Kentucky. Mouse colonies used in this study were maintained in a temperature- and humidity-controlled environment with food and water provided ad libitum. The animals were euthanized by CO2 inhalation followed by cervical dislocation, and their testes dissected. For each animal, one testis was fixed in Bouins (catalog no. M7831-88; EMD Millipore; Billerica, MA) for 6 h at room temperature. The testis was then embedded in paraffin, sectioned at 4 μm with 20 μm between each cross-section, and placed on Superfrost Plus slides (catalog no. 12-550-15; Fisher Scientific; Hampton, NH) for histological and staging analyses. The second testis was snap-frozen on dry ice and immediately stored for microarray analyses or intratesticular testosterone measurements.

WIN 18,446 and RA treatment

Spermatogenesis was synchronized in experimental animals as previously described [12, 23, 24]. Two days postpartum (dpp) male mice were treated with 100 μg/g body weight WIN 18,446 for seven consecutive days. The following day, at 9 dpp, mice were injected with 200 μg all-trans RA (catalog no. R2625-50MG; Sigma-Aldrich; St. Louis, MO). For control animals, 2 dpp male mice were treated with 1% gum tragacanth for 7 days followed by a dimethyl sulfoxide injection. Experimental and control animals were then left to recover and euthanized 42–50 days after injection. Testes were dissected and used for microarray analyses, testosterone measurements, immunohistochemistry (IHC), or analyses of synchronized spermatogenesis.

Immunoprecipitation and RNA extraction

After WIN 18,446/RA treatment, testes from RiboTag-positive/Cyp17icre-positive experimental male mice from different time points were homogenized and immunoprecipitation (IP) was performed as previously described [27]. Briefly, one testis per time point was homogenized in 3% weight/volume of homogenization buffer. For the total testis RNA samples, RNA from 50 μl of homogenized testis was extracted using the Qiagen RNeasy Mini Kit (catalog no. 74104; Qiagen; Hilden, Germany). The rest of the homogenized testis was used for IP sampling. For the IP RNA samples, 5 μl of HA-tag antibody (catalog no. MMS-101R, BioLegend; San Diego, CA) coupled to 400 μl of Dynabeads Potrein G (catalog no. 1004D; Life Technologies; Carlsbad, CA) were added to the homogenized testis and incubated overnight at 4°C. The following day, the beads coupled to anti-HA antibody and bound to the HA-tagged polysomes were washed in high salt buffer and RNA extracted using the Qiagen RNeasy Mini Kit. The extracted IP and total RNA samples were quantified using a Nano Drop 1000 Spectrophotometer. The RNA samples were then stored at –80°C until use.

Immunohistochemistry

IHC was performed in testis sections using the mouse anti-HA (1:1000; catalog no. MMS-101R; BioLegend; San Diego, CA) and rabbit anti-stimulated by retinoic acid 8 (STRA8) (1:1000; made in house) (Supplemental table 1). Similar procedures were followed as described previously [29]. Briefly, slides were washed in xylene and a graded ethanol series (100%, 95%, and 75%) followed by antigen retrieval in 10 mM citrate buffer (pH 6) and incubated at 100°C for 7.5 min. Then, testis sections were treated for 5 min with 3% hydrogen peroxide. Blocking solution of 5% goat serum/0.1% bovine serum albumin/phosphate-buffered saline (PBS) (136 mM NaCl/2.7 mM KCl/10.1 mM Na₂HPO₄/1.8 mM KH₂PO₄) was applied for 30 min at room temperature, then the sections were incubated with primary antibody diluted in blocking solution overnight at room temperature. The next day, slides were washed and incubated with biotinylated goat anti-mouse (2 μg/ml; catalog no. B7151; Sigma-Aldrich; St. Louis, MO) diluted in blocking solution or biotinylated goat anti-rabbit (Histostain SP Bulk kit; ref. 956143B; Vector Laboratories; Burlingame, CA) followed by horseradish peroxidase (HRP) streptavidin (ref. 5A-5704; Vector Laboratories; Burlingame, CA). Slides were developed with metal enhanced 3,3΄-diaminobenzidine tetrahydrochloride substrate kit (catalog no. 34065; Thermo Fisher Scientific, Waltham, MA) according to the manufacturer's instructions and stained with Harris Hematoxylin solution (1:3 dilution; ref. HHS32-1L; Sigma-Aldrich; St. Louis, MO) for 10 s. The slides were then washed by a graded series of ethanol and xylene and mounted with DPX (360294H; VWR International; Radnor, PA). For analyses, a minimum of three technical duplicates and biological replicates were performed.

Immunofluorescence

Immunofluorescence using the mouse anti-HA (1:1000; catalog no. MMS-101R; BioLegend; San Diego, CA), goat anti-3-Beta-hydroxysteroid dehydrogenase (3βHSD) (2 μg/ml; sc-30820; Santa Cruz Biotechnology; Dallas, TX), rabbit anti-early growth response 1 (EGR1) (1:100; orb34304; Biorbyt, San Francisco, CA), rabbit anti-cytochrome P450, family 21(CYP21) (1:100; orb13362; Biorbyt; San Francisco, CA), rabbit anti-extended synaptotagmin-like protein 3 (ESYT3) (1:100; HPA039200; Sigma-Aldrich; St. Louis, MO) or rabbit anti-STRA8 (1:1000; made in house) was performed (Supplemental table 1). Slides with paraffin tissue sections were washed twice in xylene followed by a graded series of ethanol washes. Then, antigen retrieval was performed with 10 mM of citrate buffer (pH 6.0) incubated at 100°C for 7.5 min. Slides were then washed with 1× PBS, incubated with blocking solution (5% goat serum or donkey serum/0.1% bovine serum albumin/PBS) for 30 min at room temperature, followed by primary antibody diluted in blocking solution and incubated in a humid chamber at room temperature overnight. The next day, slides were washed with 1× PBS and incubated for 1 h at room temperature with secondary antibodies diluted in blocking solution (Alexa Fluor 568 donkey anti-goat IgG (1:1000; A11057; Invitrogen; Carlsbad, CA), Alexa Fluor 488 donkey anti-rabbit IgG (1:1000; A21206; Invitrogen; Carlsbad, CA), or Alexa Fluor 488 donkey anti-mouse IgG (1:1000; A21202; Carlsbad, CA)). Finally, slides were washed with 1× PBS and mounted with fluoroshield mounting medium with 4΄,6-diamidino-2-phenylindole (ab104139; Abcam; Cambridge, United Kingdom). At least three technical duplicates and biological replicates were used for analyses.

Adult mouse testis staging and analysis

Controls (N = 14) and synchronized adults RiboTag-positive/Cyp17icre-positive (N = 21) and C57BL/6-129 (N = 34) testis cross-sections were sectioned at 4 μm with 50 μm between each cross-section. At least 200 tubules per testis were staged as described by Rusell et al. [7]. The midpoint of synchrony and synchrony factor for each sample was calculated based on methods described previously [12, 30, 31]. Samples with a synchrony factor more than two were used for analysis. Samples were separated into different groups based on the stages present in each testis. Additionally, the stage enrichment was confirmed by calculating the window width for each sample. Previous studies have shown the accuracy of using the calculated midpoint of synchrony, the synchrony factor, and window width to identify synchronized testes and determine at which point along the cycle the synchronized sample represents [12, 30, 31].

Quantitative real-time polymerase chain reaction

Total and IP mRNAs from synchronous samples (N = 8 for each) were used to generate cDNA using the iScript cDNA synthesis kit (catalog no. 170-8891; Bio-Rad Laboratories; Hercules, CA) based on manufacturer's instructions. Polymerase chain reaction (PCR) (two-step) was performed as previously described [32] using Fast SYBR Green Master Mix (ref. 4385612; Thermo Fisher Scientific; Waltham, MA) and primers for germ cell-specific transcript Stra8 (primers: 5΄-GTTTCCTGCGTGTTCCACAAG-3΄ and 5΄-CACCCGAGGCTCAAGCTTC-3΄), Sertoli cell-specific transcript SRY (sex determining region Y)-box 9 (Sox9) (primers: 5΄-CGCGGAGCTCAGCAAGACTCTG-3΄ and 5΄-TGTCCGTTCTTCACCGACTTCCTC-3΄), PTM-specific transcript actin, alpha 2, smooth muscle, aorta (Acta2) (primers: 5΄-GTTCAGTGGTGCCTCTGT-3΄ and 5΄-GGGATCCTGACGCTGAAG-3΄); Leydig cell-specific transcript insulin-like 3 (Insl3) (primers: 5΄TGCAGTGGCTAGAGCAGAGA-3΄ and 5΄-GGACACAGACCCAACAGGTC-3΄); and internal control ribosomal protein S2 (Rps2) (primers: 5΄-CTGACTCCCGACCTCTGGAAA-3΄ and 5΄- GAGCCTGGGTCCTCTGAACA-3΄) to normalize mRNA expression levels. The samples were run in triplicate and the data were analyzed using Analysis of Variance (ANOVA) (Excel; Microsoft) and Student t-test (Excel; Microsoft). Expression levels were deemed to be significantly different between samples with P values less than 0.05.

Microarray analysis

Total and IP mRNA samples (100 ng) were amplified and labeled using NuGen Ovation labeling kit and hybridized to Affymetix GeneChip Mouse Gene 1.0 ST Arrays. Microarray data were normalized using the Robust Multiarray Averages algorithm and analyzed using GeneSpring Multi-Omic Analysis (Version 12.6.1; Agilent Technologies; Santa Clara, CA). Duplicate samples were used in every group for the IP and total samples. For the analysis, genes with raw score greater than 200 were considered to be expressed. Statistical significance was determined by ANOVA (P < 0.05; 5% false discovery rate multiple test correction) and Student t-test (P < 0.05) for all comparisons between groups. Additionally, genes were considered to cycle if expression increased or decreased by at least 2-fold between two stages. Leydig cell-enriched genes were identified if the ratio of the signal in the IP to the total mRNA samples was at least 5 (N = 5). Functional annotation was determined using The National Center for Biotechnology Information and Database for Annotation, Visualization and Integrated Discovery (DAVID) Bioinformatics Resources 6.7 [33, 34]. All microarray data were deposited in National Center for Biotechnology Information gene expression and hybridization array data repository with the accession number GSE102214.

Testosterone measurements

Testosterone measurements within control (N = 12) and synchronized (N = 29) C57BL/6-129 adult testis samples were performed via radioimmunoassay (RIA) coupled with chromatography. Measurements were performed by the Endocrine Technology Support Core Lab at the Oregon National Primate Research Center (Beaverton, OR). The intra-assay coefficient of variation (CV) for testosterone was 6.7% and recovery was 53%. ANOVA and Student t-test (P < 0.05 is significant) was used for the statistical analysis.

Leydig cell isolation

Leydig cells isolation was performed from whole testis as described by Chang et al. with minor modifications [35]. Testes from adult C57BL/6-129 mice were dissected and collected in Dulbecco Modified Eagle Medium (DMEM 1×) (ref: 11965-092; Thermo Fisher Scientific; Waltham, MA) and the tunica removed. Seminiferous tubules were dissociated and added to 10 ml of DMEM with deoxyribonuclease I (200 μg/ml) and collagenase IA (0.5 mg/ml). Tubules were then incubated for 20 min in a shaking water bath (80 oscillations/min) at 35°C. The 10 ml containing the seminiferous tubules and interstitial cells were added to 40 ml of 5% Percoll with pH 8.5–9.5 (P1644; Sigma-Aldrich, St. Louis, MO) diluted in 1× Hank balanced salt solution (ref: 14175-095; Thermo Fisher Scientific; Waltham, MA) and incubated at room temperature for 20 min. The top 35 ml of the Percoll solution was aliquoted in a 50 ml tubes containing Dulbecco phosphate buffered saline (1× DPBS; ref: 14190-144; Thermo Fisher Scientific; Waltham, MA) and centrifuged at 500× g for 10 min at 4°C. The supernatant was removed, and the pellet resuspended in a radioimmunoprecipitation lysis buffer. The tubules collected from this protocol were also resuspended in the lysis buffer.

Western blot

Western blots were performed as previously described for primary rabbit antibodies EGR1 (1:300; catalog no. orb34304; Biorbyt; San Francisco, CA), CYP21 (1:300; catalog no. orb13362; Biorbyt; San Francisco, CA), steroidogenic acute regulatory protein (STAR) (1:300; sc-25806; Santa Cruz Biotechnology; Dallas, TX), and ESYT3 (1:200; HPA039200; Sigma-Aldrich; St. Louis, MO) with minor modifications [26]. Fifty micrograms of adult mouse testis, isolated Leydig cells, or tubules lysates were loaded onto sodium dodecyl (lauryl) sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels (catalog no. 465-1084; Bio-Rad Laboratories; Hercules, CA) for detecting EGR1, CYP21, STAR, ESYT3 or no antibody for control. The protein was then transferred to a nitrocellulose membrane and blocked with 5% skim milk in Tris-buffered saline (1× TBS-PM) for at least 1 h at room temperature. Primary antibody diluted in 1× TBS-PM with 0.1% Tween-20 was added to the nitrocellulose membrane and incubated at 4°C overnight. The next day, the membrane was washed four times with 1× TBS containing 0.1% Tween-20 for 30 min at room temperature. Secondary goat anti-rabbit IgG-HRP conjugate (1:2500; Cat. No. 170-6515; Bio-Rad Laboratories; Hercules, CA) diluted in TBS-PM with 0.1% Tween-20 was added to the membrane and incubated at room temperature for 2 h. The membrane was then washed four times with 1× TBS containing 0.1% Tween-20 for 30 min at room temperature. The membrane was then sprayed thoroughly with HyGLO Quick Spray (catalog no. E2400; Denville Scientific; Holliston, MA) chemiluminescence HRP antibody detection reagent. The membranes for EGR1, STAR, and ESYT3 were exposed for 2 min and the membrane for CYP21 for 1 min and imaged on Fujifilm LAS-4000.

Results

Leydig cell-specific transcripts were isolated across the different stages of the seminiferous epithelium

Previous publications have suggested that Leydig cells secrete factors needed for different events during spermatogenesis process and that their function might be influenced by surrounding tubules [36–39]. However, a lack of tools has made it difficult for investigators to study Leydig cell-specific transcript patterns across one cycle of the seminiferous epithelium in vivo. To circumvent this, we synchronized spermatogenesis within RiboTag-positive/Cyp17icre-positive animals using our WIN18,446/RA treatment method, and then performed IP isolation of Leydig cell-specific transcripts and total whole testis RNA extraction [12, 23, 24, 27]. The synchronized treatment generated adult RiboTag-positive/Cyp17icre-positive experimental male mice with testes containing tubules enriched for Stages XII-II/III (Group 1), II-VI (Group 2), VI-VIII (Group 3), VIII-X (Group 4), and X-XII (Group 5) (Figure 1). To verify that the HA-tagged polyribosomes were expressed specifically in the Leydig cells of the experimental animals, we performed an IHC for the HA-tag on testis cross-sections from adult RiboTag-positive/Cyp17icre-positive mice. HA-tag was detected within cells in the interstitial space (Figure 2A). The HA-tag positive interstitial cells were confirmed to be Leydig cells, as they also expressed the Leydig cell marker 3βHSD (Figure 2B–D). Quantitative real-time PCR was utilized to assess the enrichment of the IP for Leydig cell transcripts. A 74-fold enrichment for Leydig cell transcripts was detected in the IP compared to the total mRNA samples and to the transcripts of other cell types presence in the testis (Figure 2E).

Figure 1.

WIN 18,446/RA treatment synchronized adult spermatogenesis. Images depict representative cross-section of synchronized adult RiboTag-positive/Cyp17icre-positive experimental mice testes stained for STRA8 (Brown staining) containing tubules enriched for Stages XII–II/III (Group 1) (A), II–VI (Group 2) (B), VI–VIII (Group 3) (C), VIII–X (Group 4) (D), or X-XII (Group 5) (E). Scale bars = 50 μm.

Figure 2.

HA-tagged polyribosomes expression in Leydig cells of the adult experimental testes. Images show IHC analysis of adult RiboTag-positive/Cyp17icre-positive experimental male testes stained for 3βHSD (Brown staining, indicated by black arrows) (A) and IF stained for HA (green, indicated by orange arrows) (B), 3βHSD (red, indicated by blue arrows) (C), and costained with HA and 3βHSD (yellow, indicated by white arrows) (D). Graph shows expression of Stra8, Acta2, Insl3, and Sox9 in the IP (black bars) and total (gray bars) samples from the adult RiboTag-positive/Cyp17icre-positive experimental male testes (E). The error bars represent SEM (***P < 0.0005).

Known Leydig cell genes do not display cyclic expression patterns

To map the Leydig cell translatome across the different stages of the seminiferous epithelium, we performed microarray analysis on total and IP mRNA samples isolated from the five different groups of the RiboTag-positive/Cyp17icre-positive synchronized testes. To determine whether previously characterized Leydig cell marker genes display cyclic expression patterns, we assessed the enrichment of the following eight genes in our five different synchronized testis IP samples: hydroxyl-delta-5-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 1 (Hsd3b1), cytochrome P450 family 11 subfamily A member 1 (Cyp11a1), Ar, Star, hydroxyl-delta-5-steroid dehydrogenase, 3 beta- and steroid delta-isomerase 6 (Hsd3b6), cytochrome P450 family 17 subfamily A member 1 (Cyp17a1), luteinizing hormone/choriogonadotropin receptor (Lhcgr), and hydroxysteroid 17-beta dehydrogenase 3 (Hsd17b3) (Figure 3A). Using the ANOVA test, no significant cyclic changes were present for any of these Leydig cell genes across the different stages of the seminiferous epithelium (Figure 3B).

Figure 3.

No cyclic known Leydig cell transcripts detected in synchronized testes. Graphs show enrichment (IP/total) of known Leydig cell transcripts (A) and raw microarray expression data of known Leydig cell genes across the cycle of the seminiferous epithelium (B). The error bars represent SEM.

Leydig cell-enriched genes cycle with different stages of the seminiferous epithelium

Since no stage-specific expression of well-known Leydig cell transcripts was observed in the IP samples, we investigated whether other Leydig cell genes were expressed in a cyclic manner across the spermatogenic cycle. First, the expression of all genes in the IP and total samples were compared across the different stages. A total of 5475 genes were enriched by at least 2-fold in the IP compared to total samples across the five different groups (Figure 4A). Of these 5475 genes, 489 were enriched by 5-fold in the IP compared to the total mRNA samples in at least one group, now designated as Leydig cell-enriched genes. As expected, the 489 genes included well-known Leydig cell genes such as LH receptor (Lhcgr), estrogen receptor 1 (Esr1), prolactin receptor (Prlr), and cytochrome 450 family 11 (Cyp11a1). Functional annotation clustering using DAVID was performed for the top 50 Leydig cell-enriched genes (Figure 4B). Most of the top 50 Leydig cell-enriched genes were associated with glycoproteins (21 transcripts), signaling (21 transcripts), secretion (11 transcripts), extracellular matrix region (11 transcripts), lipid metabolism (5 transcripts), oxidoreductases (8 transcripts), hydrolases (10 transcripts), and mitochondrial (9 transcripts) (Table 1).

Figure 4.

Heat map and enrichment of Leydig cell-specific transcripts. Heat map depicts raw expression levels of 5475 transcripts in the IP and total samples across the different stages of the cycle of the seminiferous epithelium (A). Graph depicts the top 50 Leydig cell-enriched transcripts (B) (N = 10). The error bars represent SEM.

Table 1.

Identification of biological process of Leydig cell-enriched transcripts.

Biological process (gene ontology terms)	Number of transcripts	P value
Glycoproteins	21	3.6E-6
Signaling	21	1.3E-6
Secretion	11	8.1E-4
Extracellular matrix region	11	3.2E-3
Lipid metabolism	5	7.3E-3
Oxidoreductases	8	1.6E-4
Hydrolases	10	2.7E-3
Mitochondrial	9	2.7E-2

To further investigate whether Leydig cell-enriched genes associated differently with specific stages of the cycle of the seminiferous epithelium, the IP transcripts were compared between groups using an ANOVA test. A total of 893 genes in the IP samples were found to cycle with at least a 2-fold change between two Stage groups (Figure 5A), with 11 genes found to overlap with the Leydig cell-enriched gene list (Figure 5B; Table 2). Cyclic Leydig cell-enriched transcripts included genes involved in steroid metabolic processes, hormone regulation, transcription regulation, oxidoreductase activity, and phosphoproteins.

Figure 5.

Cyclic expression of 11 Leydig cell-enriched transcripts across one spermatogenic cycle. Heat map depicts expression of transcripts in IP samples across the different stages of the cycle of the seminiferous epithelium (A). Venn diagram shows that 11 Leydig cell-enriched transcripts that cycle in the IP samples (B).

Table 2.

Biological function of 11 Leydig cell-enriched genes that expresses cyclic pattern across the 12 stages of the seminiferous epithelium.

Gene symbol	Gene description	Function
Pdk4	Pyruvate dehydrogenase kinase, isoenzyme 4	ATP binding, pyruvate dehydrogenase (acetyl-transferring) kinase activity, and transferase activity
Cyp21a1	Cytochrome P450, family 21, subfamily a, polypeptide 1	Heme binding, iron ion binding, lipid binding, monooxygenase activity, oxidoreductase activity, steroid 21-monooxygenase activity, steroid binding, and steroid hydroxylase activity
Serpina6	Serine (or Cysteine) peptidase inhibitor, clade A, member 6	Lipid binding, serine-type endopeptidase inhibitor activity, and steroid binding
Irs1	Insulin receptor substrate 1	SH2 domain binding, insulin receptor binding, phosphatidylinositol 3-kinase binding, protein binding, transmembrane receptor protein tyrosine kinase adaptor activity, and signal transducer activity
Fam13a	Family with sequence similarity 13, member A	Phosphoprotein, molecular function
Cyp7a1	Cytochrome P450, family 7, subfamily a, polypeptide 1	Cholesterol 7-alpha-monooxygenase activity, heme binding, iron ion binding, metal ion binding, monooxygenase activity, and oxidoreductase activity.
Bcl6	B cell leukemia/lymphoma 6	DNA binding, RNA polymerase II regulatory region sequence-specific DNA binding, and transcription repressor activity.
Loxl4	Lysyl oxidase-like 4	Metal ion binding, oxidoreductase activity, and scavenger receptor activity
Txnip	Thioredoxin interacting protein	Enzyme inhibitor activity, protein binding, and ubiquitin protein ligase binding
Miox	Myo-inositol oxygenase	NADP binding, metal ion binding, and oxidoreductase activity
Bhlhe40	Basic helix-loop-helix family, member e40	DNA binding, RNA polymerase II transcription factor activity, protein binding, and transcription repressor activity

Interestingly, the two genes involved in steroid metabolic processes, Cyp21a1 and Cytochrome P450, family 7, subfamily a, polypeptide 1 (Cyp7a1), and genes involved in regulation of hormones, insulin receptor substrate 1 (Irs1) and Serine (or Cysteine) peptidase inhibitor, clade A, member 6 (Serpina6), have a similar gene expression patterns across one cycle of the seminiferous epithelium, with their highest expression in Stages XII-II/III (Figure 6A). This same pattern was also observed for the two transcription regulation genes B cell leukemia/lymphoma 6 (Bcl6) (Figure 6B) and Thioredoxin interacting protein (Txnip) (Figure 6C). In contrast, the transcription regulator Basic helix-loop-helix family, member e40 (Bhlhe40), was significantly highly expressed at Stages X-XII (Figure 6B). The expression of the oxidoreductase genes, Lysyl oxidase-like 4 (Loxl4) and Myo-inositol oxygenase (Miox), was also highest at Stages XII-II/III (Figure 6D). Pyruvate dehydrogenase kinase, isoenzyme 4 (Pdk4) and Family with sequence similarity 13, member A (Fam13a1) expressions were highest at Stages XII-II/III in comparison to other stages of the cycle (Figure 6E).

Figure 6.

Cyclic Leydig cell-enriched transcripts across one spermatogenic cycle. Graphs show the raw microarray expression pattern of Cyp21a1, Cyp7a1, Serpina6, Irs1(A), Bcl6, Bhlhe40 (B), Txnip (C), Loxl4, Miox (D), Pdk4, and Fam13a1 (E) plotted across the different stages of the seminiferous epithelium.

CYP21A1 signal intensity varies among Leydig cells

Previous studies have suggested a relationship between the localization of Leydig cells within the interstitium and the surrounding seminiferous epithelium. Interestingly, researchers have detected differences in mRNAs and protein expression between different Leydig cells [38–40]. To characterize the protein expression for one of the Leydig cell-enriched cyclic genes, we performed a western blot using Leydig cells lysate to ensure the antibody specificity (Supplemental Figure 1) and immunofluorescence in adult nonsynchronous testis cross-sections for CYP21A1. We costained for the Leydig cell marker, 3βHSD, and CYP21A1 to determine whether Leydig cells expressed CYP21A1. As expected, the 3βHSD (Figure 7A) and CYP21A1 (Figure 7B) expressions colocalized in Leydig cells (Figure 7C). High-magnification pictures of Leydig cells located in three different sections were taken to compare the signal intensity of 3βHSD (Figure 7D, G, J) and CYP21A1 (Figure 7E, H, K) among Leydig cells in an asynchronous testis cross-sections. Interestingly, it was revealed that the signal intensity of CYP21A1 varied among Leydig cells. However, there was stable expression of 3βHSD in Leydig cells throughout the testes cross-sections. The different intensity expression was more obvious in the colocalization of 3βHSD and CYP21A1 (Figure 7F, I, L). These results suggest that a heterogeneous Leydig cell population might exist in the mouse testis.

Figure 7.

Localization of CYP21A1 in asynchronous testes of adult male mice. Pictures show representative cross-section of adult wild testis stained for 3βHSD (A) (red staining), CYP21A1 (B) (green staining), and colocalized 3βHSD and CYP21A1 (C) (yellow staining) via immunofluorescence. Figures D–L depict higher magnification pictures of Leydig cells located in #1 (D–F), #2 (G–I), and #3 (J–L) of the testis cross-section stained for 3βHSD (D, G, J) (red staining), CYP21 (E, H, K) (green staining), and colocalized 3βHSD and CYP21A1 (F, I, L) (yellow staining). White arrows indicate low signal intensity of CYP21A1 and blue arrows for high signal intensity of CYP21A1. Scale bars = 50 μm.

Esyt3 expression cycle across the different stages of the seminiferous epithelium

In addition to the 11 cyclic Leydig cell-enriched genes, there were 297 transcripts detected in the IP samples that cycle across different stages of the seminiferous epithelium. However, those genes did not appear in the microarray analysis of the total samples. One of those genes was Esyt3, which is known to be involved in lipid transport and anchoring of the endoplasmic reticulum to the cell membrane. Esyt3 is highly expressed in Stages VIII-X (Figure 8). Further investigation regarding the protein expression of ESYT3 in cross-sections of asynchronous testes revealed that the Leydig cell marker 3βHSD (Figure 9A) and ESYT3 (Figure 9B) colocalized in the Leydig cells (Figure 9C). Interestingly, high magnification of sections #1, #2, and #3 shows that 3βHSD (Figure 9D, G, J) is expressed consistently among all Leydig cells; however, the ESYT3 (Figure 9E, H, K) signal intensity varied in different Leydig cells as shown by the colocalization expression of 3βHSD and ESYT3 (Figure 9F, I, L). To verify the ESYT3 expression in Leydig cells, we isolated Leydig cells from adult mouse testes and performed a western blot. Confirmation of Leydig cell isolation was confirmed by detection of STAR in the lanes with Leydig cells and whole testis lysate, and no detection in lane ran with tubules lysate (Figure 9M). Similarly, ESYT3 protein was not detected in the tubules lysate, but was detected in the lanes containing Leydig cells and whole testis lysates (Figure 9N). However, even though 50 μg of Leydig cells and whole testis lysate were loaded in the lanes and developed for the same time, the intensity of the band in the Leydig cells lysate lane was darker than the band in the lane containing whole testis lysate. Additionally, the western blot demonstrated the specificity of both antibodies by detecting a band for STAR at 30 kDa and a band for ESYT3 between 70 and 90 kDa.

Figure 8.

Expression of Esyt3 across one spermatogenic cycle. Graph shows the plotted expression data of Esyt3 across the different stages of the seminiferous epithelium.

Figure 9.

Characterization of ESYT3 in asynchronous whole testes and in isolated Leydig cells from adult male mice. Pictures show representative of immunofluorescence performed on cross-section of adult asynchronous testes for 3βHSD (A) (red staining), ESYT3 (B) (green staining), and colocalization of 3βHSD and ESYT3 (C) (yellow staining). Higher magnification pictures of location #1 (D–F), #2 (G–I), and #3 (J–L) were taken of the testis cross-section stained for 3βHSD (D, G, J) (red staining), ESYT3 (E, H, K) (green staining), and colocalized 3βHSD and ESYT3 (F, I, L) (yellow staining). White arrows indicate low signal intensity of ESYT3 and blue arrows for high signal intensity of ESYT3. Scale bars = 50 μm. Western blots show the specificity and detection of STAR in Leydig cells, whole testis, and tubules lysates (M), and ESYT3 in Leydig cells, whole testis, and tubules lysates (N).

Leydig cell transcripts change their association with ribosomes across the different stages of the seminiferous epithelium

In the murine testis, studies have shown that the translation efficiency of genes changes during germ cell development [25, 27, 41, 42] and the RiboTag mouseline has been used to investigate changes in ribosome-associated mRNA transcripts [27, 42]. Using this methodology to our advantage, we investigated the ratio of IP to total mRNA for each transcript in all different stage groups and calculated the standard deviation between the different groups to identify transcripts that change (5-fold or higher) their association with ribosomes when at least two stage groups were compared. There were a total of nine genes in the IP samples that dramatically changed their association with ribosomes in a stage-specific manner (Table 3). The IP/total ratio for the nine genes was plotted across the 12 different stages of the seminiferous epithelium. The association with ribosomes of Bhlhe40 and tumor necrosis factor, alpha-induced protein 8 (Tnfaip8) significantly increased by 7- and 6-fold, respectively, in Stages X-XII (Figure 10A). Additionally, the ratio of IP/total of one of a gene associated with proteolysis, Klk1b16, was highest in Stages VI-VII (Figure 10B). Egr1 increased its association with ribosomes dramatically in Stages II-VI compared to the other nine stages (Figure 10C). In contrast, nuclear receptor subfamily 4, group A, member 1 (Nr4a1) increased its association with ribosomes 2-fold at Stages X-XII (Figure 10C). Interestingly, four genes, perilipin 2 (Plin2), glutathione S-transferase, alpha 1(Gsta1), low density lipoprotein receptor-related protein 5 (Lrp5) and serine (or cysteine) peptidase inhibitor, clade A, member 3A (Serpina3a), displayed a similar pattern of stage-specific polysome association. Each was found to have a significantly enhanced association with polysomes in Stages X-III (Figure 10D).

Table 3.

Description of the nine Leydig cell transcripts that changes their association with ribosomes across one spermatogenic cycle.

Gene symbol	Description
Serpina3a	Serine (or cysteine) peptidase inhibitor , clade A, member 3A (enzyme regulator activity)
Gsta1	Glutathione S-transferase, alpha 1 (catalytic activity)
Plin2	Adipose differentiation related protein (binding)
Egr1	Early growth response 1 (transcription regulator activity)
Nr4a1	Nuclear receptor subfamily 4, group A, member 1 (transcription regulator activity)
Klk1b16	Similar to Kalllkrein 1-related peptidase b16 precursor (catalytic activity)
Lrp5	Low-density lipoprotein receptor-related protein 5 (molecular transducer activity)
Bhlhe40	Basic helix-loop-helix family, member e40 (transcription regulator activity)
Tnfaip8	Tumor necrosis factor, alpha-induced protein 8 (enzyme regulator activity)

Figure 10.

Leydig cell genes change their association with ribosomes across one spermatogenic cycle. Graphs depict the raw microarray expression of Bhlhe40, Tnfaip8 (A), Klk1b16 (B), Nr4a1, Egr1 (C), Plin2, Gsta1, Lrp5, and Serpina3a (E) across the different stages of the seminiferous epithelium.

EGR1 signal intensity varied among Leydig cells

The increased association of Egr1 with ribosomes in Stages II-VI implied that enhanced translation of EGR1 protein may occur during those stages. To investigate this, a western blot was used to verify the specificity of the EGR1 antibody and that the EGR1 is expressed in the Leydig cells (Supplemental Figure 1). Additionally, an immunofluorescence was performed in wildtype asynchronous adult testis cross-sections using 3βHSD (Figure 11A) and EGR1 (Figure 11B). As expected, the Leydig cell marker 3βHSD colocalized with EGR1, demonstrating that EGR1 is expressed in Leydig cells (Figure 11C). EGR1 expression was also detected in spermatocytes and in Sertoli cells inside the seminiferous epithelium (Figure 11B, E, H, K). Comparison of the detected signal intensity of 3βHSD and EGR1 among Leydig cells localized in three different section of the testis cross-sections revealed that the signal intensity of 3βHSD (Figure 11D, G, J) was constant, while the signal intensity of EGR1 (Figure 11E, H, K) was higher in certain Leydig cells in comparison to Leydig cells localized in other regions such as region of the interstitium which was more visible by the colocalization of 3βHSD and EGR1 (Figure 11F, I, L). The intensity signal of EGR1 among Leydig cells could vary depending on the association of germ cells in the seminiferous epithelium surrounding it.

Figure 11.

Localization of EGR1 in the wildtype testes of adult male mice. Images depict immunofluorescence for 3βHSD (A) (red staining), EGR1 (B) (green staining), and colocalization of 3βHSD and EGR1 (C) (yellow staining) in representative cross-section of adult wild testis. Higher magnification of location #1 (D–F), #2 (G–I), and #3 (J–L) are shown for better comparison of Leydig cells stained for 3βHSD (D, G, J) (red staining), EGR1 (E, H, K) (green staining), and colocalized 3βHSD and EGR1 (F, I, L) (yellow staining). White arrows indicate low signal intensity of EGR1 and blue arrows for high signal intensity of EGR1. Scale bars = 50 μm.

Testosterone levels do not vary across the cycle of the seminiferous epithelium

The identification of cyclic Leydig cell gene expression leads us to wonder whether testosterone production also varied across the cycle of the seminiferous epithelium. Adult C57BL/6-129 testes were synchronized using the WIN18,446/RA treatment and grouped into Stages I-II/III, IV-V, VI-VII, VIII, IX-X, and X-XII. Testosterone levels were then measured in synchronous and wildtype asynchronous control testes and plotted against across one cycle of the seminiferous epithelium (Figure 12). Supporting previously published data [21, 22], there were no significant changes in testosterone levels across the different stages of the spermatogenic cycle, indicating that testosterone levels do not fluctuate with the spermatogenic cycle.

Figure 12.

Testosterone concentration is steady across one cycle of the seminiferous epithelium. Graph shows testosterone levels treated with either vehicle control (solid red line) or WIN 18,446/RA synchronous protocol (black line). Testosterone levels in pg/mg of tissue (Y axis) were plotted across the different stages of the seminiferous epithelium (X axis). ANOVA test indicated that testosterone levels do not change during the different stages of the spermatogenic cycle. The dotted red line represents the SEM of the control data since the control testes contain all the stages of the seminiferous epithelium. The vertical black line represents the SEM for the testis with synchronous spermatogenesis during the different stages of the seminiferous epithelium.

Discussion

This study, for the first time, has mapped the Leydig cell-specific transcripts across the different stages of the seminiferous epithelium in vivo. Early studies utilizing testis cultures have hypothesized that a paracrine relationship exists between Leydig cell function and the seminiferous epithelium [36–40]. This idea was supported by published studies illustrating that Leydig cell function was regulated by unknown factors secreted by the tubules in coculture experiments [37–39, 43, 44]. More recently, reports about Leydig cells have focused on the role of testosterone in germ cell development, but very little is known about the cell–cell communication between Leydig cells and the different cell types inside of the seminiferous epithelium. The complex organization of Leydig cells in the testis and the asynchronous nature of spermatogenesis have made it difficult for investigators to study Leydig cell function across one spermatogenic cycle using routine methods. Therefore, in this report we took advantage of our WIN18,446/RA treatment to synchronize spermatogenesis and the RiboTag methodology to investigate Leydig cell-specific transcripts across one cycle of the seminiferous epithelium in vivo.

The WIN 18,446/RA treatment protocol has allowed us to overcome the complexity of the testis by synchronizing spermatogenesis and generating testes containing tubules with only two to three different stages. Utilizing the WIN 18,446/RA treatment in combination with the RiboTag methodology [12, 23, 24, 27], we were able to immunoprecipitate Leydig cell-specific transcripts during the different stages of the seminiferous epithelium and performed microarray analyses to further investigate the Leydig cell transcriptome in relation to the different spermatogenic stages. The specificity of our methodology was verified by immunohistochemical detection of the HA-tag specifically in Leydig cells and a 74-fold enrichment for the Leydig cell-specific transcript Insl3 and no significant enrichment for any other testis cell type marker genes in the IP compared to the total RNA samples. In our microarray analyses, we set a high fold change (5-fold) for the IP/total ratio to generate our Leydig cell-enriched list. Most of the Leydig cell-enriched transcripts were well-known Leydig cell genes, such as Lhcgr, Cyp11a1, Star, and Esr1. Our Leydig cell-enriched transcripts list had similar top 50 genes as the Leydig cell gene list generated by Sanz et al. [42], demonstrating the efficiency and consistency of isolating Leydig cell-specific transcripts using the RiboTag methodology and the Cyp17icre mouse line [19, 27, 42].

In this study, we aimed to investigate a possible paracrine relationship between the Leydig cells and the different stages of the seminiferous epithelium. A recent report demonstrated that Wilms tumor gene, Wt1, might be regulating paracrine factors in Sertoli cells needed for normal Leydig cell function, as deletion of Wt1 within Sertoli cells resulted in decreased testosterone levels [45]. In addition, it was reported that paracrine factors secreted by Sertoli cells, such as activin, Desert hedgehog, basic fibroblast growth factor, and platelet-derived growth factor regulate the proliferation and differentiation of the stem Leydig cells [46]. Thus, to investigate whether Leydig cells function in a cyclic manner, we identified Leydig cell-enriched transcripts that change their expression pattern across one spermatogenic cycle. From the Leydig cell-enriched transcript list, there were 11 genes that showed a cyclic pattern. Of specific interest are Cyp21a1 and Serpina6, which are involved in glucocorticoid synthesis and transport. Both genes have similar cyclic expression patterns, peaking at Stages XII-III. These data suggest that glucocorticoid might be playing an important role during those particular stages. Glucocorticoid receptor has been detected in zygotene and early pachytenes spermatocytes during Stages XIII–III [47] in the rat testis, which are homologous to Stages XII–III in the mouse. Therefore, it is possible that increased Leydig cell expression of Cyp21a1 and Serpina6 results in increased production of the glucocorticoids needed for regulating specific events during Stages XII-III. Additionally, spermatogonia are most likely to undergo apoptosis in certain stages, [48] and there are published data to show that increased glucocorticoids promote the apoptosis of spermatogonia [49]. As a result, it is possible that the glucocorticoids needed to promote apoptosis in some spermatogonia are secreted by Leydig cells in a stage-specific manner. Additionally, Gizang-Ginsberg et al. showed that proopiomelanocortin (Pomc) mRNA, which stimulates the release of cortisol, was also localized only in Leydig cells surrounded by tubules in Stages IX to XII of the cycle in the mouse [40]. Therefore, it is possible that the glucocorticoids are playing a role in the regulation of spermatogenesis during specific stages. To further investigate this, corticosteroid levels should be measured across the cycle of the seminiferous epithelium and determine whether these levels peak at Stages XII-III, where Cyp21a1 and Serpina6 are most highly expressed.

Interestingly, another gene known to be involved in immune function, Bhlhe40, was also significantly highly expressed at Stages X-XII [50]. Additionally, Irs1 showed a similar cyclic pattern, peaking at Stages XII-III and VI-VIII. Published data have demonstrated that IRS1 regulates AR activity and stability in breast cancer cells [51]. It is quite possible that IRS1 is regulating AR activity in Sertoli cells during Stages VI-VIII, where AR is maximally expressed [10, 13, 14, 15]. Most importantly, this study identified 11 genes that might be playing a role in the cell–cell communication between Leydig cells and the seminiferous epithelium. Therefore, in vivo studies using transgenic and knockout animals will need to be performed to further investigate the purpose of the Leydig cell-enriched cyclic genes and whether there is stage-specific crosstalk between Leydig cells and the seminiferous epithelium. Additionally, a more in-depth analysis has to be performed to investigate further those genes that were expressed in the IP and not in total samples such as ESYT3. It is possible, as shown by western blot, that the presence of other cell types diluted the Esyt3 in the total samples.

Furthermore, we investigated Leydig cell genes that changed their association with ribosomes across the different stages of the seminiferous epithelium. The RiboTag technique allows researchers to measure the amount of mRNAs associated with ribosomes [27]. We identified nine transcripts whose association with polysomes changed with the cycle. Of particular interest were Egr1and Nr4a1, as their transcripts displayed an increased association with ribosomes in Stages II-VI and X-XII, respectively. These two genes are known to be regulated by LH [42]. Sanz et al. demonstrated that Egr1 and Nr4a1 increased dramatically in Leydig cells after 1 h of LH treatment [42]. It is unknown whether LH enhances translation of Egr1 and Nr4a1, but the increase in association with ribosomes for these two transcripts might be the result of increased LH levels during those stages of the spermatogenic cycle. Also, in vitro published data have shown that glucocorticoids downregulate testosterone synthesis by inhibiting Star transcription through NR4A1 [52]. Interestingly, Nr4a1 association with ribosomes decreases in stages when the Cyp21a1 and Serpina6 are highly expressed. However, further studies need to be performed to distinguish whether the transcripts associated with the ribosomes are actively being translated or repressed.

To further investigate and characterize the proteins encoded by Leydig cell-enriched cyclic genes, genes in Leydig cells that change their association with ribosomes, and cyclic genes in the IP samples that were not detected in the total samples, we performed immunofluorescence for CYP21A1, ESYT3, EGR1, and the Leydig cell marker, 3βHSD, utilizing cross-section of testes with asynchronous spermatogenesis. We observed varied signal intensity for these three proteins between different Leydig cells, while the Leydig cell marker, 3βHSD, was expressed in a stable manner. These results suggest that there is heterogeneity among the Leydig cell population in the adult mouse testis. However, further studies such as single-cell RNA-seq have to be performed to investigate further the function of the heterogeneous Leydig cell population and its correlation with the complex organization of Leydig cells in the interstitium. These data support other studies which demonstrated that there are cyclic regulatory processes in the Leydig cells of the murine testis. Gizang-Ginsberg et al. showed that proopiomelanocortin mRNAs were localized only in Leydig cells surrounded by tubules in Stages IX to XII of the cycle in the mouse [40]. In addition, it was revealed that the Leydig cell morphology changes depending on their location in the testis. Leydig cells localized closer to tubules in Stages VII-VIII were larger than Leydig cells found around tubules in Stages IX-VI [53]. These combined observations indicate that an interstitial-tubular relationship exists, and that the Leydig cell function could influence the cycling of seminiferous epithelium or vice versa.

The importance of Leydig cells in the production and secretion of testosterone needed for normal spermatogenesis is well known. Therefore, the identification of cyclic Leydig cell function suggested that the testosterone levels might fluctuate with the spermatogenic cycle. To date, several published studies have investigated testosterone levels across one spermatogenic cycle utilizing mechanically separated rat seminiferous tubules or VAD/retinoids-treated animals [20–22]. However, these studies have contradictory results. Therefore, using the WIN 18,446/RA treatment protocol we were able to measure the levels of testosterone during the different stages of the seminiferous epithelium. Confirming previous studies, we found out that the levels of testosterone are constant across one spermatogenic wave [21, 22]. Even though the levels of testosterone varied among animals containing the same stages of tubules, this variation was constant across the different stages of the seminiferous epithelium. However, in order to compare it to previous rat studies, we have to consider the species differences. In the mouse, there is a lack of androgen-binding protein (ABP) resulting in most of the testosterone to be free and active. In the rat, however, most of the testosterone is bound to ABP, resulting in less free intratesticular testosterone levels. The production of testosterone in the testis has been associated with the amount of ABP [54]. Therefore, in the mouse, the Leydig cells might be producing constant levels of testosterone because of the high amount of free intratesticular testosterone and the low levels of testosterone needed for normal spermatogenesis [19, 54].

Using two methodologies, the RiboTag and WIN 18,446/RA treatment, this study has, for the first time, identified Leydig cell-enriched genes that are expressed in a cyclic pattern across one cycle of the seminiferous epithelium. Our results support the hypothesis that Leydig cell function correlates with the different stages of the cycle. However, it is unknown whether these cyclic genes are regulating germ cell development inside the seminiferous tubules or whether factors secreted by the seminiferous tubules regulate Leydig cell function in a cyclic manner. Additionally, it is possible that the Leydig cells have their own innate cycle independent from the germ cell development. This report identified potential candidates that may be playing a role in cell–cell communication between the Leydig cells and the cells inside the seminiferous epithelium. Further studies regarding the function of the Leydig cell-enriched cyclic genes will need to be performed to better understand their cyclic pattern and their contribution to the different events during spermatogenesis.

Supplementary data

Supplementary data are available at BIOLRE online.

Supplemental Figure 1. Figures show western blots performed for CYP21 and EGR1 in adult mouse isolated Leydig cells lysates.

Notes

Conference Presentation: Presented in part at the 49th Annual Meeting of the Society for the Study of Reproduction, July 16–20, 2017, San Diego, California.

Edited by Dr. Jeremy P. Wang, MD, PhD, University of Pennsylvania.