Undifferentiated spermatogonia regulate Cyp26b1 expression through NOTCH signaling and drive germ cell differentiation

Abstract

Cytochrome P450 family 26 subfamily B member 1 (CYP26B1) regulates the concentration of all−trans retinoic acid (RA) and plays a key role in germ cell differentiation by controlling local distribution of RA. The mechanisms regulating Cyp26b1 expression in postnatal Sertoli cells, the main components of the stem cell niche, are so far unknown. During gonad development, expression of Cyp26b1 is maintained by Steroidogenic Factor 1 (SF-1) and Sex-Determining Region Y Box-9 (SOX9), which ensure that RA is degraded and germ cell differentiation is blocked. Here, we show that the NOTCH target Hairy/Enhancer-of-Split Related with YRPW Motif 1 (HEY1), a transcriptional repressor, regulates germ cell differentiation via direct binding to the Cyp26b1 promoter and thus inhibits its expression in Sertoli cells. Further, using in vivo germ cell ablation, we demonstrate that undifferentiated type A spermatogonia are the cells that activate NOTCH signaling in Sertoli cells through their expression of the NOTCH ligand JAGGED-1 (JAG1) at stage VIII of the seminiferous epithelium cycle, therefore mediating germ cell differentiation by a ligand concentration-dependent process. These data therefore provide more insights into the mechanisms of germ cell differentiation after birth and potentially explain the spatiotemporal RA pulses driving the transition between undifferentiated to differentiating spermatogonia.—Parekh, P. A., Garcia, T. X., Waheeb, R., Jain, V., Gandhi, P., Meistrich, M. L., Shetty, G., Hofmann, M.-C. Undifferentiated spermatogonia regulate Cyp26b1 expression through NOTCH signaling and drive germ cell differentiation.

The active metabolite of vitamin A, all-trans retinoic acid (RA), plays a critical role in embryonic development and in a wide variety of biologic processes after birth including the maintenance of skin and epithelial cells, regulation of apoptosis, maintenance and regulation of hematopoiesis, and germ cell differentiation (1–5). Both an excess or lack of RA is detrimental; therefore, regulation of its concentration has to be tightly controlled in a spatiotemporal manner. The cytochrome P450 family 26 (CYP26) family of enzymes, which belong to the cytochrome P450 superfamily of proteins, regulate RA concentration by functioning as hydroxylases specifically inactivating RA to its hydroxylated forms. In mammals, the CYP26 family contains 3 enzymes: CYP26 subfamily A member 1 (CYP26A1), CYP26B1, and CYP26C1 (6). Cyp26a1- and Cyp26b1–null mutant mice display abnormal CNS development and die shortly after birth (7–9), whereas Cyp26b1-null mutants also lack male germ cells (10). Cyp26c1- null mutants display no abnormal features (9). Because of their importance in the control of RA availability, expression and activity of the CYP26 enzymes have to be precisely regulated (11). Regulation of Cyp26a1 expression is by far the best understood; because its promoter is highly responsive to RA itself (12), an autocrine negative feedback regulation has been suggested in tissues such as liver, cranial ganglia, and otic vesicle (11, 13). In other tissues, however, RA is produced and works in a paracrine manner on adjacent cells, and the regulation of Cyp26a1 expression is less well understood (6, 14, 15). Further, the mechanisms regulating Cyp26b1 gene expression are still poorly characterized because the proximal regulatory region of the Cyp26b1 gene lacks the retinoid response elements found in Cyp26a1 (11, 13). However, evidence of Cyp26b1 up-regulation by sex-determining region Y box 9 (SOX9) and steroidogenic factor 1 (SF-1) in the male fetal gonad has been recently presented (16).

During mouse development, migrating primordial germ cells arrive at the genital ridges at around embryonic day (E) 12.5 (17). Between E12.5 and E14.5, male primordial germ cells differentiate into prospermatogonia and experience mitotic arrest in an asynchronous manner. In contrast to female fetal germ cells that undergo meiosis before birth in response to elevated RA levels, prospermatogonia do not enter meiosis because the Sertoli cells in the male gonads produce CYP26B1, which degrades RA to form 4-OH-RA and 18-OH-RA (10, 18, 19). Shortly after birth, prospermatogonia reenter the cell cycle and migrate to the basal part of the seminiferous epithelium to become spermatogonial stem cells (SSCs), or A_single spermatogonia, that are the foundation of spermatogenesis (20, 21). These cells either self-renew to maintain the pool of SSCs or differentiate into transitory A_single spermatogonia that will give rise, through mitosis, to 2 daughter cells that remain connected by intercellular bridges and are called A_paired spermatogonia (22–24). These cells divide and form chains of A_aligned spermatogonia. A_single, A_paired, and A_aligned germ cells are collectively called undifferentiated type A spermatogonia (A_undiff). A_paired and A_aligned spermatogonia are also called progenitors because they dramatically increase the number of germ cells. The A_aligned spermatogonia will then differentiate into A₁ to A₄ spermatogonia (A_diff), type B spermatogonia, and spermatocytes, which will start the meiotic process (25). Germ cells and Sertoli cells are enclosed within seminiferous tubules, and Sertoli cells are the main component of the stem cell niche. Observation of seminiferous tubules in cross-sections of adult testes reveals different associations of germ cells at various steps of differentiation. The cellular makeup of these associations is very specific; therefore, they have been divided in stages in many mammalian species (26), with 12 stages in the mouse (27). Interestingly, transition between A_undiff and differentiating germ cells, meiotic initiation, and start of spermatid elongation all occur in stages VII/VIII in the mouse.

RA is particularly critical for the transition between A_undiff and differentiating germ cells because rats and mice deprived of dietary RA can only produce A_undiff spermatogonia and are sterile (28). Further, RA activity is critical for initiation of the meiotic process and is also necessary for postmeiotic spermatid maturation (5, 29–31). As the availability of RA within the stem cell niche increases after birth, CYP26B1 in Sertoli cells must be down-regulated to allow spermatogonial differentiation and the first wave of spermatogenesis to take place. Further, low levels of CYP26B1 are still expressed and functionally relevant in adult Sertoli cells and must be modulated to allow the transition from A_undiff to A₁ spermatogonia at stage VIII of the seminiferous epithelium (31, 32). However, the mechanisms that modulate CYP26B1 enzyme production in the postnatal testis are poorly understood.

Our previous studies indicated a possible inverse relationship between Cyp26b1 expression and NOTCH activity in immature Sertoli cells (33, 34). In the present study, we provide direct evidence that NOTCH signaling in Sertoli cells down-regulates Cyp26b1 expression through the Hairy/Enhancer-of-Split Related with YRPW Motif 1 (HEY1) transcriptional repressor in the adult testis. Further, we demonstrate that A_aligned spermatogonia, through their expression of the NOTCH receptor JAGGED-1 (JAG1), are responsible for Cyp26b1 down-regulation in Sertoli cells, possibly triggering their own differentiation into A₁ spermatogonia.

MATERIALS AND METHODS

Mouse lines and breeding schemes

Mice were housed in accordance with National Institutes of Health (Bethesda, MD, USA) guidelines. The Institutional Animal Care and Use Committees at the University of Texas M. D. Anderson Cancer Center and Texas A&M University, Institute of Biosciences and Technology approved all experimental procedures and protocols. To specifically target/express genes following Cre-mediated recombination (Cre) in Sertoli cells, we used transgenic anti-Mullerian hormone-Cre [Tg(Amh-Cre)1Flor] mice (or Amh-Cre mice) (35) provided by Dr. Paul Cooke (University of Florida, Gainesville, FL, USA). To express TdTomato following Cre-mediated recombination in Sertoli cells, we used Gt(Rosa)26Sor^{tm9(CAG-tdTomato)/Hze} mice obtained from The Jackson Laboratory (Bar Harbor, ME, USA) that we bred to Amh-Cre mice to produce Amh^Cre/+;Rosa26^{TdTomato/TdTomato} mice, also called Amh-Cre;Rosa26^RFP/RFP mice. Additionally, we used AMH-Cre;ROSA^YFP/YFP mice that specifically express YFP in Sertoli cells (34). To detect active NOTCH signaling in vivo, we used the Transgenic Notch Reporter-green fluorescent protein (TNR-GFP) mouse line, which contains a Recombination Signal-Binding Protein for Immunoglobulin κJ Region (RBPJ) response element with 4 RBPJ-binding sites and a minimal Simian Virus 40 (SV40) promoter, followed by a GFP coding sequence (36). To detect fluorescence-activated cell sorter (FACS)-separated Sertoli cells with active or inactive NOTCH signaling, we bred the TNR-GFP line with the Amh-Cre;ROSA26^RFP/RFP line to obtain TNR-GFP;Amh-Cre;ROSA26^RFP/RFP mice. Genotyping primers are found in Supplemental Table S1.

Sertoli cell isolation and cultures

Primary Sertoli cells were isolated from 21-d-old or adult (>3 mo-old) mice testes using a 2-step digestion protocol (34, 36). After digestion, Amh-Cre;Rosa26^RFP/RFP Sertoli cells were FACS-sorted using a 580/30-nm band pass filter for TdTomato on a BD Influx cell sorter (BD Biosciences, San Jose, CA, USA). Similarly, Amh-Cre; Rosa^YFP/YFP Sertoli cells were FACS-sorted using a 542/27-nm band pass filter for YFP. Alternatively, Sertoli cells were purified using datura stramonium agglutinin lectin (DSA)-coated dishes (38). They were then immediately plated on Matrigel-coated dishes or culture plates prepared according to previously published procedures (37). Sertoli cells were cultured for 24 or 48 h in DMEM/F12 complemented with GlutaMAX and 5000 U/ml penicillin/streptomycin but without serum. Matrigel was obtained from BD Biosciences. All other tissue culture supplies and reagents were obtained from ThermoFisher Scientific (Waltham, MA, USA) or VWR (Sugar Land, TX, USA).

Because primary adult Sertoli cells are notoriously difficult to consistently transfect, we used human embryonic kidney (HEK)-293T cells for luciferase assays. HEK293T cells were obtained from the M. D. Anderson Characterized Cell Line Core Facility. Cells were maintained in DMEM/F12 supplemented with GlutaMAX, 10% fetal bovine serum (FBS), and 5000 U/ml penicillin-streptomycin.

Gene expression analysis

For Sertoli cells cultured in 96-well plates, RNA extraction, reverse transcription, and quantitative PCR (qPCR) were performed using the TaqMan Gene Expression Cells-to-CT Kit (Thermo Fisher Scientific) according to manufacturer’s instructions. For Sertoli cells grown in 6-well or 24-well plates, RNA was isolated using the Aurum Total RNA Mini kit (Bio-Rad, Hercules, CA, USA) and first strand cDNA synthesis was performed with the iScript cDNA Synthesis Kit (Bio-Rad). Quantitative RT-PCR (qRT-PCR) was performed using the SsoAdvanced Universal Probes Supermix (Bio-Rad) using a Bio-Rad CFX384 Touch Detection System. In certain instances, a preamplification step was conducted for an unbiased, target-specific amplification of cDNA using the SsoAdvanced PreAmp Supermix (Bio-Rad). The TaqMan assays used for specific transcripts are listed in Supplemental Table S2. Two genes, eukaryotic initiation factor 3 (Eif3) or 40S ribosomal protein S3 (Rps3), were used as internal control genes to normalize the expression value of each gene in all the analyses. The 2^−ΔΔCt method was employed to calculate the relative quantitative fold changes. Experiments were carried out with at least 3 biologic replicates per group, and each biologic replicate was carried out in triplicate per plate. Statistical analysis including mean and sem was determined for the biologic replicates after the raw critical threshold values Ct of technical replicates were averaged.

Western blotting

Proteins were isolated from mouse C18-4 cells (39–41) using a standard lysis buffer. MCF7 human breast cancer cells and HeLa human cervical cancer cells were used as positive controls. MCF7 cells were obtained from American Type Culture Collection (HTB-22; ATCC, Manassas, VA, USA), and HeLa cells were obtained from the M. D. Anderson Characterized Cell Line Core Facility. Proteins were run on 4–12% polyacrylamide gels (Bio-Rad) and blotted on nitrocellulose membranes according to standard protocols. After transfer, blots were blocked with 5% milk in Tris-buffered saline for 1 h at room temperature and probed overnight at 4°C with a rabbit antibody against mouse and human JAG1 (ab7771; Abcam, Cambridge, MA, USA) and a goat antibody against ACTB as loading control (ab8229; Abcam). Blots were incubated with VRDye 490 Goat anti-Rabbit IgG (P/N 926-49020; Li-Cor Biosciences, Lincoln, NE, USA) and IRDye 680RD Donkey anti-Goat IgG (P/N 925-68074; Li-Cor Biosciences). Fluorescent bands were revealed using an Odyssey Fc Imager (Li-Cor Biosciences). Cell culture experiments and Western blots were repeated at least 3 times. All antibodies are listed in Supplemental Table S3.

In vitro JAG1 activation of NOTCH signaling in Sertoli cells

To immobilize JAG1, Pierce Protein A/G Coated 96-Well Plates (Thermo Fisher Scientific) were incubated overnight at 4°C with recombinant rat JAG1-Fc chimera protein (R&D Systems, Minneapolis, MN, USA) diluted in PBS to desired concentrations. Control plates were incubated with PBS only. The next day, the plates were washed, and a 1:5 dilution of Matrigel (BD Biosciences) in DMEM/F12 was added on top of the immobilized ligand for 30 min at 4°C. Excess Matrigel was removed and the plates were incubated at 37°C for another 30 min to allow Matrigel to solidify. Fifty thousand RFP-positive Sertoli cells (21-d-old) were seeded into each well and cultured serum-free in DMEM/F12 in a humidified 5% CO₂ incubator for 48 h. Total RNA was isolated and qRT-PCR analysis was performed according to standard protocols (see above).

Down-regulation of Hey1 expression

RFP-positive Sertoli cells were seeded into Pierce Protein A/G 96-Well Plates coated with JAG1-Fc chimera protein and Matrigel as previously described. The cells were then transfected with Hey1 small interfering RNAs (siRNAs), mock, or scrambled siRNAs (Thermo Fisher Scientific) using RNAiFect Kit according to the manufacturer’s instructions (Qiagen, Valencia, CA, USA). Total RNA was isolated after 48 h to measure the levels of gene expression by qRT-PCR. TaqMan probes are listed in Supplemental Table S2.

Construction of dual luciferase plasmids

Analysis of the 2500 bp mouse Cyp26b1 proximal promoter using Eukaryotic Promoter Database and the University of California–Santa Cruz Genome Browser (https://genome.ucsc.edu/cgi-bin/hgGateway) revealed 7 putative HEY-binding enhancer boxes (E-boxes) (CANNTG) but no N-box element sequences. Two adjacent sequences were cloned into the pGL3 basic vector (Promega, Madison, WI, USA). Promoter sequence 1 (−1783 to −1194 bp) contains 1 E-box at −1699 to −1694 bp, and promoter sequence 2 (−1105 to −155 bp) contains 3 E-boxes at −1021 to −1016, −677 to −672, and −431 to −426 bp. Primer Blast (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) was used to design primers, and restriction analysis with Webcutter (http://www.firstmarket.com/cutter/cut2.html) identified Kpn1 and Xho as appropriate restriction sites. The promoter sequences were amplified using mouse tail DNA as templates and subcloned into pGL3 basic using T4 DNA ligase (Promega) and the same restriction sites. Stellar competent cells (Clontech Laboratories, Mountain View, CA, USA) were used to amplify the plasmids. All the constructs were verified by sequence analysis in sense and antisense orientations. Sequencing was carried out using unidirectional sequencing primers.

Dual luciferase reporter assay

Experiments were done with HEK-293 cells because of the notoriously low transfection efficiency in primary Sertoli cells. Cells were cultured in a humidified incubator at 37°C and 5% CO₂ with DMEM supplemented with 10% FBS, 5000 U/ml penicillin and streptomycin (Thermo Fisher Scientific) and incubated until 80% confluence. They were then transfected with 10 µg/ml pGL3-Basic reporter construct using FuGENE HD transfection reagent (Promega) together with 10 µg/ml of Renilla luciferase reporter vector pRL-TK (Promega) to provide an internal control of the transfection efficiency. Each treatment was performed in triplicates for a total of 3 independent experiments. The culture medium was removed, and cells were collected 48 h after transfection and lysed. Renilla and firefly luciferase activities were directly measured using the Dual Luciferase Reporter Assay System (Promega) and a LumiStar Galaxy luminometer (BMG-Labtech, Cary, NC, USA) according to the Promega kit instructions. Firefly luciferase activity was normalized to Renilla luciferase activity. The empty pGL3-Basic vector served as a blank control. Triplicate samples were used for each point in 3 independent experiments. These experiments were repeated using optimized transfection conditions for SF7 Sertoli cells.

Chromatin Immunoprecipitation-qPCR

Chromatin immunoprecipitation (ChIP)-PCR was performed using adult primary Sertoli cells isolated and maintained as previously described. Cross-linking was performed with 1% formaldehyde followed by quenching with 0.125 M glycine. Cross-linked cell nuclei were sonicated using a Bioruptor ultrasonicator (Diagenode, Denville, NJ, USA) to obtain 250–600 bp DNA fragments. The sonication buffer (lysis buffer) contained 0.1% SDS, 10 mM EDTA, and a standard protease inhibitor cocktail (Halt; Thermo Fisher Scientific) in 50 mM Tris-HCl (pH = 8.1). After sonication, SDS was diluted to a final concentration of 0.05% by dilution with lysis buffer without SDS. The desired fragment sizes were confirmed on agarose gels, and 5 µg chromatin was incubated overnight with a hairy and enhancer-of-split (HES1) antibody (rabbit monoclonal antibody 11988; Cell Signaling Technology, Danvers, MA, USA) or a HEY1 antibody (rabbit pAb 19929-1-AP; Protein Tech, Rosemont, IL, USA) or with rabbit IgG (Jackson ImmunoResearch, West Grove, PA, USA) at 4°C with rotation (Supplemental Table S3). Immunoprecipitation was performed for 1 h with protein-G magnetic beads (Dynabeads; Thermo Fisher Scientific). The bound ChIP complexes were washed sequentially for 10 min each using the following buffers: low salt buffer containing 0.1% SDS, 1% Triton X-100, 2 mM EDTA, 150 mM NaCl in 20 mM Tris-HCl, pH 8.1, followed by a high salt buffer with 0.1% SDS, 1% Triton X-100, 2 mM EDTA, and 500 mM NaCl in 20 mM Tris-HCl, pH 8.1, and finally a LiCl wash buffer containing 0.25 M LiCl, 1% NP-40, 1% deoxycholate, 1 mM EDTA in 10 mM Tris-HCl, pH 8.1. The beads were then washed thrice with TE buffer and extracted 3 times with 1% SDS, 0.1 M NaHCO₃, 0.1 mg/ml proteinase K, and 0.6 M NaCl. The DNA was released by reverse cross-linking at 65°C for 6 h and 95°C for 10 min followed by magnetic removal of the beads. qPCR was done using iTAq Universal SYBR Green Supermix (Bio-Rad). Primer pairs flanking 4 HEY1-binding sites in the mouse Cyp26b1 proximal promoter region were used. As negative control, a primer pair recognizing a region of the Cyp26b1 promoter without HEY-binding sites was chosen. Results were normalized to IgG controls. qPCR primer pairs are listed in Supplemental Table S4.

Down-regulation of JAG1 expression

JAG1 down-regulation was carried out with short hairpin RNAs (shRNAs) cloned into GIPZ lentiviral vectors expressing TurboGFP (RMM4532-EG16449; GE Dharmacon, Lafayette, CO, USA). They were obtained from the shRNA and ORFeome Core Facility at M. D. Anderson Cancer Center. HEK293T cells were grown in T75 flasks in 10 ml of DMEM (10% FBS) and transfected 1 d after plating with second-generation lentiviral packaging plasmids. p-cytomegalovirus (CMV)-VSV-G (3 µg) (Addgene, Cambridge, MA, USA), pCMV-dR8.2 dvpr (15 µg) (Addgene), and the shRNA lentiviral vector (15 µg) were combined at a ratio of 0.2:1.0:1.0, using Lipofectamine 2000 per the manufacturer’s instructions (Thermo Fisher Scientific). Eighteen hours after transfection, fresh medium was added, and cells were cultured for another 48 h before the culture supernatant containing lentiviruses was collected and filtered. One day after plating, C18-4 germ cells (see below) were transduced in T75 flasks with 5 ml of the lentivirus-containing supernatant mixed with 5 ml DMEM/F12 (10% FBS) and polybrene (TR-1003-G; Millipore, Billerica, MA, USA) at a concentration of 8 μg/ml for 18 h. Medium was then changed, and the cells were cultured for an additional 48 h before cells positive for the shRNA/TurboGFP insertion were selected by FACS. Down-regulation of Jag1/JAG1 expression was verified by qPCR and Western blotting, and cell clones with a reduction of >80% JAG1 expression were utilized for further experiments.

C18-4 cell line and coculture experiments

The C18-4 cell line, which has many characteristics of A_undiff spermatogonia (39), was used for coculture experiments with 21-d-old primary Sertoli cells. C18-4 cells were maintained in DMEM/F12 supplemented with GlutaMAX, 10% FBS, and 5000 U/ml penicillin-streptomycin as previously described (39–42). Amh-Cre; ROSA^YFP/YFP primary Sertoli cells were FACS-isolated and cocultured with C18-4 cells at a ratio of 5 germ cells/Sertoli cell. Sertoli cells alone were used as controls. After 48 h, Sertoli cells were FACS-isolated, total RNA extracted, and gene expression quantified by qRT-PCR (see below). Alternatively, cocultures were treated with or without N-[(3,5-difluorophenyl)acetyl]-l-alanyl-2-phenyl]glycine-1,1- dimethylethyl ester (DAPT) (Tocris Bioscience, Bristol, United Kingdom; Thermo Fisher Scientific), a NOTCH-signaling inhibitor that prevents NOTCH intracellular domain (NICD) release and expression of the target genes Hes1 and Hey1 (33, 43). DAPT was solubilized and diluted in DMSO and added to the cells to a final concentration of 25 μM in 0.1% DMSO. Control-treated cells received the same amount (0.1%) of DMSO vehicle. After 48 h, Sertoli cells were FACS-isolated, total RNA extracted and subjected for qRT-PCR. TaqMan probes are listed in Supplemental Table S2.

In vivo germ cell depletion

To evaluate the role of different types of germ cells in activating NOTCH signaling in Sertoli cells, we treated 8-wk-old Amh-Cre;Rosa26^RFP/RFP mice with a moderate dose of 20 mg/kg busulfan in DMSO (50% in H₂O) injected intraperitoneally (MilliporeSigma, Burlington, MA). Testes were collected from treated (n = 10) and control mice (DMSO only, n = 10) 4, 8, 16, and 24 d after busulfan treatment, and Sertoli cells were FACS-sorted as previously described (37). Expression of the NOTCH target genes Hes1 and Hey1 as well as Cyp26b1 were quantified by qRT-PCR. To test whether Sertoli cells themselves were affected by the treatment, we repeated these experiments by treating WBB6F1/J-KitW/ KitW-v/J mice that do not contain any germ cells (n = 10/group) (The Jackson Laboratory). These Sertoli cells were isolated by differential plating on DSA (MilliporeSigma) using the method of Scarpino et al. (38). Total RNA was isolated and evaluated for differential gene expression with qRT-PCR. TaqMan probes are listed in Supplemental Table S2.

Histologic analysis

Contralateral testes from in vivo germ cell depletion experiments (d 8 after busulfan treatment) were fixed in Bouin’s fixative, embedded in methacrylate, and sectioned at 4 µm. Quantitative germ cell comparison between control and treated testes was done by determining the ratio of each generation of premeiotic germ cells to Sertoli cells (with nucleoli present) in seminiferous tubules. Testes from 3 busulfan-treated and 2 control mice were analyzed. At least 2 round or oval tubule cross-sections at each of the 12 stages of the cycle of the seminiferous epithelium were counted in 1 histologic section from each testis. The averages and se (between mice) of these ratios were calculated.

Quantification and statistical analysis

GraphPad Prism 6.01 software (GraphPad Software, La Jolla, CA, USA) and MS Excel (Microsoft, Redmond, WA, USA) were used for data processing and analysis. All data points generated from isolated primary cells and cell lines are the results of at least 3 technical repeats per experiment and 3 independent experiments. qRT-PCR data are the results of 4–6 technical repeats and 3 independent experiments. Numerical data are presented as means ± sem. qRT-PCR data obtained from treated mice were generated from n = 10 mice/cohort (mice randomly assigned to treatments). For mouse studies, we used a power analysis to estimate the appropriate number of animals needed (sample size). With n = 10 mice/group, we had over 80% power to detect a 30% difference in gene expression between treatment groups (CV = 25%, α = 0.05). The significance of differences in continuous variables between treated and control mice were evaluated by a Student’s t test. For ChIP-qPCR, additional 2-way ANOVA was used for multiple comparisons between independent experiments followed by Tukey post hoc test. Values of P < 0.05 were considered statistically significant.

RESULTS

Cyp26b1 down-regulation is associated with activated NOTCH signaling in a subset of adult Sertoli cells

Our previous studies indicated that constitutive over-activation of NOTCH signaling drove early germ cell differentiation in the fetus leading to a Sertoli cell-only phenotype shortly after birth (33, 44). Conversely, knockout of the pathway through ablation of RBPJ, a signaling mediator downstream of all NOTCH receptors, led to a 30% increase of the number of germ cells in the adult testis (34). Gene expression analysis of fetal and postnatal mutant Sertoli cells indicated that expression of Cyp26b1 was inversely related to NOTCH activity and expression of the NOTCH target genes Hes1 and Hey1, which are transcriptional repressors (34). To test if there is an inverse relationship between NOTCH signaling and Cyp26b1 expression in wild-type adult Sertoli cells, we used the TNR-GFP mouse model (36) bred to Amh-Cre;Rosa^RFP/RFP mice. Sertoli cells were FACS-sorted according to their expression of RFP alone or RFP+GFP when RBPJ is activated, and qRT-PCR was performed. Results shown in Fig. 1 indicate that, at this age, NOTCH activated and inactivated Sertoli cells coexist (Fig. 1A). In addition to high GFP expression, NOTCH activity is demonstrated by higher Hes1 and Hey1 (Fig. 1B) expression in comparison with GFP-negative Sertoli cells. Further, high expression of Hes1 and Hey1 correlates with a significant down-regulation of Cyp26b1 in GFP-expressing cells, indicating that in wild-type adult Sertoli cells, there is also an inverse relationship between Hes1/Hey1 and Cyp26b1 expression.

Figure 1.

Cyp26b1 down-regulation is associated with activated NOTCH signaling in a subset of adult Sertoli cells. A) FACS-sorting gating criteria. Cells from the seminiferous epithelium of Amh-Cre; ROSA^RFP/RFP; TNR-GFP mice (3-mo old) were isolated using standard methods. They were further isolated by FACS. Isolation of RFP + cells (NOTCH activity-negative Sertoli cells), RFP+/GFP + cells (NOTCH activity-positive Sertoli cells), GFP+/RFP− cells (peritubular cells), and GFP−/RFP− cells (germ cells). B) Gene expression analysis of Sertoli cells with active or inactive NOTCH signaling. NOTCH-signaling active Sertoli cells express levels of Hes1 and Hey1 that are significantly higher than NOTCH-inactive Sertoli cells, whereas Cyp26b1 expression is inversely correlated. Sox9, sex-determining region Y box 9. For each gene, the values were normalized to the expression level in NOTCH-inactive cells. Results are given as means ± sem. Results were obtained from 3 independent experiments. **P < 0.01, ***P < 0.005.

Both HES1 and HEY1 are able to down-regulate Cyp26b1 expression

To demonstrate that HES1 or HEY1 or both are able to down-regulate Cyp26b1 gene expression, we used a standard dual luciferase reporter assay. Experiments were performed with human HEK293T cells because of very low rates of transfection of primary Sertoli cells (2–3%). HEK293T cells were transfected with 2 different mouse Cyp26b1 promoter regions driving luciferase. The first promoter sequence (1: −1783 to −1194 bp) contained 1 E-box allowing binding of HEY1 protein (Fig. 2A). The second promoter sequence (2: −1105 to −155 bp) contained an additional 3 E-boxes (Fig. 2A). We observed a significant reduction of luciferase enzyme activity (50%) only after the cells were cotransfected with a pCMV vector driving expression of HES1 or HEY1 gene/protein (Fig. 2B). Despite the fact that no N-box was found within the proximal promoter, we attribute the influence of HES1 to its ability to heterodimerize with HEY1 or to its weak affinity to E-boxes, as previously demonstrated (45–47). This experiment strongly suggests that Cyp26b1 down-regulation depends on canonical NOTCH activity carried out via HES1 and HEY1 transcriptional repressors.

Figure 2.

HES1 and HEY1 transcription factors down-regulate Cyp26b1 expression in a dual luciferase reporter assay. A) Analysis of the mouse Cyp26b1 promoter with respect to HES1 and HEY1 putative binding sites. Rigorous manual analysis of the mouse Cyp26b1 proximal promoter sequence (−2500 to +1 bp) retrieved from the Eukaryotic Promoter Database (EPD) indicated absence of N-boxes (CACNAG) that could allow binding of HES1. It contains 7 E-boxes (CANNTG) allowing putative binding of HEY1. The TATA box is at −31 to −26 bp from the transcription start (+1 bp). Boxes 1 and 2 indicate Cyp26b1 promoter sequences subcloned into the pGL3 Basic Luciferase reporter vector. B) Both HES1 and HEY1 can inhibit Cyp26b1 expression despite absence of N-boxes. Two different nucleotide sequences containing E-boxes upstream of the transcription start site 1 (TSS1) of the mouse Cyp26b1 proximal promoter were cloned into pGL3-Basic vector and cotransfected with Hes1- or Hey1-pCMV expression vectors into HEK-293 cells. Expression of HES1 and HEY1 both significantly down-regulated luciferase activity, showing that HES1 probably heterodimerizes with HEY1, or recognizes E-boxes (45, 46). Luc, luciferase. For each graph, results were obtained from 3 independent experiments and are given as means ± sem. **P < 0.01.

HEY1 directly binds to the Cyp26b1 promoter

We next tested if there is a physical interaction between HEY1 and E-boxes within the mouse Cyp26b1 promoter using ChIP followed by qPCR analysis in primary Sertoli cells. qPCR was performed on ChIPed DNA after immunoprecipitation with HES1 and HEY1 antibodies. The Cyp26b1 promoter region −1800 to +1 contains 4 E-boxes that we probed using the appropriate primers. Results shown in Fig. 3 demonstrate that the HEY1 antibody immunoprecipitated DNA-bound HEY1 at all putative binding sites, whereas the HES1 antibody could not immunoprecipitate any DNA. No binding was detected in regions that did not contain any E-boxes. The relative amount of immunoprecipitated DNA compared with input DNA after qPCR analysis was around 4%. These results demonstrate that the canonical NOTCH target HEY1 consistently binds E-boxes at the Cyp26b1 proximal promoter to down-regulate its expression. These data are in accordance with our previous results that demonstrated an increase of HEY1 expression between d 0 and d 7 after birth (34), which corresponds to an increase of RA production, an increase of STRA8 expression and the first wave of spermatogonial differentiation.

Figure 3.

ChIP-qPCR analysis of the mouse Cyp26b1 promoter showing binding of HEY1. A) Additional Cyp26b1 promoter analysis. HEY1-binding sites (E-boxes) are shown after the transcription start. Each binding site (1–5) is recognized by primer pairs [primer pair 1 (PP1) to PP5] used in the qPCR experiment following ChIP. B) ChIP-qPCR analysis of the mouse Cyp26b1 promoter. Analysis shows quantification of enriched DNA after ChIP using a mouse HEY1 antibody. PP1 probes a region containing 2 E-boxes at +11 to +137 bp. PP2 probes a region at −486 to −382 bp, PP3 probes a region at −746 to −583 bp, PP4 probes a region at −1119 to −982 bp, and PP5 probes a region containing 2 E-boxes between −1900 and −1652 bp. The control primers recognize a region between −1632 and −1551 bp devoid of E-boxes. All PPs are listed in Supplemental Table S4. Results were obtained from 3 independent experiments and are given as means ± sem. *P < 0.05, **P < 0.01, ***P < 0.005.

JAG1-expressing germ cells down-regulate Cyp26b1 expression through NOTCH signaling in Sertoli cells

Histologic sections of 8-d-old testis confirmed that premeiotic germ cells express high amounts of JAG1, the main NOTCH pathway ligand in the seminiferous epithelium (48) (Fig. 4A). To demonstrate a direct effect of germ cells on Cyp26b1 expression in vitro, we cocultured C18-4 cells, which resemble A_undiff and express JAG1 (Fig. 4B), together with freshly isolated Amh-Cre;Rosa^RFP/RFP primary Sertoli cells (P21) for 48 h. The cell line was chosen over primary germ cells for better reproducibility. We then FACS-sorted Sertoli cells from these cocultures and measured their expression of the NOTCH target genes Hes1 and Hey1, as well as the expression of Cyp26b1. Figure 4C demonstrates that C18-4 cells are able to up-regulate NOTCH signaling and Hes1 and Hey1 expression while simultaneously down-regulating Cyp26b1. Further, inhibition of NOTCH signaling with DAPT restored Cyp26b1 expression. Therefore, these experiments demonstrate that JAG1-expressing germ cells could likely activate NOTCH signaling and down-regulate Cyp26b1 in Sertoli cells.

Figure 4.

Effect of JAG1-expressing germ cells on Cyp26b1 expression by Sertoli cells in vitro. A) Expression of JAG1 by A_undiff. Mouse spermatogonia cells (postnatal d 8) express high levels of JAG1, as seen by immunofluorescence microscopy (green, arrows). Sertoli cells are negative (asterisk). Germ cell nuclei were specifically stained with a TRA98 antibody (red). B) C18-4 spermatogonia express JAG1. Western blot analysis indicated that mouse JAG1 (MW = 134.16, UniProt) is expressed by C18-4 cells, which are immortalized Type A_undiff spermatogonia. The antibody also recognized human JAG1 (MW = 133.8, UniProt), as well as a putative glycolysated form in MCF7 breast cancer cells and HeLa cells, which were used as positive controls. Effective down-regulation of JAG1 expression (and degradation products) by 2 shRNA lentiviral probes in C18-4 cells (352,710 and 352,711) is also shown. C) Effect of JAG1 on Hes1, Hey1, and Cyp26b1 in Sertoli cell-germ cell cocultures. Primary Sertoli cells were FACS-sorted from Amh-Cre;Rosa^RFP/RFP testes of 21-d-old mice and cultured for another 48 h in 24-well plates and different conditions. DAPT, a NOTCH inhibitor, induced down-regulation of Hes1 and Hey1, as well as up-regulation of Cyp26b1 in Sertoli cells grown without germ cells. When cultured on top of growing C18-4 cells that express JAG1, Sertoli cells increased their expression of Hes1 and Hey1 expression and down-regulated Cyp26b1. However, addition of DAPT to these cocultures returned gene expression to control values. Dhh, Desert Hedgehog. Results are given as means ± sem. *P < 0.05, **P < 0.01.

Cyp26b1 expression is modulated by JAG1

Because JAG1 is highly expressed in undifferentiated spermatogonia in vivo (47) (Fig. 4A), we tested if various concentrations of this NOTCH ligand would influence Cyp26b1 expression in primary Sertoli cells in vitro. We therefore cultured postnatal Sertoli cells with increasing concentrations of immobilized JAG1 and quantified the expression of the canonical NOTCH targets Hes1 and Hey1, as well as the expression of the potential HES/HEY target Cyp26b1. As shown in Fig. 5A, JAG1 triggers a dose-dependent up-regulation of Hes1 and Hey1 expression, which is concomitant with a down-regulation of Cyp26b1. In addition, degradation of Hey1 through RNA interference (RNAi) decreases the effects of JAG1 and restores Cyp26b1 expression to control levels (Fig. 5B). These experiments therefore demonstrate that Cyp26b1 expression is indeed down-regulated by the canonical NOTCH-signaling pathway and its effector HEY1.

Figure 5.

Cyp26b1 expression is dependent on the NOTCH ligand JAG1. A) Influence of JAG1 on Hes1, Hey1, and Cyp26b1 expression. Sertoli cells (21-d old) were cultured for 48 h with different concentrations of JAG1 immobilized onto the bottom of 96-well plates prior coating with Matrigel. Expression of Hes1 and Hey1, quantified by qRT-PCR, was dose-dependent and increased as the concentration of JAG1 increased, whereas expression of Cyp26b1 is inversely related. HeyL, hairy/enhancer-of-split related with YRPW motif-like. B) JAG1-induced decrease in CYp26b1 expression in Sertoli cells is ablated through Hey1 siRNA-mediated knockdown. Freshly isolated Sertoli cells (21-d old) were cultured in the presence or absence of immobilized JAG1 prior to treatment with control (scrambled) or Hey1 siRNAs. Hey1 is increased and Cyp26b1 is decreased by JAG1. However, knockdown of Hey1 produces the opposite effect, confirming that HEY1 downregulates Cyp26b1 expression. Rr, Rat recombinant. For each graph, results were obtained from 3 independent experiments and are given as means ± sem. *P < 0.05, ***P < 0.005.

A_aligned spermatogonia down-regulate Cyp26b1 expression through NOTCH signaling in Sertoli cells in vivo

To test which germ cells truly activate NOTCH signaling in vivo, we performed germ cell–specific ablation followed by time-dependent recovery after a single moderate dose of busulfan (20 mg/kg) in adult mice. It is well known that germ cell depletion affects Sertoli cell transcript levels (49) and kinetics of germ cell recovery after busulfan treatments have been previously described (50–54). We therefore treated Amh-Cre;Rosa^RFP/RFP adult mice with 20 mg/kg busulfan and isolated Sertoli cells by FACS at 4, 8, 16, and 24 d after busulfan injections. Expression data of NOTCH targets at 4, 16, and 24 d after treatments were not significantly different from the control mice without busulfan (data not shown). However, a significant decrease of Hes1, Hey1, and HeyL together with an increase in Cyp26b1 was achieved with Sertoli cells isolated 8 d after treatment (Fig. 6A). Busulfan treatment of W/Wv mice, which are devoid of germ cells but harbor normal Sertoli cells (55), did not cause any significant changes in expression of the Hes/Hey and Cyp26b1 genes (Fig. 6B). This indicates that the presence of a subset of germ cells is critical for modulating NOTCH signaling in Sertoli cells. Further, our histology data shown in Fig. 7A indicated that at d 8 after busulfan treatment, the germ cell populations that are the most depleted are In- to B spermatogonia (5% of control) and preleptotene/zygotene spermatocytes (5% of control), although other populations were depleted as well (A_aligned = 17% of control, A₁–A₄ = 17% of control, leptotene-zygotene = 7% of control, early pachytene = 39% of control). No changes were observed in the numbers of pachytene spermatocytes (Fig. 7B) or later stages (not quantified). Figure 7C expresses the ratio of germ cells/Sertoli cells in treated vs. nontreated testes. Further, the expression of germ cell–specific markers, obtained from FACS-isolated germ cells from the contralateral testes, indicated that A_single and A_paired spermatogonia had recovered from the busulfan treatment at this time point because there were no significant changes in Id4 and Nanos2 (Fig. 7D). The reduction of the numbers of A_aligned and differentiating spermatogonia was confirmed by reduced levels of c-Kit, Plzf, Dazl, and Sohlh1, and the maintenance of numbers of pachytene spermatocytes and spermatids was confirmed by the levels of Tekt1. There was also a significant decrease in Jag1 expression, which is normally expressed in all spermatogonia, albeit at different levels (Fig. 7E) (48).

Figure 6.

Busulfan treatment induces temporal down-regulation of NOTCH signaling in Sertoli cells in vivo. A) Ablation of premeiotic germ cells influences NOTCH signaling in Sertoli cells. Mice (Amh-Cre; ROSA^RFP/RFP) were treated intraperitoneally with a 20 mg/kg dose of busulfan to temporally eliminate germ cells. Control mice were treated intraperitoneally with DMSO vehicle only. Sertoli cells were isolated from testes at different time points after injection using FACS to quantify gene expression by qRT-PCR. Eight days after injection, a significant decrease of Hes1, Hey1, and HeyL expression together with an increase in Cyp26b1 was observed in comparison to Sertoli cells treated with DMSO only. Glial cell line–derived neurotrophic factor (Gdnf) was used as a gene expression control (48). B) Sertoli cells in testes permanently devoid of germ cells do not respond to busulfan treatment. W/Wv mice, which are devoid of germ cells, were treated with a 20 mg/kg dose of busulfan or with DMSO vehicle. Sertoli cells were isolated at different time points after injection using lectin-coated dishes to study gene expression. No significant changes in NOTCH-related gene expression were observed, indicating the importance of germ cells in regulating NOTCH signaling in Sertoli cells. Number of mice in experimental and control groups, n = 8. For each graph, results are given as means ± sem. *P < 0.05, **P <0.01, ***P < 0.005.

Figure 7.

Effects of busulfan on subsets of germ cells and kinetics of germ cell recovery. A) Quantitation of spermatogonia 8 d after busulfan treatment. Amh-Cre; ROSA^RFP/RFP mice were treated with 20 mg/kg busulfan or with DMSO (controls), and testes retrieved 8 d after treatment, when NOTCH signaling in Sertoli cells was significantly down-regulated. Testes were evaluated using histology. Results demonstrate that there is a significant decrease in the number of spermatogonia/Sertoli cell at this time point. B) Quantitation of spermatocytes 8 d after busulfan treatment. Mice were treated and testes evaluated as above. Results demonstrate that there is a significant decrease in the number of preleptotene to zygotene spermatocytes in comparison to controls but no change in the number of pachytene spermatocytes at this time point. C) Ratio of germ cells/Sertoli cells in treated vs. nontreated testes. Results recapitulate graphs A and B, but they also clearly indicate that at stage VIII of the seminiferous epithelium, there is loss of A_undiff and preleptotene/zygotene spermatocytes that might be determinant for down-regulation of NOTCH signaling in Sertoli cells. D) Germ cell gene expression analysis at d 8 after busulfan treatment. Germ cells (entire population) were isolated from busulfan-treated and control testes at d 8 after treatment using FACS and analyzed by qPCR. Results indicate depletion of undifferentiated (Oct4, Gfra1, Plzf) and differentiating spermatogonia (Kit, Sohlh1). SSC (Id4, Nanos2) numbers have recovered at this time point. E) Expression analysis of Jag1 at d 8 after busulfan treatment. Germ cells (entire population) were isolated and analyzed as above. Results demonstrate depletion of Jag1 expression concomitant with the decrease of c-Kit and Gfra1 expression. This indicates that Jag1 is expressed by undifferentiated and differentiating spermatogonia, which drive NOTCH signaling in Sertoli cells. A–D) Results were obtained from n = 8 busulfan-treated mice and n = 8 control mice. E) Results were obtained from n = 10 busulfan-treated mice and n = 5 control mice. Avg, average; Bus, busulfan; Cont/Contl, control; Int, intermediate; Pl/prelept, preleptotene; Zyg/Zygot, zygotene. For each graph, results are given as means ± sem. *P < 0.05, **P < 0.01.

Although these cellular changes occurred throughout the cycle of the seminiferous epithelium (Fig. 7C), the levels of CYP26B1 are most important around stage VIII, which is the stage at which the germ cells depend on RA for progression (31). Our results clearly show that around stage VIII of the seminiferous epithelium, there is loss of A_aligned and preleptotene/zygotene spermatocytes. Because preleptotene/zygotene spermatocytes do not express JAG1 (48), we can assume that the germ cells that normally activate NOTCH signaling in Sertoli cells around stage VIII of the seminiferous epithelium cycle belong only to the A_aligned spermatogonia population. Thus, this subset of germ cells down-regulates Cyp26b1 around stage VIII and allows RA activity at the transition between A_undiff and A₁ spermatogonia, preleptotene and leptotene spermatocytes, and round and elongating spermatids.

DISCUSSION

Because both an excess and lack of RA are detrimental for most tissues, regulation of its concentration by CYP26 enzymes needs to be appropriately controlled (56–60). In the seminiferous epithelium, the major CYP26 enzyme that hydrolyzes RA is CYP26B1 (18, 32). However, despite its importance for RA degradation in the testis, the molecular control of spatiotemporal CYP26B1 expression is still poorly understood. In the male embryonic gonad, RA is synthesized by Sertoli cells but is readily degraded by CYP26B1 in order to maintain proper testis development and prospermatogonia in an undifferentiated state (58). However, induction of the spermatogenetic program shortly after birth absolutely requires RA (61–63), therefore implying down-regulation of CYP26B1 production. This is accompanied by an increase of RA provided to the germ cells, which then express STRA8 and differentiate toward meiosis. In the present study, we demonstrate that the canonical NOTCH signaling pathway, specifically through the transcriptional repressor HEY1, directly down-regulates Cyp26b1 expression in adult Sertoli cells. Previous studies had demonstrated that FOXL2 is an antagonist of Cyp26b1 expression in developing ovaries, ensuring prenatal exposure to RA for entry into meiosis and initiation of the oogenic pathway (16). Our own studies of embryonic testes using a model of NOTCH signaling constitutively activated in Sertoli cells demonstrated a decrease of Cyp26b1 expression and premature prospermatogonial differentiation (33, 44). However, a conclusive link to NOTCH signaling, in particular in the postnatal and adult testis, was missing. We also previously demonstrated that tissue-specific inactivation of NOTCH signaling in Sertoli cells (Amh-Cre; Rbpj^fl/fl mice) caused an increase in the number of undifferentiated spermatogonia in the first round of spermatogenesis and beyond (34). Although this could be the consequence of an increase in GDNF (48), an increase of CYP26B1 expression might also cause delayed differentiation.

The nature and extent of cell-cell contacts is an important factor that shapes the NOTCH response. NOTCH is one of several signaling pathways that are critical components of the stem cell niche in different organs, including the hematopoietic system, gut, mammary gland, and muscles. However, which cells produce the ligand and which cells receive the signal is not always evident and depends on the organ and species studied. In the Caenorhabditis elegans germ line, the distal tip somatic cell creates the niche that provides the NOTCH ligand LAG-2 and its homolog APX-1, while the germline stem cells express the NOTCH receptor GLP-1 (64). In this case, as long as the cells are in physical contact, NOTCH signaling will inhibit germ cell differentiation (65). By comparison, in Drosophila, overexpression of NOTCH ligands in the germline, or activated NOTCH in the somatic cells, results in extra niche cells that are able to support more germ cells through expression of TGFB1 (66). Similar to Drosophila, our data in the mouse testis show that germ cells are at the signal sending end, but not the receiving end, of the NOTCH pathway (Fig. 8A). In addition, our previous data indicated that in the mouse testis, the main NOTCH ligand presented by germ cells is likely JAG1 (48). This suggests that germ cells, by inducing NOTCH activation in Sertoli cells through JAG1, trigger their own differentiation through down-regulation of CYP26B1 production. Conversely, when germ cells do not produce functional JAG1, the NOTCH pathway is not activated, CYP26B1 is present, and the germ cells remain undifferentiated.

Figure 8.

Proposed role of NOTCH signaling on CYP26B1 expression in Sertoli cells. A) Ligand-mediated activation of NOTCH signaling in Sertoli cells. When the NOTCH ligand JAG1, expressed at the membrane of A_undiff, binds to the NOTCH receptor at the surface of Sertoli cells, a series of proteolytic cleavages (S2 and S3 cleavages) occur that cleave the NOTCH intracellular domain (NICD). NICD translocates into the nucleus and binds to the transcription factor RBPJ at the promoter of target genes. This allows the recruitment of the nuclear protein MAML and the histone acetyltransferase protein p300, which results in the transcription of the canonical NOTCH target genes Hes and Hey, which are transcriptional repressors. Adapted from Leong and Karsan (69). B) Regulation of CYP26B1 expression by Sertoli cells. CYP26B1 is produced by Sertoli cells and normally degrades RA. However, as the number of A_undiff increases, in particular A_aligned spermatogonia, more JAG1 ligand is available to activate NOTCH signaling in Sertoli cells. Activated NOTCH will down-regulate the expression of Cyp26b1, which allows RA to trigger the transition from A_undiff to A_diff.

In the adult testis, germ cell responses to RA are occurring at stage VIII of the seminiferous epithelium, and it has been proposed that pulses of RA are triggered around this stage by somatic cells and germ cells to allow proper germ cell differentiation and maturation (5, 31, 67). The modulation of these pulses is yet unclear. In the adult testis, CYP26B1 is produced by peritubular myoid cells (68) and at a lower level by Sertoli cells and germ cells. Elimination of its activity within both germ cells and Sertoli cells induces loss of germ cells and severe subfertility, demonstrating a clear function for this enzyme in the adult seminiferous epithelium (32). Although the question of which cells produce RA in the adult seminiferous epithelium is still debated, it has been recently demonstrated that Sertoli cells produce RA necessary for the transition between A_undiff and A_diff spermatogonia, whereas midpachytene spermatocytes produce the RA necessary for induction of the meiotic prophase, spermatid elongation, and spermiation. There is still disagreement over the source of RA necessary for the initiation of meiotic prophase at the leptotene stage (31). Because all premeiotic germ cells express some levels of JAG1 (48) and can potentially induce NOTCH signaling in Sertoli cells and down-regulate Cyp26b1 expression, we sought to identify the subset of adult germ cells that initiate the signal, in particular at stage VIII. By performing busulfan ablation of specific germ cell types in the adult, we demonstrate in this study that premeiotic germ cells, which express JAG1, are able to activate NOTCH in Sertoli cells; therefore, the extent of the response may depend on the total number of JAG1 ligand molecules presented. Further, by precisely assessing the identity and numbers of germ cells per Sertoli cell at each stage of the seminiferous epithelium in the presence or absence of busulfan and by comparing with Sertoli cell Cyp26b1 expression, we demonstrate for the first time that the germ cell population that is the most likely to present enough JAG1 signal to decrease Cyp26b1 expression and allow RA function at stage VIII of the seminiferous epithelium cycle is the pool of accumulating A_aligned spermatogonia. This may answer the elusive question of how RA is modulated spatiotemporally. This implies a negative feedback loop wherein germ cells regulate their own homeostasis (Fig. 8B). Lastly, it is important to note that SSCs and their direct progeny also express some levels of CYP26B1 to possibly prevent early differentiation (32), which might counteract the effects of NOTCH signaling on RA availability in these cells. Further understanding of how Cyp26b1 expression is regulated in subsets of germ cells across the stages of the seminiferous epithelium is therefore an important future endeavor.

In summary, we have determined that the SSC niche is maintained through negative feedback regulation of the A_undiff germ cells. These cells express the NOTCH ligand JAG1, enhancing NOTCH receptor activation in Sertoli cells. This up-regulates the HEY1 transcriptional repressor and down-regulates Cyp26b1 expression (Fig. 8B). The accumulation of A_aligned spermatogonia up until stage VII–VIII of the seminiferous epithelium cycle triggers down-regulation of Cyp26b1 in Sertoli cells, contributing to maintaining RA levels to induce the transition between A_undiff and A₁ spermatogonia. Altogether, our results demonstrate that the effects of JAG1 on NOTCH signaling in Sertoli cells in vivo are strictly dependent on ligand concentration.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.

Glossary

A_diff

A1 to A4 spermatogonia

APX-1

Anterior pharynx in excess protein-1

A_undiff

undifferentiated type A spermatogonia

AMH

anti-Mullerian hormone

ChIP

chromatin immunoprecipitation

CMV

cytomegalovirus

Cre

Cre-recombinase (P1 bacteriophage)

CYP26B1

cytochrome P450 family 26 subfamily B member 1

DAPT

N-[(3,5-difluorophenyl)acetyl]-l-alanyl-2-phenyl]glycine-1,1- dimethylethyl ester

DAZL

Deleted in azoospermia-like

DHH

Desert Hedegehog

E-box

enhancer box

FACS

fluorescence-activated cell sorter

FBS

fetal bovine serum

GDNF

Glial cell line-derived neurotrophic factor

GFP

green fluorescent protein

GFRA1

GDNF Family Receptor α1

GLP-1

(Abnormal) Germ Line Proliferation-1

HEK

human embryonic kidney

HES1

Hairy and Enhancer-of-Split 1

HEYL

Hairy/Enhancer-of-Split Related with YRPW motif-like

HEY1

Hairy/Enhancer-of-Split Related with YRPW Motif 1

JAG1

JAGGED-1

KIT

KIT proto-oncogene receptor tyrosine kinase

LAG-2

(Lin-12 And Glp-1)-2

N-box

N-box element

NANOS2

Nanos C2HC-type zinc finger 2

NICD

NOTCH intracellular domain

OCT4

octamer-binding transcription factor 4 (POU5F1)

PLZF

ZBTB16, zinc finger and BTB domain containing 16

qPCR

quantitative PCR

qRT-PCR

quantitative RT-PCR

RA

retinoic acid

RBPJ

Recombination Signal-Binding Protein for Immunoglobulin κJ Region

RFP

red fluorescent protein

SF-1

Steroidogenic Factor 1

shRNA

short hairpin RNA

siRNA

small interfering RNA

SOHLH1

Spermatogenesis and oogenesis specific basic helix-loop-helix 1

SOX9

Sex-Determining Region Y Box 9

SSC

spermatogonial stem cell

STRA8

Stimulated by Retinoic Acid Gene 8

SV40

Simian Virus 40

SYCP3

Synaptonemal Complex Protein 3

TEKT1

Tektin 1

TGFB1

Transforming growth factor, β1

TNR-GFP

transgenic NOTCH reporter-green fluorescent protein

W/Wv

KitW/KitWv, Kit dominant spotting/viable dominant spotting

YFP

yellow fluorescent protein

AUTHOR CONTRIBUTIONS

P. A. Parekh contributed to conception and design, the acquisition and analysis of data, drafting and revision of the manuscript, and final approval; T. X. Garcia contributed to conception and design, the acquisition and analysis of data, revision of the manuscript, and final approval; R. Waheeb, V. Jain, P. Gandhi, M. L. Meistrich, and G. Shetty contributed to conception and design, the acquisition and analysis of data, revision of the manuscript, and final approval; and M.-C. Hofmann contributed to conception and design, the analysis of the data, revising and finalizing the manuscript and figures, final approval, and funding of research.