BIOLOGY OF REPRODUCTION 78, 537–545 (2008)Published online before print 21 November 2007.DOI 10.1095/biolreprod.107.064337

Expression of Stimulated by Retinoic Acid Gene 8 (Stra8) and Maturation of MurineGonocytes and Spermatogonia Induced by Retinoic Acid In Vitro1

Qing Zhou,3 Ying Li,3 Rong Nie,3 Patrick Friel,3 Debra Mitchell,3 Ryan M. Evanoff,3 Derek Pouchnik,3

Brent Banasik,3 John R. McCarrey,4 Christopher Small,3 and Michael D. Griswold2,3

School of Molecular Biosciences,3 Washington State University, Pullman, Washington 99164Department of Biology,4 University of Texas, San Antonio, Texas 78249

ABSTRACT

Vitamin A deficiency in the mouse results in an arrest in theprogression of undifferentiated spermatogonia to differentiatingspermatogonia. The supplement of retinol to vitamin-A-deficientmice reinitiates spermatogenesis in a synchronous mannerthroughout the testes. It is unclear whether the effects ofretinoids are the result of a direct action on germ cells or areindirectly mediated through Sertoli cells. The expression ofStimulated by retinoic acid gene 8 (Stra8), which is required forspermatogenesis, is directly related to the availability of retinoicacid (RA). Analysis of gene expression by microarrays revealedmoderate levels of Stra8 transcript in gonocytes and high levelsin A and B spermatogonia. Stra8 mRNA levels were greatlyreduced or absent in germ cells once they entered meiosis. Thisstudy examined the effect of retinoic acid on cultured neonataltestes and isolated gonocytes/spermatogonia in vitro. THY1+ andKIT+ germ cells were isolated by magnetic-activated cell sortingfrom the testes of mice of different ages. Isolated germ cells werecultured and treated with either vehicle (ethanol) or RA withoutfeeder cells. We found that 1) Stra8 is predominantly expressedin premeiotic germ cells, 2) RA stimulates gonocyte DNAreplication and differentiation in cultured neonatal testes, 3) inthe absence of feeder cells, RA directly induces the transition ofundifferentiated spermatogonia to differentiating spermatogoniaby stimulating Stra8 and Kit gene expression, 4) RA dramaticallystimulates Stra8 expression in undifferentiated spermatogoniabut has a lesser impact in differentiating spermatogonia, 5)endogenous Stra8 gene expression is higher in differentiatingspermatogonia than in undifferentiated spermatogonia andcould mediate the RA effects on spermatogonial maturation,and 6) RA stimulates a group of genes involved in themetabolism, storage, transport, and signaling of retinoids.

differentiation, gonocytes, in vitro, retinoic acid, spermatogenesis,spermatogonia, spermatogonial differentiation, Stra8

INTRODUCTION

In many adult male mammals, sperm production is acontinuous process in which millions of sperm are generatedeach day, throughout life. Sperm production is truly one of the

most dynamic cell-producing systems in the body, and itdepends on the active proliferation and differentiation ofspermatogonial stem cells (SSC) [1]. SSC originate fromprimordial germ cells, which, in turn, become gonocytes oncethey migrate to the genital ridges and become enclosed bydifferentiating Sertoli cells [2]. In rodents, gonocytes divide fora few days and then arrest in the G0/G1 phase of the cell cycle.The gonocytes arrested in mitosis migrate from the center ofthe cords to the peripheral region and reach the basementmembrane. In mice, this migration is completed within the firstfew days after birth, and the gonocytes resume proliferationand gonocytes/prespermatogonia differentiate into SSC in thesame period [1, 2].

In rodents, SSC are single cells located on the basementmembrane of the seminiferous tubules and are called A-single(As) spermatogonia. The SSC divide either into two new singlecells or into a pair of spermatogonia (Apr) that do not completecytokinesis and stay connected by an intercellular bridge. TheApr spermatogonia divide further to form chains of 4, 8, andoccasionally, up to 32 A-aligned (Aal) spermatogonia. The Aal

spermatogonia go through a differentiation step and become A1

spermatogonia [3]. In rodents, there are five divisionsfollowing A1 formation, forming successively A2, A3, A4, In(intermediate), and B spermatogonia. The B spermatogoniadivide into spermatocytes. Each step of differentiation exceptthe transition of Aal to A1 is associated with a mitotic division.As spermatogonia are considered the only true SSC, eventhough As, Apr, and Aal are termed ‘‘undifferentiatedspermatogonia,’’ whereas spermatogonia from A1 to B aretermed ‘‘differentiating spermatogonia’’ [3].

The massive production of mature sperm is dependent onsuccessive mitotic divisions and associated cell differentiationof spermatogonia. A few factors involved in the regulation ofspermatogonial differentiation have been identified [4],including retinoic acid (RA) [5, 6], cyclin D2 (CCND2) [7],KIT/KITL [8, 9], and DAZL [10]. It has been known since1925 that vitamin A is required for normal spermatogenesis.Spermatogenesis ceases and only A spermatogonia are found invitamin-A-deficient (VAD) mouse testes. Supplementingvitamin A (retinol) to VAD mice reinitiates spermatogenesisin a synchronous fashion throughout the testis [11]. Haneji etal. [12] showed that both retinol and RA could inducedifferentiation of type A spermatogonia in cultured mousecryptorchid testes. This conclusion was based on themorphological identification of In and type B spermatogoniaafter 9 days of culture with retinoids [12]. However, it is notknown whether the action of RA in inducing spermatogonialdifferentiation is a direct effect on germ cells or an indirecteffect mediated through Sertoli cells, because both spermato-gonia and Sertoli cells possess nuclear receptors for retinoids[2, 13]. Moreover, the molecular mechanism of RA action inspermatogonia is obscure. In a recent study, RA was

1Supported by NIH grants R01 HD 10808 and U54 HD 042454. Themicroarray data have been deposited in the National Center forBiotechnology Information Gene Expression Omnibus database (seriesno. GSE9521).2Correspondence: Michael D. Griswold, 531 Fulmer Hall, School ofMolecular Biosciences, Washington State University, Pullman, WA99164-4660. FAX: 509 335 9688; e-mail: mgriswold@wsu.edu

Received: 17 July 2007.First decision: 10 August 2007.Accepted: 2 November 2007.� 2008 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

537

demonstrated to induce the expression of Stra8, a factoressential for the onset of meiosis in both male and femalegonads [14]. The knockout of the Stra8 gene blocks entry intomeiosis in both embryonic ovaries and pubertal testes [14, 15].Thus, the expression of Stra8 appears to be a sensitive indicatorof the availability of RA and entry of cells into meiosis.

Several different methods to collect spermatogonia enrichedin SSC have been established, with one of the most effectiveways being magnetic-activated cell sorting (MACS) directedagainst cell surface markers [16]. THY1 is likely a commonsurface marker for all undifferentiated spermatogonia, andTHY1-positive selection using MACS results in significantenrichment of SSC [16]. This technique allows the examinationof factors that potentially regulate the proliferation anddifferentiation of spermatogonia in a controlled, in vitrosystem [17]. Our study utilized short-term cultured neonatalmouse testes and specific MACS subpopulations of spermato-gonia using Stra8 and Kit as endpoints to test two objectives: 1)determine if RA can induce the in vitro differentiation of earlygerm cells in the absence of somatic cells and 2) identify thesubpopulation of spermatogonia in which RA induces Stra8expression and spermatogonial differentiation.

MATERIALS AND METHODS

Animal Care and Treatment

All animal experiments were approved by Washington State UniversityAnimal Care and Use Committees and were conducted in accordance with theguiding principles for the care and use of research animals of the NationalInstitutes of Health. A BL/6–129 mouse colony was maintained in atemperature- and humidity-controlled room with food and water provided adlibitum. To investigate the induction of STRA8 protein expression by RA inneonatal gonocytes, 2-day-postpartum (dpp) male mice were injectedsubcutaneously with all-trans-RA (90% oil, 10% ethanol) at a dose of 17.5lg/mouse or injected with vehicle as control. Testes were fixed 24 h later infreshly prepared 4% (w/v) paraformaldehyde for immunohistochemical stainingof STRA8 protein. Four mice were used in each treatment group.

Testicular Cell Types and RNA Preparation

Enriched germ cell isolates for microarray analysis were prepared viagravity sedimentation from CD1 mice as previously described [18]. Thesesamples included type A and B spermatogonia, pachytene spermatocytes, andround spermatids from animals 8, 30, and 60 days of age, respectively. Thepurity of all the isolated and cultured cells was determined by morphology. Thepurity of type A and B spermatogonia was .85%, and the purity of all othergerm cell types was .95%. The quality control of the total RNA samples wasperformed as previously described [19]. Two replicates were used in microarrayanalysis.

Microarray Processing

The following method was used for microarray processing of 2 dpp/5 dppgonocytes, type A/type B spermatogonia, pachytene spermatocytes, roundspermatids, and Kit-deficient (W/Wv) mutant testes. Ten micrograms of totalRNA from each sample was used to create the target for the microarray. Thelabeled cRNA was hybridized to Mouse Genome 430 2.0 arrays (Affymetrix,Santa Clara, CA) and stained in accordance with the manufacturer’s standardprotocol. The same procedure was performed on duplicate samples. The arrayswere stained utilizing the Affymetrix GeneChip Fluidics Station 400 andscanned using a GeneArray Scanner 2500A (Agilent Technologies, Palo Alto,CA). The resulting data were viewed and analyzed using GeneSpring software(Agilent Technologies, Santa Clara, CA). All reactions and microarrayhybridization procedures were performed in the Laboratory for Biotechnologyand Bioanalysis I at Washington State University. The different cell type datawas normalized within GeneSpring using the default/recommended normali-zation methods. These include setting of signal values below 0.01 to 0.01, totalchip normalization to the 50th percentile. These normalizations allowed for thevisualization of data based on relative abundance at any cell type or organrather than compared to a specific control value.

A different method was used for microarray analysis of RA-treated 2 dppTHY1þ gonocytes. Approximately 75 ng total RNA from each sample wasused to generate cDNA using the Ovation RNA Amplification System V2(NuGEN, San Carlos, CA) according to the manufacturer’s instructions. Arrayquality was assessed, and images were quantified using GeneChip OperatingSoftware v1.2 (Affymetrix). The data were viewed and analyzed usingGeneSpring software. The criteria for RA-regulated genes included a raw signalvalue of no less than 50 in either control or treated samples, a 2-fold or higherchange in signal between the control and treated samples, and the consistentchanges in duplicates.

Neonatal Testis Organ Culture and Treatments

Neonatal testes from 2 dpp mice were detunicated, cut into four to six smallpieces per testis, placed on Millicell CM filters (Millipore, Bedford, MA)floating on the surface of medium and covered with drops of medium(Dulbecco modified Eagle medium/F12/10% fetal bovine serum) and culturedfor 24 h. The cultured testes were treated with all-trans-RA (ATRA) or vehicle(ethanol). The final concentration of ATRA, 9 cis-RA, 4-oxo-RA, and retinolacetate (ROL) in the culture medium was 0.7 lM [14], and the finalconcentration of ketoconazole was 0.94 lM. After 24 h, tissues were harvested,and RNA was isolated using the PicoPure system (Arcturus Bioscience,Mountain View, CA). For studies involving RNA, three vehicle-treated andthree ATRA-treated cultured testes were used. For studies of 5-bromo-20-deoxyuridine (BrdU) incorporation, the vehicle control group contained fivetestes, and the RA-treated group contained six cultured testes. The finalconcentration of BrdU in the culture medium was 10 lM. The testis tissueswere fixed in Bouin solution after treatment.

Immunohistochemical Double Staining of BrdU and GermCell Nuclear Antigen 1

Tissues were fixed in Bouin solution and embedded in paraffin. Tissueswere dewaxed and incubated in 2N HCl at 378C for 30 min. The stainingprocedure was performed as previously described, with minor modifications[19]. Briefly, mouse BrdU antibody (1:200 dilution, Vision BioSystems,Norwell, MA) was applied first and incubated at 48C overnight. After threewashes in PBS, tissues were incubated with secondary antibody (tetramethyl-rhodamine-isothiocyanate-conjugated donkey anti-mouse IgG, 1:100; JacksonImunoResearch Laboratory, West Grove, PA) at 378C for 1 h. Tissues wereblocked with 10% donkey serum and incubated with rat germ cell nuclearantigen 1 (GCNA1) antibody (1:50 dilution; gift from Dr. George Enders of theUniversity of Kansas) at 378C for 2 h, and then secondary antibody (1:100dilution; fluorescein-isothiocyanate-conjugated donkey anti-rat IgG; JacksonImunoResearch Laboratory, West Grove, PA) was applied at 378C for 1 h.Sections incubated in PBS without primary antibody were used as negativecontrols for color development on the same slide. Images were captured with anOlympus OLY-200 digital camera (Olympus America, Inc., Melville, NY)using Olympus MagnaFire Camera Imaging and Control (version 1.0; OlympusAmerica) and compiled using Adobe PhotoShop 7.0 (Adobe Systems, SanJose, CA). The number of GCNA1-positive cells in each testis section wascounted first. The GCNA1 and BrdU staining images were overlapped inPhotoShop, and double positive cells were counted. Finally, the percentage ofBrdU-positive germ cells per total germ cell number was calculated. TheStudent t-test was used for statistical analysis, and P , 0.05 was considered tobe statistically significant.

Production of Anti-STRA8 Serum

To generate STRA8 antibody, a peptide termed P3 and corresponding to theC-terminus (C L N Q E P E P P D D) of mouse STRA8 protein was synthesizedand conjugated with carrier protein maleimide activated keyhole limpethemocyanin (mcKLH, Pierce, Rockford, IL) following manufacturer’sinstructions. At Day 0, preimmune serum was collected from a New Zealandrabbit, then 1 mg of conjugated protein in 0.7 ml of PBS mixed with 0.7 ml ofFreund complete adjuvant was injected into the rabbit subcutaneously. At Day28, the rabbit was given a booster of 0.5 mg conjugated protein in Freundincomplete adjuvant. The STRA8 antiserum was harvested 2 wk after thebooster injection.

Western Blot Analysis, Immunohistochemistry, andAntibody Competition

The specificity of the STRA8 antibody was tested in three ways: a Westernblot analysis using P19 cells, immunohistochemistry performed on normaladult mouse testes and on Stra8-deficient mouse testes, and antibody

538 ZHOU ET AL.

competition with P3 peptide (corresponding to the C-terminus of STRA8protein and used in STRA8 antibody generation) in Western blot andimmunohistochemistry. Briefly, P3 peptide (300 lg/ml) was incubated togetherwith STRA8 antibody overnight at 48C, then the mixture was used as primaryantibody as described later. Protein from the murine embryonic carcinoma cellline P19 cells (ATCC, Manassas, VA) was used for Western blot analysisbecause Stra8 was first identified in this cell line [20]. P19 cells were culturedbased on the instructions of ATCC and were treated with either vehicle ethanolor 1 lM RA (final concentration in the culture medium) for 24 h. Protein washarvested in 23 Laemmli buffer and then separated by 12% SDS-PAGE andtransferred to nitrocellulose membrane (0.45 lm). The membrane was blockedin PBS containing 5% nonfat milk and 0.025% Tween 20 for 1 h at roomtemperature and incubated with STRA8 antiserum (1:500) overnight at 48C.Total proteins from MSC1 cells (a Sertoli cell line) were used as negativecontrol. After incubation, membranes were rinsed with PBS with Tween 20 andincubated with horseradish peroxidase conjugated goat anti-rabbit IgG for 1 hat room temperature. Reactive proteins were visualized and exposed to films byWestern lighting chemiluminescence reagent (Perkin Elmer, Boston, MA). Forimmunohistochemistry, serial sections of the mouse testis incubated withpreimmune serum (1:500) were used as negative controls.

Immunohistochemistry of STRA8 Protein in Neonatal Testes

Neonatal testes were removed 24 h after RA injection and fixed in freshlyprepared 4% (w/v) paraformaldehyde. Antigen retrieval in 0.01 M sodiumcitrate, pH 6.0, was done using microwave radiation and a trypsin (1 mg/ml)digestion for 7 min at 378C. For immunohistochemical staining, ZYMED LAB-SA SYSTEM (Zymed Laboratory, South San Francisco, CA) was used,following the manufacturer’s instructions. Briefly, tissue sections wereincubated with 10% goat serum followed by incubation with STRA8 antiserum(1:2000) or preim (1:2000) at 48C overnight. The slides were incubated withdiluted biotinylated secondary antibody for 45 min, followed by the applicationof streptavidin-horseradish peroxidase. The diaminobenzidine solution wasapplied until brown color developed. Staining was observed with a NikonMicrophot-FX (Meridian Instrument Company, Kent, WA) microscope.Photographs were taken with an Olympus OLY-200 digital camera usingOlympus MagnaFire Camera Imaging and Control version 1.0 (OlympusAmerica).

Real-Time RT-PCR

A two-step real-time RT-PCR was used to measure the expression ofcandidate genes as previously described [19]. Each RNA sample was analyzedin triplicate with primers for each gene. Stra8 primers amplify a 151-bp product(primers: 50-GTTTCCTGCGTGTTCCACAAG-30 and 50-CACCCGAGGCT-CAAGCTTC-30); control Rps2 primers amplify a 112-bp product (primers: 50-CTGACTCCCGACCTCTGGAAA-30 and 50-GAGCCTGGGTCCTCTGAA-CA-30); Kit primers amplify a 100-bp product (primers: 50-GATGGCG-TTCCTCGCCT-30 and 50-GCCCGAAATCGCAAATCTTT-30). Expression ofStra8 and Kit were normalized to Rps2 expression. The Student t-test was usedto analyze the results from neonatal testes, and a pair-wise t-test was used toanalyze the results of all experiments using isolated cells. Data represent mean6 SEM.

MACS of Gonocytes/Spermatogonia

Testes were collected from 2 dpp, 4 dpp, or 8 dpp mouse pups into Hanksbalanced salt solution, and the tunica were removed. MicroBeads with mouseCD117 (KIT) and CD90 (THY1) antibodies and MS columns (Miltenyi Biotec,

Auburn, CA) were used in the isolation according to manufacturer’sinstructions. The culture was performed based on Kubota and Brinster’smethod with modifications [16]. For experiments using 8 dpp testes, the cellswere always subjected to THY1þ MACS first, then the THY1– cells weresubjected to KITþ MACS to obtain two separate cell populations. Isolated cellswere cultured under feeder-cell-free, serum-free, and growth-factor-freeconditions for 8 or 24 h with only vehicle (ethanol) or ATRA (0.7 lM finalconcentration) or ROL (0.7 lM final concentration). For each cell isolation,twelve to fifteen 2 dpp, ten to twelve 4 dpp, and six to ten 8 dpp mice were usedto ensure a high enough number of cells for the subsequent experiments. Eachtreatment group contained at least three independent repeats.

STA-PUT Gonocyte Isolation

Testes from 2 dpp mice were enzymatically digested as described byO’Brien et al. [21] with modifications. Gonocytes or spermatogonia wereidentified by phase contrast microscopy and pooled. The identity of the cellswas confirmed by GCNA1 immunocytochemistry, and 80%–90% of the cellsobtained by STA-PUT were GCNA1 positive (data not shown).

Immunocytochemistry of POU5F1 (OCT4) and GCNA1

Spermatogonia isolated from 8 dpp testes by MACS (THY1 or KITpositive) were fixed in 4% paraformaldehyde (with 0.25% glutaraldehyde) forPOU5F1 staining and in Bouin solution for GCNA1 staining. After a few PBSwashes, cells were incubated with 3% H2O2 for 20 min at room temperatureand then blocked with 10% normal goat serum (for POU5F1) or 10% normalrabbit serum (for GCNA1) for 30 min (goat anti-rabbit IgG for POU5F1staining, Zymed; rabbit anti-rat IgM for GCNA1 staining, Cortex Biochem, SanLeandro, CA). After PBS washes, primary antibodies, either rabbit polyclonalanti-POU5F1 (1:100; Abcam, Cambridge, MA) or GCNA1 (1:50), were addedand incubated at 378C for 2 h, followed by a PBS rinse and biotin-conjugatedsecondary antibody incubation at 378C for 1 h. Diaminobenzidine was used assubstrate for peroxidase after streptavidin conjugation. The cells subjected toGCNA1 staining were counterstained with hematoxylin (Sigma-Aldrich, St.Louis, MO), and cells stained for POU5F1 were counterstained with nuclearfast red (Vector, Burlingame, CA).

RESULTS

Stra8 Expression in Several Types of Premeiotic Germ Cells

The gene expression profile of Stra8 mRNA was obtainedby using microarray results from isolated germ cell samples(Fig. 1). Stra8 mRNA was expressed in 2 dpp and 5 dppgonocytes, type A spermatogonia, and type B spermatogonia.The expression level of Stra8 mRNA decreased significantly inmeiotic and postmeiotic germ cells with the lowest levels ofexpression observed in pachytene spermatocytes and roundspermatids. In the adult W/Wv mutant testis, Stra8 expressionwas at background levels.

RA Stimulation of DNA Replication in Vitamin-A-SufficientNeonatal Gonocytes

It has been shown that retinoids can induce DNA replicationand proliferation in arrested A spermatogonia in VAD adult

FIG. 1. Stra8 mRNA expression in differ-ent germ cell types by microarray analysis.Values represent raw signal of Stra8 6 SEM.2 dpp, 2 d post partum gonocytes; 5 dpp, 5d post partum gonocytes; Type A, type Aspermatogonia; Type B, type B spermato-gonia; Pachy, pachytene spermatocytes;Round, round spermatids; W/Wv, adultW/Wv mutant testes.

RA-INDUCED IN VITRO DIFFERENTIATION OF GONOCYTES 539

rats [5]. The effect of RA on gonocytes/prespermatogonia invitamin-A-sufficient, normal neonatal testes was examined.Neonatal testes (2 dpp) were cultured for 24 h with or withoutRA, fixed in Bouin solution, and embedded in paraffin. Tissuesections were double-stained with GCNA1 and BrdU antibod-ies (Fig. 2). The BrdU-positive germ cells were counted and thepercentage of BrdU-positive gonocytes/spermatogonia per totalgerm cell number calculated. The percentage of BrdU-positivegonocytes/spermatogonia per total germ cell number increased17% (44% 6 6% vs. 61% 6 5%; P , 0.01) in the testestreated with RA, compared to vehicle-treated testes.

RA Induction of Stra8 and Kit mRNA Expression in CulturedVitamin-A-Sufficient Neonatal Testes

Previous studies demonstrated that males and females sharean identical meiotic initiation pathway in which RA inducesStra8 expression in both embryonic ovaries and adult testes,and Stra8 is required for the transition into meiosis in bothmale and female germ cells [14, 15]. The action of RA on theexpression of Stra8 mRNA in 2 dpp testes was examined inorgan culture by real-time RT-PCR. ROL (0.7 lM), RA (0.7lM), and 9-cis-RA (0.7 lM), one major retinoic acidmetabolite 4-oxo-RA and a pan-P450 enzyme inhibitorketoconazole (0.9 lM) were utilized in this study (Fig. 3A).In 2 dpp mouse testes cultured on floating membranes, RAsignificantly increased the expression of Stra8 17-fold 24 hrafter treatment (P , 0.01), and 9-cis-RA treatment increasedStra8 expression 8-fold (P , 0.05). Retinol acetate and 4-oxo-RA increased Stra8 expression 7.8-fold and 8.7-fold, respec-tively. Ketoconazole interferes with the degradation of RA byinhibiting the enzymatic activity of CYP26B1 [14]. Ketocona-zole alone did not make a significant change in Stra8expression. However, ketoconazole seems to reduce themagnitude of the effect of RA. KIT has been demonstratedto be a marker for differentiating spermatogonia [8, 9]. Wefound that the expression of Kit transcript was increased 4-foldon average (P , 0.05) after RA treatment for 24 h (Fig. 3B).

RA Induction of STRA8 Protein Expression in Vitamin-A-Sufficient Neonatal Testes

Western blot analysis confirmed that the STRA8 antibodyrecognized a 44–46 Kd protein product in P19 cells with orwithout RA treatment (Fig. 4, lanes 3 and 4) and did not detectany signal in MSC1 cells (Fig. 4, lane 1) as expected. TheSTRA8 band was completely abolished when antibody wasincubated overnight with P3 peptide (Fig. 4, lanes 6 and 7),which was used to generate this antibody. STRA8 proteinexpression was mostly undetectable in the gonocytes of 2 dpp(starting age) testes treated by oil vehicle (for 24 h) byimmunohistochemistry with only 0.03% of the tubules onaverage containing STRA8-positive germ cells (Fig. 5A).However, in 24 h RA-treated 2 dpp testes, STRA8 protein wasinduced in the cytoplasm of some, but not all, gonocytes, andthe percentage of tubules containing STRA8-positive germcells increased to 41% (P , 0.05; n¼ 4; Fig. 5B). Preimmuneserum was used as negative control and did not show anydetectable positive staining. The positive immunostaining ofSTRA8 in spermatogonia in testes of 2 dpp mice was abolishedby incubating antibody with P3 peptide overnight (data notshown). No immunohistochemical staining was detected usingthis antibody on testes from Stra8-deficient mice [15] (data notshown).

RA Induction of Stra8 Expression in Cultured Gonocytes/Prespermatogonia Without Feeder Cells

Previous studies have shown that RA can induce Stra8expression in cultured embryonic testes and ovaries and also inVAD adult testes [14]. Whether this RA induction is a directeffect on germ cells or an indirect effect mediated throughtesticular somatic cells, in particular Sertoli cells, is stillunknown. Neonatal gonocytes were isolated by THY1-positiveselection from 2 dpp and 4 dpp testes. THY1 antigen isconsidered to be a surface marker for male germline stem cells,and THY1-positive selection was used extensively to isolate

FIG. 2. Immunohistochemical doublestaining of GCNA1 and BrdU in culturedneonatal testes. A) Immunostaining ofGCNA1 (green). B) Immunostaining of BrdU(red). C) 40,6-diamidino-2-phenylindolecounterstaining of cell nuclei (blue). D)Overlay of images A, B, and C. Nuclei ofproliferating germ cells appear yellow.Original magnification 3100 (A–D) and3200 (inset).

540 ZHOU ET AL.

spermatogonial stem cells for germ cell transplantationexperiments [16, 22]. The purity of the cell population wasexamined by immunocytochemical staining of GCNA1, andapproximately 90% of the cells used in the followingexperiments were GCNA1 positive. The isolated THY1þ

gonocytes/spermatogonia were cultured under feeder-cell-free,serum-free, and growth-factor-free conditions and treated withretinoids for 8 h. In THY1þ gonocytes isolated from 2 dpptestes, RA and ROL increased the Stra8 expression level 31-fold (P , 0.01) and 17-fold (P , 0.01), respectively (Fig. 6A).In THY1þ gonocytes/spermatogonia isolated from 4 dpp testes,RA and ROL stimulated Stra8 expression 20-fold (P , 0.01)and 16-fold (P , 0.01), respectively (Fig. 6B). Beforeestablishment of THY1 sorting methods, STA-PUT sedimen-tation was used to separate different types of germ cells, basedon their sizes and shapes [18]. To test whether gonocytes/prespermatogonia isolated by STA-PUT would confirm the RAinduction of Stra8 expression observed in THY1-positive cells,isolated gonocytes (85%–90% purity as confirmed by GCNA1staining) from 2 dpp testes were subjected to identical RA/ROLtreatments. Stra8 mRNA expression increased 47-fold by RA(P , 0.05) and 25-fold by ROL in the 8-h treatment (Fig. 6C).

RA Induction of Stra8 Expression in Undifferentiated andDifferentiating Spermatogonia

Premeiotic male germ cells include primordial germ cells,embryonic gonocytes, prespermatogonia in neonatal testes,spermatogonia, and preleptotene spermatocytes. In the sper-matogonial population, As, Apr, Aal (undifferentiated sper-matogonia), and differentiating spermatogonia, including A1–4,AIn, and B, have been described and studied [3]. Weinvestigated whether RA could induce Stra8 expression inspermatogonia, and, if so, in which specific cell type does RAinduce Stra8 expression? Eight-day-old mouse testes containundifferentiated and differentiating spermatogonia. KIT is asurface protein preferentially expressed in differentiating

spermatogonia [8]. THY1-positive sorting has been used toisolate cells enriched with SSC [22], which we believe ispreferentially expressed in undifferentiated spermatogonia.Therefore, we isolated THY1þ and KITþ spermatogonial cellsfrom 8 dpp testes. The isolated cells were subjected to theTHY1þ and KITþ MACS sequentially to minimize potentialvariation. POU5F1 (previously known as OCT4) was used as amarker for undifferentiated spermatogonia [23]. Using immu-nocytochemical staining for GCNA1 and POU5F1, we foundthat 80%–90% of the spermatogonial cells isolated from 8 dpptestes by THY1þ selection were GCNA1þ/POU5F1þ, imply-ing that the majority of these cells are undifferentiatedspermatogonia. Meanwhile, 40%–50% of the KITþ spermato-gonial cells isolated by this means were also POU5F1 positiveand classified as differentiating spermatogonia. The expressionof Stra8 was shown to be 15-fold higher (P , 0.05) indifferentiating (KITþ) spermatogonia than in undifferentiated(THY1þ) spermatogonia. In order to determine if there was a

FIG. 4. Western blot analyses and STRA8 antibody. Western blot resultsof STRA8 (against P3 peptide) antibody with total proteins from P19 cellsand MSC1 cells before and after antibody competition with P3 peptide.Lanes 1, 2, 3, and 4 were exposed to STRA8 antibody only, and lanes 6, 7,and 8 were exposed to anti-STRA8 serum preincubated with P3 peptide.Lane 1: MSC1; lane 2: prestained molecular weight marker 1; lane 3: P19þ vehicle (ethanol); lane 4: P19 þ RA; lane 5: prestained molecularweight marker 2; lane 6: P19 þ RA; lane 7: P19 þ vehicle; lane 8: MSC1.

FIG. 3. Induction of Stra8/Kit expressionby retinoids in cultured neonatal testesexamined by real-time RT-PCR. Culture/treatment duration is 24 h. Values representrelative fold change of Stra8 or Kit 6 SEM.A) Induction of Stra8 expression by differenttypes of retinoids and pan-P450 inhibitorketoconazole in cultured 2 dpp testes. B)Induction of Kit expression by RA incultured 2 dpp testes. Asterisk (*) indicates astatistically significant difference (P , 0.05)between the labeled sample and controlsample. C, vehicle control (ethanol); 9-CIS,9-cis-RA; 4-OXO, 4-oxo-RA; KETO, keto-conazole.

RA-INDUCED IN VITRO DIFFERENTIATION OF GONOCYTES 541

difference between undifferentiated and differentiating sper-matogonia with respect to their abilities to respond to RAstimulation, the isolated spermatogonia were cultured for 8 h inthe absence of feeder cells with or without RA. Stra8expression was increased 15.7-fold (P , 0.05) in THY1þ

spermatogonia in 8 h, whereas there was only a 2.7-foldincrease in Stra8 expression in KITþ spermatogonia (Fig. 7).

RA Induction of Cell Differentiation in CulturedUndifferentiated Spermatogonia

It has been shown that retinoids induce differentiation ofarrested A spermatogonia in VAD rat testes and culturedcryptorchid testes. It is unknown whether retinoids can inducethe differentiation of murine As, Apr, or Aal spermatogonia inthe absence of somatic cells in vitro. To this end, theexpression of Kit after RA treatment under culture conditionswithout feeder cells was examined. Consistent, but notstatistically significant, up-regulation of Kit after 8 h of RAtreatment was observed in THY1þ spermatogonial cells (datanot shown). Because the S phase of A spermatogonia lasts

about 25 h, and differentiation (after a mitotic division) into A2

spermatogonia occurred only 48 h after retinoid stimulation invivo under VAD conditions [5], the RA treatment duration wasextended to 24 h, and 3.6-fold up-regulation (P , 0.05) of Kitexpression in THY1þ spermatogonial cells was detected.

Altered Expression of Genes Involved in Vitamin AMetabolism and Signaling in RA-treated 2 dpp THY1þ

Gonocytes Revealed by Microarray Profiling

Using the restrictions as described in Materials andMethods, a signal level of at least 50 in at least one of thesamples and the expression levels being consistently up- ordown-regulated by 2-fold in the two duplicates, 80 transcriptswere up-regulated, and 97 transcripts were down-regulated byRA in 8 h in THY1þ gonocytes. A group of genes involved inthe transport, storage, metabolism, and signaling of retinoidswas significantly up-regulated by RA stimulation, includingStra6, Lrat, Cyp26a1, Cyp26b1, Por, Dhrs3, Rarb, and Stra8(Table 1). A complete list of the transcripts will be availablethrough the Griswold Lab home page (http://www.wsu.edu/

FIG. 5. Immunohistochemical detection of STRA8 protein induced by RA in cultured neonatal testes. A) Immunostaining of STRA8 after 24 h oil-treated(starting at age 2 dpp mouse testis). B) Immunostaining of STRA8 after 24 h RA treated (starting at age 2 dpp mouse testis). Inset shows a dividing STRA8-positive gonocyte (arrow). Original magnification 3200 (A, B) and 3400 (inset).

FIG. 6. Induction of Stra8 expression by RA/ROL (8 h) in cultured neonatal gonocytes without feeder cells examined by real-time RT-PCR. Valuesrepresent relative fold change of Stra8 expression 6 SEM. A) Induction of Stra8 expression by RA/ROL in THY1þ 2 dpp gonocytes cultured without feedercells for 8 h. B) Induction of Stra8 expression by RA/ROL in THY1þ 4 dpp gonocytes cultured without feeder cells for 8 h. C) Induction of Stra8 expressionby RA/ROL in 2 dpp gonocytes isolated by STA-PUT and cultured without feeder cells for 8 h. Culture medium did not contain serum or any growth factor.Asterisk (*) indicates a statistically significant difference (P , 0.05) between the labeled sample and control sample. C, vehicle control.

542 ZHOU ET AL.

griswold/microarray) and the National Center for Biotechnol-ogy Information Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/projects/geo/query/acc.cgi?acc¼GSE9521).

DISCUSSION

In this study we found that 1) Stra8 was predominantly, ifnot exclusively, expressed in premeiotic germ cells, 2) RAartificially induced Stra8 and Kit expression and increasedgonocyte DNA replication in cultured 2 dpp testes, and 3) RAinduced STRA8 protein in the cytoplasm of some, but not all ofthe 2 dpp gonocytes, 4) RA can directly stimulate Stra8expression in vitro in premeiotic male germ cells withoutsupport from somatic cells, 5) RA induced expression of thespermatogonial differentiation marker gene, Kit, in the samefeeder-cell-free culture environment, 6) RA dramaticallystimulated Stra8 expression in undifferentiated spermatogoniabut had a lesser impact on differentiating spermatogonia, 7) theendogenous level of Stra8 expression is higher in differenti-ating than in undifferentiated spermatogonia, and 8) RAsignificantly stimulated the expression of a group of genesinvolved in the metabolism, storage, transport, and signaling ofretinoids.

Based on these findings, a detectable level of gonocyte/spermatogonia differentiation can be achieved in vitro by RAstimulation in the absence of somatic support. In addition,Stra8 is very likely a critical component mediating the RAregulation of spermatogonial differentiation. Previous studiessuggest that, in the development of early spermatogonia, RAinduces the transition of either Aal to A1 or A1 to A2

spermatogonia [5, 11]. RA induction of Stra8 expression isprobably a vitally important trigger to the subsequent

differentiation steps and commitment toward the formation ofprimary spermatocytes. It is worth noting that different types ofprespermatogonia, including embryonic gonocytes and earlyneonatal gonocytes, were capable of responding to RAinduction as well [14].

The recent findings from several studies revealed that RA isthe signal triggering the onset of meiosis in embryonic ovaries[14, 24], and Stra8 is the downstream factor required for thiscommitment [14]. In the study of Koubova et al. [14], the sameregulatory pathway was also confirmed in adult VAD testes. Inthe male germline, a series of developmental events occurwithin the first 10 days after birth. The male germ cells do notenter meiosis until 8–10 dpp, as evidenced by the firstappearance of preleptotene spermatocytes. This current studydemonstrates RA induction of Stra8 mRNA and STRA8protein in cultured neonatal testes prior to meiotic initiation,which suggests that the first round of meiosis very likely sharesthe same mechanisms in embryonic ovaries and adult testes.Further, the strong induction of DNA replication by RAindicates that RA might promote the differentiation of theseearly gonocytes, because almost every step of spermatogonialdevelopment is associated with a mitotic division [3].However, the possibility that RA also stimulates the self-renewal of these gonocytes cannot be ruled out. RA inductionof Kit expression in these neonatal testes ultimately substan-tiates the hypothesis that RA triggers the differentiation of 2dpp gonocytes, because KIT is proven to be a marker fordifferentiating spermatogonia.

Several lines of evidence suggest that neonatal gonocytesrepresent a heterogeneous population. In the early 1970s,Zamboni and Merchant [25] observed bridges connecting someof the gonocytes, indicating that these cells had alreadydifferentiated. Later, de Rooij [1, 2] proposed that at thebeginning of spermatogenesis some gonocytes give rise to SSC(As), and others directly differentiate to A1 spermatogonia. Amore recent result by Yoshida et al. [23], analyzing the uniqueexpression of Ngn3, strongly indicates that the first round ofspermatogenesis initiates directly from gonocytes, and theysuggest that the state of undifferentiated spermatogonia is notan inevitable step but a developmental option. These recentdata show that if an appropriate signal is given, the directtransition of gonocytes to A1 spermatogonia can be induced inan in vitro culture system. This report suggests that the signalinducing differentiation is RA. In this study, RA inducedSTRA8 protein in some of the 2 dpp gonocytes but not in all ofthem. This is a strong indication of heterogeneity of neonatalgonocytes. It is possible that some of these neonatal gonocytescould directly enter the phase of differentiating spermatogonia,and others would commit to becoming SSC. Koubova et al.[14] demonstrated that RA could artificially induce Stra8 inembryonic testes. However, the period of incubation of theembryonic testes was 2 days, and the possibility of serial,accelerated differentiation events, including the conversion of

TABLE 1. Transcripts involved in retinoid metabolism, storage, transport and signaling that are up-regulated by RA greater than 2-fold in 2 dpp THY1þ

gonocytes.

Genes GenBank accession no. Fold change Function

Lrat BB518473 12 Retinoids absorption and storageCyp26a1 NM_007811 11 Metabolism of RAStra8 NM_009292 9 Onset of meiosisRarb BB266455 9 Nuclear receptor of RADhrs3 NM_011303 6 Metabolism of all-trans-retinalStra6 NM_009291 5 Membrane receptor for retinol binding proteinCyp26b1 AW049789 3 Metabolism of RAPor NM_008898 2 Metabolism of RA

FIG. 7. RA induction of Stra8 expression in THYþ and KITþ spermato-gonia in 8 h without feeder cells. Values represent relative fold change ofStra8 expression 6 SEM. Asterisk (*) indicates a statistically significantdifference (P , 0.05) between the RA-treated sample and control sample.THY1þ, THY1þ spermatogonia; KITþ, KITþ spermatogonia.

RA-INDUCED IN VITRO DIFFERENTIATION OF GONOCYTES 543

embryonic gonocytes to undifferentiated spermatogonia occur-ring after RA stimulation, could not be excluded.

The results from cultured neonatal testes generated twomajor observations. First, in culture systems with intact somaticsupport, RA induces Kit expression, a marker of differentiatingspermatogonia. Second, it also induces Stra8, a marker for theonset of meiosis. These two observations question whethersomatic cells are required for RA induction of gonocytedifferentiation in vitro. Development of a method to induce invitro differentiation of SSC has been widely pursued, and somestudies show some progression of spermatogenesis in a culturesystem. However, the starting materials used in previousreports are mostly preleptotene or leptotene spermatocytes [26–28]. Our study is the first report of spermatogonial differen-tiation induced by one extrinsic factor in vitro. RA is a factorthat can induce differentiation in many organs and cells. It hasbeen demonstrated that RA could induce the differentiation ofAal to A1 in VAD animals and cultured cryptorchid adult testes[5, 12]. However, whether the RA effect on early germ cells isdirect or indirect has not been determined, because receptorsfor retinoids are found in both germ cells and somatic cells.Retinoids play pivotal regulatory roles in Sertoli and Leydigcells and also interstitial macrophages. Our results obtainedfrom cultured THY1þ gonocytes/spermatogonia in a feeder-cell-free condition demonstrate, for the first time, that in vitrodifferentiation of undifferentiated early germ cells could beinitiated by RA without somatic cells and that the differenti-ation of these cells can be validated by the induction ofexpression of two markers, Kit and Stra8. Whether thedifferentiation triggered by RA in this somatic-free systemcan lead to formation of late differentiating spermatogonia (Inand B spermatogonia) and primary spermatocytes is not clear,because germ cells under this feeder-cell-free and serum-freecondition could survive only up to 48 to 72 hours (data notshown). It is noteworthy that in both cultured THY1þ

gonocytes and neonatal testes, RA-stimulated germ cellsshowed a higher percentage of cell death after the first 24 h(data not shown). This strongly suggests that once RA triggersthe differentiation of these germ cells, the somatic cells,particularly the mature Sertoli cells, become vital to theirsurvival and are likely to be important for the continuingprogression of the differentiation pathway.

If Stra8 is required for the onset of male meiosis, when doesRA act during spermatogonial development and whichsubpopulation of spermatogonia is responsible for Stra8expression? Analysis of Stra8-deficient testes demonstratedthat Stra8 is essential for normal progression into meioticprophase [14, 15]. The findings of robust induction of Stra8/Kitexpression after 8 h using 2 dpp, 4 dpp, and 8 dpp THY1þ

gonocytes/spermatogonia suggest that induction occurs duringgonocyte/spermatogonial maturation. There is additional evi-dence supporting this conclusion: 1) previous studies from deRooij and others [3] suggest that some gonocytes directlydifferentiate to A1 in the first round of spermatogenesis and 2)because the first appearance of KIT was found in late Aal to A1,and it was confirmed to be a marker of differentiatingspermatogonia (A1-B), the magnitude and temporal sequencesof induction of Stra8 vs. Kit expression in THY1þ gonocytesimplies that RA induction of Stra8 expression might be earlierthan induction of Kit expression. So far, data demonstrate thatRA initially induces Stra8 expression, and gonocyte andspermatogonia differentiation occurs. However, some ques-tions arise from this study that prevent the exclusion of thepossibility of a second induction at B-preleptotene-leptotene.For instance, which subpopulation of differentiating spermato-

gonia (late Aal, A1, or B spermatogonia) contributes to thehigher level of Stra8 in KITþ cells?

Microarray profiling of RA-treated 2 dpp THY1þ gonocytesnot only confirmed the induction of Stra8 expressiondemonstrated by real-time RT-PCR in this study, but alsorevealed that a group of genes involved in multiple aspects ofRA action were significantly up-regulated by RA. Theinduction of Stra6, which was originally identified as an RAresponsive gene in P19 carcinoma cells [29], further validatesthe effectiveness of RA treatment on these gonocytes.Recently, STRA6 was found to be a specific membranereceptor of retinol binding protein, representing a majorphysiological mediator of cellular vitamin A uptake [30]. Sofar, these gonocytes are the only known noncancer cells tostimulate Stra6 expression in response to RA. The increasedStra6 expression indicates a facilitated uptake of retinol intothe cells, and the physiological relevance of this up-regulationin gonocytes needs to be studied. In the meantime, anaccelerated RA degradation in these gonocytes upon the RAtreatment was indicated by the dramatic increase of expressionof three genes: Cyp26a1, Cyp26b1, and Por. CYP26A1 andCYP26B1 are two of the three P450 enzymes in the CYP26subfamily to oxidize RA for further metabolism and inactiva-tion [31]. Cytochrome P450 oxidoreductase (POR) coded byPor is an NADPH-dependent protein essential for transferringelectron to microsomal cytochrome P450 [31]. Lack of PORwould impair the activity of all three CYP26 enzymes [31]. Inaddition to the increased metabolism of RA, the synthesis ofRA is likely attenuated, as indicated by the induction of Lratand Dhrs3. Lecithin-retinol acyltransferase (LRAT) is thepredominant enzyme in the body responsible for the physio-logic esterification of retinol [32]. Increased Lrat indicates apotentially increased storage of retinyl ester and reduction ofsupply of retinol for the production of RA. Dehydrogenase/reductase member 3 (DHRS3) regenerates retinol from all-trans-retinal, which is an intermediate in the production of all-trans-RA [33]. The increased DHRS3 could mean a potentialreduction of retinal for the synthesis of RA. Other RA-regulated genes in the gonocytes are being analyzed andstudied. The data obtained from this microarray experimentwill provide a fertile basis for the hypothesis-driven studies ofRA-regulated gonocyte differentiation.

In conclusion, this study demonstrates that a limited level ofin vitro differentiation of gonocytes/spermatogonia could beachieved after RA stimulation without somatic involvement. Inaddition, RA induction of Stra8 expression in the undifferen-tiated spermatogonia was demonstrated, implying its importantrole in spermatogonial differentiation.

REFERENCES

1. Aponte PM, van Bragt MP, de Rooij DG, van Pelt AM. Spermatogonialstem cells: characteristics and experimental possibilities. Apmis 2005;113:727–742.

2. de Rooij DG. Proliferation and differentiation of spermatogonial stemcells. Reproduction 2001; 121:347–354.

3. de Rooij DG, Russell LD. All you wanted to know about spermatogoniabut were afraid to ask. J Androl 2000; 21:776–798.

4. de Rooij DG, de Boer P. Specific arrests of spermatogenesis in geneticallymodified and mutant mice. Cytogenet Genome Res 2003; 103:267–276.

5. van Pelt AM, van Dissel-Emiliani FM, Gaemers IC, van der Burg MJ,Tanke HJ, de Rooij DG. Characteristics of A spermatogonia andpreleptotene spermatocytes in the vitamin A-deficient rat testis. BiolReprod 1995; 53:570–578.

6. Gaemers IC, van Pelt AM, van der Saag PT, de Rooij DG. All-trans-4-oxo-retinoic acid: a potent inducer of in vivo proliferation of growth-arrested A spermatogonia in the vitamin A-deficient mouse testis.Endocrinology 1996; 137:479–485.

7. Beumer TL, Roepers-Gajadien HL, Gademan IS, Kal HB, de Rooij DG.

544 ZHOU ET AL.

Involvement of the D-type cyclins in germ cell proliferation anddifferentiation in the mouse. Biol Reprod 2000; 63:1893–1898.

8. Schrans-Stassen BH, van de Kant HJ, de Rooij DG, van Pelt AM.Differential expression of c-kit in mouse undifferentiated and differenti-ating type A spermatogonia. Endocrinology 1999; 140:5894–5900.

9. de Rooij DG, Okabe M, Nishimune Y. Arrest of spermatogonialdifferentiation in jsd/jsd, Sl17H/Sl17H, and cryptorchid mice. Biol Reprod1999; 61:842–847.

10. Schrans-Stassen BH, Saunders PT, Cooke HJ, de Rooij DG. Nature of thespermatogenic arrest in Dazl �/� mice. Biol Reprod 2001; 65:771–776.

11. Siiteri JE, Karl AF, Linder CC, Griswold MD. Testicular synchrony:evaluation and analysis of different protocols. Biol Reprod 1992; 46:284–289.

12. Haneji T, Maekawa M, Nishimune Y. Retinoids induce differentiation oftype A spermatogonia in vitro: organ culture of mouse cryptorchid testes. JNutr 1983; 113:1119–1123.

13. Akmal KM, Dufour JM, Kim KH. Retinoic acid receptor alpha geneexpression in the rat testis: potential role during the prophase of meiosisand in the transition from round to elongating spermatids. Biol Reprod1997; 56:549–556.

14. Koubova J, Menke DB, Zhou Q, Capel B, Griswold MD, Page DC.Retinoic acid regulates sex-specific timing of meiotic initiation in mice.Proc Natl Acad Sci U S A 2006; 103:2474–2479.

15. Baltus AE, Menke DB, Hu YC, Goodheart ML, Carpenter AE, de RooijDG, Page DC. In germ cells of mouse embryonic ovaries, the decision toenter meiosis precedes premeiotic DNA replication. Nat Genet 2006; 38:1430–1434.

16. Kubota H, Avarbock MR, Brinster RL. Growth factors essential for self-renewal and expansion of mouse spermatogonial stem cells. Proc NatlAcad Sci U S A 2004; 101:16489–16494.

17. Oatley JM, Avarbock MR, Telaranta AI, Fearon DT, Brinster RL.Identifying genes important for spermatogonial stem cell self-renewal andsurvival. Proc Natl Acad Sci U S A 2006; 103:9524–9529.

18. Shima JE, McLean DJ, McCarrey JR, Griswold MD. The murine testiculartranscriptome: characterizing gene expression in the testis during theprogression of spermatogenesis. Biol Reprod 2004; 71:319–330.

19. Zhou Q, Shima JE, Nie R, Friel PJ, Griswold MD. Androgen-regulatedtranscripts in the neonatal mouse testis as determined through microarrayanalysis. Biol Reprod 2005; 72:1010–1019.

20. OuladAbdelghani M, Bouillet P, Decimo D, Gansmuller A, Heyberger S,Dolle P, Bronner S, Lutz Y, Chambon P. Characterization of a premeioticgerm cell-specific cytoplasmic protein encoded by Stra8, a novel retinoicacid-responsive gene. J Cell Biol 1996; 135:469–477.

21. O’Brien DA. Stage-specific protein synthesis by isolated spermatogenic

cells throughout meiosis and early spermiogenesis in the mouse. BiolReprod 1987; 37:147–157.

22. Oatley JM, Brinster RL. Spermatogonial stem cells. Adult Stem Cells2006; 419:259–282.

23. Yoshida S, Sukeno M, Nakagawa T, Ohbo K, Nagamatsu G, Suda T,Nabeshima Y. The first round of mouse spermatogenesis is a distinctiveprogram that lacks the self-renewing spermatogonia stage. Development2006; 133:1495–1505.

24. Bowles J, Knight D, Smith C, Wilhelm D, Richman J, Mamiya S, YashiroK, Chawengsaksophak K, Wilson MJ, Rossant J, Hamada H, Koopman P.Retinoid signaling determines germ cell fate in mice. Science 2006; 312:596–600.

25. Zamboni L, Merchant H. The fine morphology of mouse primordial germcells in extragonadal locations. Am J Anat 1973; 137:299–335.

26. Rassoulzadegan M, Paquis-Flucklinger V, Bertino B, Sage J, Jasin M,Miyagawa K, van Heyningen V, Besmer P, Cuzin F. Transmeioticdifferentiation of male germ cells in culture. Cell 1993; 75:997–1006.

27. Weiss M, Vigier M, Hue D, Perrard-Sapori MH, Marret C, Avallet O,Durand P. Pre- and postmeiotic expression of male germ cell-specificgenes throughout 2-week cocultures of rat germinal and Sertoli cells. BiolReprod 1997; 57:68–76.

28. Staub C, Hue D, Nicolle JC, Perrard-Sapori MH, Segretain D, Durand P.The whole meiotic process can occur in vitro in untransformed ratspermatogenic cells. Exp Cell Res 2000; 260:85–95.

29. Bouillet P, Sapin V, Chazaud C, Messaddeq N, Decimo D, Dolle P,Chambon P. Developmental expression pattern of Stra6, a retinoic acid-responsive gene encoding a new type of membrane protein. Mech Dev1997; 63:173–186.

30. Kawaguchi R, Yu JM, Honda J, Hu J, Whitelegge J, Ping PP, Wiita P, BokD, Sun H. A membrane receptor for retinol binding protein mediatescellular uptake of vitamin A. Science 2007; 315:820–825.

31. Ribes V, Otto DME, Dickmann L, Schmidt K, Schuhbaur B, HendersonC, Blomhoff R, Wolf CR, Tickle C, Dolle P. Rescue of cytochrome P450oxidoreductase (Por) mouse mutants reveals functions in vasculogenesis,brain and limb patterning linked to retinoic acid homeostasis. Dev Biol2007; 303:66–81.

32. O’Byrne SM, Wongsiriroj N, Libien J, Vogel S, Goldberg IJ, Baehr W,Palczewski K, Blaner WS. Retinoid absorption and storage is impaired inmice lacking lecithin: retinol acyltransferase (LRAT). J Biol Chem 2005;280:35647–35657.

33. Haeseleer F, Huang J, Lebioda L, Saari JC, Palczewski K. Molecularcharacterization of a novel short-chain dehydrogenase/reductase thatreduces all-trans-retinal. J Biol Chem 1998; 273:21790–21799.

RA-INDUCED IN VITRO DIFFERENTIATION OF GONOCYTES 545