Developmental Biology 356 (2011) 370–382

Contents lists available at ScienceDirect

Developmental Biology

journal homepage: www.elsevier.com/developmentalbiology

Rex1 (Zfp42) null mice show impaired testicular function, abnormal testis morphology, and aberrant expression

Naira C. Rezende a,b,1, Mi-Young Lee a,1, Sébastien Monette c, Willie Mark d, Ailan Lu a,2, Lorraine J. Gudas a,b,⁎ a Department of Pharmacology, Weill Cornell Medical College of Cornell University, 1300 York Avenue, New York, NY 10065, United States b BCMB Graduate Program, Weill Cornell Medical College of Cornell University, 1300 York Avenue, New York, NY 10065, United States c Center of Comparative Medicine and Pathology, Weill Cornell Medical College of Cornell University, 1300 York Avenue, New York, NY 10065, United States d Developmental Biology Program, Sloan-Kettering Institute, 1275 York Avenue, New York, NY 10065, United States article info abstract

Article history: Rex1 (Zfp42), GeneID 132625, is a gene whose expression is closely associated with pluripotency/multipotency Received for publication 21 January 2011 in both mouse and human embryonic stem cells. To study the function of the murine Rex1 gene in vivo,wehave Revised 18 May 2011 used cre/lox technology to create Rex1(floxed) mice and mice deficient in Rex1 gene function. Rex1−/−males are Accepted 19 May 2011 characterized by an age-associated decrease in sperm counts, abnormal sperm morphology, and mild testicular Available online 27 May 2011 atrophy. We characterized global patterns of in primary germ cells by microarray and identified the growth hormone responsive gene, GRTP1, as a transcript present at a 4.5 fold higher level in wild type (WT) Keywords: −/− Testis compared to Rex1 mice. We analyzed immature germ cell (Dazl), proliferating (PCNA), and Sertoli cell −/− Spermatogenesis populations, and quantitated levels of apoptosis in Rex1 as compared to WT testes. We evaluated the Germ cells expression of previously reported to correlate with Rex1 expression, such as STAT3, phospho-STAT3, Stem cells p38, and phospho-p38 in the testis. We report a distinct cellular localization of total STAT3 in Rex1−/− Sertoli cells affected testes. Our data suggest that loss of Rex1 leads to impaired testicular function. Sox9 © 2011 Elsevier Inc. All rights reserved.

Introduction of vitamin A (retinol) (Hosler et al., 1989, 1993). Subsequently, we characterized Rex1 expression during mouse development and Understanding the cell signaling networks commonly used in showed that Rex1 mRNA is primarily detected in derived defining the pluripotent state is critical to understanding tissues and in the of the blastocyst (ICM), the source of identity and differentiation. However, more must be learned about embryonic stem cells in culture (Rogers et al., 1991). In adult mice, the functional role of a gene whose expression is associated with Rex1 mRNA is expressed in a subset of male germ cells, spermatocytes pluripotency, namely Rex1, in the context of adult and embryonic actively undergoing (Rogers et al., 1991). Rex1 (mRNA stem cell biology. The Rex1 (Zfp42) gene (GeneIDs: mouse, 22702; and protein) is selectively expressed in dividing cells of the human human, 132625), a member of the zinc finger protein family that also testes and ovaries (Kristensen et al., 2008), suggesting a conserved includes YY1 (Kim et al., 2007), was identified by our laboratory functional role. Rex1 transcripts are also detected in a variety of (Hosler et al., 1989) and is widely used as one of seven markers of pluripotent cells, normal primary human keratinocytes (NHEK), human and murine embryonic stem (ES) cells (Brivanlou et al., 2003). human prostate epithelial cells (PrEC), renal parenchymal tissue, and We first identified Rex1 as an abundant transcript expressed in some human neoplastic cell lines, such as MDA-MB-468 mammary F9 teratocarcinoma stem cells; the transcription of Rex1 rapidly carcinoma cells, SCC-15 head and neck squamous carcinoma cells, and decreased upon addition of all-trans retinoic acid (RA), an active form NTERA2 teratocarcinoma cells (Kristensen et al., 2010; Mongan et al., 2006; Mongan and Gudas, 2007; Raman et al., 2006). Interestingly, we also showed that Rex1 expression increases in mouse digits during Abbreviations: ES, embryonic stem; mES, mouse ; RA, all-trans tissue regeneration/wound healing (Agrawal et al., 2010; Wang et al., retinoic acid; hMSCs, human mesenchymal stem cells; RT-PCR, reverse transcription PCR; HET, heterozygote; KO, knockout; H & E, Hematoxylin and Eosin. 2010). ⁎ Corresponding author at: Department of Pharmacology, Weill Cornell Medical Most of what is functionally known about Rex1 comes from results College of Cornell University, 1300 York Avenue, Room E-409, New York, NY 10065, obtained in cell differentiation studies, and/or analysis of the United States. Fax: +1 212 746 8858. regulating its expression. First, our laboratory created Rex1 knockout E-mail address: [email protected] (L.J. Gudas). − − (Rex1 / ) F9 teratocarcinoma stem cell lines by homologous 1 These authors contributed equally to this work. −/− 2 Current address: inGenious Targeting Laboratory, Inc., 25 Health Science Drive, recombination, and demonstrated that F9 Rex1 cells fail to LIHTI Building, Suite 118, Stony Brook, NY 11790, United States. differentiate completely into visceral (VE) in aggregate

0012-1606/$ – see front matter © 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2011.05.664 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 371 culture after RA treatment (Thompson and Gudas, 2002). We further the BglII sites of plasmid pgRex1-1.38 (3′ of Rex1 exon 4) (Fig. 1A). We demonstrated that Rex1 acts as a negative regulator of STAT3 during then ligated vectors pgRex1-1.37 and pgRex1-1.38 at their HindIII differentiation of F9 cells (Xu et al., 2008), identifying a possible restriction sites (Fig. 1A). Subsequently, we subcloned the PGK- mechanism behind the phenotype in F9 cells. We analyzed changes neo/MCI-TK cassette (pNEOTKLOX plasmid) into the NsiI site within in gene expression during the differentiation of Rex1−/− murine intron 3′ of Rex1. This strategy created a targeting vector of embryonic stem cell (mES) lines, and showed that Rex1 affects approximately 12.7 kb (Fig. 1A); the 5′ region of this targeting vector differentiation and cell cycle progression in mES (Scotland et al., 2009). contains 2.49 kb homologous to Rex1 intron 3, 3.86 kb of selection More recently, Bhandari et al. (2010) investigated the relationship markers flanked by LoxP sites (PGK-neo, MCI-TK), 2.28 kb of the Rex1 between Rex1 expression and p38 mitogen-activated protein kinase gene (including all of exon 4), a LoxP site, and a 3′ arm consisting of a (MAPK) activation in human mesenchymal stem cells (hMSCs), and 4.07 kb genomic region 3′ of Rex1 exon 4. reported that Rex1 regulates hMSC proliferation/differentiation This construct was linearized and introduced into the CJ7 mouse by suppressing the p38 MAPK signaling pathway. At the molecular ES cell line (originally derived from 129S1/SvImJ) by electroporation level, murine Rex1 expression is reportedly regulated by the using a Bio-Rad electroporator at a voltage of 230 mV and a pluripotency genes Oct4, and Nanog (Ben-Shushan et al., 1998; capacitance of 500 μF. Drug resistant clones were isolated after Hosler et al., 1993; Rosfjord and Rizzino, 1994; Shi et al., 2006). Not G418 selection (300 μg/ml active G418), and a total of 116 cell clones surprisingly, Rex1 is one of a few select genes shown to be re- were screened by Southern Blot analysis. Eight positive clones from expressed in induced pluripotent stem cells (iPS cells) (Chan et al., cells that had undergone homologous recombination were obtained. 2009). iPS cells hold enormous potential not only for regenerative One clone, # 66, was injected into C57BL/6 blastocysts to generate medicine, but also for studies in developmental biology, drug chimeric mice. Germ-line transmission was confirmed by mating development, and the derivation of patient specific stem cell lines. chimeras to wild type C57BL/6 mice. Seven Rex1floxed/+ female mice Analysis of the of the human Rex1 gene by our laboratory were chosen to mate with CAG-cre male mice, which are on the showed that human prostate cancer cells exhibited lower Rex1 C57BL/6J background, to excise exon 4 of the Rex1 gene and/or the promoter activity than normal prostate epithelial cells in culture PGK-neo-MCI-TK cassette. The Cre gene in these mice is under the (Lee et al., 2010). control of the cytomegalovirus immediate early enhancer linked to Spermatogenesis, the proliferation and maturation of germ cells the chicken beta-actin promoter so that the Cre gene is expressed leading to formation of spermatozoa, occurs from puberty to death early and ubiquitously (Sakai and Miyazaki, 1997). Recombination in male mammals. The testis is composed of germ cells as well as was detected by PCR-based genotyping (the primers used for this somatic cells, such as Sertoli cells and Leydig cells (Meng et al., 2000; genotyping are in Supplemental Table 2), and heterozygous mice O'Shaughnessy et al., 2009; Yoshinaga et al., 1991). Testicular germ (Rex1+/−) were inter-crossed to generate the Rex1−/− null mice. To cells consist of spermatogonial stem cells which can undergo self- determine the genotypes by PCR, two primer sets, P1F and P2R, and renewal as well as differentiate into spermatocytes, spermatids, and P1F and P4R were used. To confirm the genotype by Southern blot, finally into spermatozoa, the mature form of sperm. Sertoli cells genomic DNA was first digested with BglII. A 0.9 kb probe, located in produce growth factors, such as glial cell line-derived neurotrophic genomic DNA 3′ of the targeting construct, was obtained by digesting factor (GDNF), that maintain germ cell functions, and endocrine pgRex1-1.46 plasmid DNA with HindIII/BglII, and hybridized to BglII hormones such as inhibin, which acts on the hypothalamic–pituitary– digested genomic DNA. To prove homologous recombination we gonadal axis (Hogarth and Griswold, 2010; Meng et al., 2000). Leydig generated a second probe targeting the genomic DNA 5′ of the cells are primarily responsible for testosterone production (Hogarth targeting construct. This second probe (0.3 kb) was obtained by and Griswold, 2010). We wanted to understand the functions of digesting pgRex1-1.51 plasmid DNA with BglII, and it was hybridized Rex1 in stem cells in vivo and thus generated Rex1−/− mice by using to EcoRI digested genomic DNA. The locations of the primer sets are a conditional knockout strategy (Cre–LoxP). We show here that depicted in Fig. 1A. For these studies, mice were partially backcrossed Rex1−/− mice are viable, fertile, and grossly similar to WT littermates. into the C57BL/6J background, and thus they represent the N3 Rex1 null mice were also generated by a conventional gene targeting generation. Mice are now at the N12 generation with respect to the strategy and while these mice were not characterized in depth, Rex1 Rex1 knockout in the C57Bl/6J background. was reported to be dispensable in the development (Masui et al., 2008). Upon further characterization we discovered that Rex1−/− Sperm counts, morphology evaluation, and testes weight measurements males showed an age associated decrease in sperm counts, abnormal sperm morphology, and mild testicular atrophy, phenotypes that Either 6–7 week or 4–6 month old male mice, both Rex1+/+ and could reflect a defect in the stem cell pool. Our data, collectively, Rex1−/−, were sacrificed by cervical dislocation. For sperm counts, indicate that Rex1 plays an important role in germ cell differentiation. our protocol was adapted from Yu et al. (2006). Briefly, the cauda Here we report experiments characterizing the germ cell compart- epididymis was removed, minced in 3 ml of PBS (pH 7.4), and ment of Rex1 null mice in detail. incubated for 5 min at room temperature by gently rocking. The samples were transferred to 15 ml tubes and centrifuged at 500×g for Materials and methods 5 min. Then, the supernatants were used to count sperm with a hemocytometer. Three independent counts were performed for each Reagents mouse and the number of sperm was averaged. Briefly, sperm from 23 WT and 17 Rex1−/− mice were counted for the 6 week age group, and Unless stated, all reagents were purchased from Sigma Aldrich sperm from 10 WT and 20 Rex1−/− mice were counted for the four to (St. Louis, MO, USA). six month age group. For sperm morphology evaluations, samples were prepared according to Ribeiro et al. (1987). Briefly, 0.9 ml of the Construction of the Rex1 targeting vector and generation of Rex1−/− supernatant, prepared as above, was transferred to 0.1 ml of 1% mice Eosin Y in a 1.5 ml tube and incubated for 30 min at 22 °C. The solution (10–15 μl) was used to prepare a smear on a slide, and this was air- We previously generated a phage library with genomic DNA for dried. Mounting medium was added and sperm morphology was the Rex1 gene (Hosler et al., 1993). To generate a targeted deletion of analyzed under the microscope. For testes weight measurements, the protein coding exon, Rex1 exon 4, we first subcloned the LoxP both testes from each animal were removed, weighed, and the sequence from the pBS64 plasmid (Baubonis and Sauer, 1993) into numbers were averaged. 372 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 A

B +/+ +/- -/-

(6.6kb) D +/+ +/- -/- (470bp)

(350bp) (4.97kb) E +/+ -/-

Rex1 (670bp) C +/+ +/- -/- F +/+ -/-

(8.8kb) 36B4 (450bp)

(5.8kb)

Fig. 1. (A) Strategy and structure of targeting vector and strategy for generating Rex1−/− mice. (B, C) Southern blot strategy verifying the accuracy of the targeting event. (B) 3′ Southern; genomic DNA from two WT, three heterozygous, and two knockout mice was digested with BglII, and the probe (0.9 kb) was prepared by digesting pgRex1-1.46, a genomic fragment outside of the targeting construct, was prepared by digesting pgRex1-1.46 with HindIII/BglII. WT mice yield a band of 4.97 kb, heterozygous mice yield a 4.97 kb band and a 6.6 kb band, and knock out animals yield a 6.6 kb band. (C) 5′ Southern; genomic DNA from two WT, three heterozygous, and two knockout mice was digested with EcoRI, and the probe (0.3 kb) was prepared by digesting pgRex1-1.51, a genomic fragment outside of the targeting construct, with BglII. WT mice yield a band of 8.12 kb, heterozygous mice yield a 8.12 kb band and a 5.6 kb band, and knock out animals yield only a 5.6 kb band. (C) Genotype by PCR: genomic DNA from mouse tails is routinely genotyped by semi- quantitative PCR. We detect the recombination event by using a three primer PCR reaction: (1) specific forward primer homologous to an intragenic genomic region (P1F), in combination with two unique reverse primers: (2) a primer complementary to the Rex1 first exon (P2R); (3) and a primer complementary to the exogenous LoxP site (P4R). WT mice yield one band of approximately 350 bp, homozygous null animals yield a band of 470 bp, and heterozygous animals yield both bands. Primer sequences are listed in Supplemental Table 2. (D, E) RT-PCR: total RNA was isolated from testis and 2 μg of RNA was used to make cDNA using reverse transcriptase (Superscript, Invitrogen). 2 μl of 1:5 diluted cDNA was used to amplify mRex1 (D), and 36B4 (E) was used as a loading control. Rex1 primer pairs used for this are called Rex1F and Rex1R and result in a band of 670 bp; this band identity was verified by sequencing. 36B4 primers are called 36B4F and 36B4R and result in a 450 bp band. All primer sequences are listed in Supplemental Table 2. N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 373

Testicular cell preparation and cell cycle analyses sections (7 μm thickness) were stained with H&E and tissue morphology was evaluated blindly by a pathologist (SM). For im- Testicular cells were prepared for cell cycle analysis following a munohistochemistry, tissue sections were deparaffinized for 5 min in protocol adapted from Malkov et al. (1998) and Browning. Briefly, Histo-Clear (National Diagnostics, Atlanta, GA) twice, and then testes were removed and decapsulated by making a small incision in rehydrated in serially diluted ethanol solutions (100%, 95%, 70%, and the testis. The contents of the testes were collected through the then H2O). Antigen retrieval was performed by cooking sections for incision into a 15 ml tube containing 5 ml ice-cold M199-BSA media 3 min in a pressure cooker containing antigen unmasking solution (9.5 g tissue culture medium M199 (Invitrogen, Carlsbad, CA), 2.2 g (pH 6.0, H-3300, Vector), and tissue sections were then blocked

NaHCO3, 1 g bovine serum albumin in 1 l distilled water, pH 7.3–7.4). for 30 min in 5% goat serum in PBST (0.05% Tween 20 in PBS). 1–2 ml of 2 mg/ml collagenase was added to this tube, and the After washing with PBST, the tissue sections were incubated contents were incubated for 40 min at 37 °C with vigorous shaking. with primary antibodies: GRTP1 (1:500, Sigma, PRS4589), Sox9 Then, the tubes were placed on ice and incubated to allow the (1:200, H-90, sc-20095, Santa-Cruz Biotechnology, Santa Cruz, CA), seminiferous tubules to settle. The supernatants were discarded and DazL (1:400, ab34139, Abcam, Cambridge, MA), PCNA (1:400, M0879, the seminiferous tubules were washed twice in 10 ml of M199-BSA Dako, Carpinteria, CA) overnight at 4 °C for Sox9, or 1 h at 22 °C for medium. 5 ml of M199-BSA medium containing 2.5 μg/ml trypsin GRTP1, DazL and PCNA. Following three washes with PBST the and 1 U/ml DNase I was added to seminiferous tubules, and they sections were incubated with secondary antibodies, either anti-rabbit were incubated for 20 min at 37 °C. The reactions were stopped by HRP for Sox9 and DazL (superpicture HRP polymer conjugated rabbit transferring the tubes to ice and the cords were disaggregated by primary, 87-9263, Invitrogen, Carlsbad, CA) or anti-mouse HRP for pipetting up and down with a Pasteur pipette several times. The PCNA (M.O.M kit, PK-2200, Vector Labs, Burlingame, CA) for 10 min resulting cell suspensions were filtered through a 50 μm nylon mesh or 30 min at 22 °C, respectively. DAB signals were generated and and washed twice with M199-BSA medium and the cells were sections were counterstained with Hematoxylin. We provided total counted. A total of 2×106 cells were used for each cell cycle analysis. Stat3 (Cell Signaling Technology, 9132), phospho-Stat3 (Cell Signaling Cell cycle analyses were performed following the manufacturer's Technology, 9145), p38 MAPK (Cell Signaling Technology, 9212, lot guide (Becton Dickinson Immunocytometry System, San Jose, CA). 16), and phospho-p38 MAPK (Cell Signaling Technology, 4631, lot 6) Cells (2×106) were resuspended in 875 μl of cold PBS containing 0.1% antibodies for automated staining at the Memorial Sloan Kettering sodium azide and fixed by adding 125 μl of cold 2% paraformaldehyde Cell Biology Core Facility. in PBS and incubating for 1 h at 4 °C. Cells were centrifuged for 5 min at 250 g at 4 °C and the supernatant was aspirated. Cells were then Germ cell (GC) RNA isolation, labeling and microarray hybridization permeabilized by incubating them in 1 ml of PBS containing 0.05% Tween 20 for 15 min at 37 °C. After washing the cells with 1 ml of PBS Germ cell RNA was isolated from whole testis following an containing 0.1% sodium azide, cells were stained with propidium adaptation of a well established protocol (Malkov et al.). Briefly, testes iodide (PI) staining solution (50 μl of 1 mg/ml of PI, 1 μl of 100 mg/ml were decapsulated and digested with 0.1 mg/ml of collagenase type II RNase A in 1 ml PBS), and cell cycle analysis was performed on a (Invitrogen) for 40 min. After collagenase digestion, the dissociated Becton-Dickinson FACS Calibur in the Weill Cornell FACS Core. seminiferous tubules were isolated by gravity sedimentation, and germ cells were further disconnected from tissue by digestion with Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) 2.5 μg/ml trypsin (Lonza) supplemented with 1 U/ml DNase (Invitro- assay gen). Single cells were filtered through a 50 μm nylon mesh, washed twice in separation medium (1) for 5 min at 300 g, and counted. Total The TUNEL assay was performed following the manufacturer's germ cell RNA (GC RNA) was extracted using TRIZOL reagent guide (TUNEL enzyme, Ref: 11767305001, TUNEL label Ref: (Invitrogen). GC RNA from two WT mice and three Rex1−/− mice 11767291910, Roche, Basel, Switzerland); paraffin sections of testis were labeled and hybridized to oligonucleotide microarrays (Gene- (5 μm-thickness) were deparaffinized for 5 min in Histo-Clear Chip Mouse Gene 1.0 ST (Affymetrix)), according to standard (National Diagnostics, Atlanta, GA) twice, and then rehydrated in a protocols at the Weill Cornell Medical College Microarray Core Facility series of serially diluted ethanol solutions (100%, 95%, 70%, and then (the GC RNA of one wt mouse was analyzed twice).

H2O). Tissue sections were then treated with proteinase K (final concentration of 10 ug/ml in 10 mM Tris–HCl, pH 7.4–7.8) for 15 min Microarray data analysis at 22 °C. A positive control was prepared by incubating a tissue section with DNase I (3 U/ml in 50 mM Tris–HCl, pH7.5, 1 mg/ml BSA) for Raw data CEL files were normalized using the Robust Multi-Array 10 min at 22 °C before proteinase K treatment. A labeling reaction (RMA-16) algorithm in GeneSpring GX 11.0.2 software (Agilent). was prepared by mixing 45 μl of TUNEL-label solution with 5 μlof Triplicate samples were grouped as wild type (wt) and homozygous TUNEL-enzyme, added to tissue sections, and incubated for 60 min null (Rex1–/–), and statistical analysis was calculated with unpaired t- at 37 °C in a humidified chamber in the dark. A negative control was tests applied to log-transformed data. prepared without TUNEL enzyme. Tissue sections were washed three times with PBS and analyzed under a fluorescence microscope using Results an excitation wavelength in the range of 450–550 nm and a detection wavelength in the range of 515–565 nm. Positive staining cells were Rex1 null animals are characterized by premature age-associated manually counted from a section of each slide (100× magnification testicular germ cell depletion, decreased sperm count, and abnormal field). sperm morphology

Histology and immunohistochemistry To investigate the functions of Rex1 in germ cell biology we have generated a mouse with a targeted deletion of the Rex1 gene (Fig. 1A). Morphology of testicular and ovarian tissue was assessed by Initial attempts by our laboratory to create Rex1 null mice failed so we examination of Hematoxylin and Eosin (H&E) stained slides. Mice used Cre/Lox recombination to produce a conditional Rex1 knockout were sacrificed by cervical dislocation and testes was removed and mouse. A LoxP site was inserted into the 3′ end of the Rex1 exon 4 fixed in 4% paraformaldehyde overnight at 4 °C. Tissues were while the selectable marker genes, the Neo-TK cassette, were flanked transferred to 70% ethanol and paraffin blocks were prepared. Tissue by LoxP sites as well (Fig. 1A). After germline transmission of the 374 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 targeted , Rex1flox/wt mice were mated with a transgenic mouse characterized by mild germ cell depletion, and a premature age- line carrying the chicken β-actin promoter (Niwa et al., 1991) associated decrease in sperm counts. directing Cre recombinase expression in all tissues (Sakai and Miyazaki, 1997). Rex1+/− mice were then identified, indicating Rex1−/− ovaries have similar numbers of follicles, corporea lutea, excision of both the Neo-TK gene and exon 4 of the Rex1 gene and ceroid pigment as compared to WT ovaries (Figs. 1B, C, D). We obtained WT, heterozygous, and homozygous mutant Rex1 mice. Rex1−/− mice are viable and fertile and there In mammals the development of female germ cells entails a appear to be no sex preferences among the progeny obtained complex and tightly regulated process. At birth, the murine female (Table 1). As expected, Rex1 null mice failed to express Rex1 message ovaries contain primary oocytes arrested at stage I of mitosis (Pepling, as assayed by semi-quantitative PCR (Figs. 1E, F). We tested two 2006). Following puberty, a number of oocytes will periodically commercially available antibodies against murine Rex1 (Millipore respond to environmental hormonal clues, resume mitosis, and MAB4316, Abcam ab28141), and two custom generated Rex1 mature further (Pepling, 2006). We previously failed to detect Rex1 antibodies (Alpha Diagnostics), but failed to detect endogenous transcript in murine ovaries (Rogers et al., 1991). Nevertheless Rex1 protein by Western blot analysis of WT mouse testicular Kristensen et al. (2008) reported human REX-1 staining specifically tissue, although the antibodies could detect mouse Rex1 over- in the cycling oocytes of gestation week 40. To investigate a role for expressed in Cos cells (data not shown). Rex1 in female germ cell development we undertook a comparison of To explore the physiological consequences of the loss of Rex1, the the number of follicles, corporea lutea and ceroid pigment in a group structure and function of the testes from Rex1 null mice were of mice across three distinct age groups, 21 days old, 6 weeks old, and examined in detail. First, testis histology of six week (WT, n=23, KO, 1 year old. We reasoned that if transient Rex1 expression in cycling n=17) and 4–6 month old (WT, n=10, KO, n=20) mice was cells affects the number of female germ cells we would be able to see a evaluated by H&E staining (Figs. 2B, D, C, E). We found that six of the quantitative difference in the number of follicles and corporea lutea Rex1−/− mutant testes showed a mild to moderate decrease in the between WT and Rex1−/− mice. In general, ovary morphology was number of germ cells in the seminiferous tubules (KO six week, n=5, grossly similar between WT and Rex1−/− mice. At 21 days old the KO 4–6 month old, n=1). None of the WT, either 6 week or 4– average number of follicles in the WT mice was 52.6 as compared to 6 month old, showed this phenotype. This phenotype was seen 75.5 in Rex1−/− mice (p=0.06) (Figs. 3A, B. Table 2). At 6 weeks old most predominantly in spermatids (both early and late), but also the average number of follicles in the WT mice was 37.6 as compared occasionally in spermatocytes. In affected testes depleted seminifer- to 24.6 in Rex1−/− mice (p=0.4), and there was an average of one ous tubules were observed multifocally, admixed with normal corpus luteum per WT mice and 1.3 per Rex1−/− mice (Figs. 3C, D. tubules. We sacrificed two 1 year old animals for testis histology, at Table 2). At 1 year old, the average number of follicles in the WT mice 13 months old (Supplemental Figs. 1A, B). Germ cell depletion is was 8.5 as compared to 7.8 in Rex1−/− mice (p=0.8), and there was known to naturally occur with increased age, and we observed mild an average of 7.25 corporea lutea per WT mice and 4.3 corporea lutea germ cell depletion, admixed with physiologically normal tubules per Rex1−/− mice (p=0.09) (Figs. 3E, F. Table 2). We note mild in both WT and Rex1−/− 1 year old male mice (Supplemental Fig. 1). lymphocytic infiltration in the ovarian bursa, which was observed in To investigate further the germ cell defect in Rex1−/− males, we both WT and Rex1−/− mice (Table 2). Overall, we conclude that Rex1 performed sperm counts in animals of various ages. The same mice expression does not affect the number or the morphology of follicles were evaluated for sperm counts (Fig. 2H) and testes weight (Fig. 2I). and corpus luteum in Rex1−/− mice. Both testes were removed from each mouse, weighed, and averaged. No significant difference was seen between six week old WT and Depletion of the germ cell compartment in Rex1−/− males cannot Rex1−/− mice, though a decline in the number of spermatozoa was be explained by defects in cell cycle distribution observed in Rex1−/− animals between four to six months of age (Fig. 2H). The average sperm count of WT mice in the four to six Previously, our laboratory utilized microarray approaches to month age group was 502×104 ±103/ml/epididymis for WT mice, identify transcripts that were differentially expressed between wild and it was 301×104 ±50/ml/epididymis in Rex1−/− mice (p=0.03). type and Rex1−/− mouse ES cells (Scotland et al., 2009). One of We also determined testis weights of six-week old WT mice and the differentially expressed transcripts was cyclin D2, which showed a Rex1−/− mice. The tissue weights were 80±4 mg and 77±7 mg, six fold decrease in Rex1−/− mES cells. To determine if alterations respectively (Fig. 2G). No statistically significant differences were in cell cycle distribution, potentially indicating a block in cell cycle observed in testis weights between 6 week old WT vs. Rex1−/− mice progression, could explain the decrease in sperm counts in Rex1−/− (p=0.324), and the same is true for 4–6 month old mice. mice, we measured the DNA content of cells in the seminiferous We studied the sperm in 9 Rex1−/− and 10 WT mice between tubules of 4–6 month old mice (Fig. 3). Comparing WT (mean±S.E) 4 and 6 months old, and noted the morphology to be different (Figs. 4A, C), and Rex1−/− mice (mean±S.E)(Figs. 4B, C), we (Figs. 2F, G). In many Rex1−/− mice most sperm had bent tails just observed no significant difference in the distribution of cells distal to the head (2G, white arrows). In addition, some sperm with 4N DNA content (19.2±1.6% vs 16.9±0.8%) and 2N DNA heads lacked cytoplasmic content, indicated by weaker staining (2G, content (9.7±0.8% vs 8.5±0.4%). A small difference in the spermatid blue arrows). The magnitude of this phenotype varied, suggesting and spermatozoa populations was observed between WT and mutant incomplete penetrance. We conclude that Rex1 null animals are mice; WT exhibited 68.9±2% of cells with 1N DNA content and Rex1 mutants 72.9±0.8% (p=0.03). These results indicate that defects Table 1 in cell cycle distribution/progression cannot explain the reduction in − −/− Genotype analysis of progeny from crosses of Rex1+/ mice. Distribution of male and sperm count in older Rex1 males. female, wild type (WT), heterozygous (HT) and knock out (Rex1−/−) progeny from − crosses between Rex1 mice. Primers P1F, P2R and P4R (sequences in Supplemental Microarray analysis of WT and Rex1−/− primary germ cells reveals Table 2) were used to genotype genomic DNA extracted from the mice tails. effects on the transcriptome Genotype Male Female

WT (Rex1+/+) 60 (0.32) 50 (0.3) To investigate if Rex1 expression influences the transcriptome in HT (Rex1+/-) 78 (0.41) 68 (0.4) testicular cells we performed microarray analyses on isolated primary -/- KO (Rex1 ) 52 (0.27) 51 (0.3) germ cells from the seminiferous tubules of six week old WT and Total 190 169 Rex−/− mice. We were not able to detect Rex1 mRNA by RT-PCR in N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 375

A)

B) C)

D) E)

F) G)

HI

Fig. 2. Measurements of testes weight, sperm counts and sperm morphology. Immature spermatids (red arrows) and mature spermatids (black arrows) are highlighted in an H&E stained wild type reference slide (A). Testis histology of six week old and 4–6 month old mice evaluated by H&E staining (WT B, D; Rex1−/− C, E). Sperm from 4–6 month old mice were collected from a cauda epididymis and stained with 1% Eosin. Stained sperm were observed under a bright field microscope and evaluated for morphological abnormalities (WT F, Rex1−/− G). The white arrows highlight the kinks in the Rex1−/− sperm, while the blue arrows highlight a lighter eosin staining pattern (G). Evaluation of the same mice for sperm counts (H), and testes weights (I). For sperm counts, a cauda epididymis was removed from each mouse and minced in 3 ml of PBS. The supernatant was used to count sperm by the trypan blue exclusion assay. Three independent counts were performed for each sample and averaged. * pb0.05, t-test. Scale 100 μm (B, C), 50 μm (D, E). isolated primary germ cells from Rex1−/− mice using primers homol- Growth Hormone Regulated TBC Protein-1, GRTP1, which shows ogous to exons three and four (Fig. 5B). Three genes showed at least a decreased mRNA expression in germ cells of Rex1−/− mice (~4.5 fold, four fold decrease in expression in Rex1 null mice. Two are unknown Fig. 5B) (Lu et al., 2001). Although primary germ cells of Rex1 null genes (Fig. 5A, and Supplemental Table 1), whereas the third is the mice express less GRTP1 mRNA than germ cells from WT mice 376 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382

Wild Type A B

21Days 21Days

C D

66Weeks Weeks

E F

1 Year Old 1 Year Old

Fig. 3. Histological evaluation of Follicles, Corporea Lutea, and Ceroid Pigment in WT vs Rex1−/− mice. Ovary histology of twenty one day, six week, and 1 year old mice was evaluated by H&E staining (WT A, C, E; Rex1−/− B, D, F). In panels A thru D black arrows highlight examples of follicles; in panels E and F black arrows highlight examples of corpus luteum. We analyzed ovaries across a range of different ages (WT, n=10, Rex1−/−, n=13 animals). Scale: Black bars denote 100 μm.

(Fig. 5C), WT and Rex1−/− mES cells in culture (+LIF) exhibit the the TUNEL signal in the testes from six week old mice (WT, n=21, KO, same GRTP1 transcript levels (data not shown). n=19). The average number of apoptotic cells observed per field This is the first time that a microarray analysis of Rex1 expressing in WT testes was 39±4 (Figs. 6A, C), while the average number versus Rex1 deficient cells taken directly from mice has been of apoptotic cells in Rex1−/− testes was 56±13 (Figs. 6B, C). We performed. We compared the results of our microarray analysis to observed a large variance in the number of cells undergoing apoptosis our previously published data from WT vs Rex1−/− null stem cells within the testes of different WT and Rex1−/− mice, but with the (Scotland et al., 2009; Xu et al., 2008) and that of others (Masui et al., exception of an outlier (Fig. 6B) most animals displayed comparable 2008) to identify themes that might suggest conserved functions. levels of apoptosis. The apoptotic cells were mostly immature germ GRTP1 transcript changes were only found in our analysis of WT vs cells located in the outer layers of the seminiferous tubule, and we Rex1−/− germ cells. The available data consist of two separately found no correlation between the number/location of immature derived mouse embryonic stem cell lines (Masui et al., 2008; Scotland germ cells and germ cell depletion. et al., 2009), a F9 teratocarcinoma stem cell line (Xu et al., 2008) and The p38 mitogen activated protein kinases are a well-known our primary germ cells (Fig. 5). class of proteins involved in responses to physiological stress (Han et al., 1994). Since human mesenchymal stem cells in culture Characterization of apoptotic cells, STAT3, phospho-STAT3(Tyr705), p38 show a direct relationship between Rex1 expression and activation and phospho-p38 positive cells in the testes of Rex1−/− mice of p38 by phosphorylation, we assessed primary germ cells from WT and Rex1−/− mice for phospho-p38 levels. We stained testes Bhandari et al. (2010) have recently shown that Rex1 acts as an for phospho-p38 (P-p38) (WT, n=20, KO, n=19 animals) (Supple- inhibitor of the pro-apoptotic p38 mitogen-activated protein kinase mental Figs. 2A, B), and for total p38 levels (WT, n=21, KO, n=19) (p38-MAPK) in human mesenchymal stem cells. The abnormal (Supplemental Figs. 2D, E). There was a high variation in staining morphology of many seminiferous tubules and the lower sperm within both WT and Rex1−/− samples. Ten Rex1−/− mice displayed counts seen in Rex1−/− mutants may reflect an increase in apoptosis. phospho-p38 expression as compared to seven WT mice. In most To test this, we performed TUNEL assays to assess the level of cases phospho-p38 was detected in Sertoli cells (Supplemental apoptosis in the testicular cells of Rex1−/− mice. We defined a 100× Fig. 2B, black arrows), and occasionally in germ cells (Supplemental magnification section of comparable slides as a field, and evaluated Fig. 2B, white arrows). There was no statistically significant N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 377

− − Table 2 Rex1 / testes displaying testicular atrophy (Fig. 6E). Although we Tabulation of the number of follicles, corporea lutea, and ceroid pigment in WT vs − − observe a positive correlation between the degree of atrophy and the Rex1 / mice. − − number of P-Stat3 (Tyr705) positive Sertoli cells in the Rex1 / Age: 6 weeks One ovary examined per animal, numbers are per ovary testes, it is unclear if activation of Stat3 by phosphorylation is the ⁎ ⁎⁎ Genotype n follicles n corpora lutea ceroid pigment cause or a consequence of atrophy. We observe nuclear and cyto- plasmic staining of total STAT3 in the Sertoli and germ cells of both WT 31 1 0 −/− −/− WT 41 1 0 WT and Rex1 mice (Figs. 6G, H). Rex1 mice are characterized WT 41 1 0 by strong cytoplasmic staining of total Stat3, particularly in the − − KO 43 2 0 seminiferous tubules of Rex1 / mice with obvious, germ cell KO 15 0 0 depletion (Fig. 6G). Our data suggest that loss of Rex1 may lead to KO 16 2 0 depletion of the germ cell compartment via indirect activation of the Age: 21 days Two ovaries examined per animal, numbers are per pair JAK/STAT3 pathway. of ovaries unless otherwise specified ⁎ ⁎⁎ Genotype n follicles n corpora lutea ceroid pigment Rex1 supports proper cell proliferation and proper distribution of the germ cell and Sertoli cell populations within the seminiferous tubules WT 55 0 0 WT 62 0 0 WT 41 0 0 The mature testis is composed of many different cell types, KO 77 0 0 including germ cells and Sertoli cells. Germ cells are closely associated KO 80 0 0 with Sertoli cells throughout their development and differentiation, KO 83 0 0 and in wild type mice Rex1 is predominantly expressed in sper- KO 62 0 0 matocytes undergoing meiosis (Rogers et al., 1991). To examine how Age: 12 months Two ovaries examined per animal, numbers are per pair of the lack of Rex1 affects testis architecture, histological sections of ovaries unless otherwise specified testicular tissues from WT and Rex1−/− mice were immunostained ⁎ ⁎⁎ Genotype n follicles n corpora lutea ceroid pigment with markers of specific cell types. First, sections were stained WT 10 5 1 with DazL, a protein known to be expressed in immature germ cells −/− WT 9 7 1 (Saunders et al., 2003). In both WT and Rex1 mice, nuclear WT 12 7 2 staining was predominantly restricted to spermatogonia and was WT 3 10 1 evenly distributed within this population along the outer edges of the KO 5 8 1 KO 10 5 3 seminiferous tubules (Figs. 6A, B). This indicates that Rex1 expression KO 3 6 1 is not necessary for maintenance of the most immature type of germ KO 13 3 3 cells, the spermatogonia. KO 12 4 2 To evaluate the proliferation of germ cells, sections were stained KO 4 2 1 with proliferating cell nuclear antigen (PCNA) (Fig. 6C). In sections − − Histological analysis of the ovaries in WT and Rex1 / mice ranging from 21 days old, from both WT and Rex1−/− mice we observed positive staining for 6 weeks old and 1 year old. At 21 days and 1 year two ovaries were evaluated per PCNA within well organized spermatogonia (Fig. 6D). Consistent with animal and the numbers listed are per pair of ovaries. At 6 weeks one ovary was −/− examined per animal, and the numbers listed are per ovary. our observation that Rex1 mice display depletion of their germ ⁎ All types of follicles were counted, except primordial follicles. cell compartment, PCNA staining was often absent from spermato- − − ⁎⁎ Numbers correspond to scale measuring the level of ceroid pigment: 0: absent, gonia within the tubules of the sections from the Rex1 / mice 1: minimal, 2: mild, 3: moderate, 4: marked. (Fig. 6D). The structural design of the testis is well defined, with germ cell differentiation and maturation occurring in well synchronized steps. Sox9 expression appears to be tightly associated with Sertoli cell development (Kent et al., 1996; Morais da Silva et al., 1996; Vidal difference between the numbers of seminiferous tubules expressing et al., 2001). Sox9 expression remains in Sertoli cells exclusively phospho-p38 between WT and Rex1−/− mice (Supplemental after testis cord formation and is used as a Sertoli cell marker Fig. 2C). Comparable total p38 levels were observed in the cytoplasm (Hemendinger et al., 2002; Morais da Silva et al., 1996). To investi- of both Sertoli and a range of immature germ cells in WT and gate the status of Sertoli cells in Rex1−/− mice, we examined Sox9 Rex1−/− (Supplemental Figs. 2D, E). In mature spermatids, p38 protein in WT and Rex1−/− testes. As expected, Sox9 staining is staining was detected in the nucleus of unaffected seminiferous observed in the nuclei of the Sertoli cells of WT mice (Fig. 6E). We tubules in WT and Rex1−/− mice (Supplemental Figs. 3D, E). also observe similar patterns of Sox9 staining in Rex1−/− mice Our laboratory previously examined the role of Rex1 in murine F9 (Fig. 6F). The staining of the immature germ cells (DazL), prolife- teratocarcinoma stem cell during differentiation, and we discovered rating germ cells (PCNA), and Sertoli cells (Sox9) within WT and that F9 cells that lack Rex1 exhibited a large increase in transcription Rex1−/− seminiferous tubules (Fig. 7), along with the observed of the supressor of cytokine signaling 3 (SOCS3) gene (Xu et al., 2008). depletion of mature spermatids(HematoxylinandEosin)and We found that this increase occurs via the signal transducer and abnormal sperm morphology in Rex1−/− mice, suggest that Rex1 activator of transcription (STAT3) binding sites located at −99 to −60 expression influences differentiation of mouse germ cells at later of the SOCS3 promoter. We further showed that overexpression of a stages. This indicates that Rex1 mRNA may be necessary for proper Stat3 dominant negative (DN) mutant diminished the increase of differentiation or maturation of germ cells. SOCS3 expression in F9 cells (Xu et al., 2008). To gain insight into possible mechanisms leading to increased apoptosis in the testes of Discussion Rex1−/− mice, we investigated the levels of active and total Stat3 protein in the testis. To this end, we stained testes for P-Stat3 (Tyr705) Rex1(Zfp42) is a zinc finger transcription factor (Hosler et al., (WT, n=25, KO, n=26) (Figs. 6D, E, F), and for total Stat3 levels 1989) and is expressed in human and murine stem cells (Boyer et al., (WT, n=27, KO, n=24) (Figs. 6G, H). Most WT and Rex1−/− testes 2006; Brivanlou et al., 2003; Loh et al., 2006; Masui et al., 2007; are negative for P-Stat3 (Tyr705) staining. However, strong P-Stat3 Mongan et al., 2006; Rogers et al., 1991). Although Rex1 is not (Tyr705) nuclear staining was observed in the Sertoli cells of the six necessary for the maintenance of ES cell pluripotency, Rex1−/− 378 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 A Rex1+/+

B Rex1-/-

C and 4N DNA content % of cells with 1N, 2N

Fig. 4. Cell cycle analysis. Testicular cells were prepared from the testes of 4–6 month old mice (WT, n=6, Rex1−/−, n=7) following the protocol described in Materials & methods. ~2×106 cells were stained with propidium iodide (PI) to analyze the cell cycle. * pb0.05, t-test. deficient ES cells do show a greater tendency to differentiate towards mutant mice displayed abnormal sperm morphology. These data could somatic lineages (Scotland et al., 2009; Toyooka et al., 2008). To indicate defects in DNA replication or segregation in understand the function of Rex1 in stem cells, we made a conditional meiosis, a phase during which differentiation/maturation of germ cells knockout mouse that lacks this gene. We show here that mutant occurs and when Rex1 is highly expressed (Kristensen et al., 2008; mice deficient in Rex1 are viable and fertile. This is the first Rogers et al., 1991). Alternatively, this phenotype could indicate that conditional Rex1 knockout mouse line to be reported. Rex1−/− mice are undergoing accelerated aging (Zhang et al., 2006). Yet Since the testis expresses a high level of Rex1 RNA (Rogers et al., another possibility is that the phenotypes can be attributed to defects in 1991), we examined Rex1−/− mutant mice testes and found mild X chromosome inactivation. X chromosome inactivation is an example to moderate germ cell depletion (Figs. 2B, C, D, E), with early/late of an epigenetic regulatory mechanism used to equalize the X linked spermatids being the population mostly affected. Additionally, Rex1−/− gene dosage. During mammalian germ cell development the X N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 379 A

Map2k7 Dctn6 primary germ cells -/-

95300002B09Rik Grtp1 in Rex1 1700104B16Rik Expression levels of detected transcripts Expression levels ENSMUSG00000074383 Defb11

Expression levels of detected transcripts in Rex1+/+ primary germ cell BC Grtp1 WTWT Rex1-/- Rex1-/- 15 **

Grtp1 (179 bp) 10 **

Rex1 (670 bp) 5

36B4 (450 bp) 0

Fig. 5. Microarray analysis. Primary germ cells were isolated from 6 week old WT and Rex1−/− mice. RNA from primary germ cells was isolated with Trizol. Normalized intensity value (Log scale) of WT (X-axis) and Rex1−/− (Y-axis) primary germ cell transcripts with a fold change N1.2 (A). Individual dots correspond to unique genes with a change in expression greater than 1.2 fold between WT and Rex1−/− primary germ cells. Semi-quantitative RT-PCR analysis of primary germ cell cDNA and mES cells (B). Validation of one of the differentially expressed genes identified in the array, GRTP1, was performed in another preparation of newly isolated primary germ cells by semi-quantitative RT-PCR (B), and real time quantitative PCR (C). (B) Levels of Grtp1, Rex1 and 36B4 transcripts shown in duplicate samples of newly isolated primary germ cells from 6 week old mice. (C) Quantitation of the signal from the samples described in (B) by real time PCR. chromosome is transiently silenced at meiosis (Namekawa et al., 2006; panied by morphological and biochemical changes (Johnston et al., 2008) Nguyen and Disteche, 2006), presumably to equalize the X linked gene and changes in regulation of the cell cycle (Kierszenbaum, 2006). The dosage during those critical steps of development. In cultured mES cells entire process can be subdivided into 12 stages. Although we have not Rex1 was recently identified as a required factor for efficient elongation analyzed each step in depth, we studied cellular pathways relevant to the of , a key non-coding transcript involved in X-inactivation (Navarro observed phenotype. We did not observe statistically significant et al., 2010), and we and others have identified key regulators of X- differences between populations with 1N, 2N and 4N DNA content inactivation as differentially expressed genes in global gene expression within the testes of 4–6 month old mice, indicating that Rex1 ablation analysis (Masui et al., 2008; Scotland et al., 2009). Kim et al. recently does not interfere with cell cycle distribution in male germ cells (Fig. 3). characterized a mouse line in which the β-geo gene trap vector was Similarly, the evaluation of apoptosis in the testis does not show a inserted into the third intron of the Rex1 gene; since the entire open statistically significant difference between WT and Rex1−/− mice, reading frame for Rex1 protein is present in the fourth exon (Hosler et although there was high variability within individual testes. The al., 1989, Hosler et al., 1993), we would predict that the entire Rex1 above phenotypes of Rex1−/− mice are variable, suggesting incomplete protein would still be produced in this β-geo gene trap mouse line (Kim penetrance. One possible explanation for this is that the ubiquitous et al., 2011). polycomb transcription factor Yin Yang 1 (YY1) may be compensating for Spermatogenesis is a highly organized process that takes place in Rex1 to some extent (Kim et al., 2007; Mongan et al., 2006). A high degree three distinct phases in the seminiferous epithelium: spermatogonial of homology exists between Rex1 and YY1, and although YY1 gene proliferation (spermatogoniogenesis), meiosis of spermatocytes, and knockout results in embryonic lethality during peri-implantation differentiation of haploid round spermatids into elongated spermatids (Donohoe et al., 1999), a conditional knockout of YY1 showed defects (spermiogenesis) (Leblond and Clermont, 1952). Each phase is accom- in spermatogenesis (Wu et al., 2009). When YY1 expression was lost in 380 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382

ABC

D EF

GH

Fig. 6. TUNEL assay and STAT3 analysis. Apoptosis of cells in the testis was evaluated by the TUNEL assay. Testes from six week old mice (wt, n=21, ko, n=19) were removed and fixed in paraffin. Blocks were sectioned into 7 μm thick sections and stained with the TUNEL assay kit following the Roche manufacturer's protocol (A, B). Positive staining was counted in an entire section of each mouse and averaged for each group to compare (C). Status of phospho-Stat3 (Tyr705) was evaluated in WT (D), and Rex1−/− (E) testes. (E) Black arrows highlight strong P-STAT3 nuclear staining in Sertoli cells. Number of P-STAT3 positive staining cells in 25 WT and 26 Rex1−/− whole testis sections (F). Status of total Stat3 was evaluated in WT (G), and Rex1−/− testes (H). Black arrows (H) highlight strong cytoplasmic staining of total STAT3 in the Rex1−/− testes with marked germ cell depletion. Scale 50 μm. the testes, testis size was reduced to a third the size of YY1 heterozygous not contain receptors for hormones (Ruwanpura et al., 2010). It is mice (YY1f/+/Cre), and the testes had abnormal and vacuolated semi- unclear if the lack of Rex1 leads to a cell intrinsic defect in germ cell niferous tubules and less pachytene spermatocytes (Wu et al., 2009). hormonal response. In mouse ES cells the key regulators of Stat3 is an essential mammalian protein with well-defined roles in mammalian X-inactivation, Tsix and Xist, were respectively identified the maintenance of mES cell pluripotency (Raz et al., 1999). Phosphory- as a 3.6 fold downregulated gene in Rex1−/− (Masui et al., 2008), lation of the STAT3 protein leads to the formation of active STAT3 and a ~2.9 fold upregulated gene in the presence and absence of LIF dimers, which translocate to the nucleus and work as active transcrip- (Scotland et al., 2009). We have previously studied in detail the tion factors (Zhang et al., 1995). Rex1−/− mice show strong cytoplasmic mechanisms underlying the transcriptional activation of the suppres- staining of total Stat3 in Sertoli cells, spermatogonia, and spermatocytes, sor of cytokine signaling 3 (SOCS3) gene by a combination treatment particularly in the seminiferous tubules with obvious atrophy/germ cell of retinoic acid, theophylline and dibutyryl cyclic AMP in F9 cells. By depletion (Fig. 5G). Characterization of the more immature germ cells promoter deletion, mutation and transient transfection analysis (DazL), proliferating germ cells (PCNA), and Sertoli cells (Sox9) we identified this transcriptional increase to be mediated by STAT3 (Figs. 6A-F) indicates that those populations are unaffected by the lack DNA binding elements in the SOCS-3 promoter. We also observed of Rex1 expression. These results, along with the abnormal sperm activation of phospho-STAT3 in the context of differentiation when a morphologies (Fig. 2G) and spermatid depletion in Rex1−/− mice, retinoic acid, theophylline and dibutyryl cyclic AMP combination indicate that Rex1 expression affects the later stages of spermatogenesis. treatment was used; we concluded that Rex1 plays an indirect role Lastly, from the microarray analysis, we identified a growth in the Janus Kinase (JAK)/signal transducer and activator pathway. hormone regulated gene, GRTP1, that was transcriptionaly down- Our results provide in vivo genetic evidence demonstrating that regulated in Rex1−/− germ cells. Although hormones such as follicle Rex1 plays an important role within the testis. We are currently stimulating hormone (FSH), luteinizing hormone (LH), and testoster- investigating if loss of Rex1 leads to defects in other adult stem cell one have all been shown to play important roles in germ cell biology, compartments and/or misregulation of adult stem cell differentiation. it is generally accepted that hormone triggered physiology effects Supplementary materials related to this article can be found online occur via signaling from the neighboring Sertoli cells, as germ cells do at doi:10.1016/j.ydbio.2011.05.664. N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382 381

Wild Type AB

DazL

CD

PCNA

EF

Sox9

Fig. 7. Evaluation of Germ cell, proliferating, and Sertoli cell populations immunohistological evaluation. Six week old mice were assessed by immunohistochemistry using antibodies to Dazl, PCNA and Sox9. (A, B) Dazl staining of both wild type and Rex1−/− mice (33 wt, 28 Rex1−/−). Black arrows highlight examples of the most immature, Dazl positive germ cells within the WT and Rex1−/− testis. (C, D) PCNA staining of WT and Rex1−/− testes (33 wt, 28 Rex1−/−). Black arrows highlight examples of proliferating germ cells. (E, F) Sox9 staining of WT and Rex1−/− testes (34 wt, 26 Rex1−/−). Black arrows highlight examples of Sertoli cells of both WT and Rex1−/− mice.

Acknowledgments Bhandari, D.R., Seo, K.W., Roh, K.H., Jung, J.W., Kang, S.K., Kang, K.S., 2010. REX-1 expression and p38 MAPK activation status can determine proliferation/differen- tiation fates in human mesenchymal stem cells. PLoS One 5, e10493. We thank all of the members of the Gudas laboratory for scientific Boyer, L.A., Plath, K., Zeitlinger, J., Brambrink, T., Medeiros, L.A., Lee, T.I., Levine, S.S., discussions, Tamara Weissman for editorial assistance, the WCMC Wernig, M., Tajonar, A., Ray, M.K., Bell, G.W., Otte, A.P., Vidal, M., Gifford, D.K., Young, R.A., Jaenisch, R., 2006. Polycomb complexes repress developmental Genomics Core Facility and the Histology Core of Weill Cornell Medical regulators in murine embryonic stem cells. Nature 441, 349–353. College for section preparation, and the Transgenic Core Facility for ES Brivanlou, A.H., Gage, F.H., Jaenisch, R., Jessell, T., Melton, D., Rossant, J., 2003. Stem cells. cell injections into blastocysts. We would like to thank Sergei Setting standards for human embryonic stem cells. Science 300, 913–916. Rudchenko and Stanka Semova for their assistance with the FACS Chan, E.M., Ratanasirintrawoot, S., Park, I.H., Manos, P.D., Loh, Y.H., Huo, H., Miller, J.D., Hartung, O., Rho, J., Ince, T.A., Daley, G.Q., Schlaeger, T.M., 2009. Live cell imaging analysis in the WCMC FACS Core Facility, and Jacqueline Bromberg distinguishes bona fide human iPS cells from partially reprogrammed cells. Nat. from Memorial Sloan Kettering Cancer Center for guidance with STAT3 Biotechnol. 27, 1033–1037. and p-STAT3 staining. This research was supported in part by Donohoe, M.E., Zhang, X., McGinnis, L., Biggers, J., Li, E., Shi, Y., 1999. Targeted disruption of mouse Yin Yang 1 transcription factor results in peri-implantation lethality. Mol. NIHR01CA043796 to L.J.G., and by Weill Cornell funds. M.Y.L was Cell. Biol. 19, 7237–7244. supported for a portion of this research by DOD grant W81XWH-06-1- Han, J., Lee, J.D., Bibbs, L., Ulevitch, R.J., 1994. A MAP kinase targeted by endotoxin and 0109. N.C.R. is a Howard Hughes Medical Institute Gilliam Fellow. hyperosmolarity in mammalian cells. Science 265, 808–811. Hemendinger,R.A.,Gores,P.,Blacksten,L.,Harley,V.,Halberstadt,C.,2002.Identification of a specific Sertoli cell marker, Sox9, for use in transplantation. Cell Transplant. 11, 499–505. References Hogarth, C.A., Griswold, M.D., 2010. The key role of vitamin A in spermatogenesis. J. Clin. Invest. 120, 956–962. Agrawal, V., Johnson, S.A., Reing, J., Zhang, L., Tottey, S., Wang, G., Hirschi, K.K., Braunhut, Hosler, B.A., LaRosa, G.J., Grippo, J.F., Gudas, L.J., 1989. Expression of REX-1, a gene S., Gudas, L.J., Badylak, S.F., 2010. Epimorphic regeneration approach to tissue containing zinc finger motifs, is rapidly reduced by retinoic acid in F9 teratocarcinoma replacement in adult mammals. Proc. Natl. Acad. Sci. U. S. A. 107, 3351–3355. cells. Mol. Cell. Biol. 9, 5623–5629. Baubonis, W., Sauer, B., 1993. Genomic targeting with purified Cre recombinase. Nucleic Hosler, B.A., Rogers, M.B., Kozak, C.A., Gudas, L.J., 1993. An octamer motif contributes to Acids Res. 21, 2025–2029. the expression of the retinoic acid-regulated zinc finger gene Rex-1(Zfp-42) in F9 Ben-Shushan, E., Thompson, J.R., Gudas, L.J., Bergman, Y., 1998. Rex-1, a gene encoding a teratocarcinoma cells. Mol. Cell. Biol. 13, 2919–2928. transcription factor expressed in the early embryo, is regulated via Oct-3/4 and Oct- Johnston, D.S., Wright, W.W., Dicandeloro, P., Wilson, E., Kopf, G.S., Jelinsky, S.A., 2008. 6 binding to an octamer site and a novel protein, Rox-1, binding to an adjacent site. Stage-specific gene expression is a fundamental characteristic of rat spermatogenic Mol. Cell. Biol. 18, 1866–1878. cells and Sertoli cells. Proc. Natl. Acad. Sci. U. S. A. 105, 8315–8320. 382 N.C. Rezende et al. / Developmental Biology 356 (2011) 370–382

Kent, J., Wheatley, S.C., Andrews, J.E., Sinclair, A.H., Koopman, P., 1996. A male-specific Pepling, M.E., 2006. From primordial germ cell to primordial follicle: mammalian role for SOX9 in vertebrate sex determination. Development 122, 2813–2822. female germ cell development. Genesis 44, 622–632. Kierszenbaum, A.L., 2006. Cell-cycle regulation and mammalian gametogenesis: a Raman, J.D., Mongan, N.P., Liu, L., Tickoo, S.K., Nanus, D.M., Scherr, D.S., Gudas, L.J., 2006. lesson from the unexpected. Mol. Reprod. Dev. 73, 939–942. Decreased expression of the human stem cell marker, Rex-1(zfp-42), in renal cell Kim, J.D., Faulk, C., Kim, J., 2007. Retroposition and evolution of the DNA-binding motifs carcinoma. Carcinogenesis 27, 499–507. of YY1, YY2 and REX1. Nucleic Acids Res. 35, 3442–3452. Raz, R., Lee, C.K., Cannizzaro, L.A., d'Eustachio, P., Levy, D.E., 1999. Essential role of STAT3 Kim, J.D., Kim, H., Ekram, M.B., Yu, S., Faulk, C., Kim, J., 2011. Rex1/Zfp42 as an epigenetic for embryonic stem cell pluripotency. Proc. Natl. Acad. Sci. U. S. A. 96, 2846–2851. regulator for genomic imprinting. Hum. Mol. Genet. Ribeiro, L.R., Salvadori, D.M., Pereira, C.A., Becak, W., 1987. Activity of ethylene oxide in Kristensen, D.M., Nielsen, J.E., Skakkebaek, N.E., Graem, N., Jacobsen, G.K., Rajpert-De the mouse sperm morphology test. Arch. Toxicol. 60, 331–333. Meyts, E., Leffers, H., 2008. Presumed pluripotency markers UTF-1 and REX-1 are Rogers, M.B., Hosler, B.A., Gudas, L.J., 1991. Specific expression of a retinoic acid- expressed in human adult testes and germ cell neoplasms. Hum. Reprod. 23, 775–782. regulated, zinc-finger gene, Rex-1, in preimplantation embryos, trophoblast and Kristensen, D.M., Nielsen, J.E., Kalisz, M., Dalgaard, M.D., Audouze, K., Larsen, M.E., spermatocytes. Development 113, 815–824. Jacobsen, G.K., Horn, T., Brunak, S., Skakkebaek, N.E., Leffers, H., 2010. OCT4 and Rosfjord, E., Rizzino, A., 1994. The octamer motif present in the Rex-1 promoter binds downstream factors are expressed in human somatic urogenital epithelia and in Oct-1 and Oct-3 expressed by EC cells and ES cells. Biochem. Biophys. Res. culture of epididymal spheres. Mol. Hum. Reprod. Commun. 203, 1795–1802. Leblond, C.P., Clermont, Y., 1952. Definition of the stages of the cycle of the seminiferous Ruwanpura, S.M., McLachlan, R.I., Meachem, S.J., 2010. Hormonal regulation of male epithelium in the rat. Ann. N. Y. Acad. Sci. 55, 548–573. germ cell development. J. Endocrinol. 205, 117–131. Lee, M.Y., Lu, A., Gudas, L.J., 2010. Transcriptional regulation of Rex1 (zfp42) in normal Sakai, K., Miyazaki, J., 1997. A transgenic mouse line that retains Cre recombinase prostate epithelial cells and prostate cancer cells. J. Cell. Physiol. 224, 17–27. activity in mature oocytes irrespective of the Cre transgene transmission. Biochem. Loh, Y.H., Wu, Q., Chew, J.L., Vega, V.B., Zhang, W., Chen, X., Bourque, G., George, J., Biophys. Res. Commun. 237, 318–324. Leong, B., Liu, J., Wong, K.Y., Sung, K.W., Lee, C.W., Zhao, X.D., Chiu, K.P., Lipovich, L., Saunders, P.T., Turner, J.M., Ruggiu, M., Taggart, M., Burgoyne, P.S., Elliott, D., Cooke, H.J., Kuznetsov, V.A., Robson, P., Stanton, L.W., Wei, C.L., Ruan, Y., Lim, B., Ng, H.H., 2006. 2003. Absence of mDazl produces a final block on germ cell development at The Oct4 and Nanog transcription network regulates pluripotency in mouse meiosis. Reproduction 126, 589–597. embryonic stem cells. Nat. Genet. 38, 431–440. Scotland, K.B., Chen, S., Sylvester, R., Gudas, L.J., 2009. Analysis of Rex1 (zfp42) function Lu, C., Kasik, J., Stephan, D.A., Yang, S., Sperling, M.A., Menon, R.K., 2001. Grtp1, a novel in embryonic stem cell differentiation. Dev. Dyn. 238, 1863–1877. gene regulated by growth hormone. Endocrinology 142, 4568–4571. Shi, W., Wang, H., Pan, G., Geng, Y., Guo, Y., Pei, D., 2006. Regulation of the pluripotency Malkov, M., Fisher, Y., Don, J., 1998. Developmental schedule of the postnatal rat testis marker Rex-1 by Nanog and Sox2. J. Biol. Chem. 281, 23319–23325. determined by flow cytometry. Biol. Reprod. 59, 84–92. Thompson, J.R., Gudas, L.J., 2002. Retinoic acid induces parietal endoderm but not Masui,S.,Nakatake,Y.,Toyooka,Y.,Shimosato,D.,Yagi,R.,Takahashi,K.,Okochi,H.,Okuda,A., primitive endoderm and visceral endoderm differentiation in F9 teratocarcinoma Matoba, R., Sharov, A.A., Ko, M.S., Niwa, H., 2007. Pluripotency governed by Sox2 via stem cells with a targeted deletion of the Rex-1(Zfp-42) gene. Mol. Cell. Endocrinol. regulation of Oct3/4 expression in mouse embryonic stem cells. Nat. Cell Biol. 9, 625–635. 195, 119–133. Masui, S., Ohtsuka, S., Yagi, R., Takahashi, K., Ko, M.S., Niwa, H., 2008. Rex1/Zfp42 is Toyooka, Y., Shimosato, D., Murakami, K., Takahashi, K., Niwa, H., 2008. Identification dispensable for pluripotency in mouse ES cells. BMC Dev. Biol. 8, 45. and characterization of subpopulations in undifferentiated ES cell culture. Meng, X., Lindahl, M., Hyvonen, M.E., Parvinen, M., de Rooij, D.G., Hess, M.W., Development 135, 909–918. Raatikainen-Ahokas, A., Sainio, K., Rauvala, H., Lakso, M., Pichel, J.G., Westphal, H., Vidal, V.P., Chaboissier, M.C., de Rooij, D.G., Schedl, A., 2001. Sox9 induces testis Saarma, M., Sariola, H., 2000. Regulation of cell fate decision of undifferentiated development in XX transgenic mice. Nat. Genet. 28 (3), 216–217. spermatogonia by GDNF. Science 287, 1489–1493. Wang, G., Badylak, S.F., Heber-Katz, E., Braunhut, S.J., Gudas, L.J., 2010. The effects of Mongan, N.P., Gudas, L.J., 2007. Diverse actions of retinoid receptors in cancer DNA methyltransferase inhibitors and histone deacetylase inhibitors on digit prevention and treatment. Differentiation 75, 853–870. regeneration in mice. Regen. Med. 5, 201–220. Mongan, N.P., Martin, K.M., Gudas, L.J., 2006. The putative human stem cell marker, Rex- Wu, S., Hu, Y.C., Liu, H., Shi, Y., 2009. Loss of YY1 impacts the heterochromatic state and 1(Zfp42): structural classification and expression in normal human epithelial and meiotic double-strand breaks during mouse spermatogenesis. Mol. Cell. Biol. 29, carcinoma cell cultures. Mol. Carcinog. 45, 887–900. 6245–6256. Morais da Silva, S., Hacker, A., Harley, V., Goodfellow, P., Swain, A., Lovell-Badge, R., Xu, J., Sylvester, R., Tighe, A.P., Chen, S., Gudas, L.J., 2008. Transcriptional activation of 1996. Sox9 expression during gonadal development implies a conserved role for the suppressor of cytokine signaling-3 (SOCS-3) gene via STAT3 is increased in F9 the gene in testis differentiation in mammals and birds. Nat. Genet. 14, 62–68. REX1 (ZFP-42) knockout teratocarcinoma stem cells relative to wild-type cells. Namekawa, S.H., Park, P.J., Zhang, L.F., Shima, J.E., McCarrey, J.R., Griswold, M.D., Lee, J.T., J. Mol. Biol. 377, 28–46. 2006. Postmeiotic sex chromatin in the male germline of mice. Curr. Biol. 16, 660–667. Yoshinaga, K., Nishikawa, S., Ogawa, M., Hayashi, S., Kunisada, T., Fujimoto, T., 1991. Navarro, P., Oldfield, A., Legoupi, J., Festuccia, N., Dubois, A., Attia, M., Schoorlemmer, J., Role of c-kit in mouse spermatogenesis: identification of spermatogonia as a Rougeulle, C., Chambers, I., Avner, P., 2010. Molecular coupling of Tsix regulation specific site of c-kit expression and function. Development 113, 689–699. and pluripotency. Nature 468, 457–460. Yu, Z., Dadgar, N., Albertelli, M., Scheller, A., Albin, R.L., Robins, D.M., Lieberman, A.P., Nguyen, D.K., Disteche, C.M., 2006. Dosage compensation of the active X chromosome 2006. Abnormalities of germ cell maturation and sertoli cell cytoskeleton in mammals. Nat. Genet. 38, 47–53. in androgen receptor 113 CAG knock-in mice reveal toxic effects of the mutant Niwa, H., Yamamura, K., Miyazaki, J., 1991. Efficient selection for high-expression protein. Am. J. Pathol. 168, 195–204. transfectants with a novel eukaryotic vector. Gene 108 (2), 193–199. Zhang, X., Blenis, J., Li, H.C., Schindler, C., Chen-Kiang, S., 1995. Requirement of serine O'Shaughnessy, P.J., Morris, I.D., Huhtaniemi, I., Baker, P.J., Abel, M.H., 2009. Role of phosphorylation for formation of STAT-promoter complexes. Science 267, androgen and gonadotrophins in the development and function of the Sertoli cells 1990–1994. and Leydig cells: data from mutant and genetically modified mice. Mol. Cell. Zhang, X., Ebata, K.T., Robaire, B., Nagano, M.C., 2006. Aging of male germ line stem cells Endocrinol. 306, 2–8. in mice. Biol. Reprod. 74, 119–124.