381 FSH regulates the formation of adherens junctions and ectoplasmic specialisations between rat Sertoli cells in vitro and in vivo

P Sluka1,2, L O’Donnell1, J R Bartles3 and P G Stanton1 1Prince Henry’s Institute of Medical Research, Level 4, Block E, Monash Medical Centre, 246 Clayton Road, Clayton, Victoria 3168, Australia 2Department of Anatomy and Cell Biology, Monash University, Clayton, Victoria 3168, Australia 3Department of Cell and Molecular Biology, Northwestern University Medical School, Chicago, Illinois 60611, USA (Requests for offprints should be addressed to P Stanton; Email: [email protected])

Abstract Spermatogenesis is dependent on the ability of Sertoli cells but ES junctions were not present. The adherens junc- to form mature junctions that maintain a unique environ- tion puncta included filaments that were oriented ment within the seminiferous epithelium. Adjacent Sertoli perpendicularly to the Sertoli cell plasma membrane, but cells form a junctional complex that includes classical were not associated with the intermediate filament adherens junctions and testis-specific ectoplasmic speciali- . When FSH was added in vitro, ES junctions sations (ES). The regulation of inter-Sertoli cell junctions formed, and adjacent adherens junction puncta fused into by the two main endocrine regulators of spermatogenesis, extensive adherens junction belts. After hormone suppres- FSH and testosterone, is unclear. This study aimed to sion in vivo, ES junctions were absent, while FSH replace- investigate the effects of FSH and testosterone on inter- ment restored ES junctions, as confirmed by electron Sertoli cell adherens junctions (as determined by immu- microscopy and confocal analysis of ES-associated pro- nolocalisation of cadherin, and actin) and ES teins. Testosterone alone did not affect adherens junctions junctions (as determined by immunolocalisation of espin, or ES in vitro or in vivo. We conclude that FSH can actin and ) in cultured immature Sertoli cells and regulate the formation of ES junctions and stimulate GnRH-immunised adult rat testes given FSH or testos- the organisation and orientation of extensive adherens terone replacement in vivo. When hormones were absent junctions in Sertoli cells. in vitro, adherens junctions formed as discrete puncta Journal of Endocrinology (2006) 189, 381–395 between interdigitating, finger-like projections of Sertoli cells,

Introduction blood–testis barrier. The formation of this barrier is essential for normal germ cell development (Steinberger & Spermatogenesis is defined as the process of male germ cell Steinberger 1975, Weber et al. 1988). development, and involves the division, Typical AJs are found between Sertoli cells and are akin reduction and morphological differentiation of diploid to those found in all classical epithelia. AJs are character- stem-line germ cells to produce haploid spermatozoa ised by the presence of transmembrane, calcium- (Steinberger & Steinberger 1975). Sertoli cells are the dependent cadherins that link to the actin via somatic cells of the testis, serving to nurture germ cells as (see Lui et al. (2003) for recent review) and are they proceed through spermatogenesis (Steinberger & visible by electron microscopy as focal opposing subsur- Steinberger 1975). Sertoli cells are polarised epithelial cells face densities on plasma membranes (Russell & Peterson that interact with one another via a unique and large 1985, Byers et al. 1993). In classical epithelia, AJs are complex of basolaterally located junctions (Vogl et al. formed initially as AJ puncta where actin filaments ter- 1993). This complex consists of various junction types, minate perpendicularly to the plasma membrane, and, including adherens junctions (AJ), desmosome-like upon appropriate stimulation, reorganise to form AJ belts junctions, tight junctions and gap junctions (Russell & in which actin filaments are oriented parallel to the plasma Peterson 1985, Vogl et al. 1993), and is further character- membrane (Krendel & Bonder 1999, Vasioukhin & Fuchs ised by the presence of testis-specific Sertoli cell structures 2001). Although the presence of inter-Sertoli cell AJs has termed ‘ectoplasmic specialisations’ (ES) (Vogl et al. been well documented (reviewed in Russell & Peterson 2000). The entire inter-Sertoli cell junctional complex 1985 and Lui et al. 2003), the regulation of AJ puncta regulates the movement of substances into and out of the and AJ belt formation in Sertoli cells has not yet been seminiferous epithelium, thus forming the basis of the described.

Journal of Endocrinology (2006) 189, 381–395 DOI: 10.1677/joe.1.06634 0022–0795/06/0189–381  2006 Society for Endocrinology Printed in Great Britain Online version via http://www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 382 P SLUKA and others · Regulation of Sertoli cell junctions

The ES is generally regarded as a testis-specific, actin- increased germ cell apoptosis in a setting of normal serum based adhesion junction (Vogl et al. 2000, Toyama et al. and testicular testosterone (Meachem et al. 1999). 2003, Mruk & Cheng 2004); however, they exhibit In order to begin to understand the relationship molecular, ultrastructural and regulatory differences from between hormones, spermatogenesis and Sertoli cell junc- classical AJs. ES junctions are found only in the testis, and tions, earlier studies assessed the effects of FSH and are a Sertoli cell cytoskeletal structure characterised by testosterone on Sertoli cell junctional structures that hexagonally packed bundles of actin filaments, oriented contain actin and vinculin in adult rats (Muffly et al. 1993, parallel to the Sertoli cell plasma membrane, and sand- 1994). The reduction of luteinizing hormone (LH) and wiched between the plasma membrane and cisternae of FSH by hypophysectomy resulted in an abnormal pattern the endoplasmic reticulum (Russell 1977, Vogl et al. of actin and vinculin staining in the seminiferous epi- 1991). The packing of actin filaments into hexagonal thelium (Muffly et al. 1993, 1994). The subsequent bundles is achieved by the actin-bundling protein espin restoration of FSH, but not testosterone (Muffly et al. (Bartles et al. 1996), while vinculin, also localised to the 1993, 1994), re-established the normal organisation of layer of actin filaments (Pfeiffer & Vogl 1991), is thought these junctional molecules, which correlated with the to link the actin filaments of the ES to transmembrane restoration of spermatogenesis. These data indicate that molecules (Pokutta & Weis 2002). Furthermore, actin FSH promotes the ability of Sertoli cells to form normal filaments at ES junctions are arranged in a regular and junctions, both with themselves and with other germ cells, unipolar manner, in contrast to classical AJs (Vogl et al. and this, in turn, is important for normal spermatogenesis 1993), and do not exhibit contractile properties (Vogl (Muffly et al. 1993, 1994). FSH has also been shown to et al. 1993). ES junctions are present at two sites within regulate the formation of integrin-containing junctions the seminiferous epithelium; between adjacent Sertoli between developing Sertoli cells (Salanova et al. 1998). cells (basal ES junctions), and between Sertoli cells and However, the regulation, mechanism of formation and elongating and elongated spermatids (apical ES junctions). interdependence of different junction types at the inter- Basal ES junctions are characterised by two opposing ES Sertoli cell junctional complex remain largely unknown. structures (Vogl et al. 2000). It is hypothesised that FSH and/or testosterone regulate The transmembrane and intercellular adhesion mol- the formation of mature adhesion junctions, including AJs ecules that comprise the putative adhesive domain of ES and ES junctions, between Sertoli cells. Hence, the aim of junctions are unclear, although 1-integrin is a likely this study was to examine the role of FSH and testosterone candidate (Mulholland et al. 2001). The nectin-afadin in the formation of classical cadherin/catenin-based AJs system may also mediate adhesion at sites of apical ES and the Sertoli cell-specific ES junction, containing actin junctions (Ozaki-Kuroda et al. 2002). In contrast, cadher- and espin, in immature rat Sertoli cells in vitro, and in adult ins, the classical transmembrane adhesion molecules found rats in vivo. The gonadotrophin-releasing hormone at AJs (Johnson & Boekelheide 2002), have been reported (GnRH)-immunised rodent model was chosen, as it is to be absent from ES junctions (Mulholland et al. 2001, analogous to current androgen-based male contraceptive Johnson & Boekelheide 2002), although a recent study methods (McLachlan et al. 2002). suggests that N-cadherin is present at sites consistent with ES junctions (Lee et al. 2004). Although it is well known that testosterone is essential Materials and Methods for spermatogenesis, a consensus is developing that follicle-stimulating hormone (FSH) is also required for Animals quantitatively normal spermatogenesis in the rodent (see McLachlan et al. (2002) and O’Donnell et al. (2005) for Male Sprague-Dawley rats (19–21 or 75–85 days old) obtained from Monash Animal Services (Monash review). Transgenic male mice with targeted disruption of  the FSH are fertile (Kumar et al. 1997), while male University, Clayton, Australia) were maintained at 20 C mice lacking a functional FSH receptor show some in a fixed 12-h light–dark cycle with free access to food impairment of fertility (Dierich et al. 1998, Krishnamurthy and water. Animal experimentation was approved by the et al. 2000). However, a closer examination of the Monash Medical Centre animal ethics committee. phenotypes of these mice revealed defects in the ability of Sertoli cells to support germ cells in FSH-null mice In vitro studies (Wreford et al. 2001) and defects to germ cells, such as elongated spermatids (Krishnamurthy et al. 2000) and Sertoli cell isolation and culture Immature Sertoli cells Sertoli cells (Grover et al. 2004), in FSH receptor-null for use in culture were isolated by previously published mice. Further support for a role for FSH in spermatogen- protocols (Cameron & Muffly 1991, Perryman et al. esis comes from a study demonstrating that short-term 1996). Briefly, twelve 19–21-day-old male Sprague (1-week) passive immunoneutralisation of circulating Dawley rats were killed by CO2 asphyxiation. Testes were FSH in normal rats decreased germ cell numbers and removed and decapsulated, and seminiferous tubules were

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 383

finely chopped. Interstitial tissue was removed by diges- were based on previously published work demonstrating tion with trypsin (2·5 mg/ml; Sigma)/DNase (50 µg/ml; that these doses maximally stimulate the binding of Sigma) dissolved in Dulbecco’s PBS (DPB; Gibco), using isolated round spermatids to a monolayer of Sertoli cells an orbital shaking air bath (25 min, 37 C, 85 r.p.m.). in vitro (Cameron & Muffly 1991, Perryman et al. 1996). Digestion was halted by the addition of 10% fetal calf In addition, the dose of FSH used maximally stimulates serum (FCS) (Trace Biosciences, Sydney, Australia)/ the in vitro production of two Sertoli cell-derived prod- DNase (10 µg/ml) in DPB. Seminiferous tubules were ucts, N-cadherin (Perryman et al. 1996) and inhibin separated from interstitial fragments by unit gravity sedi- (Lampa et al. 1999), and the dose of testosterone used is mentation, followed by a series of wash/sediment cycles in similar to the levels measured in the 20-day-old rat testis 10% FCS/DNase (10 µg/ml), and three washes in DNase (Killian et al. 2003). (20 µg/ml) in DPB. Seminiferous tubules were then resuspended with 1% BSA (Sigma)/DNase (10 µg/ml) in Sertoli cell culture fixation After culture, Sertoli cell DPB and processed in a glass tissue homogeniser to monolayers were fixed with paraformaldehyde (3·7% w/v) generate a single-cell suspension. Cells were filtered in DPB for 10 min at room temperature, followed imme- through 80 µm mesh, pelleted by centrifugation (100 g for diately by permeabilisation with Triton X-100 (Sigma; 5 min), washed twice with DPB, and resuspended with 0·1% w/v) in DPB for 5 min at room temperature. After 10 ml culture media (see below). fixation, Sertoli cell monolayers were washed twice with Isolated Sertoli cells were plated (1106 cells/ml per cold DPB (4 C) and stored in DPB at 4 C until well; 0·5106 cells per cm2) into 24-well culture plates processed for immunocytochemistry. All incubations for (Costar, Acton, MA, USA) containing glass cover-slips immunocytochemistry were performed in a humidified (13 mm diameter; Knittel Gläser, Braunschweig, Germany) chamber at room temperature, and all solutions, unless previously coated with Matrigel (1·97 mg/ml in DMEM; otherwise stated, were diluted in 0·01 M PBS, 0·154 M 20 µl/well, excess removed; BD Biosciences Bedford, NaCl and 0·1% (w/v) BSA (PBS/BSA). MA, USA). This plating concentration of Sertoli cells results in the seeding of Sertoli cells at a final density of 2 In vivo studies approximately 20 Sertoli cells/10 000 µm , as assessed by stereological counting of Sertoli cell nuclei after culture In vivo hormone treatments Testis tissue from rats (data not shown). (75–85 days old) undergoing various hormone treatments After plating, cells were incubated for 72 h at 37 Cin were used from a previous study (Meachem et al. 1998), in a humidified 5% CO2/95% air incubator, and then which 24 rats received 100 µl proprietary GnRH immu- hypotonically shocked with 10% culture media in water nogen preparation (BA-1666–4; Biotech, Sydney, for 45 s to remove contaminating germ cells from the Australia) every month for 3 months. This treatment Sertoli cell monolayer, and replaced in culture medium. results in the suppression of testicular testosterone levels to Sertoli cells were then allowed to recover for a period of 4% of control, serum FSH to close to the limits of  24hat37C, 5% CO2/95% air. Culture media were detection of the assay (1 ng/ml), and testis weights to 19% replaced, hormones added, and cells cultured for a of control (Meachem et al. 1998). Control animals  further 48 h at 37 C, 5% CO2/95% air, after which they received monthly injections of adjuvant alone. were fixed and processed for immunocytochemistry as After 3 months of GnRH immunisation, six animals described below. were killed immediately (GnRH-immunised group), and Culture media consisted of DMEM:F12 (mixed the remaining animals (six per group) received one of the 1:1 v/v; Trace Biosciences, Sydney, Australia) supple- following: mented with 1non-essential amino acids (Trace v daily injections of hrFSH (Gonal-F, Serono, Biosciences), -glutamine (1 mM, Trace Biosciences), Australia; 25 IU/kg per day s.c.) for 7 days (+FSH group) sodium bicarbonate (0·12% w/v, Trace Biosciences) and v administration of 6 cm silastic implants filled with penicillin/streptomycin/fungizone (200 U/ml, 200 µg/ testosterone powder plus daily injections of normal sheep ml, 500 ng/ml respectively; CSL Limited, Melbourne, serum (2 mg/kg per day s.c.) for 7 days (+T6 group) Australia) with final pH 7·2–7·4. Prior to use, BSA (1% v administration of 6 cm silastic testosterone implants w/v; Sigma), HEPES buffer (10 mM) and penicillin/ plus daily injections of polyclonal sheep antiserum against streptomycin/fungizone (as above) were added to the FSH (2 mg/kg per day s.c.) for 7 days (+T6+FSH culture media. antibody group). This FSH antibody has been shown to immunoneutral- In vitro hormone treatments Where indicated, human ise over 90% of circulating FSH in a normal adult rat and recombinant (hr)FSH (Puregon, N.V. Organon, Oss, The thus prevent the testosterone-dependent restoration of Netherlands) and testosterone (Sigma) were added to final circulating FSH levels in GnRH-immunised animals given concentrations of 2·35 IU/ml and 28 ng/ml (107 M) testosterone (Meachem et al. 1998). The testis weights, respectively. Both FSH and testosterone concentrations hormone levels and spermatogenic cell populations in www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 384 P SLUKA and others · Regulation of Sertoli cell junctions

these animals are described elsewhere (Meachem et al. cell junctions. Given the numerous cadherin family mem- 1998). Six additional animals were used for electron bers found in the testis (some 24 cadherin superfamily microscopy studies: two control animals, two ‘GnRH- members were identified in the testis in a recent study immunised group’ animals and two ‘+ FSH group’ (Johnson et al. 2000)), the polyclonal anticadherin anti- animals. body used was chosen in order to visualise all classic For vinculin immunohistochemical localisation, ad- cadherins at inter-Sertoli cell junctions. ditional archival testis tissue (Bouin’s fixed) was obtained Primary antibodies were detected with either Alexa from a variety of other treatment groups (n=4 per group), Fluor 488-goat antimouse IgG (green fluorescence; di- as described elsewhere (Meachem et al. 1998); these luted 1:100; Molecular Probes, Eugene, CA, USA) for included GnRH-immunised rats given higher doses of espin, -catenin, vinculin, and vimentin, or Alexa Fluor testosterone (24 cm) with or without the FSH antibody 568-goat antirabbit IgG (red fluorescence; diluted 1:100; for 7 days, control rats given the FSH antibody for 7 days Molecular Probes) for pan-cadherin. to acutely suppress FSH, and rats given testosterone (3 cm) Cover-slips containing immature Sertoli cell monolay- + 0·4 cm oestradiol (TE) for 9 weeks to suppress serum ers were processed for immunocytochemistry of the above LH, testicular testosterone (FSH levels are not suppressed) junctional molecules as follows. Cover-slips were cut into with or without FSH antibody for the last 7 days of eighths with a diamond-tipped pen, and washed twice for treatment. 5 min with PBS/BSA. Non-specific binding sites were blocked with CAS Block (Zymed, San Francisco, CA, Tissue fixation Testes for immunohistochemistry were USA) containing 10% normal sheep serum for 20 min, perfusion fixed in situ with Bouin’s fixative, as described after which primary antisera in PBS/BSA were added and previously (Meachem et al. 1997), and testes for electron incubated overnight. After two washes with PBS/BSA at microscopic studies were perfused in 5% glutaraldehyde in 5 min each, secondary antisera were applied and incubated 0·1 M cacodylate (O’Donnell et al. 2000). After perfusion for 1 h. Finally, cover-slips were washed twice for 5 min fixation, Bouin’s-fixed testes were removed, immersed in with PBS/BSA, counterstained simultaneously with DAPI Bouin’s fixative for approximately 5 h, and then kept in (4,6-diamidino-2-phenylindole dihydrochloride; 100 nM;  70% ethanol at 4 C until use. Testis wedges were Molecular Probes) and TO-PRO-3 (10 µM; Molecular embedded in low-melting-temperature ribboning polyes- Probes) for 5 min, rinsed three times with PBS/BSA, and ter wax (Oke & Suarez-Quian 1993), as described else- mounted with FluorSave Reagent (Calbiochem, San Di- where (O’Donnell et al. 2000). Sections (7 µm) were cut ego, CA, USA). For double-label experiments, primary  on a cryostat set at 0 C, floated onto a water bath set at and secondary antisera were incubated on cover-slips sim-  32 C, collected onto slides coated with 2% ultaneously. 3-aminopropyltriethoxy-silane (Sigma), and allowed to  dry for 48–72 h at 4 C. Glutaraldehyde-fixed tissues were Testis sections Antibodies for analysis of junctions in vivo prepared for electron microscopy, as described previously were as follows: vinculin (as above, 1:5000), espin (O’Donnell et al. 2000). (affinity-purified rabbit polyclonal antibody; 1:50 for im- munofluorescence; 1:1000 for conventional light immu- nocytochemistry (Bartles et al. 1996)), actin (mouse Immunocytochemistry monoclonal, 1:250 (clone C4; ICN Biomedicals, Aurora, Sertoli cell cultures Probes for analysis of junctions OH, USA), pan-cadherin (as above, 1:25 000) and vimen- in vitro were as follows: filamentous actin (phalloidin- tin (as above, 1:1000). TRITC; 10 µg/ml; Sigma), espin (mouse monoclonal Details of immunocytochemical procedures are de- IgG; 1:100; BD Transduction Laboratories, Franklin scribed elsewhere (O’Donnell et al. 2000). All incubations Lakes, NJ, USA), -catenin (mouse monoclonal IgG; and dilutions were in PBS (0·01 M PBS, 0·154 M NaCl 1:500; BD Transduction Laboratories), vinculin (mouse (pH 7·4) and no sodium azide). Briefly, sections were monoclonal IgG; 1:500; clone hVIN-1; Sigma), pan- dewaxed, subjected to antigen retrieval (either micro- cadherin (rabbit polyclonal antibody; 1:500; Sigma), and waved in 0·01 M citrate buffer (pH 6·0) for 10 min or vimentin (mouse monoclonal IgG; 1:100; clone V9; incubated in 0·02% trypsin in 0·1% CaCl2 for 40 min), Sigma). Negative controls consisted of PBS/BSA for incubated in 0·3% H2O2, blocked with 300 mM glycine phalloidin-TRITC, normal rabbit serum (Dako, Carpin- and then 0·1% Triton X-100, and incubated in CAS teria, CA, USA) for pan-cadherin, and normal mouse IgG Block (Zymed, San Francisco, CA, USA) containing 10% (Serotec, Oxford, UK) for espin, -catenin, vinculin and normal serum (sheep or rabbit) for 20 min. Sections were vimentin, diluted to concentrations equivalent to the incubated with primary antibodies either overnight or primary probe. for 2 h at room temperature. For conventional light The aim of the current study was to visualise cadherin- microscopy, sections were incubated for 1 h with an containing junctions as a whole rather than to identify appropriate biotinylated second antibody followed the involvement of individual cadherins in inter-Sertoli by streptavidin-horseradish peroxidase ABC Complex

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 385

(Vectastain Elite, Vector Laboratories, Burlingame, CA, and actin localisation in cultures treated with testosterone USA). The pink-coloured substrate used was Vector VIP alone (Fig. 1B and G, arrowheads) did not differ from (Vector Laboratories), and the sections were counter- cultures without hormones (Fig. 1A and F), indicating stained with Mayer’s haematoxylin (Sigma) prior to de- that this dose of testosterone does not modulate AJ hydration and mounting in DPX, as described previously function in immature Sertoli cells. In contrast, the nature (O’Donnell et al. 2000). For improved detection of and extent of cadherin and -catenin colocalisation, and pan-cadherin in the testis, sections were subjected to therefore AJ formation, was potentiated by FSH and FSH Tyramide Signal Amplification (TSA kit; NEN Life Sci- plus testosterone treatment, where AJs were observed as ence Products, Boston MA, USA) prior to the secondary extensive continuous belts around the periphery of these antibody steps, according to the manufacturer’s instruc- cells (Fig. 1C, D, H and I, arrows). tions. Details of the espin and actin dual-label immun- The colocalisation of cadherin and catenin at sites of ofluorescence procedures and confocal microscopy are inter-Sertoli cell contact could also represent inter- given elsewhere (O’Donnell et al. 2000). Sertoli cell desmosome-type (intermediate filament- Specificity of all primary antibodies was verified by associated) junctions, as there is evidence for substitution with an equivalent dilution of normal (Mulholland et al. 2001) and against (Lee et al. 2004) mouse serum (vinculin or actin), preimmune rabbit the association of cadherin with intermediate-filament immunoaffinity-purified IgG (espin), normal rabbit serum related junctions in the testis. Therefore, colocalisation (PanCad) or normal mouse IgG (vimentin). In all cases, no of vimentin (intermediate filaments) and cadherin was significant staining or fluorescence was observed for con- performed (Fig. 1K–O). In all cultures, staining with trol antisera in any seminiferous tubules. vimentin produced a pattern characteristic of intermedi- ate filaments, which were concentrated around the cell Confocal microscopy Confocal microscopy was per- nucleus and produced a diffuse network throughout the formed with a Nikon Diaphot 300 microscope equipped cell cytoplasm. Occasional intermediate filaments were with a krypton/argon laser linked to a BioRad MRC1000 seen to extend to contact intercellular boundaries confocal unit (BioRad) and a personal computer with perpendicularly; however, colocalisation with cadherin COMOS software (BioRad). Laser lines were at 488, 568, at the cell boundary was not apparent in any treatment and 594 nm for excitation of Alexa Fluor 488, Alexa Fluor (Fig. 1K–N). The intermediate filament staining pattern 568 and Phalloidin-TRITC, and TO-PRO-3 respect- did not change with FSH and/or testosterone treatment ively. A 60 oil immersion objective was used, and red, (Fig. 1K–N). The lack of vimentin and cadherin green and far-red channels were scanned sequentially. colocalisation indicates that the cadherin-associated Colour channels were merged with the Confocal Assistant adhesion junctions stimulated by FSH contain actin v 4·02 software package Todd Clark Brelje. (Fig. 1F–I), but not intermediate filaments. In the adult rat testis, extensive ES junctions are formed Electron microscopy Seminiferous tubules were exam- between Sertoli cells and can be assessed by the colocali- ined at 60 kV with a JEOL 1200EX transmission electron sation of actin and espin (O’Donnell et al. 2000). There- microscope. Areas of inter-Sertoli cell contact were ident- fore, the colocalisation of these two molecules was assessed ified under low power (3000–7000) and junctional in immature Sertoli cells in vitro (see Fig. 1P–T). Although structures were examined at higher power (15 000– Sertoli cells in vitro expressed actin and espin, these 20 000). did not colocalise at sites of inter-Sertoli cell contact in the absence of hormones (Fig. 1P, asterisks) or in cultures treated with testosterone alone (Fig. 1Q, Results asterisks). After treatment with FSH or FSH plus testos- terone, extensive colocalisation of actin and espin was Formation of ES and reorganisation of adherens junction belts observed around the periphery of many cells (Fig. 1R and is stimulated by FSH, but not testosterone, in immature S, arrows), suggesting ES junctions formed under these Sertoli cells in vitro conditions. However, not all actin was colocalised with Confocal analysis of cadherin and -catenin was used to espin, as confocal analysis revealed fibres of both actin and assess the formation of cadherin-based inter-Sertoli cell espin oriented parallel to the base of the cells (Fig. 1P–S). AJs in vitro (Fig. 1A–E). As shown in Fig. 1A (forked Both FSH and FSH plus testosterone treatment appeared arrowheads), cadherin and -catenin colocalised at regions to increase the level of espin staining within Sertoli cells of inter-Sertoli cell contact in the absence of hormones (compare Fig. 1P, Q, Z and Z1 with Fig. 1R, S, Z2 and in vitro, suggesting that AJs had formed. These AJs were Z3). The pattern of espin and actin staining in cultures observed as discrete focal adherens plaques oriented per- treated with testosterone alone did not differ from that in pendicularly to the cell boundary, and were observed to cultures without hormones (compare Fig. 1Q with Fig. 1P). occur at sites of cytoskeletal actin filament termination The colocalisation of espin and cadherin was also (Fig. 1F and G, closed arrowheads). Cadherin, -catenin investigated to explore the relationship between ES junctions www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 386 P SLUKA and others · Regulation of Sertoli cell junctions

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 387 and cadherins (Fig. 1Z–Z4). In the absence of hormones FSH stimulates the formation of basal ES junctions and the or with testosterone alone, cadherin did not colocalise orientation of inter-Sertoli cell junctions in the adult rat testis with espin (Fig. 1Z and Z1, arrowheads) suggesting that unstimulated Sertoli cells form focal AJ puncta, but not ES The colocalisation of espin and actin by dual-label junctions. In contrast, treatment with FSH or FSH plus immunofluorescence confocal microscopy was used to testosterone resulted in extensive espin and cadherin investigate the regulation of inter-Sertoli cell ES junctions colocalisation at the periphery of cells (Fig. 1Z2 and Z3, in adult testes in vivo. In control animals (where FSH, LH arrows), suggesting that AJ belts and extensive ESs formed and testosterone levels are normal), espin and actin co- at similar sites between Sertoli cells. localised in a scalloped pattern parallel to the basement Given that the arrangement of the cytoskeletal mol- membrane that is characteristic of inter-Sertoli cell ES ecules actin and vinculin has been shown to be regulated junctions (Fig. 2A, arrowheads). The colocalisation of by FSH (Cameron & Muffly 1991, Muffly et al. 1994), the espin and actin was observed in all stages of the seminif- colocalisation of these molecules was also performed (Fig. erous epithelium; however, the nature of this staining 1U–Y). In the absence of hormones or with testosterone differed between stages. During stages I–VIII, a scalloped alone, vinculin was concentrated at the tips of basal actin staining pattern was seen above spermatogonia and early filaments, suggesting localisation at cell–matrix junctions spermatocytes, and below the pachytene spermatocyte- (Fig. 1U and V, open arrowheads), as well as diffuse spermatid layers. During stages IX–X, staining was ob- localisation within the Sertoli cell cytoplasm (Fig. 1U and served both above and below leptotene spermatocytes, V). Upon FSH and FSH plus testosterone stimulation, while, during stages X–XIV, staining was observed below vinculin became colocalised with actin at the periphery of leptotene and zygotene spermatocytes. many cells (Fig. 1W and X, arrows), and the intensity of When FSH and LH levels, and thus testicular testoster- cytoplasmic vinculin staining appeared to increase (com- one levels, were suppressed by long-term immunisation of pare Fig. 1W and X with Fig. 1U and V). The localisation adult rats against GnRH, limited colocalisation of espin of vinculin at the tips of basal actin filaments remained and actin was seen despite the presence of both molecules unchanged (Fig. 1W, open arrowheads). within the Sertoli cell cytoplasm (Fig. 2B). Espin fibres An interesting observation was that regions of the not colocalised with actin were visible and were arranged Sertoli cell monolayer that contained a low local density of in a pattern that was perpendicular to the basement Sertoli cells (for example, near the edge of the cover-slip, membrane of the tubule (Fig. 2B, arrows). Administration where the Sertoli cell density was approximately 10 Sertoli of hrFSH for 7 days to GnRH-immunised animals cell nuclei per 10 000 µm2) formed AJ puncta; however, induced colocalisation of espin and actin into staining they did not display the formation of extensive AJ belts or patterns characteristic of ES junctions (Fig. 2C, arrow- ES junctions (as evidenced by immunostaining for heads). This restoration of ES junctions in Sertoli cells was cadherin/catenin and espin/actin), despite stimulation not apparent when testosterone alone (+T6 + FSH with FSH (data not shown). This suggests that the antibody group) was present for 7 days (Fig. 2D), FSH-stimulated formation of AJ belts and ES junctions is the epithelium having a similar appearance to that of density-dependent. GnRH-immunised animals (Fig. 2B). When both FSH

Figure 1 FSH promotes the formation of adherens junction belts and ES junctions, but not desmosome-like junctions, between Sertoli cells in vitro. Sertoli cells isolated from immature rats were cultured either in the absence of hormones (medium alone), testosterone (28 ng/ml), FSH (2·35 IU/ml) or both FSH and testosterone. Dual-label immunofluorescence was performed and colocalisation analysed by confocal microscopy. Immunostaining was observed as red or green fluorescence, with colocalisation of antigens indicated by yellow staining. Cadherin (red) and -catenin (green) localisation (A–E) and actin (red) and -catenin (green) localisation (F–J) were performed to assess the formation of adherens junctions (AJs). Forked arrowheads in (A–D) and (F–H), indicate discrete foci oriented perpendicular to the cell plasma membrane, indicating the formation of discrete AJ puncta; closed arrowheads in (F) and (G), indicate sites of cytoskeletal actin attachment to AJ puncta; arrows in (C) and (D), and (H) and (I), indicate colocalisation of cadherin–catenin and catenin–actin respectively, in extensive zones oriented parallel to the cell plasma membrane, indicating the formation of extensive AJ belts. Vimentin (green) and cadherin (red) localisation (K–O) was performed to assess formation of desmosome-like, or intermediate filament-associated, junctions. Espin (green) and actin (red) localisation (P–T) and vinculin (green) and actin (red) localisation (U–Y) were performed; colocalisation of these molecules (yellow) observed with confocal microscopy suggests the formation of ES junctions. Cultures were also immunostained for espin (green) and cadherin (red) (Z–Z4) to assist with visualisation of cell borders. Arrows in (R) and (S), and (W) and (X), indicate colocalisation of actin–espin and actin–vinculin, respectively, into extensive ES structures at the cell boundary. Asterisks in (P) and (Q), and (U) and (V), indicate absence of ES formation, as evidenced by the absence of espin, actin and vinculin staining at cell boundaries. Open arrowheads in (U–W) indicate colocalisation of vinculin and basal actin filaments associated with cell–matrix, not cell–cell, junctions. Forked arrowheads in (Z) and (Z1), indicate absence of espin and therefore ES junction at discrete cadherin-containing AJ puncta. Arrows in (Z2) and (Z3), indicate colocalisation of cadherin-espin in extensive structures at the cell boundary. Cell nuclei were counterstained with TO-PRO-3 iodide (blue). Negative controls for each immunostaining method are shown in E, J, O, T, Y and Z4. Red staining of cell nuclei is an artefact caused by non-specific binding of the secondary antibody to the cell nucleus. All micrographs were taken at the same magnification, bar=30 m. T, testosterone. www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 388 P SLUKA and others · Regulation of Sertoli cell junctions

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 389 and testosterone were restored in GnRH-immunised animals for 7 days (+T6 group), colocalised espin and actin were seen in a pattern identical to FSH-treated animals (data not shown). Electron microscopic analysis of ES junctions in these tissues confirmed the findings of the espin/actin colocali- sation. Extensive ES junctions were visible between Sertoli cells in normal adult rats (Fig. 3A). This junction is characterised by hexagonally packed actin bundles sand- wiched between the Sertoli cell plasma membrane and parallel cisternae of endoplasmic reticulum, which exhibit ribosomes on their membrane closest to the inside of the cell. Within the domain of the ES junction, focal oppos- ing subsurface densities associated with the plasma mem- brane of each cell were also visible (Fig. 3A, arrows). After Figure 3 Electron microscopic analysis of inter-Sertoli cell GnRH immunisation, extensive ES junctions were not junctions in vivo. (A) Inter-Sertoli cell junctions in normal adult rat seen between Sertoli cells (Fig. 3B), focal areas of ES testis. (B) Inter-Sertoli cell junctions, in the absence of ES junctions being rarely observed. Instead, focal junctions, junctions, in testes from GnRH-immunised rats. In (A) the which appeared morphologically similar to desmosome- arrowheads indicate two opposing Sertoli cell plasma membranes, and the arrows indicate two Sertoli cell plasma membranes that like junctions (Vogl et al. 1993), were seen between are closely associated; focal opposing densities on the plasma adjacent Sertoli cells in GnRH-immunised animals (Fig. membranes are visible. Extensive ES junctions are visible as 3B, arrows). When FSH was given to GnRH-immunised indicated by actin bundles (a) sandwiched between the Sertoli cell animals, extensive ES junctions, similar to those in Fig. plasma membrane and endoplasmic reticulum (er). Bar=200 nm. 3A, were again apparent within the epithelium (data not In (B) the arrowheads indicate two opposing Sertoli cell plasma membranes, and the arrows indicate an example of the focal shown). adhesion junctions that were seen. No associated ES junctions The immunolocalisation of vinculin, pan-cadherin and were visible. Bar=500 nm. vimentin was then examined to compare the localisation of ES, cadherin-based AJs and desmosome-like junctions, respectively, in control, GnRH-immunised and FSH- described for the localisation of actin and espin above, replaced rat testes. In control rats, vinculin (an ES junction although cadherin staining at inter-Sertoli cell junctions marker) was present in ES junctions between adjacent was much less extensive than staining of ES junction Sertoli cells (Fig. 2E, arrowheads), in ES junctions be- markers (compare Fig. 2E with 2H). Vimentin (a marker tween Sertoli cells and spermatids (Fig. 2E, ‘est’), and in of intermediate filament-related junctions or desmosome- the central Sertoli cell cytoplasm (Fig. 2E). Cadherin like junctions) was apparent in the perinuclear region and (PanCad – an AJ marker) was visible at focal locations central cytoplasm (Fig. 2K, arrow) of the Sertoli cell in a consistent with staining at inter-Sertoli cell junctions (Fig. stage-dependent manner, at sites of Sertoli cell contact 2H, arrowheads, shows an example of cadherin staining at with the basement membrane, and around some basally inter-Sertoli cell junctions in the characteristic scalloped located germ cells, as previously described (Vogl et al. pattern). The stage specificity of cadherin staining and its 1993, Zhu et al. 1997). Vimentin staining was also relationship to developing germ cells were identical to that apparent in a scalloped pattern (Fig. 2K, arrowhead)

Figure 2 FSH is involved in the restoration of ectoplasmic specialisation junctions and the organisation of inter-Sertoli cell junctions in vivo. (A, E, H and K) Normal adult rat testis; (B, F, I and L) testis from GnRH-immunised rat; (C, G, J and M) testis from GnRH-immunised rat that was given daily injections of hrFSH for 7 days to restore FSH levels; (D) testis from GnRH-immunised rat that was given 6 cm testosterone-filled implants together with daily injections of an FSH antibody; this treatment partially restores testicular testosterone levels but prevents a concomitant rise in FSH. (A–D) were stained with espin (red) and actin (green) by dual-label immunofluorescence and analysed by confocal microscopy; colocalisation is indicated by the yellow staining. The arrowheads in (A) and (C) show espin and actin colocalised in ES junction structures in a characteristic scalloped pattern. The arrows in (B) and (D) refer to fibres of espin staining that are not colocalised with actin, and that are oriented perpendicular to the basement membrane. (E–G) show vinculin immunostaining as indicated by pink staining. (H–J) show pink immunostaining for cadherin by a broad-specificity (pan-cadherin) rabbit polyclonal antibody that was raised against the cytoplasmic tail common to all classic cadherins. (K–M) show pink immunostaining for vimentin. An example of a negative control slide (in this case rabbit IgG substituted for pan-cadherin antibody) is shown in the inset in (H). (E–M) were counterstained with haematoxylin. Arrowheads in E, G, H, J, K and M indicate immunostaining in a scalloped pattern parallel to the basement membrane, consistent with inter-Sertoli cell junctions; arrows in (I) indicate cadherin staining in an abnormal orientation, being perpendicular to the basement membrane; arrows in (K–M) indicate vimentin staining in Sertoli cell cytoplasm. SC, Sertoli cell nuclei; est, vinculin in ES structures opposite elongated spermatids. The basement membrane of the seminiferous tubule is toward the bottom of each micrograph. Bars=10 m; (A–D) were taken at the same magnification, as were (E–M). www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 390 P SLUKA and others · Regulation of Sertoli cell junctions

Table 1 Presence (+) or absence (–) of vinculin localisation in ES structures between Sertoli cells in vivo as a function of treatments which result in the presence or absence of FSH (n=4 animals per treatment*)

Presence of FSH Absence of FSH

Treatment Vinculin localin Treatment Vinculin localin Control + Control+FSH Abb + TEb + TE+FSH Abc –/+f GnRH+hrFSH + GnRH – GnRH+T6+control Ab + GnRH+T6+FSH Ab – GnRH+T24+control Abe + GnRH+T24+FSH Abd –/+f

* Bouin’s-fixed, wax-embedded testis sections were obtained from a previous study (see Meachem et al. (1998) for details). a Control animals were administered an FSH antibody for 7 days to suppress acutely FSH in normal animals. b Animals given low-dose T (3 cm)+0·4 cm oestradiol (TE) for 9 weeks to suppress serum LH, testicular T; however, FSH levels were not significantly affected (Meachem et al. 1998). c Animals given TE for 9 weeks as above; however, FSH antibody was then administered for 1 week to suppress acutely FSH (Meachem et al. 1998). d Animals GnRH immunised as described in Materials and Methods, and then given higher dose T (T=24 cm) for the final 7 days treatment, along with FSH antibody to block the concomitant rise in FSH that occurs upon T treatment (Meachem et al. 1998). e GnRH+T24 as described above, but with control antibody administered; these animals thus have normal levels of FSH in the final 7 days of treatment (Meachem et al. 1998). f One out of four animals showed a vinculin pattern consistent with normal ES; the remaining animals showed disorganized vinculin staining as per GnRH-immunised animals. Ab, antibody; T, testosterone.

consistent with the known localisation of intermediate indicative of ES junctions (Fig. 2E and G), or disorganised filament-related junctions between Sertoli cells (Vogl et al. vinculin localisation (Fig. 2F) was scored in the various 1993, Zhu et al. 1997). in vivo treated groups. When FSH was present in rats, In GnRH-immunised animals that lacked FSH and LH, normal localisation of vinculin at inter-Sertoli cell ES the ordered arrangement of vinculin at ES junctions was junctions was observed (Table 1). When FSH was acutely lost (Fig. 2F). Cadherin immunostaining (Fig. 2I) was withdrawn from control rats for 7 days, normal vinculin apparent in a linear pattern perpendicular to the basement localisation was seen. However, when FSH was with- membrane. Heavy vimentin staining (Fig. 2L) was evident drawn from TE-treated animals, which have low LH and in Sertoli cell cytoplasm and the perinuclear region, near testosterone (but normal ES junctions (O’Donnell et al. the basement membrane of the tubule and around basally 2000)), most animals showed a lack of normal ES pattern. located germ cells. A higher dose of testosterone (24 cm implants) in the The organisation of vinculin in ES structures was absence of FSH could not restore vinculin localisation in restored by 7 days of FSH treatment (Fig. 2G, arrow- ES structures in three out of four rats, although one rat did heads), but not by testosterone (6 cm implants), in the show normal vinculin localisation. In general, in GnRH- absence of FSH (not shown). When FSH was adminis- immunised rats, FSH, but not testosterone, could restore tered to GnRH-immunised rats for 7 days, there was no normal vinculin localisation, and absence of FSH was visible increase in intensity of cadherin or vimentin associated with absence of normal vinculin localisation. immunostaining (Fig. 2J and M respectively). However, cadherin immunostaining was again visible in focal regions parallel to the basement membrane in a localisation Discussion consistent with inter-Sertoli cell junctions (Fig. 2J). Vimentin immunostaining suggested a more ordered This study investigated the hormonal regulation of inter- arrangement of intermediate filaments, with a scalloped Sertoli cell junction formation in vitro and in vivo. Immuno- pattern over basal germ cells more evident (Fig. 2M) staining for markers of ES junctions, cadherin, catenin and compared with GnRH-immunised testes (Fig. 2L). Ani- intermediate filaments allowed an appreciation of the mals administered 6 cm implants of testosterone plus an formation of several different junction types that are found FSH antibody (i.e. testosterone in the absence of FSH) at sites of inter-Sertoli cell contact, these being ES, AJ, and showed similar cadherin and vimentin immunostaining to desmosome-like junctions respectively. In the absence GnRH-immunised animals (data not shown). of FSH and testosterone in vitro, immature Sertoli cells The organisation of vinculin at inter-Sertoli cell ES were able to form focal AJs, seen as short discrete plaques junctions was then investigated in a large number of (AJ puncta), but not ES junctions or desmosome-like animals receiving various treatments that manipulate junctions. Administration of FSH alone potentiated the either FSH and/or testosterone (Table 1). The presence of formation of extensive junctional plaques containing vinculin in the normal scalloped pattern arrangement cadherin, catenin and the ES markers espin, actin

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 391 and vinculin. In contrast, testosterone alone did not promote extensive junction formation, suggesting that the formation of morphologically mature junctions between immature Sertoli cells in vitro is primarily dependent on FSH. These data suggest that FSH stimulates the forma- tion of ES junctions and extensive AJs between immature Sertoli cells in vitro. The ability of FSH to stimulate inter-Sertoli cell junction formation, particularly ES formation, was also confirmed in vivo by a GnRH immunisation model. After suppression of FSH, LH and testicular testosterone levels in adult rats for several months, basal ES junctions were disassembled, and ES was absent from the seminiferous epithelium, as assessed by electron microscopy and ab- sence of colocalised actin and espin, which has previously been shown to be an excellent marker of ES junctions in the testis (O’Donnell et al. 2000). Immunostaining for cadherin and vimentin suggested that AJs and intermediate filament-related junctions (desmosome-like junctions) may be present between Sertoli cells; however, their orientation appeared altered. Electron microscopy re- vealed the presence of desmosome-like junctions between Sertoli cells in GnRH-immunised rats, suggesting that some inter-Sertoli cell junctional types remain during hormone suppression in adult rats, but that ES junctions are completely absent. Restoration of FSH alone in vivo was able to re-establish ES junctions between Sertoli cells. In addition, immuno- staining of cadherin and vimentin suggested that FSH stimulated the correct orientation of AJs and desmosome- like junctions between Sertoli cells, such that the immu- nostaining pattern was similar to control animals. Taken together, these results suggest that FSH stimulates the formation of extensive inter-Sertoli cell adherens junc- tions, and that the ES junction in particular is stimulated by FSH. Figure 4 Model of inter-Sertoli cell junction formation. In the It is well known that immature Sertoli cells in vitro are absence of FSH, adjacent Sertoli cells are in contact at interdigitating finger-like projections. At these sites of contact, AJ highly responsive to FSH; however, it is generally con- puncta form, containing cadherin and -catenin, and are linked to sidered that FSH responsiveness declines as Sertoli cells the actin cytoskeleton. Both AJ puncta and actin filaments are mature (reviewed in Griswold (1993)). Nevertheless, an oriented perpendicularly to the cell’s plasma membrane. ES effect of FSH on inter-Sertoli cell junctions was observed junctions do not form in the absence of FSH; however, molecules both in immature Sertoli cells in vitro and in adult Sertoli that are located at ES junctions (espin and vinculin) are seen in the cytoplasm of Sertoli cells. Cell–matrix junctions cells in vivo. Previous studies in our laboratory using the (hemidesmosomes) containing vinculin linked to basal actin experimental in vivo model utilised in the current study filaments link the Sertoli cell to the basement membrane. After the identified the importance of FSH in the restoration addition of FSH, actin filament dynamics at Sertoli cell finger-like of spermatogenesis after gonadotrophin suppression projections push the plasma membranes of the Sertoli cells together to form extensive sites of inter-Sertoli cell contact, where (Meachem et al. 1998). Stereological analysis of germ cell a junctional complex is formed containing AJ belts, ES junctions numbers showed that long-term gonadotrophin suppres- and tight junctions. AJ belts form from fused AJ puncta that were sion caused arrest of spermatogenesis primarily at the level once located at adjacent finger-like projections. Actin filaments of spermatogonial and spermatocyte development, and that are now arranged parallel to the Sertoli cell plasma membrane, the restoration of these cell numbers, and subsequently and are packaged into bundles by espin and vinculin, which move from the cytoplasmic pool into inter-Sertoli cell ES junctions. more mature cell types, was dependent on FSH rather than Linking elements connect adjacent Sertoli cell plasma membranes testosterone (Meachem et al. 1998). The current study at sites of ES junction formation, and also connect AJ belts and thus suggests that the re-establishment of mature inter- tight junctions to the actin cytoskeleton. If FSH is removed, ES Sertoli cell junctions by FSH coincides with the restora- junctions are disassembled, and the Sertoli cell junctional complex tion of germ cell development (Meachem et al. 1998). is pushed toward that observed without FSH. www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 392 P SLUKA and others · Regulation of Sertoli cell junctions

Consistent with this, many others have suggested that the junctional components. It was noted that higher doses of formation of mature inter-Sertoli cell junctions is required testosterone could potentially mimic the effects of FSH on for spermatogenesis to proceed beyond meiosis, since all ES junctions, since one out of four GnRH-immunised post-meiotic cells reside above the blood–testis barrier rats showed normal vinculin orientation upon treatment (Byers et al. 1993). with higher-dose testosterone (24 cm implants) plus the Although FSH is not required for fertility or the FSH antibody. A previous study has also shown that restoration of qualitatively normal spermatogenesis, it is higher doses of testosterone can overcome a lack of FSH now becoming clear that FSH has an important role in in the restoration of germ cell populations (Meachem et al. establishing the Sertoli cell population and supporting its 1998). Such synergy is commonly noted in the testis with function (reviewed in Allan & Handelsman (2005)). regard to the hormonal regulation of spermatogenesis Despite the fact that various roles for FSH in the seminif- (reviewed in McLachlan et al. (2002), Meachem et al. erous epithelium are emerging (see Allan & Handelsman (1998), O’Donnell et al. (2005) and Saito et al. (2000)). (2005) and O’Donnell et al. (2005)), reports on inter- The in vitro model used in this study revealed interesting Sertoli cell junctions in various transgenic models of FSH, information on the series of events leading to the LH and androgen action deficiency are limited. However, hormone-dependent formation of AJ and ES junctions ultrastructural studies in FSH receptor-deficient between Sertoli cells. AJs are ubiquitous cell–cell adhesion (FORKO) mice show various defects in Sertoli cell junctions that are characterised by transmembrane cad- ultrastructure, particularly fluid-filled cytoplasm indicative herin molecules that link to the filamentous actin cyto- of disturbed fluid dynamics (Grover et al. 2004). Although skeleton via the catenin family of linker proteins (for ES structures were observed in FORKO mice, some reviews, see Lui et al. (2003), Nagafuchi (2001) and Yap abnormalities were noted, including dilated endoplasmic et al. (1997)). In epithelial cells, AJs are initially formed as reticulum associated with ES (Grover et al. 2004). These AJ puncta at sites of cell–cell contact, found at the tips of studies in transgenic mice support a role for FSH in interdigitating filopodial extensions. The polymerisation promoting normal inter-Sertoli cell junctions in vivo. of actin at AJ puncta serves to push adjacent puncta to It was of interest that testosterone was not able to closer proximity such that they can merge to form restore ES junctions in the absence of FSH, suggesting that extensive AJ belts (Vasioukhin & Fuchs 2001, Vasioukhin inter-Sertoli cell junctions are particularly FSH sensitive et al. 2000). Sertoli cell AJs may also connect to the in vivo, as well as in vitro. The observation that AJ and ES recently discovered nectin/afadin adhesion complex via junction formation was unaffected by testosterone in both related links of these two junction complexes to -catenin our in vitro and in vivo models was somewhat surprising. (Takai & Nakanishi 2003). We identified the formation of The dose of testosterone used in vitro approximates the AJs in Sertoli cells in vitro by the colocalisation of cadherin normal testicular testosterone concentrations in the 20- and -catenin. In the absence of FSH, AJ puncta were day-old testis (Killian et al. 2003), and the dose of observed, whereas the addition of FSH stimulated the testosterone used in vivo has been shown to maintain reorganisation of AJ puncta into extensive AJ belts. The qualitatively normal spermatogenesis (O’Donnell et al. formation of both AJ puncta and AJ belts in vitro have 1999). Although the current and previous (O’Donnell previously been demonstrated in the mouse TM4 Sertoli et al. 2000) studies indicate that testosterone is not primarily cell line (Sandig et al. 1997) and in mouse keratinocytes involved in the regulation of AJs or ES junctions, further (Vasioukhin et al. 2000); however, the progression of AJs work in our (T J Kaitu’u, P Sluka, C F Foo & P G in these cells from puncta to belts did not require FSH, Stanton, personal communication) and other laboratories probably because these cell types are functionally different (Janecki et al. 1991, Florin et al. 2005, Meng et al. 2005) from primary Sertoli cell cultures. That the junctions suggests that testosterone regulates the formation of revealed by immunostaining for cadherin and catenin are another blood–testis barrier component, inter-Sertoli cell actin-related AJs, not intermediate filament-related tight junctions, although FSH may also have a role in tight desmosome junctions, is confirmed by the colocalisation junction permeability, as seen in the hamster (Bergmann of cadherin and actin and the absence of colocalisation of 1987, Tarulli et al. 2006) and more recently in the rat cadherin and vimentin (a component of intermediate (T J Kaitu’u, P Sluka, C F Foo&PGStanton, personal filaments) at junction sites. communication; Janecki et al. (1991)). In addition, it is The specific members of the cadherin superfamily that possible that the normally high levels of testicular testos- are found at inter-Sertoli cell AJs remains unknown. The terone in untreated rats may support the maintenance of pan-cadherin antibody used in this study is known to AJ and ES junctions to some degree; however, our detect all classical cadherin forms (E-, N- and P-cadherin) findings suggest that FSH will be important in the (Geiger et al. 1990), and a similar antibody was used to restoration of these junctions after gonadotrophin suppres- detect cadherin at inter-Sertoli cell junctions by immuno- sion. We propose that FSH and testosterone cooperate histochemistry, and to immunoprecipitate -catenin from and may even synergise in the initiation and maintenance seminiferous tubules (Mulholland et al. 2001). More of inter-Sertoli cell junctions by regulating the various specifically, N-cadherin has been localised to sites of

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 393 inter-Sertoli cell contact in the adult rat testis (Wine & (Yonemura et al. 1995, Krendel & Bonder 1999). Chapin 1999, Johnson & Boekelheide 2002, Lee et al. Although AJs, as defined by cadherin immunostaining, 2003, 2004). Although mRNA for other classic cadherins were present between Sertoli cells in adult GnRH immu- have been detected in whole adult testis extracts by nised rats in vivo, the normal orientation of these junctions RT–PCR (Johnson et al. 2000, Lee et al. 2003, Munro & appeared dependent on FSH. Thus, it is possible that the Blaschuk 1996), they have been localised to sites other formation and orientation of mature inter-Sertoli cell than inter-Sertoli cell junctions (Byers et al. 1994, Johnson junctions is affected by FSH, possibly via effects on actin & Boekelheide 2002, Lee et al. 2003, Wine & Chapin reorganisation. Previous studies in adult hypogondo- 1999). In addition to classical cadherins, a large number of trophic rats support the proposition that FSH is important protocadherins and cadherin-related proteins have also for the correct organisation of actin in Sertoli cells (Muffly been found in testis; however, their localisation has not yet et al. 1993, 1994). been determined (Johnson et al. 2000). Thus, it is likely By combining the observations made in this study with that N-cadherin is present at inter-Sertoli cell AJs, but the data from the literature, we propose a model of FSH presence of other cadherin-related proteins cannot be stimulation of inter-Sertoli cell junctions in vitro (Fig. 4). conclusively excluded. In this model, the presence of FSH facilitates the reor- In this study, the ES has been defined as hexagonally ganisation of inter-Sertoli cell AJs from puncta to exten- packed bundles of actin filaments sandwiched between the sive junctions, along with reorientation of the associated Sertoli cell plasma membrane and cisternae of the endo- actin cytoskeleton and cell membrane. Furthermore, FSH plasmic reticulum. A number of molecular components of stimulates the recruitment of espin and vinculin, which ES junctions have been identified, including espin (Bartles serve to cross-link actin filaments into the characteristic et al. 1996) and vinculin (Pfeiffer & Vogl 1991), both of hexagonal arrangement found in ES junctions. Based on which are codistributed with actin bundles in ES junc- this model, results from the in vivo studies presented here tions. Espin links actin filaments into hexagonal bundles suggest that gonadotrophin suppression causes disassembly (Chen et al. 1999, Bartles 2000), while vinculin binds of ES junction components, breakdown of AJ belts into filamentous actin, along with other cell–cell junction AJ puncta, and rearrangement of actin filaments from components (Menkel et al. 1994, Huttelmaier et al. 1997), being arranged parallel to the plasma membrane to termi- serving to link the actin cytoskeleton to the functional nating perpendicularly at AJ puncta. adhesion molecules. Our observations on FSH stimulation In conclusion, we have shown that FSH stimulates the of actin and vinculin organisation is consistent with data formation of mature inter-Sertoli cell junctions, particu- from Muffly and colleagues, who investigated the pattern larly via effects on the formation of ES junctions and of actin and vinculin localisation in seminiferous epithelial extensive AJs. The effects in immature Sertoli cells in vitro sheets to suggest that ES junction formation opposite were confirmed in adult testes in vivo, and the restoration spermiogenic cells is dependent on FSH in vivo (Cameron of these junctions correlated with the restoration of et al. 1998, Muffly et al. 1993, 1994) and in vitro (Cameron spermatogenesis. The in vitro results indicate that focal AJs &Muffly 1991). (or AJ puncta), but not ES junctions, form in the absence The fact that ES junctions in vitro were observed only at of hormones, and that FSH promotes the formation of ES sites of AJ belt formation, and never at sites of AJ puncta, junctions and extensive AJ belts. We believe that forma- suggests that the parallel arrangement of actin filaments tion of these junctions is important for mature Sertoli cell found at AJ belts (Krendel & Bonder 1999) is important function and facilitates the establishment of a functional for ES junction formation. It is possible that bundling of epithelium capable of supporting spermatogenesis. The actin filaments by espin into the hexagonal arrangement in vitro model used in this study provides an interesting characteristic of ES junctions may occur only after actin model to study the hormone-induced molecular events filaments are arranged parallel to the plasma membrane. leading to inter-Sertoli cell junction formation. This model of sequential junction formation has been proposed to exist between AJs and tight junctions by Takai and Nakanishi (2003). Once ES junctions are formed, an Acknowledgements interdependency of junctions may exist, where the for- mation of ES junctions may serve to stabilise, and indeed We thank Dr Sarah Meachem for providing in vivo testis facilitate the organisation of associated cytoskeletal and tissue, and Dr David M Robertson for helpful discussions architectural proteins, thereby strengthening the junc- during the course of this work. tional complex, which includes AJs, and other junctions such as desmosome-like junctions and tight junctions. It was noted that during remodeling of AJ puncta into AJ belts, Funding actin filaments were reorganised to become integral com- ponents of AJ belts oriented parallel to the plasma mem- This study was supported by Program Grant no. 983212 brane, consistent with current models of actin dynamics and 241000 from the National Health and Medical www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access 394 P SLUKA and others · Regulation of Sertoli cell junctions

Research Council of Australia (NH&MRC). The authors Johnson KJ, Patel SR & Boekelheide K 2000 Multiple cadherin declare that there is no conflict of interest that would superfamily members with unique expression profiles are produced in rat testis. Endocrinology 141 675–683. prejudice the impartiality of this scientific work. Killian J, Pratis K, Clifton RJ, Stanton PG, Robertson DM & O’Donnell L 2003 5 alpha-reductase isoenzymes 1 and 2 in the References rat testis during postnatal development. Biology of Reproduction 68 1711–1718. Allan CM & Handelsman DJ 2005 In vivo FSH actions. In Sertoli Cell Krendel MF & Bonder EM 1999 Analysis of actin filament bundle Biology, pp 171–197. Eds MK Skinner & MD Griswold. San dynamics during contact formation in live epithelial cells. Cell Diego, CA, USA: Elsevier Academic. Motility and the Cytoskeleton 43 296–309. Bartles JR 2000 Parallel actin bundles and their multiple actin- Krishnamurthy H, Danilovich N, Morales CR & Sairam MR 2000 bundling proteins. Current Opinion in Cell Biology 12 72–78. Qualitative and quantitative decline in spermatogenesis of the Bartles JR, Wierda A & Zheng L 1996 Identification and follicle-stimulating hormone receptor knockout (FORKO) mouse. characterization of espin, an actin-binding protein localized to the Biology of Reproduction 62 1146–1159. F-actin-rich junctional plaques of Sertoli cell ectoplasmic KumarTR,WangY,LuN&Matzuk MM 1997 Follicle stimulating specializations. Journal of Cell Science 109 (Pt 6) 1229–1239. hormone is required for ovarian follicle maturation but not male Bergmann M 1987 Photoperiod and testicular function in Phodopus fertility. Nature Genetics 15 201–204. sungorus. Advances in Anatomy, Embryology and Cell Biology Lampa J, Hoogerbrugge JW, Baarends WM, Stanton PG, Perryman KJ, 105 1–76. Grootegoed JA & Robertson DM 1999 Follicle-stimulating Byers SW, Pelletier R-M & Suarez-Quian CA 1993 Sertoli cell hormone and testosterone stimulation of immature and mature junctions and the seminiferous epithelium barrier. In The Sertoli Sertoli cells in vitro: inhibin and N-cadherin levels and round Cell, pp 431–446. Eds LD Russell & MD Griswold. Clearwater, spermatid binding. Journal of Andrology 20 399–406. FL, USA: Cache River Press. Lee NP, Mruk DD, Conway AM & Cheng CY 2004 Zyxin, axin, Byers SW, Sujarit S, Jegou B, Butz S, Hoschutzky H, Herrenknecht K, and Wiskott-Aldrich syndrome protein are adaptors that link the MacCalmanC&Blaschuk OW 1994 Cadherins and cadherin- cadherin/catenin protein complex to the cytoskeleton at adherens associated molecules in the developing and maturing rat testis. junctions in the seminiferous epithelium of the rat testis. Journal of Endocrinology 134 630–639. Andrology 25 200–215. CameronDF&Muffly KE 1991 Hormonal regulation of spermatid Lee NPY, Mruk DD, Lee WM & Cheng CY 2003 Is the binding. Journal of Cell Science 100 (Pt 3) 623–633. cadherin/catenin complex a functional unit of cell–cell actin-based Cameron DF, Muffly KE & Nazian SJ 1998 Development of Sertoli adherens junctions in the rat testis? Biology of Reproduction cell binding competency in the peripubertal rat. Journal of Andrology 68 489–508. 19 573–579. Lui WY, Mruk DD, Lee WM & Cheng CY 2003 Adherens junction ChenB,LiA,WangD,WangM,ZhengL&Bartles JR 1999 Espin dynamics in the testis and spermatogenesis. Journal of Andrology contains an additional actin-binding site in its N terminus and is a 24 1–14. major actin-bundling protein of the Sertoli cell-spermatid McLachlan RI, O’Donnell L, Meachem SJ, Stanton PG, de Kretser DM, ectoplasmic specialization junctional plaque. Molecular Biology of the PratisK&Robertson DM 2002 Identification of specific sites of Cell 10 4327–4339. hormonal regulation in spermatogenesis in rats, monkeys, and man. Dierich A, Sairam MR, Monaco L, Fimia GM, Gansmuller A, Recent Progress in Hormone Research 57 149–179. LeMeur M & Sassone-Corsi P 1998 Impairing follicle-stimulating Meachem SJ, Wreford NG, Robertson DM & McLachlan RI 1997 hormone (FSH) signaling in vivo: targeted disruption of the FSH Androgen action on the restoration of spermatogenesis in adult rats: receptor leads to aberrant gametogenesis and hormonal imbalance. effects of human chorionic gonadotrophin, testosterone and PNAS 95 13612–13617. flutamide administration on germ cell number. International Journal Florin A, Maire M, Bozec A, Hellani A, Chater S, Bars R, Chuzel F of Andrology 20 70–79. & Benahmed M 2005 Androgens and postmeiotic germ cells Meachem SJ, Wreford NG, Stanton PG, Robertson DM & regulate claudin-11 expression in rat Sertoli cells. Endocrinology McLachlan RI 1998 Follicle-stimulating hormone is required for 146 1532–1540. the initial phase of spermatogenic restoration in adult rats following Geiger B, Volberg T, Ginsberg D, Bitzur S, SabanayI&HynesRO gonadotropin suppression. Journal of Andrology 19 725–735. 1990 Broad spectrum pan-cadherin antibodies, reactive with the Meachem SJ, McLachlan RI, Stanton PG, Robertson DM & C-terminal 24 amino acid residues of N-cadherin. Journal of Cell Wreford NG 1999 FSH immunoneutralization acutely impairs Science 97 (Pt 4) 607–614. spermatogonial development in normal adult rats. Journal of Griswold MD 1993 Actions of FSH on mammalian Sertoli cells. In Andrology 20 756–762; discussion 755. The Sertoli Cell, pp 493–508. Eds LD Russell & MD Griswold. Meng J, Holdcraft RW, Shima JE, Griswold MD & Braun RE 2005 Clearwater, FL, USA: Cache River Press. Androgens regulate the permeability of the blood–testis barrier. Grover A, Sairam MR, Smith CE & Hermo L 2004 Structural and PNAS 102 16696–16700. functional modifications of Sertoli cells in the testis of adult follicle-stimulating hormone receptor knockout mice. Biology of Menkel AR, Kroemker M, Bubeck P, Ronsiek M, Nikolai G & Reproduction 71 117–129. Jockusch BM 1994 Characterization of an F-actin-binding domain Huttelmaier S, Bubeck P, Rudiger M & Jockusch BM 1997 in the cytoskeletal protein vinculin. Journal of Cell Biology Characterization of two F-actin-binding and oligomerization sites 126 1231–1240. in the cell-contact protein vinculin. European Journal of Biochemistry Mruk DD & Cheng CY 2004 Sertoli–Sertoli and Sertoli–germ cell 247 1136–1142. interactions and their significance in germ cell movement in the Janecki A, Jakubowiak A & Steinberger A 1991 Regulation of seminiferous epithelium during spermatogenesis. Endocrine Reviews transepithelial electrical resistance in two-compartment Sertoli cell 25 747–806. cultures: in vitro model of the blood–testis barrier. Endocrinology Muffly KE, Nazian SJ & Cameron DF 1993 Junction-related Sertoli 129 1489–1496. cell cytoskeleton in testosterone-treated hypophysectomized rats. Johnson KJ & Boekelheide K 2002 Dynamic testicular adhesion Biology of Reproduction 49 1122–1132. junctions are immunologically unique. II. Localization of classic Muffly KE, Nazian SJ & Cameron DF 1994 Effects of follicle- cadherins in rat testis. Biology of Reproduction 66 992–1000. stimulating hormone on the junction-related Sertoli cell

Journal of Endocrinology (2006) 189, 381–395 www.endocrinology-journals.org

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access Regulation of Sertoli cell junctions · P SLUKA and others 395

cytoskeleton and daily sperm production in testosterone-treated immature mouse Sertoli cells. Cell Motility and the Cytoskeleton hypophysectomized rats. Biology of Reproduction 51 158–166. 36 149–163. Mulholland DJ, DedharS&VoglAW2001 Rat seminiferous Steinberger E & Steinberger A 1975 Spermatogenic function of the epithelium contains a unique junction (ectoplasmic specialization) testis. In Handbook of Physiology, pp 325–345. Eds RO Greep & with signaling properties both of cell/cell and cell/matrix junctions. EB Astwood. Baltimore, MD, USA: Waverley Press. Biology of Reproduction 64 396–407. TakaiY&Nakanishi H 2003 Nectin and afadin: novel organizers of Munro SB & Blaschuk OW 1996 A comprehensive survey of the intercellular junctions. Journal of Cell Science 116 17–27. cadherins expressed in the testes of fetal, immature, and adult mice Tarulli GA, Stanton PG, LerchlA&Meachem SJ 2006 Adult Sertoli utilizing the polymerase chain reaction. Biology of Reproduction cells are not terminally differentiated in the Djungarian hamster: 55 822–827. effect of FSH on proliferation and junction protein organization. Nagafuchi A 2001 Molecular architecture of adherens junctions. Biology of Reproduction (In press). Current Opinion in Cell Biology 13 600–603. Toyama Y, Maekawa M & Yuasa S 2003 Ectoplasmic specializations O’Donnell L, Pratis K, Stanton PG, Robertson DM & McLachlan RI in the Sertoli cell: new vistas based on genetic defects and testicular 1999 Testosterone-dependent restoration of spermatogenesis in toxicology. Anatomical Science International 78 1–16. adult rats is impaired by a 5 alpha-reductase inhibitor. Journal of Vasioukhin V & Fuchs E 2001 Actin dynamics and cell–cell adhesion Andrology 20 109–117. in epithelia. Current Opinion in Cell Biology 13 76–84. O’Donnell L, Stanton PG, Bartles JR & Robertson DM 2000 Sertoli Vasioukhin V, Bauer C, Yin M & Fuchs E 2000 Directed actin cell ectoplasmic specializations in the seminiferous epithelium of polymerization is the driving force for epithelial cell–cell adhesion. the testosterone-suppressed adult rat. Biology of Reproduction 63 Cell 100 209–219. 99–108. Vogl AW, Pfeiffer DC & Redenbach DM 1991 Ectoplasmic O’Donnell L, Meachem SJ, Stanton PG & McLachlan RI 2005 The (‘junctional’) specializations in mammalian Sertoli cells: influence endocrine regulation of spermatogenesis. In The Physiology of on spermatogenic cells. Annals of the New York Academy of Sciences Reproduction, pp 1017–1070. Eds E Knobil & JD Kneill. San Diego, 637 175–202. CA, USA: Elsevier. Vogl AW, Pfeiffer DC, Redenbach DM & Grove BD 1993 Sertoli Oke BO & Suarez-Quian CA 1993 Localization of secretory, cell cytoskeleton. In The Sertoli Cell, pp 39–86. Eds LD Russell & membrane-associated and cytoskeletal proteins in rat testis using an MD Griswold. Clearwater, FL, USA: Cache River Press. improved immunocytochemical protocol that employs polyester Vogl AW, Pfeiffer DC, Mulholland D, KimelG&GuttmanJ2000 wax. Biology of Reproduction 48 621–631. Unique and multifunctional adhesion junctions in the testis: Ozaki-Kuroda K, Nakanishi H, Ohta H, Tanaka H, Kurihara H, ectoplasmic specializations. Archives of Histology and Cytology Mueller S, Irie K, Ikeda W, Sakai T, Wimmer E et al. 2002 Nectin 63 1–15. ff couples cell–cell adhesion and the actin sca old at heterotypic Weber JE, Turner TT, Tung KS & Russell LD 1988 Effects of testicular junctions. Current Biology 12 1145–1150. cytochalasin D on the integrity of the Sertoli cell (blood–testis) Perryman KJ, Stanton PG, Loveland KL, McLachlan RI & barrier. American Journal of Anatomy 182 130–147. Robertson DM 1996 Hormonal dependency of neural cadherin Wine RN & Chapin RE 1999 Adhesion and signaling proteins in the binding of round spermatids to Sertoli cells in vitro. spatiotemporally associated with spermiation in the rat. Journal of Endocrinology 137 3877–3883. Andrology 20 198–213. ff Pfei er DC & Vogl AW 1991 Evidence that vinculin is co-distributed Wreford NG, Rajendra Kumar T, Matzuk MM & de Kretser DM with actin bundles in ectoplasmic (‘junctional’) specializations of 2001 Analysis of the testicular phenotype of the follicle-stimulating mammalian Sertoli cells. Anatomical Record 231 89–100. hormone beta-subunit knockout and the activin type II receptor PokuttaS&WeisWI2002Thecytoplasmic face of cell contact sites. knockout mice by stereological analysis. Endocrinology Current Opinion in Structural Biology 12 255–262. 142 2916–2920. Russell L 1977 Observations on rat Sertoli ectoplasmic (‘junctional’) Yap AS, Brieher WM & Gumbiner BM 1997 Molecular and specializations in their association with germ cells of the rat testis. functional analysis of cadherin-based adherens junctions. Annual Tissue and Cell 9 475–498. Reviews in Cell Development Biology 13 119–146. Russell LD & Peterson RN 1985 Sertoli cell junctions: morphological Yonemura S, Itoh M, NagafuchiA&Tsukita S 1995 Cell-to-cell and functional correlates. International Review of Cytology adherens junction formation and actin filament organization: 94 177–211. similarities and differences between non-polarized fibroblasts and Saito K, O’Donnell L, McLachlan RI & Robertson DM 2000 polarized epithelial cells. Journal of Cell Science 108 (Pt 1) 127–142. Spermiation failure is a major contributor to early spermatogenic Zhu LJ, Zong SD, Phillips DM, Moo-Young AJ & Bardin CW 1997 suppression caused by hormone withdrawal in adult rats. Changes in the distribution of intermediate filaments in rat Sertoli Endocrinology 141 2779–2785. cells during the seminiferous epithelium cycle and postnatal Salanova M, Ricci G, Boitani C, Stefanini M, De Grossi S & development. Anatomical Record 248 391–405. Palombi F 1998 Junctional contacts between Sertoli cells in normal and aspermatogenic rat seminiferous epithelium contain Received in final form 13 January 2006 alpha6 beta1 integrins, and their formation is controlled by follicle-stimulating hormone. Biology of Reproduction 58 371–378. Accepted 24 January 2006 Sandig M, Kalnins VI & Siu CH 1997 Role of NCAM, cadherins, Made available online as an Accepted Preprint and microfilaments in cell–cell contact formation in TM4 30 January 2006

www.endocrinology-journals.org Journal of Endocrinology (2006) 189, 381–395

Downloaded from Bioscientifica.com at 09/26/2021 10:24:32AM via free access