Male sex determination in the majority of mammals commences after the induction of Sry expression in the XY genital ridge (Koopman et al.,2001). This determination leads to major morphological changes within the embryonic male gonad in a short period of time, during which there are few obvious changes occurring in the female gonad. While some progress has been made in studying the signalling and physical interactions that occur between the different cell lineages present in the male gonad, and how they influence each other's differentiation and/or proliferation (reviewed by Ross and Capel,2005), much remains to be learned regarding the histogenesis of the testes.
The main obvious event marking the development of the male gonad is the formation of the testis cords. The various cell lineages required for cord formation differentiate from cells that have migrated into the XY gonad from both the neighboring mesonephros and the coelomic epithelium (Karl and Capel,1998). Enclosed within the cords are the germ cells and their supporting cell lineage, the Sertoli cells, which envelop and adhere to the germ cells. A basement membrane of extracellular matrix (ECM) protein surrounds the testis cords, whose constituents are secreted by both the Sertoli cells and the peritubular myoid cells (PMC), which form a single cell layer on the outer edge of the cords. These ECM proteins include the structural proteins laminin, collagens, and fibronectin (Richardson et al.,1995). ECM proteins have long been known to play an important role in the induction and maintenance of the various cell lineages that differentiate in the male gonad. In vitro experiments have shown that Sertoli, Leydig, and germ cells require the presence of ECM to maintain their in vivo morphology and phenotype (Hadley et al.,1985; Vernon et al.,1991; Diaz et al.,2002).
In addition to the importance of ECM proteins that contribute to structures such as basement membrane, it is likely that a separate class of matrix proteins, the matricellular proteins, also have a role to play in gonadal morphogenesis. Matricellular proteins are known to act as regulators of cell function and cell adhesion/de-adhesion and modulators of growth factor and cytokine activity (reviewed by Bornstein and Sage,2002). Among the better-characterized matricellular proteins are osteopontin (OPN), thrombospondin-1 (TSP1), and osteonectin.
OPN is a secreted cell adhesion glycoprotein that can bind to the cell surface by means of multiple receptors, including members of the integrin receptor family. In addition to its role as a secreted protein, OPN has also been shown to be present intracellularly. Intracellular OPN in osteoclasts is associated with CD44 receptor and is proposed to play a role in cell migration and in the formation of cellular processes (Suzuki et al.,2002).
OPN has been implicated in multiple processes, including embryonic development, wound healing, immune response, tumor formation, and bone resorption and calcification (Denhardt and Noda,1998; Giachelli and Steitz,2000; Higuchi et al.,2004). It has been proposed as a prognostic marker for cancers as it is required for progression of many cancers (Furger et al.,2001; Rittling and Chambers,2004). It is unknown how OPN acts to enhance malignancy, and it may be by inducing several signalling pathways to increase cell proliferation. OPN has been found also to promote cell migration and adhesion (Xuan et al.,1995; Das et al.,2003; Zhu et al.,2004). It is abundant in the bone matrix where it is thought to mediate adhesion of osteoclasts to resorbing bone (Reinholt et al.,1990).
Despite the many proposed functions for osteopontin, mice that lack this protein develop normally, although there is some evidence of altered vascular function, wound healing, and bone resorption (Liaw et al.,1998; Yoshitake et al.,1999; Myers et al.,2003). The lack of dramatic phenotype of OPN knockout mice may be due to redundancy with related proteins.
OPN has been classed as a member of the SIBLING (Small Integrin Binding LIgand N-linked Glycoprotein) family. All SIBLING genes are clustered on human chromosome 4 and mouse chromosome 5, respectively. In addition to OPN, members of this family include bone sialoprotein (BSP), dentin matrix protein (DMP), dentin sialophosphoprotein (DSPP), and matrix extracellular phosphoglycoprotein (MEPE; Fisher et al.,2001). Despite differences in their primary sequence, these genes share a similar exon–intron structure and encoded protein domain structure. Three SIBLING proteins, BSP, OPN and DMP, have been shown to bind to and activate matrix metalloproteases (MMPs; Fedarko et al.,2004). These genes also have been found to have overlapping expression patterns in bone and teeth, suggesting that they may share some biological functions.
Previous studies have detected OPN expression in the adult rat testis using Northern blotting hybridization. It was concluded that OPN could be a Sertoli cell product as it was not detected in RNA from interstitial cells or isolated germ cell preparations (Siiteri et al.,1995). However, in a second study using immunohistochemistry, OPN protein was detected in association with early spermatogonia but was not detected in Sertoli cells in postnatal and adult rat testis (Luedtke et al.,2002). Neither study investigated OPN expression in the developing embryonic testis.
In the present study, we investigated the expression of OPN in the developing embryonic testis at both the mRNA and protein level. We found that OPN mRNA was detected male-specifically from an early stage in testis differentiation. We find that Sertoli cells expressed osteopontin mRNA and that the resulting protein product is predominately localized within the cytoplasm of Sertoli cells during embryonic development of the testis.
Osteopontin Is Expressed by Sertoli Cells
To investigate the timing and sex-specificity of osteopontin expression during gonadal development, male and female genital ridges were isolated from embryos aged between 11.5 and 14.5 days post coitum (dpc) and in situ hybridization (ISH) carried out using a probe specific for the OPN transcript. OPN expression was sex-specific in the gonad at 11.5 dpc (Fig. 1). Expression increased as the XY gonad developed, with strongest staining visible in the testis cords. No expression was detected in the XX gonads at corresponding stages. Thus, OPN expression is male-specific in the embryonic gonad from 11.5 dpc, immediately after sex determination. Strong expression of OPN was also found in the Müllerian ducts of both sexes at 12.5 dpc, and in the mesonephric tubules of both sexes at later stages (Fig. 1).
To determine what cell types (Sertoli or germ cells) express OPN within testis cords, reverse transcriptase-polymerase chain reaction (RT-PCR) and whole-mount ISH was performed using gonads from We/We mutant mice. These gonads lack germ cells due to a mutation in the c-kit gene, resulting in a failure of germ cells to proliferate and migrate to the gonad. OPN mRNA levels in We/We were similar to those in wild-type (WT) littermates as determined by both RT-PCR and ISH (Fig. 2). By comparison, Sox9, a known Sertoli cell product, was also detected at similar levels in We/We and WT gonads, whereas expression of synaptonemal complex protein gene 3 (Scp3), expressed by male germ cells at this stage (DiCarlo et al.,2000), was absent in We/We samples.
To confirm that Sertoli but not germ cells expressed OPN, ISH was carried out on sections of 13.5 dpc gonads. OPN mRNA staining was typical of a transcript expressed by Sertoli cells. The nuclei of these cells are located toward the edges of the cords, and their cellular membranes project into the cords to surround germ cells, producing a distinctive stellate staining pattern. In contrast, a probe for the germ cell marker Oct4, showed ring-like staining in the cytoplasms of rounded cells in the lumen of the cords (Fig. 2C). Together, these data suggest that OPN is not expressed in germ cells, that Sertoli cells are responsible for the observed expression in developing testis cords, and that the presence of germ cells is not required for expression of OPN by Sertoli cells.
Osteopontin Protein Localization in the Developing Male Gonad
OPN protein has been shown to be either intracellular or secreted (Craig et al.,1988; McKee and Nanci,1996; Sodek et al.,2002; Suzuki et al.,2002); therefore, immunofluorescence was carried out to determine its location in the embryonic gonad. Sections of 12.5 and 15.5 dpc gonads were double-labeled with an osteopontin antibody and markers for either the ECM protein laminin, the Sertoli cell nuclear marker SOX9, or the germ and endothelial cell marker platelet endothelial cell adhesion molecule-1 (PECAM-1). OPN protein was located within the cords and did not overlap with laminin expression, which is found in the basement membrane (Fig. 3A).
At 12.5 dpc, OPN protein was intracellular, found in large clusters concentrated toward the periphery of Sertoli cells; this pattern of staining is typical for intracellular OPN as demonstrated previously in several cell types and has come to be referred to as “perimembranous” (Zohar et al.,1997; Sodek et al.,2002; de Oliveria et al.,2003). At a later stage (15.5 dpc), OPN protein was more diffuse throughout the cytoplasm of the Sertoli cells (Fig. 3B). These staining characteristics suggest that the protein is retained in the cytoplasm of the Sertoli cells, but we are unable to formally exclude the possibility that the OPN protein is in the secretory pathway. Furthermore, more detailed studies involving higher resolution studies on cultured cells would be required to fully address this issue.
Some protein was detected at the surface of the germ cells, as shown by costaining with PECAM-1 (Fig. 3C), which is expressed on the plasma membrane of germ cells (Wakayama et al.,2003).
In the tubules of the mesonephros, the majority of OPN was secreted and found in the lumen, whereas laminin was localized to the basement membrane that surrounds the mesonephric tubules (Fig. 4).
Osteopontin Null Testis
Mice deficient for OPN are fertile (Liaw et al.,1998), but embryonic testes of these mice have not been examined previously for subtle defects. OPN null and WT embryos were isolated at 13.5 dpc, just after testis cord formation, and sections were stained with antibodies detecting markers for Sertoli (AMH and SF-1), Leydig (SF-1), and germ cells (Vasa and OCT4). No morphological or cellular defects were identified in the OPN null testis, with germ cells enclosed by Sertoli cells and cords forming similarly to those in the WT testes (Fig. 5). This finding suggests that OPN may either not be required for testis development or may act redundantly with other similar proteins.
Two Other Members of the SIBLING Family Are Expressed Male-Specifically in the Embryonic Testis
Osteopontin has been shown to have some overlapping functions with two other SIBLING family members, BSP and DMP (Jain et al.,2002; Fisher et al.,2004; Ogbureke and Fisher,2004). To investigate if these genes are also expressed in the developing testis, RT-PCR was carried out with RNA isolated from male and female 13.5 dpc gonads. As shown in Figure 6A, expression of Dmp and Bsp could be detected only in XY gonad samples. ISH carried out on the same stage gonads revealed that, like OPN, the expression of these genes was restricted to the testis cords; very little (Bsp) or no (Dmp) expression was detected in female gonads of the same stage (Fig. 6B). These results demonstrate that other SIBLING members are also up-regulated male-specifically in the developing gonad.
To determine which cell type expresses Bsp and Dmp in the testis cords, RT-PCR on RNA from We/We gonads and section ISH (SISH) analysis were carried out (Fig. 6C,D). Bsp and Dmp expression levels were similar in We/We gonads, indicating that their expression does not require the presence of germ cells (Fig. 6C). In SISH, Bsp and Dmp were expressed by cells located toward the edge of the testis cord and exhibited a staining pattern typical of Sertoli cells (Fig. 6D). Taken together, these results show that Sertoli cells express Bsp and Dmp.
The present data demonstrate that Sertoli cells express OPN mRNA from a very early stage in testis development. In addition, the gene is expressed transiently in the Müllerian duct and mesonephric tubules of both sexes. Analysis of embryonic gonads from OPN null mice indicated that there were no major morphological defects. Two other SIBLING family members, BSP and DMP, were also found to be expressed male-specifically, suggesting that their functions may overlap in the testis and they may act redundantly in testis development.
By using an antibody to OPN, we were able to show that this protein is present in the cytoplasm of Sertoli cells. Intracellular OPN (iOPN) has also been found in fibroblastic, metastatic, osteoclast, and macrophage cells (Sodek et al.,2002; Suzuki et al.,2002; Kury et al.,2004). Recent work has suggested that intracellular OPN may have a role in the cell process formation and cell migration: OPN colocalized with CD44, largely at the periphery of cell processes extended by migrating/spreading osteoclasts cells (Suzuki et al.,2002) and can modulate CD44-dependent chemotaxis of macrophages (Zhu et al.,2004). In the developing testis, iOPN may be involved in the migration of Sertoli cells within the gonad and its reorganization into cords. Another possible role could be the formation of cytoplasmic processes in Sertoli cells that project into the cords to surround and nurse the germ cells. Matricellular proteins such as SPARC and Tenascin-C modulate cell de-adhesion, controlling the restructuring of focal adhesions between cells (Murphy-Ullrich,2001). In the testis, the cytoplasmic processes of Sertoli cells associate with the germ cells by anchoring junctions (reviewed by Siu and Cheng,2004). This association is particularly important in sperm maturation, when these junctions are restructured/disassembled during spermatogenesis. It may be that matricellular proteins like OPN are involved as modulators of Sertoli-germ cell adhesion, a possibility suggested by our immunofluorescence study that showed that OPN protein is located within the cytoplasmic processes and at the plasma membrane of germ cells.
We found that OPN is expressed and secreted into the tubules of the developing mesonephros. In males, mesonephric tubules later form the vasa deferentia, and OPN may play a role in epididymal development and maturation by regulation of MMP activity. MMPs are important regulators of tissue remodeling that act by degrading the extracellular matrix. It has been shown previously that osteopontin can enhance the proteolytic activity of MMP3 (Fedarko et al.,2004), and mesonephric tubules and Sertoli cells of the XY embryonic gonad express MMP3 mRNA (M. Wilson, data not shown). OPN has also been shown to enhance secretion and/or activation of MMP2 (Philip et al.,2001; Philip and Kundu,2003). MMP2 is expressed in the XY Müllerian ducts and is likely to be involved in their regression (Roberts et al.,2002).
In summary, the mesonephric tubules and the Sertoli cells of the testis cords express OPN. The presence of an intracellular form of OPN within the Sertoli cells suggests that it may be involved in regulating migration, cell process formation, or adhesion of these cells during testis development.
Mice Staging and Dissection
Timed matings were carried out using Swiss Quackenbush mice. Gonadal tissue was collected between 11.5 and 14.5 dpc, where 0.5 dpc was defined as noon on the day of discovery of the copulation plug. At 11.5 dpc, embryos were sexed by PCR for the Y-linked gene Zfy (Hogan et al.,1994). XY gonadal samples were also dissected from homozygous We mutant mice (Tan et al.,1990) at 13.5 dpc.
The construction and genotyping of the OPN mutant stain has been described previously (Liaw et al.,1998). Embryos were isolated at 13.5 dpc from mutant and WT (C57BL/6) mice for embedding and sectioning.
Osteopontin probe was generated by PCR from 13.5 dpc male gonad cDNA using the following primer pair: 5′-ACTGAGGTCAAAGTCTAGGA-3′ and 5′-CCTCTTCTTTAGTTGACCTC-3′. The resulting PCR product was cloned into pGEMT-easy (Promega) and sequenced. Antisense RNA probe was synthesized with T7 polymerase and a sense probe using SP6 RNA polymerase (Roche Biochemicals). Probes for Oct4 (Schöler et al.,1990) and Sox9 (Wright et al.,1995) were synthesized as described previously. The whole-mount ISH procedure has been published previously (Wilkinson and Nieto,1993).
Embryo gonads where isolated and fixed in 4% paraformaldehyde, washed with PBS, incubated overnight in PBS/30% sucrose at 4°C, and then embedded in OCT. SISH was carried out as described previously (Schepers et al.,2003).
For analysis of cryosections, samples were prepared as described above. Twelve-micrometer cryosections were washed in PBSTx (PBS with 0.1% Triton X-100) and blocked with 10% heat inactivated horse serum. Slides were then incubated with the appropriate combination of antibodies. Goat anti-OPN antibody was purchased from R&D systems (AF808) and used at 1:100 dilution. Rabbit anti-SOX9 was used at a 1:300 dilution and rat anti–PECAM-1 (purchased from BM Pharmingen) at a 1:400 dilution. Rabbit anti-Laminin (1:1,000, Sigma) was also used in combination with anti-osteopontin. After incubation overnight at 4°C, slides were washed three times with PBSTx and incubated with the appropriate secondary antibody diluted 1:500 in PBSTx. Secondary antibodies (anti-rat Alexa488, anti-rabbit Alexa594, anti-mouse Alexa488, anti-rabbit Alexa488, and anti-goat Alexa594) were purchased from Molecular Probes and used at a 1:200 dilution. Conventional negative controls were performed; no staining similar to that of anti-osteopontin antibody was seen when either the primary antibody was omitted or when an isotype control antibody was used. Slides were mounted and photographed by confocal microscopy using the Zeiss LSM 510 Meta Confocal microscope.
For immunofluorescence on paraffin sections, embryos were embedded in paraffin and sectioned at 5 μm. Paraffin sections were dewaxed in Xylene and rehydrated through an enthanol:water series (100%, 95%, 90%, 70%, 50%, 0%). Slides were microwaved for 10 min at 70% power in antigen unmasking solution (Vector Laboratories) and then subjected to antibody staining as described above. Rabbit anti–SF-1 antibody (1:1,000 dilution) was a kind gift from K. Morohashi (National Institute for Basic Biology, Okazaki, Japan). Goat anti-AMH antibody (sc-6886) and mouse anti-OCT4 (sc-5279) were purchased from Santa-Cruz Biotechnology and used as a 1:100 dilution. Rabbit anti-MVH (Fujiwara et al.,1994) was prepared at a 1:3,000 dilution.
Total RNA was isolated from embryonic gonads using Trizol reagent (InVitrogen) according to the manufacturer's instructions. Conditions for the cDNA synthesis and PCR reaction and the sequences for the oligonucleotides for Sox9, Hprt, and Scp3 were published previously (Smith et al.,2004). The sequences of the oligonucleotide pairs were as follows: BSP, 5′-ACACTTACCGAGCTTATGAGG-3′ and 5′-TTGCGCAGTTAGCAATAGCAC-3′; DMP, 5′-TGACAATGACTGTCAGGACGG-3′ and 5′-GGCTTTGCTACTGTGGAACCT-3′. Osteopontin primers are described above.
The authors thank James Smith and Dagmar Wilhelm for critical reading of this manuscript. P.K. is a Professorial Research Fellow of the Australian Research Council. L.L. was supported by NIH grants HL070865-01A1 and RR 15555-04S1