Abstract
The role of fibroblast growth factor (FGF) 2 and FGF9 as mediators of cell–cell interactions between Sertoli cells (SCs) and peritubular cells (PCs) was investigated. Using RT-PCR, we demonstrated that SCs and PCs recovered from 20-day-old rats expressed several of the seven FGF receptors (FGFRs), and more specifically the FGFR1 IIIc. FGF2 and FGF9 did not elicit any morphological changes in primary cultures of SCs, nor did they alter the number of SCs in culture. By contrast, changes in shape were observed in FGF2- and FGF9-treated PCs. In addition, FGF2 but not FGF9 enhanced significantly and dose-dependently the number of PCs in culture, indicating that FGF2 was a survival factor for these cells. It was also mitogenic because it enhanced the [3H]thymidine labeling index in PCs. We next examined the effects of FGF2 and FGF9 in a coculture system using 20-day-old rat SCs and PCs, and in an organotypic culture system using XY rat embryonic gonads. In both models, FGF2 and FGF9 were found to promote cellular interactions as evidenced by the extent of cellular reorganization in the coculture system, and cord morphogenesis and growth in the organotypic culture system. A key feature in SC–PC interactions is the synthesis and remodeling of the basement membrane which is co-elaborated by the two cell types. Since basement membrane homeostasis depends upon the coordinated activity of proteinases and inhibitors, the proteinases and inhibitors produced by PCs and SCs degrading or opposing degradation of the major components of the basement membrane were further studied. Specifically, we monitored the metalloproteinases (MMP)- 2 and -9 and the tissue inhibitors -1, -2 and -3, the plasminogen activators (PAs) and the PA inhibitor-1, using zymography for the proteinases and Western blots for the cognate inhibitors. Cocultures received FGF or an analog of cAMP in order to prevent cellular reorganization. We found that FGF2 was unique in inducing MMP-9 in coculture. Also, the enhanced levels of the PA inhibitor-1 and the 30 kDa band glycosylated form of tissue inhibitor-3 correlated with the enhanced SC–PC reorganization. It was concluded that FGF2 and FGF9 are morphogens for the formation of testicular cords.
Introduction
The morphogenesis of seminiferous tubules may depend upon interactions among Sertoli cells (SCs) and peritubular cells (PCs) that are homologous to epithelial–mesenchymal cell interactions. SCs are of the epithelial cell type; they form tubules and provide structural and nutritional support for the developing germinal cells. PCs are of the mesenchymal cell type; they surround the SCs and form the exterior wall of the seminiferous tubule. PCs are separated from the SCs by a basement membrane. This complex extracellular matrix (ECM) is produced co-operatively by SCs and PCs, and is essential in generating and maintaining the structural integrity of the tubule and in promoting the structural differentiation of the cells (Dym 1994).
Epithelial–mesenchymal cell interactions are regulated by a complex array of regulatory agents including the fibroblast growth factors (FGFs). FGFs belong to a large family of heparin-binding polypeptides, and 22 members and four high affinity receptors (FGFRs) encoding trans-membrane protein tyrosine kinase receptors have been characterized. The four FGFR genes encode seven functionally distinct variants by alternative splicing, and these variants exhibit differential binding affinities toward the FGF ligands. FGFs are involved in an array of biological processes during development and in adult life, including regulation of cell proliferation, migration and differentiation. There is evidence of a functional redundancy between some FGFs and FGFRs. Furthermore, most FGFs can signal through multiple FGFRs. Consequently, targeted deficiency of an FGF ligand or one of its receptors may induce mild to severe defects. Examples have been shown, on the one hand by fgf1–fgf2 single or double (Miller et al. 2000), fgfr3 (Colvin et al. 1996) or fgfr4 (Weinstein et al. 1998) knockout mice which are viable and fertile, and on the other hand by the lethality observed in the fetal or in the early postnatal period in mice lacking either one of fgf 4, 8, 9, 10, 15 and 18 or fgfr1 and fgfr2 (Goldfarb 1996, Itoh & Ornitz 2004).
Within the testis, FGF2 and FGF9 at least have emerged as being important regulators throughout testicular development. FGF2 was first isolated and characterized from bovine testes (Ueno et al. 1987), and its presence was next confirmed in the testes of different species, including humans, cows, mice and rats. Evidence then accumulated showing that FGF2 exerted mitogenic and non-mitogenic functions in the rat testis, where several FGFRs have been localized (Mullaney & Skinner 1992, Han et al. 1993, Le Magueresse-Battistoni et al. 1994, 1996, Van Dissel-Emiliani et al. 1996, Cancilla & Risbridger 1998, Cancilla et al. 2000). The importance of FGF9 in testis functions was shown more recently by the finding that most fgf9 −/− XY reproductive systems appeared grossly female at birth. Fgf9 −/− mice also exhibited lung hypoplasia and died at birth (Colvin et al. 2001). Detailed studies have shown that FGF9 supports the normal formation of testis cords (Colvin et al. 2001, Schmahl et al. 2004), downstream of Sry, the sex-determining gene in mammals (Swain & Lovell-Badge 1999). Colvin et al.(2001) and Schmahl et al.(2004) suggested that FGF9 could be involved in the interactions between the fetal Sertoli cells and the adjacent mesonephros from which the peritubular cells arise (Buehr et al. 1993).
These observations prompted us to investigate whether FGF2 or FGF9, which are testicular products, acted as morphogenic, survival or mitogenic factors in primary cultures of SCs and PCs. Prepubertal rat testes were used as a source of cells and we examined by RT-PCR the pattern of FGFRs present in these cells. We also questioned whether FGF2 and FGF9 acted as morphogens in the formation of testis cords in two different culture systems: a coculture system using prepubertal SCs and PCs, and an organotypic culture system using XY rat embryonic gonads. Finally, we monitored the pattern of expression of several proteinases and proteinase inhibitors produced by testicular cell types (Fritz et al. 1993, Longin et al. 2001, Siu et al. 2003). Our interest stems from several considerations. It has been demonstrated that in SC–PC cocultures, cord-like structures are allowed to assemble de novo depending on the production and deposition by the two cell types of ECM proteins within a basement membrane (Tung & Fritz 1980, 1987). Proteinases and proteinase inhibitors regulate, at least in part, the homeo-stasis of ECM proteins (Vu & Werb 2000). The major components of the testicular basement membrane, i.e. collagen IV, fibronectin (FN) and laminin (Richardson et al. 1995) are putative substrates for matrix metallo-proteinases (MMPs) and plasminogen activators (PAs) (Vu & Werb 2000). PCs and/or SCs produce several MMPs and tissue inhibitors of metalloproteinases (TIMPs) as well as the PAs and the inhibitor-1 (PAI-1) (Fritz et al. 1993, Longin et al. 2001, Siu et al. 2003), which may well take part in basement membrane turnover and synthesis.
Materials and Methods
Organotypic cultures and immunohistochemistry
Gonads with attached mesonephroi were isolated from 13.5 days postcoitum (dpc) embryos. This age corresponds to the very onset of sexual differentiation when gonads cannot be discriminated on a morphological basis. Embryos were therefore sexed extemporaneously scoring the presence of the barr bodies in amniotic membranes, and explants were cultured for 1 day on steel grids previously coated with 2% agar, as described elsewhere (Magre & Jost 1984). Grids were placed in organ culture dishes with 0.8 ml medium necessary to wet the grid. CMRL 1066 (Life Technologies, Grand Island, NY, USA) was used as basal culture medium. FGF2 or FGF9 (TEBU, Le Perray-en-Yvelines, France) at a dose of 50 ng/ml were added to some cultures. Controls were cultures without FGF2 or FGF9. After 1 or 3 days of culture, explants were fixed for 1 h at 4 °C in 2% paraformal-dehyde in 0.1 M phosphate buffer, pH 7.4. This was followed by washes in 0.1 M phosphate-buffered 0.15 M saline (PBS), pH 7.4, and immersion in a series of cold sucrose solutions. Specimens were frozen and sectioned in a cryostat at 5 μm. To identify gonadal structures, the distribution of anti-Müllerian hormone/Müllerian-inhibiting substance (AMH/MIS), FN and keratin 8 was examined using indirect immunoperoxidase detection as described (Mazaud et al. 2002). Antibodies used were a rabbit anti-AMH/MIS raised against bovine AMH/MIS (kindly provided by Dr B Vigier, Unité Biologie du Développement et Reproduction INRA, Jouy en Josas, France), a rabbit anti-human plasma FN serum (6071 SA; Life Technologies) and a monoclonal mouse anti-keratin 8 (LE 41, kindly provided by Professor E B Lane, Cancer Research UK, University of Dundee, Dundee, UK).
Cell preparations
SCs and PCs were isolated from 20-day-old Sprague–Dawley rats and cultured at 32 °C in a humidified atmosphere of 5% CO2 as previously described (Le Magueresse-Battistoni et al. 1998). Experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals. Briefly, decapsulated testes were minced, then subjected to sequential enzymatic treatment with collagenase-dispase (0.05%) (Roche Biochemicals, Meylan, France), hyaluronidase (0.1%) (Sigma Chemical Co.) and collagenase-dispase (0.05%) at 32 °C for 30 min each. Cells were allowed to sediment by gravity between enzymatic treatment and washes. The final SC suspension was filtered through a 60μm nylon screen. At that point, the SC preparations were contaminated with 10–15% of germ cells (spermatogonia and early spermatocytes) and 1% PCs (Le Magueresse-Battistoni et al. 1998). PCs were collected in the supernatant of the first collagenase-dispase digestion. Cells were filtered through a 20μm nylon screen and cultured at 32 °C in the culture medium (Dulbecco’s modified Eagles’ medium–Ham’s F-12; Life Technologies) supplemented with 10% fetal calf serum (FCS). An average of 20 Petri dishes (diameter 60 mm) could be prepared when starting with ten testes. Two hours after plating, culture dishes were rinsed well to remove serum and non-adhering cells, mainly the contaminating germ cells.
Cell culture and coculture
SCs (1 × 106 viable cells) were either cultured alone (SC culture) or cocultured (SC–PC) so that SCs were placed on top of PCs at a ratio of 3:1 (SC:PC) as described (Skinner & Moses 1989). PCs were also cultured alone (PC cultures). Twenty-four hours after seeding, SC and SC–PC culture dishes were exposed to a hypotonic treatment to remove the 10–15% of germ cells which contaminate the SC preparations (Galdieri et al. 1981). Culture dishes were replenished with culture media supplemented or not (control) with FCS (10%), dibutyryl cAMP (bu2cAMP) (1 mM), platelet-derived growth factor (PDGF) (50 ng/ml; TEBU), FGF9 or FGF2 (50 ng/ml, unless otherwise specified). Morphological changes were observed daily by phase contrast microscopy. On day 5 of culture, cells were scraped from the dishes and RNA or proteins were extracted. Culture and coculture media were collected for protein analysis.
MTS test
SC and PC number was examined in 96 multiwell plates (10 000 cells/well) after a 4-day treatment with increasing doses of FGF2 or FGF9 ranging from 3 to 100 ng/ml, FCS (10%) or Bu2cAMP (1 mM). Cells were quantitated using the celltiter 96 R Aqueous non-radioactive cell proliferation assay according to manufacturer’s intructions (Promega, France). The assay is based on the ability of viable cells to reduce the yellow tetrazolium salt, MTS, to water-insoluble, dark blue formazan crystals, which can be spectrophotometrically measured at 490 nm. Usually four replicate wells were used for each group. The control included untreated cells, whereas medium alone was used as a blank.
Measurement of DNA synthesis
DNA synthesis was measured as [3H]thymidine (Amersham Pharmacia Biotech Europe GMBH, Orsay, France) incorporation into trichloroacetic acid (TCA)-insoluble material. PCs were plated in 96-well plates (10 000 cells/well). On the next day, culture medium was changed and FGF2 (1.5, 15, 30 and 50 ng/ml) or FGF9 (50 ng/ml) was added per well together with 0.5 μCi/ml [3H]thymidine (50 Ci/mM). Control cells received only 0.5 μCi/ml [3H]thymidine. After a 24-h incubation period, cells were washed twice with ice-cold PBS and twice with ice-cold 5% TCA to remove unincorporated radioactivity. Cells were solubilized by adding 0.25 M NaOH and the cell lysate was transferred to scintillation fluid and counted by a spectrometer (Beckman-Coulter, Villepinte, France). Each measurement consisted of a four-well replica.
RNA extraction and RT-PCR
The procedure for RNA extraction and RT-PCR has been described previously (Longin et al. 2001, Guyot et al. 2003). Total RNA was extracted from whole testes, liver and kidney recovered from 20-day-old Sprague–Dawley rats. Total RNA was also extracted from the SCs and PCs isolated from 20-day-old Sprague–Dawley rats and cultured for 5 days. Specific primers for RT-PCR were designed with the help of the Gene-Jockey sequence processor (Biosoft, Cambridge, Cambs, UK). The optimal temperature of annealing for each pair of primers was defined (Table 1). Negative controls contained water instead of cDNA. PCR with no RT reactions gave no band, eliminating the possibility of a genomic DNA contamination in the RNA preparations. Amplified cDNAs were visualized in a 1.5% agarose gel stained with ethidium bromide. A DNA ladder (Promega, Charbonniè res, France) was loaded on each gel, and PCR products were sequenced by Biofidal (Lyon, France).
Gelatin zymography
Culture and coculture media were concentrated ten times using Centriprep (cut-off at 10 kDa; Amicon, Beverly, MA, USA). Proteins (10–20 μg/lane measured by BCA protein assay (Pierce-Interchim, Montluçon, France)) were electrophoresed at 4 °C on 10% polyacrylamide gels containing 1 mg/ml gelatin (Sigma Chemical Co.) in the absence of any reducing agent (Longin et al. 2001, Guyot et al. 2003). Following electrophoresis, SDS was removed from the gel by exchange in Triton X-100 (two washes of 30 min at room temperature in 2.5% Triton X-100 followed by two washes of 20 min in distilled water). The gel was subsequently incubated at 37 °C for 48 h in 100 mM Tris–base, pH 7.6, containing 15 mM CaCl2. In these conditions, gelatinases present in the samples renatured and autoactivated. White zones of lysis indicating gelatin-degrading activity were revealed by staining with Coomassie Brillant blue R-250.
Plasminogen zymography
Cells scraped from the culture and coculture dishes were lysed in a Tris-buffer solution (0.5 M, pH 8.2) containing 0.5% Triton X-100 and proteins (20 μg), and then electrophoresed at 4 °C on 8% polyacrylamide gels in the absence of any reducing agent. Following electrophoresis, SDS was removed from the gel by exchange in Triton X-100 (two washes of 30 min at room temperature in 2.5% Triton X-100 followed by three washes of 30 min in distilled water). The gel was subsequently placed on a casein–agar–plasminogen underlay essentially as described (Tolli et al. 1995, Odet et al. 2004). Clear lytic zones indicating plasminogen-degrading activity appeared after a 24-h incubation at 37 °C, and gels were scanned.
Western blot analysis
SDS-PAGE and Western blotting were carried out as reported earlier (Longin et al. 2001, Guyot et al. 2003), using either (i) tenfold concentrated cell culture media for the analysis of TIMPs-1 and -2 and PAI-1 or (ii) cell lysates prepared using PBS containing 1% NP-40 and 5 mM EDTA for the analysis of TIMP-3. Briefly, proteins were separated by electrophoresis performed on 10% (15% for TIMP-2) polyacrylamide gels under reducing conditions. Precision protein standards (Biorad, Hercules, CA, USA) were loaded onto each gel for accurate estimation of the Mr of the bands. The proteins were then electrophoretically transferred to a PVDF membrane (Biorad). After treatment with a blocking solution (5% skim milk in Tris-buffered saline; TBS) for 3 h, the membrane was incubated overnight with the primary antibody at 4 °C. Primary antibodies were a rabbit anti-PAI-1 (working dilution 1:200) from American Diagnostics (Greenwich, CT, USA), a rabbit anti-TIMP-1 (working dilution, 1:2000), a rabbit anti-TIMP-2 (working dilution, 1:1000) and a rabbit anti-TIMP-3 (working dilution, 1:1000) (all three from Chemicon International, Temecula, CA, USA). The PVDF membrane was washed and incubated with peroxidase-conjugated goat anti-rabbit IgG antibody (Dako, Trappes, France). Different experiments were performed to ensure equal loading of the samples. Gels processed for Western blot analysis were systemically stained with red ponceau before proteins were transferred. In the case of TIMP-3, PVDF membranes hybridized with the anti-TIMP-3 antibody were stripped and reprobed with a rabbit anti-β-actin antibody (working dilution, 1:1000; Sigma Chemical Co).
Data analysis
All experiments were repeated three to five times using independent cell preparations. A representative experiment from each series is presented. Data are the means±s.e.m. The significance of the results was determined by ANOVA followed by the Mann–Whitney U test. Differences are accepted as significant at P<0.05.
Results
Morphological aspect of SC and PC cultures exposed for 4 days to FGF2 and FGF9
After a few days in culture, SCs formed a confluent monolayer (Fig. 1). No gross morphological changes could be observed between SCs cultured for 5 days in control culture medium and SCs cultured in the presence of FGF2 or FGF9 (50 ng/ml) from day 1 to day 5 of culture. This was in contrast with the morphological aspect of the PCs which changed depending on the treatment performed. For example, FCS-treated PC cultures displayed a more flattened fibroblast-like appearance than did control cells, and bu2cAMP-treated cells remained round and exhibited neuronal-like cytoplasmic expansions. This finding extended previous results (Ailenberg et al. 1990). FGF9-treated cells appeared more retracted than control or FCS-treated cells even though they had spread and showed slender long processes. FGF2-treated cells adopted a highly retracted cell shape with fewer and shorter cytoplasmic expansions than the FGF9-treated cells (Fig. 1).
FGF2 but not FGF9 is a survival and mitogenic factor for PCs
FGF2 and FGF9 were added to primary cultures of SCs and PCs at increasing doses ranging from 3 to 100 ng/ml, and cell number was determined after a 96-h incubation period using the MTS viability test. In preliminary experiments (not shown), we ensured that FCS could enhance PC but not SC proliferation as reported (Skinner et al. 1989). We found that FGF2 and FGF9 did not alter SC number (not shown). By contrast, FGF2 but not FGF9 dose-dependently increased the number of PCs. First significant effects were seen with FGF2 (6 ng/ml) (+22%, P<0.05, n=4). Maximal effects were obtained with FGF2 (50 ng/ml) (+48%, P < 0.05, n=4). Bu2cAMP had no effect (Fig. 2A). To discriminate between cell survival and cell proliferation, we examined [3H]thymidine incorporation. FCS was used as a positive control (not shown). Data presented indicated that FGF2 enhanced the DNA labeling index of PCs after a 24-h incubation period. First significant effects were seen for FGF2 (15 ng/ml) (+38%; P<0.05) and +83% enhancement (P<0.05) was observed with FGF2 (50 ng/ml). FGF9 (50 ng/ml) had no effect (Fig. 2B).
RT-PCR screening of FGFRs and FGF9
The nature of the receptors present on the cultured PCs as well as on the cultured SCs and in the whole testes of 20-day-old rats was determined using RT-PCR. Specific pairs of primers (Table 1) were designed against FGFRs 1–3 isoform IIIb or IIIc and FGFR4. RT-PCR studies were also conducted using primers directed against 18S to ensure that equal amounts of material were used. RNA extracted from PCs gave RT-PCR bands for FGFRs 1 IIIc, 2 IIIb and IIIc, a very weak band for FGFR3 IIIc, and no band for the IIIb isoform of FGFRs 1 or 3, or for FGFR4. RNA extracted from SCs gave a PCR band for receptors 1–3, but not for FGFR4. An RT-PCR band was found in the whole rat testes for the seven receptors. With regard to the ligands, FGF9 was present in SC cultures but not in PC cultures (Fig. 3), whereas FGF2 has been shown to be a PC as well as an SC product in prepubertal testes (Mullaney & Skinner 1992). A PCR band for FGF9 was also detected in RNA from rat fetal testes. RNA recovered from kidney was used as a positive control for FGFR4.
Morphological aspect of SC-PC cocultures exposed for 4 days to different treatments
When SCs are seeded on top of PCs they reaggregate and form tubule-like structures resembling seminiferous cords of the differentiating testis (Tung & Fritz 1980). In the present study, we observed that the extent of reorganization varied depending on the factor added in the culture medium (Fig. 4). We found that FGF2 (50 ng/ml) promoted the formation of condensed nodules of aggregated SC. Cocultures exposed to FGF9 (50 ng/ml) also showed a reorganization, and its effect was comparable with that elicited by PDGF (50 ng/ml) known to stimulate cord formation (Uzumcu et al. 2002, Brennan et al. 2003, Ricci et al. 2004). However, the nodules were smaller than in the FGF2-treated cocultures and the PCs could still develop long cytoplasmic expansions between the adjacent nodules. The control cocultures and the FCS-treated cocultures were characterized by the presence of very small aggregates of SCs and whirls of PCs (Fig. 4). Cocultures treated with FGF2 (10 ng/ml) formed fewer tubule-like structures than cocultures treated with FGF2 (50 ng/ml), and cocultures treated with FGF9 (10 ng/ml) were not different in their aspect from control cocultures (not shown). The extent of reorganization was time-dependent, and large aggregates could be seen after a period of 8 days in control cocultures. However, as the nodules increased in size they exhibited a reduced ability to remain attached to the plastic dishes and finally they were detaching when culture media was changed (e.g. in the FGF2-treated cocultures older than 8 days). Consistent with previous data (Tung & Fritz 1987), no aggregation of SCs was observed in the presence of bu2cAMP (1 mM; Fig. 4).
Effect of FGF2 and FGF9 on cord formation in rat fetal gonads
We also examined whether FGF2 and FGF9 (50 ng/ml) acted as morphogens for rat embryonic testes. Typically when gonads at the very onset of testis differentiation (13.5 dpc in the rat) remain associated with their mesonephroi, seminiferous cords differentiate in vitro (Magre & Jost 1984). These cords are identifiable using keratin 8 and FN to identify the epithelial and mesenchymal cells respectively. After a 24-h exposure, explants have grown, cords have differentiated and SCs could be immunostained for AMH/MIS in the three culture conditions. It is noteworthy that the extent of growth was more pronounced and the well-differentiated cords were higher in number in the FGF2-treated explants as compared with the FGF9-treated explants. Also, FGF9-treated explants were more developed than the explants in control cultures (Fig. 5). After 3 days of culture, all explants had further enlarged. FGF2- and FGF9-treated explants could no longer be discriminated on a morphological basis, and control explants remained less developed than the FGF2-or the FGF9-treated explants (data not shown).
Zymographic analysis of SC and SC–PC cultures and cocultures
Two groups of gelatinolytic bands are classically found with testicular extracts: the pro- and active form of MMP-9 migrating at 92 and 84 kDa, and the pro, intermediate and active forms of MMP-2 migrating as a triplet around 72, 66 and 62 kDa (Hoeben et al. 1996, Longin et al. 2001). Gelatinases are mostly secreted proteins, and we analyzed concentrated media of SCs and SC–PC cultured for 96 h (day 1 to day 5 of culture). We found that bu2cAMP stimulated MMP-2 synthesis and activity, extending previous observations on SC and PC cultures (Fritz et al. 1993, Hoeben et al. 1996). FGF9 had no effect on the activity/synthesis of the MMPs-2 and -9 in SC–PC coculture but a slight decrease was detected in SC culture. PDGF slightly stimulated the two MMPs in coculture (Fig. 6). In fact, the most striking effect was observed in the FGF2-treated cocultures in which MMP-9 and, to a lesser extent MMP-2 were stimulated (both the pro- and active forms). By contrast, MMP-9 was not enhanced in FGF2-treated SC cultures and MMP-2 was weakly stimulated (Fig. 6A).
Two types of PAs have been described in the testis, the urokinase-type (uPA) and the tissue-type (tPA) which migrate around 45 and 70 kDa respectively. Cell lysates were examined, and we first ensured that bu2cAMP stimulated tPA and inhibited uPA as reported (Fritz et al. 1993, Tolli et al. 1995). In these conditions, we found that FGF2 but not FGF9 enhanced the tPA lytic band. It is noteworthy that the tPA lytic band is barely detectable by comparison with the urokinase lytic band (Fig. 6B).
Western blot study of cognate inhibitors
Three PAIs, i.e. PAI-1, -2 and -3 have been described. However, PAI-2 is barely detectable in the testis and PAI-3 is a Leydig cell product in the immature rat testis (Odet et al. 2004). PAI-1 is a doublet of 46 and 49 kDa produced by PCs and SCs. It is downregulated by bu2cAMP and upregulated by FGF2 in both cell types (Le Magueresse-Battistoni et al. 1996, 1998). Results presented in Fig. 7 extend these previous data, and demonstrate that in SC–PC cocultures FGF9 and PDGF stimulated PAI-1 to the same extent as did FGF2.
TIMPs can inhibit more or less all classes of MMPs (Vu & Werb 2000). TIMPs-1 and -2 are secreted proteins in contrast to TIMP-3 which binds tightly to the ECM (Woessner 2001). TIMPs-1 and -2 antigen levels increased in coculture in the presence of bu2cAMP, extending previous data on SC cultures (Ulisse et al. 1994, Gronning et al. 2001, Longin & Le Magueresse-Battistoni 2002). The growth factors exerted a smaller (TIMP-1) or an equivalent (TIMP-2) effect as compared with the effects elicited by bu2cAMP (Fig. 7). TIMP-3 occurs as a minor 25 kDa unglycosylated protein and a major 30 kDa glycosylated protein in the cell lysates of untreated cocultures. Treatment of cocultures with FGF2 or FGF9 resulted in an increase of the 30 kDa glycosylated form of TIMP-3 (Fig. 7). The protein profile observed in the bu2cAMP-treated cocultures was very different as compared with the untreated or growth factor-treated cocultures, in that a high molecular weight band migrating around 60 kDa was strongly labeled in addition to the 30 kDa protein. β-actin was used to ensure equal loading of the cell lysates proteins (Fig. 7).
Discussion
The purpose of this study was to investigate further which paracrine factors of testicular origin may mediate SC–PC interactions in the rat testis. The choice of FGF9 and FGF2 stems from previous findings. FGF9 is expressed in mice fetal testes where it promotes testis cord formation, and its genetic deletion results in male-to-female sex reversal (Colvin et al. 2001, Schmahl et al. 2004). An RT-PCR band of the right sequence for FGF9 was detected in total RNA recovered from rat fetal testes as well as from 20-day-old rat SCs (this study). FGF2 immunostaining is restricted to PCs and Leydig cells in the rat fetal testis (Gonzales et al. 1990). FGF2 is also produced in the rat postnatal testis where it exerts mitogenic and regulatory functions on somatic and/or germinal cells (Mullaney & Skinner 1992, Han et al. 1993, Le Magueresse-Battistoni et al. 1994, 1996, Van Dissel-Emiliani et al. 1996, Cancilla & Risbridger 1998, Cancilla et al. 2000).
In a first series of experiments, we investigated whether FGF2 and FGF9 acted as morphogenic, survival or mitogenic factors in primary cultures of SCs and PCs recovered from 20-day-old rats. We observed that both FGFs were morphogenic for PCs but not for SCs. In addition, FGF2 but not FGF9 was a survival and a mitogenic factor for PCs in vitro. Dose–response studies indicated that maximal effects of FGF2 were seen with the dose of 50 ng/ml. This finding was consistent with previous publications examining the mitogenic or regulatory function of FGF2 (Hoeben et al. 1999, Van Der Wee & Hofmann 1999), FGF9 (Colvin et al. 2001) or PDGF (Uzumcu et al. 2002, Puglianiello et al. 2004, Ricci et al. 2004) on testicular cells. Hence, this dose of 50 ng/ml was chosen for the next experiments for each of the growth factors studied. We also found that FGF2 was not a mitogen for SCs, consistent with the fact that prepubertal SCs have a very low mitotic index (Skinner et al. 1989).
The next step was to determine the nature of the FGFRs present on SCs and PCs, because seven FGFR variants exhibiting differential affinity toward the FGF ligands have been described. Specifically, FGF2 preferentially binds FGFR1 IIIc, 3 IIIc and FGFR4, whereas FGF9 prefers FGFR2 IIIc, 3 IIIc and FGFR4 (Goldfarb 1996, Itoh & Ornitz 2004). RT-PCR analysis was conducted because specific antibodies necessary to discriminate the isoforms IIIb and IIIc of the FGFRs 1, 2, and 3 are still lacking. In adition, immunohistochemical data have previously shown the presence of the four FGFRs within the rat testis (Cancilla & Risbridger 1998). We found that cultured SCs exhibited signals for the FGFRs 1–3, and the isoform IIIc of the FGFR1 was the most abundant. The isoform IIIc of the FGFR1 was also the most abundant in cultured PCs. The signal for FGFR3 IIIc was weak in PCs versus the signal in SCs. Finally, we found no signal for FGFR4 in PCs and SCs, contrasting with previous immunohistochemical data showing FGFR4 in PCs (Cancilla & Risbridger 1998). This discrepancy may result from the two different experimental conditions, in situ versus cells cultured for 5 days in a chemically defined medium, in the present study. Therefore, it may be hypothesized that FGFR1 IIIc is the receptor mediating the mitogenic and morphogenic actions of FGF2 on PCs, and that the lack of effect of FGF9 on PC proliferation reflects the low abundance of FGFR3 IIIc in these cells. We are currently addressing this issue.
We then questioned whether FGF2 and FGF9 acted as morphogens for the formation of testis cords. Two different culture systems were used, a coculture system using the SCs and PCs recovered from prepubertal rat testes and an organotypic culture system using XY rat embryonic gonads attached to their mesonephroi. We observed that both FGFs were morphogens in SC–PC coculture in that they promoted reorganization into cord-like structures. In order to evaluate the potency of each FGF, other co-cultures received PDGF, which promotes cord formation (Uzumcu et al. 2002, Brennan et al. 2003, Puglianielli et al. 2004, Ricci et al. 2004), or an analog of cAMP to block cellular reorganization (Tung & Fritz 1987). FGF9- and PDGF-treated cocultures could not be distinguished on a morphological basis, whereas FGF2 was more efficient in promoting cellular reorganization after a 4-day exposure. Whether the differential effect of the FGF2 and FGF9 in cocultures reflects the differential abundance of the FGFRs remains plausible. However, it appears unlikely that the morphogenic effect of FGF2 in coculture resulted only from its mitogenic action on PCs because FCS stimulates PC proliferation. Nonetheless, FCS-treated co-cultures showed little reorganization as compared with FGF2-treated cocultures. Using rat embryonic gonads as an alternate model, we have also described a morphogen action for the two FGFs and a primary effect of FGF2 over FGF9 after a 24-h but not after a 72-h exposure. However, the expression pattern of FGFR1 IIIc and FGFR3 IIIc in the embryonic rat testes is presently unknown.
It has also been described that heparan sulfate proteoglycans (HSPGs) are low affinity receptors for FGFs and are required to efficiently activate FGFRs (Itoh & Ornitz 2004). Moreover, tissue-specific expression and modification of HSPGs have been shown to be responsible for the tissue-specific action of FGFs in other systems (Friedl et al. 1997, Park et al. 2000, Allen & Rapraeger 2003). Therefore, monitoring the expression pattern of the HSPGs in SC–PC cocultures as well as in the organotypic cultures treated by FGF2 and FGF9 should be instrumental in further elucidating the mechanisms of action of the two FGFs.
It should also be pointed out that several growth factors are morphogens during embryonic testis development, i.e., the hepatocyte growth factor (HGF) (Ricci et al. 2002), the PDGF (Uzumcu et al. 2002, Brennan et al. 2003, Puglianielli et al. 2004, Ricci et al. 2004) and the neurotropins (Cupp et al. 2000, 2003). Such an apparent redundancy may explain the lack of a severe testicular phenotype (at least with respect to embryonic testis development) in mice deleted for c-Met, the HGF receptor (Ricci et al. 2002), for PDGF receptor-β (Soriano 1994) and for trkA and trkC neurotropin receptors (Cupp et al. 2002). FGF2 null mice also seem to have a normal testicular function and reproduction (Miller et al. 2000) although no detailed studies have been reported so far. The notable exception to date is FGF9 (Colvin et al. 2001). It is therefore really intriguing that FGF2 was a better morphogen than FGF9 in our study. Whether this is a species-related phenomenon, i.e. rat versus mouse, should be examined.
Finally, we examined whether proteinases and their cognate inhibitors might be downstream targets in the FGF-signaling cascade. Interestingly, FGF2-treated co-cultures had a unique signature of proteinase expression. Indeed, FGF2 increased the overall synthesis and activity of the two gelatinases, particularly that of MMP-9, whereas bu2cAMP could induce MMP-2 but not MMP-9. Furthermore and in addition to MMP-9, we found that TIMP-3 and PAI-1 antigen levels, but not those of TIMPs-1 and -2, correlated with the degree of reorganization in cocultures. Indeed, the 30 kDa glycosylated form of TIMP-3 was the major band in control and growth factor-treated cocultures, whereas the 60 kDa band predominated in the bu2cAMP-treated cocultures. TIMP-3 occurs normally as 25 kDa unglycosylated and 30 kDa glycosylated proteins. In addition, high molecular weight bands may also be labeled. Because TIMP-3 is an ECM-associated protein, the high molecular weight proteins may represent ECM ligands that form SDS stable complexes with TIMP-3 as reported in the rat central nervous system (Jaworski & Fager 2000). It remains that the nature of such bu2cAMP-induced ECM ligand is presently unknown. With regard to PAI-1, bu2cAMP-treated cocultures had decreased antigen levels whereas they were enhanced in the growth factor-treated co-cultures. It is also worth noting that, apart from being a fast-acting and a specific inhibitor of the PAs, PAI-1 exhibits multiple activities including the ability to regulate the attachment–detachment of a cell from the extracellular matrix components by inactivating integrins (Czekay et al. 2003). However, it is not yet known whether this activity prevails in our coculture model.
In summary, we have demonstrated that FGF2 and FGF9 are morphogens for the formation of the testis cords in two different in vitro models. FGF2 and FGF9 may therefore participate to the formation of testis cords in the rat embryonic testis as demonstrated recently for FGF9 in the mouse embryonic testis (Colvin et al. 2001). Such a morphogen action may also be important in the prepubertal testes which is characterized by important restructuring. Specifically, tight junctions are formed between neighboring SCs thus creating the blood–testis barrier, and cords develop a lumen becoming tubules. Because there is evidence that ECM components may regulate the expression of tight junction proteins (Siu et al. 2003), it may well be hypothesized that FGF2 and FGF9 mediate mesenchymal–epithelial interactions between PCs and SCs and selectively regulate the expression pattern of specific proteinases and inhibitors and thus, subsequently, affect the basement membrane remodeling.
List of sequences of the designed specific primers for PCR studies
| Primers (5′ → 3′) | Size (bp) | Optimal temperature (°C) | |
|---|---|---|---|
| The optimal temperature for annealing (°C) and the size of the expected PCR fragments (bp) are reported. s, sense; as, antisense; FGF, fibroblast growth factor; FGFR, FGF receptor; IIIb and IIIc, FGFR isoforms IIIb and IIIc respectively; 185, 185 ribosomal RNA. | |||
| Studied genes (accession number) | |||
| FGF9 (NM_012952) | s: TGTGGACAGTGGACTCTACC as: ATCGGCAAGTCCTCATCCAAGC | 463 | 58 |
| FGFR1 IIIb (AF176552) | s: TGGGAGCATTAACCACACCTA as: CTTTTGCCACAGGTCTGGTGA | 382 | 64 |
| FGFR1 IIIc (NM_024146) | s: TGGGAGCATTAACCACACCTA as: GCACCTCCATTTCCTTGTGG | 271 | 66.3 |
| FGFR2 IIIb (AF294649) | s: GGTAAATAGCTCCAATGCAGAAG as: TGTTTGGGCAGGACAGTGAGCCA | 138 | 58 |
| FGFR2 IIIc (NM_010207) | s: CCCATCCTCCAAGCTGGACTGCCT as: AACTGTCAACCATGCAGAG | 314 | 57 |
| FGFR3 IIIb (X59672) | s: CTGAAGACACAGGGGCTCCTT as: CCCGTGAACACGCAGCCAAA | 639 | 66.5 |
| FGFR3 IIIc (AF277717) | s: CTGAAGACACAGGGGCTCCTT as: TCTGTCCTGCACAATGTCACC | 559 | 71 |
| FGFR4 (M91599) | s: TGGACAACAGCAACATCGGA as: GTTCTTGTGTCTTCCGATTAGC | 527 | 65 |
| 18S (X00686) | s: GGAATAATGGAATAGGACCG as: TCTGTCAATCCTGTCCGTGTCC | 435 | 61 |
Morphological aspect of primary cultures of SCs (SC culture) and PCs (PC culture) recovered from 20-day-old rats and cultured for 4 days from day 1 to day 5 of culture in control culture medium (control) or in the presence of FGF2 (50 ng/ml) (+FGF2) or FGF9 (50 ng/ml) (+FGF9). After a few days in culture, SCs formed a monolayer, and no evident changes could be visualized by phase-contrast microscopy between the control SC monolayer and the FGF2- and FGF9-treated SCs. Bars, 100 μm for SCs. In contrast, the cellular appearance of PCs was highly dependent on the culture conditions, i.e. the presence of FGF2 or FGF9. FCS (10%) (+FCS) and bu2cAMP (1 mM) (+bu2cAMP) were also tested. Bars, 60 μm for PCs.
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
Effects of FGF2 and FGF9 on (A) PC number and (B) on the [3H]thymidine labeling index. (A) PCs were plated at 10 000 cells per 96 multiwell plates, and FGF2 and FGF9 were added at increasing concentrations ranging from 3 to 100 ng/ml on the next day. Other cells received 1 mM bu2cAMP. After a 4-day exposure, cells were quantitated using the celltiter 96 R Aqueous non-radioactive cell proliferation assay according to the manufacturer’s intructions. (B) DNA synthesis was measured as [3H]thymidine incorporation into TCA-insoluble material. PCs were plated at 10 000 cells in 96 multiwell plates and stimulated with FGF2 (1.5, 15, 30 or 50 ng/ml) and/or FGF9 (50 ng/ml) for 24 h in the presence of [3H]thymidine. Control cells received only [3H]thymidine. For (A) and (B) three experiments were performed and a representative experiment is shown. The results represent the means±s.e. of four dishes. P<0.05 and not significant (n.s.) as compared with control values.
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
RT-PCR screening of FGF9 and various FGFRs in 20-day-old whole rat testis (wT), in SCs or in PCs recovered from 20-day-old rat testis and cultured for 5 days. The set of primers is listed in Table 1. Total RNA was also isolated from rat fetal testes (fT) and from kidney. RT-PCR studies were also conducted using primers directed against 18S to ensure that equal amounts of material were used. A DNA ladder was included in each gel for accurate determination of the size of the PCR band (shown on the left).
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
Morphological changes elicited in SC–PC cocultures exposed for 4 days to FCS (10%), bu2cAMP (1 mM), PDGF, FGF9 or FGF2 (all at 50 ng/ml). SCs were placed at the top of the PCs at the optimal ratio of SC:PC 3:1. On the next day (day 1 of culture), fresh culture medium was added together with FCS, bu2cAMP, PDGF, FGF2 or FGF9, at the concentrations indicated. Control cocultures received only fresh culture media. Cocultures were examined by phase-contrast microscopy, and morphological changes were photographed on day 5 of culture. Asterisks indicate zones of reorganization containing the aggregates of SCs. Bar, 100 μm.
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
Explants from 13.5 dpc male urogenital ridges were cultured in control culture medium or in culture medium supplemented with FGF9 or FGF2 (50 ng/ml). One day later, explants were prepared for immunohistochemistry, and keratin 8, AMH/MIS and FN were revealed using indirect immunoperoxidase on serial sections. SCs are positive for keratin 8 and AMH/MIS, i.e. they appear dark grey whereas mesenchymal cells are unlabeled for keratin 8 and AMH/MIS. The FN immunostaining is present in the mesenchymal cells (they appear dark grey) but absent within the cords which are white. Note that the gonadal territory marked by a dotted line is more developed in the FGF2-treated explants than in the FGF9-treated explants. These latter contained also more cords than the explants cultured in control conditions. Bar, 100 μm.
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
(A) Gelatinase and (B) plasminogen activity in SC cultures and SC–PC cocultures exposed for 4 days to FGF2, FGF9, PDGF (all at 50 ng/ml) or bu2cAMP (1 mM). Control, untreated cultures. (A) Culture and coculture media concentrated and processed for gelatin zymography with 20 (culture) and 10 (coculture) μg proteins loaded onto SDS-PAGE gel. White lytic bands correspond to MMP-9 and MMP-2, as indicated on the right. (B) Cells scraped from the culture and coculture dishes and lysed. Proteins levels were measured and 20 μg proteins were loaded onto SDS-PAGE gel. White lytic bands correspond to tPA and uPA, as indicated on the right. Experiments were repeated three times and a representative experiment is shown in (A) and in (B).
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
Western blot analysis of PAI-1, TIMP-1, TIMP-2, TIMP-3 and β-actin. Cell culture media (15 μg proteins) were used for the analysis of PAI-1 and TIMPs-1 and -2. Cell lysates (20 μg proteins) were used for the analysis of TIMP-3 and β-actin to ensure equal loading. The size of the bands for each inhibitor and β-actin is indicated on the right. Experiments were repeated three times and a representative experiment is shown.
Citation: Journal of Endocrinology 187, 1; 10.1677/joe.1.06146
We are indebted to Drs B Vigier and E B Lane for the gift of the AMH/MIS and keratin 8 (LE 41) antibodies respectively. We thank Dr M G Forest for her contribution in carefully reading the manuscript. This work was supported by the Ministère de l’Aménagement du Territoire et de l’Environnement (MATE AC014 G to B L M-B). R E R is funded by Agence Française de Sécurité Sanitaire des Aliments (AFSSA). F O and S M are funded by Ministère de la Recherche et de la Technologie (MRT) and Organon (Azko Nobel, France) respectively. The authors declare that there is no conflict of interest that would prejudice the impartiaility of this scientific work.
Footnotes
R El Ramy is now at AFSSA-LERMVD, Toxicologie Alimentaire, Fougères BP 90203, Javene 35133, France
References
Ailenberg M, Tung PS & Fritz IB 1990 Transforming growth factor beta elicits changes and increases contractility of testicular peritubular cells. Biology of Reproduction 42 499–509.
Allen BL & Rapraeger AC 2003 Spatial and temporal expression of heparan sulfate in mouse development regulates FGF and FGF receptor assembly. Journal of Cell Biology 11 637–648.
Brennan J, Tilmann C & Capel B 2003 Pdgfr-a mediates testis cord organization and fetal Leydig cell development in the XY gonad. Genes and Development 17 800–810.
Buehr M, Gu S & McLaren A 1993 Mesonephric contribution to testis differentiation in the fetal mouse. Development 117 273–281.
Cancilla B & Risbridger GP 1998 Differential localization of fibroblast growth factor receptor-1, -2, -3, and -4 in fetal, immature, and adult rat testes. Biology of Reproduction 58 1138–1145.
Cancilla B, Davies A, Ford-Perris M & Risbridger GP 2000 Discrete cell- and stage-specific localisation of fibroblast growth factors and receptor expression during testis development. Journal of Endocrinology 164 149–159.
Colvin JS, Bohne BA, Harding GW, McEwen DG & Ornitz DM 1996 Skeletal overgrowth and deafness in mice lacking fibroblast growth factor receptor 3. Nature Genetics 12 390–397.
Colvin JS, Green RP, Schmahl J, Capel B & Ornitz DM 2001 Male-to-female sex reversal in mice lacking fibroblast growth factor 9. Cell 104 875–879.
Cupp AS, Kim GH & Skinner MK 2000 Expression and action of neurotropin-3 and nerve growth factor in embryonic and early postnatal rat testis development. Biology of Reproduction 63 1617–1628.
Cupp AS, Tessarollo L & Skinner MK 2002 Testis developmental phenotypes in neurotropin receptor trkA and trkC null mutations: role in formation of seminiferous cords and germ cell survival. Biology of Reproduction 66 1838–1845.
Cupp AS, Uzumcu M & Skinner MK 2003 Chemotactic role of neurotropin 3 in the embryonic testis that facilitates male sex determination. Biology of Reproduction 68 2033–2037.
Czekay RP, Aertgeerts K, Curriden SA & Loskutoff DJ 2003 Plasminogen activator inhibitor-1 detaches cells from extracellular matrices by inactivating integrins. Journal of Cell Biology 160 781–791.
Dym M 1994 Basement membrane regulation of Sertoli cells. Endocrine Reviews 15 102–115.
Friedl A, Chang Z, Tiemey A & Rapraeger AC 1997 Differential binding of fibroblast growth factor-2 and -7 to basement membrane heparan sulfate: comparison of normal and abnormal human tissues. American Journal of Pathology 150 1443–1455.
Fritz IB, Tung PS & Ailenberg M 1993 Proteases and antiproteases in the seminiferous tubules. In The Sertoli Cell, pp 217–235. Eds LD Russell & MD Griswold. Clearwater, Florida: Cache River Press.
Galdieri M, Ziparo E, Palombi F, Russo MA & Stefanini MJ 1981 Pure Sertoli cell cultures: a new model for the study of somatic-germ cell interactions. Journal of Andrology 5 249–259.
Goldfarb M 1996 Functions of fibroblast growth factors in vertebrate development. Cytokine and Growth Factor Review 7 311–325.
Gonzales AM, Buscaglia M, Ong M & Baird A 1990 Distribution of basic fibroblast growth factor in the 18-day old rat fetus: localization in the basement membranes of diverse tissues. Journal of Cell Biology 110 753–765.
Gronning LM, Wang JE, Ree AH, Haugen TB, Tasken K & Tasken KA 2001 Regulation of tissue inhibitor of metalloproteinases-1 in rat Sertoli cells: induction by germ cell residual bodies, interleukin-1 alpha and second messengers. Biology of Reproduction 62 1040–1046.
Guyot R, Magre S, Leduque P & Le Magueresse-Battistoni B 2003 Differential expression of tissue inhibitor of metalloproteinases type 1 (TIMP-1) during mouse gonad development. Developmental Dynamics 227 357–366.
Han IS, Sylvester SR, Kim KH, Schelling ME, Venkateswaran S, Blanckaert VD, McGuiness MP & Griswold MD 1993 Basic fibroblast growth factor is a testicular germ cell product which may regulate Sertoli cell function. Molecular Endocrinology 7 889–897.
Hoeben E, Van AI, Swinnen JV, Opdenakker G & Verhoeven G 1996 Gelatinase A secretion and its control in peritubular and Sertoli cell cutures: effects of hormones, second messengers and inducers of cytokine production. Molecular and Cellular Endocrinology 118 37–46.
Hoeben E, Swinnen JV, Heyns W & Verhoeven G 1999 Heregulins or neu differentiation factors and the interactions between peritubular myoid cells and Sertoli cells. Endocrinology 140 2216–2223.
Itoh N & Ornitz DM 2004 Evolution of the Fgf and Fgfr gene families. Trends in Genetics 20 563–569.
Jaworski DM & Fager N 2000 Regulation of tissue inhibitor of metalloproteinase-3 (Timp-3) mRNA expression during rat CNS development. Journal of Neurosciences Research 61 396–408.
Le Magueresse-Battistoni B, Wolff J, Morera AM & Benahmed M 1994 Fibroblast growth factor receptor type 1 expression during rat testicular development and its regulation in cultured Sertoli cells. Endocrinology 135 2404–2411.
Le Magueresse-Battistoni B, Pernod G, Kolodié L & Benahmed M 1996 Plasminogen activator inhibitor-1 regulation in cultured rat peritubular cells by basic fibroblast growth factor and transforming growth factor alpha. Endocrinology 137 4243–4249.
Le Magueresse-Battistoni B, Pernod G, Sigillo F, Kolodié L & Benahmed M 1998 Plasminogen activator inhibitor-1 is expressed in cultured rat Sertoli cells. Biology of Reproduction 59 591–598.
Longin J & Le Magueresse-Battistoni B 2002 Evidence that MMP-2 and TIMP-2 are at play in the FSH-induced changes in Sertoli cells. Molecular and Cellular Endocrinology 189 25–35.
Longin J, Guillaumot P, Chauvin MA, Morera AM & Le Magueresse-Battistoni B 2001 MT1-MMP in rat testicular development and the control of Sertoli cell proMMP-2 activation. Journal of Cell Sciences 114 2125–2134.
Magre S & Jost A 1984 Dissociation between testicular organogenesis and endocrine cytodifferentiation of Sertoli cells. PNAS 81 7831–7834.
Mazaud S, Guigon CJ, Lozach A, Coudouel N, Forest MG, Coffigny H & Magre S 2002 Establishment of the reproductive function and transient fertility of female rats lacking primordial follicle stock after fetal gamma-irradiation. Endocrinology 143 4775–4787.
Miller DL, Ortega S, Bashayan O, Basch R & Basilico C 2000 Compensation by fibrobast growth factor 1 (FGF1) does not account for the mild phenotypic defects observed in FGF2 null-mice. Molecular and Cellular Biology 20 2260–2268.
Mullaney BP & Skinner MK 1992 Basic fibroblast growth factor (bFGF) gene expression and protein production during pubertal development of the seminiferous tubule: follicle-stimulating-hormone-induced Sertoli cell bFGF expression. Endocrinology 131 2928–2934.
Odet F, Guyot R, Leduque P & Le Magueresse-Battistoni B 2004 Evidence for similar expression of protein C inhibitor and the urokinase-type plasminogen activator system during mouse testis development. Endocrinology 145 1481–1489.
Park PW, Reizes O & Bernfield M 2000 Cell surface heparan sulfate proteoglycans: selective regulators of ligand-receptor encounters. Journal of Biological Chemistry 275 29923–29926.
Puglianiello A, Campagnolo L, Farini D, Cipollone D, Russo MA & Siracusa G 2004 Expression and role of PDGF-BB and PDGFR-beta during testis morphogenesis in the mouse embryo. Journal of Cell Science 117 1151–1160.
Ricci G, Catizone A & Galdieri M 2002 Pleiotropic activity of hepatocyte growth factor during embryonic mouse testis development. Mechanisms of Development 118 19–28.
Ricci G, Catizone A & Galdieri M 2004 Embryonic mouse testis development: role of platelet derived growth factor (PDGF-BB). Journal of Cellular Physiology 200 458–467.
Richardson LR, Kleinman HK & Dym M 1995 Basement membrane gene expression by Sertoli and peritubular myoid cells in vitro in the rat. Biology of Reproduction 52 320–330.
Schmahl J, Kim Y, Colvin JS, Ornitz DM & Capel B 2004 Fgf9 induces proliferation and nuclear localization of FGFR2 in Sertoli precursors during male sex determination. Development 131 3627–3636.
Siu MKY, Lee WM & Cheng CY 2003 The interplay of collagen IV, tumor necrosis factor-a, gelatinase B (matrix metalloprotease-9), and tissue inhibitor of metalloproteases-1 in the basal lamina regulates Sertoli cell-tight junction dynamics in the rat testis. Endocrinology 144 371–387.
Skinner MK & Moses HL 1989 Transforming growth factor beta gene expression and action in the seminiferous tubule: peritubular cell–Sertoli cell interactions. Molecular Endocrinology 4 625–634.
Skinner MK, Takacs K & Coffey RJ 1989 Transforming growth factor-alpha gene expression and action in the seminiferous tubule: peritubular cell–Sertoli cell interactions. Endocrinology 124 845–854.
Soriano P 1994 Abnormal kidney development and hematological disorders in PDGF beta-receptor mutant mice. Genes and Development 8 1888–1896.
Swain A & Lovell-Badge R 1999 Mammalian sex determination: a molecular drama. Genes and Development 13 755–767.
Tolli R, Monaco L, Di Bonito P & Canipari R 1995 Hormonal regulation of urokinase-and tissue-type plasminogen activator in rat Sertoli cells. Biology of Reproduction 53 193–200.
Tung PS & Fritz IB 1980 Interactions of Sertoli cells and myoid cells in vitro.Biology of Reproduction 23 207–217.
Tung PS & Fritz IB 1987 Morphogenetic restructuring and formation of basement membranes by Sertoli cells and testis peritubular cells in co-culture: inhibition of the morphogenetic cascade by cyclic AMP derivatives and by blocking direct cell contact. Developmental Biology 120 139–153.
Ueno N, Baird A, Esch F, Ling N & Guillemin R 1987 Isolation and partial characterization of basic fibroblast growth factor from bovine testis. Molecular and Cellular Endocrinology 49 189–194.
Ulisse S, Farina AR, Piersanti D, Tiberio A, Cappabianca L, D’Orazi G, Jannini EA, Malykh O, Stetler-Stevenson WG, D’Armiento M, Gulino A & Mackay AR 1994 Follicle-stimulating hormone increases the expression of tissue inhibitors of metalloproteinases TIMP-1 and TIMP-2 and induces TIMP-1 AP-1 site binding complex(es) in prepubertal Sertoli cells. Endocrinology 135 2479–2487.
Uzumcu M, Dirks KA & Skinner MK 2002 Inhibition of platelet-derived growth factor actions in the embryonic testis influences normal cord development and morphology. Biology of Reproduction 66 745–753.
Van Der Wee K & Hofmann MC 1999 An in vitro tubule assay identifies HGF as a morphogen for the formation of seminiferous tubules in the postnatal mouse testis. Experimental Cell Research 252 175–185.
Van Dissel-Emiliani FM, De Boer-Brouwer M & De Rooij DG 1996 Effect of fibroblast growth factor-2 on Sertoli cells and gonocytes in coculture during the perinatal period. Endocrinology 137 647–654.
Vu TH & Werb Z 2000 Matrix metalloproteinases: effectors of development and normal physiology. Genes and Development 14 2123–2133.
Weinstein M, Xu X, Ohyama K & Deng CX 1998 FGFR-3 and FGFR-4 function cooperatively to direct alveogenesis in the murine lung. Development 125 3615–3623.
Woessner JF Jr 2001 That impish TIMP: the tissue inhibitor of metalloproteinases-3. Journal of Clinical Investigation 108 799–800.
