Gap junctions are conglomerates of intercellular channels formed by a family of homologous proteins termed connexins. They allow transfer of ions and second messengers such as Ca2+ (Saéz et al., 1989) and cAMP (Murray and Fletcher, 1984; Tsien and Wiengart, 1974) between adjacent cells. Gap junctions have been implicated in the control of development (Juneja et al., 1999) and regulation of hormone release in secretory glands (Chanson et al., 1998; Munari-Silem and Rousset, 1996; Oyoyo et al., 1997; Pérez-Armendariz et al., 1995).
The testis is the gland required for masculinization and male reproduction. In mice, testicular weight increases 60-fold from birth to adulthood as a result of the development of two cellular compartments: the tubule and the interstitium (Vergoween et al., 1993). The interstitium contains seven different cell types: mesenchymal, peritubular or myoid, macrophages, endothelial, perivascular, lymphatic endothelial and Leydig cells (LC) (Vergoween et al., 1991). Leydig cells secrete testosterone, the main testicular hormone required for masculinization. In the prepubertal animal, mesenchymal cells represent the largest interstitial cell population. With the onset of puberty, their relative number decreases whereas LC number increases to become the most abundant interstitial cell type in the mature animal (Hardy et al., 1989; Vergoween et al., 1993).
Leydig cells have gap junctions (reviewed, Russell, 1996) and are extensively dye-coupled, both in vivo (Risley et al., 1992) and in freshly dispersed preparations (Kawa, 1987; Pérez-Armendariz et al., 1994; Varanda and de Carvalho, 1994). Cx43 is the only connexin described in the interstitium so far. It has been localized in peritubular cells and in rodent (Batias et al., 1999; Pérez Armendariz et al., 1994; Pelletier, 1996; Risley et al., 1992) and human (Steger et al., 1999) Leydig cells. In mouse, functional studies support the idea that Cx43 is the main connexin communicating between adult Leydig cells: a) biophysical properties of macroscopic and microscopic currents recorded from double whole cell voltage clamp of isolated pairs of LC are similar to those recorded from deficiently connected cell pairs transfected with Cx43 (Pérez-Armendariz et al., 1994); b) decrease in Cx43 expression in LC clumps cultured for 36 hr occurs associated with an 80% decay in dye coupling (Pérez-Armendariz et al., 1996); and c) time-dependent decay in Cx43 and dye coupling are prevented by maintaining LC cultures in the presence of luteinizing (LH) or chorionic gonadotrophic (CG) hormones or dibutyryl-cAMP (Pérez-Armendariz et al., 1996).
LH and CG are the main physiological regulators of LC function. They activate the same receptor, stimulate adenylate cyclase and produce an increase in intracellular cAMP levels, [cAMP]i. Elevation of [cAMP]i importantly enhances steroidogenesis and testosterone secretion (reviewed by Segaloff and Ascoli, 1993). Because serum LH and testosterone levels change during postnatal development (Okamoto et al., 1982; Ogasawara et al., 1984), it is possible that Cx43 expression in LC is developmentally regulated. In this respect, it has been described that Cx43-immunofluorescent (IF) labeling in interstitial cells with lipid droplets, presumably LC, is reduced in immature compared with mature testes in guinea pig (Pelletier, 1996) and rat (Risley et al., 1992). Because Cx43 may be expressed in the various seven interstitial cells there is no evidence that changes in interstitial Cx43 labeling occur exclusively in Leydig cells.
In rodent tubular cells, mRNA for 12 different connexins has been identified (see Batias et al., 2000), eight in Sertoli and nine in germinal cells (Risley, 2000). A specific spatial distribution has been found for Cx26, Cx31, Cx32, Cx33 and Cx43 using IF studies in testis from adult animals (Batias et al., 1999; Mok et al., 1999; Pérez-Armendariz et al., 1994; Pelletier, 1996; Risley et al., 1992; Tan et al., 1996). Several lines of evidences support the idea that Cx43 is required for adequate control of fertility. It has recently been shown that alterations in spermatogenesis, in both human (Steger et al., 1999) and mouse (Batias et al., 1999), are associated with decreased Cx43 expression in adult Sertoli cells. In addition, the number of germinal cells in fetal testis from Cx43 knock-out mice is significantly reduced (Juneja et al., 1999). Moreover, substitution of the coding region of the Cx43 gene by coding sequences of Cx40 or Cx32 is not sufficient to prevent reduction in germinal cells, and still results in infertile males due to the complete absence of secondary spermatogonia (Plum et al., 2000). In mouse, spermatogenesis starts shortly after birth, however, to date there is no information about Cx43 expression in testicular cells at this developmental stage.
In the present study, we investigated whether Cx43 is developmentally regulated in LC by comparing the spatial distribution and temporal correlation of Cx43 with 3βHSD expression, as well as with intratesticular testosterone content. Additionally, we investigated if Cx43 is expressed and regulated in Sertoli and germinal cells at early postnatal developmental times.
CD1 mice were maintained and handled in accordance with the International Guiding Principles for Biomedical Research Involving Animals, as promulgated by the Society for the Study of Reproduction. Female breeders showing a vaginal plug the day after mating were considered to be at day 0.5 of pregnancy. To record the time of partum, females were inspected at 6 hr intervals (9, 15, 21 hr) at day 18 after mating. To prevent large variation in mouse weight, litter size was adjusted to 8 to 9 pups. Animals were sacrificed at various days after birth (0, 3, 5, 7, 14, 21, 24, 26, 28, 35, 60) at the mean delivery time (12, 18 hr) with a variation of ±1 hr at most. Animals were anesthetized with CO2 and then killed by decapitation. Testes were micro-dissected, embedded in TissueTek, immersed in cryo-preservative and frozen in liquid nitrogen.
Two different rabbit polyclonal antisera against unique peptide sequences of Cx43 were used. An affinity-purified antibody was obtained from Zymed (San Francisco, CA; Cat. No. 71-0700). A serum, F100, was generated in our laboratory according to the procedure of Traub et al. (1994) against a highly purified synthetic peptide corresponding to aa. 346–363 (Biosynthesis Inc., Lewisville, TX) of mouse Cx43 carboxy terminus [Cx43(346–363)]. A BLAST search revealed that this peptide is homologous with Cx43 sequences from other species, but not with any other protein or cloned connexin. One monoclonal antibody, MAB3068, against a synthetic peptide corresponding to another specific sequence (aa 252–270) of Cx43 (Chemicon, Temecula, CA) was also used. Polyclonal antibodies against Mullerian inhibitory hormone (MIH) (Vigier et al., 1985) and steroidogenic enzyme 3βHSD (Dupont et al., 1993; Mason et al., 1993) were kindly donated by Drs. N. Josso, V. Luu-The, and I. Mason, respectively. A monoclonal IgM antibody against a mouse germ cell nuclear antigen (GCNA1) was a gift from Dr. G.C. Enders (Enders and May, 1994).
Immunocytochemical Localization of Cx43, 3βHSD, MIH and GCNA1 Antigen
Cryosections of 10–12 μm were processed as previously described (Pérez-Armendariz et al., 1994). Briefly, sections were fixed in 70% ethanol at −20°C; rinsed with phosphate buffer saline (PBS), pH 7.3, 1 mM CaCl2, and preincubated in 2% albumin IgG-free, 0.1% Triton X-100 in PBS. For single labeling, tissue sections were incubated with antibody 71-0700 (1:100), F100 (1:250) or anti-3βHSD (1:100) for 1 hr at room temperature. After washing in PBS, tissue sections were incubated with FITC-goat anti-rabbit IgG for 1 hr at RT, washed and mounted using a fluorescence protector medium (Vecta Shield, Burlingame, CA), observed under phase contrast and fluorescence microscopy (Olympus-1X70) and photographed. For double labeling experiments, tissue sections were incubated overnight at 4°C with polyclonal antibodies against 3βHSD or MIH (1:100) or with undiluted supernatant containing monoclonal antibody (IgM) against GCNA1. Then, sections were washed and incubated for 1 hr with FITC-labeled goat anti-rabbit IgG or anti-mouse IgM (Zymed), respectively. Sections were washed again and incubated for 2 hr with MAB3068 (IgG) against Cx43 (1:100). After washing, tissue sections were incubated with rhodamine-conjugated goat anti-mouse IgG (Pierce, Rockford, IL) for 1 hr. The same field of the section was analyzed sequentially under fluorescein- and rhodamine-excitation filters. Monochromatic digitized images at each excitation wavelength were acquired using a Hammatsu digital camera (ORCA) and Metamorph software (Universal Imaging, West Chester, PA). Using the same software, images were superposed and displayed in pseudocolor. Green was used to represent fluorescence emitted by cells positive for 3βHSD or MIH. Blue and red were used to code fluorescence of GNCA1 and Cx43 labeling, respectively. Micrographs of images shown for each developmental age are representative of observations obtained from at least six different animals.
Freshly dissected testis, heart and transfected HeLa cells were homogenized in cold buffer containing 250 mM NaCl, protease inhibitors (200 μg/ml soybean trypsin inhibitor, 1 mg/ml benzamidine, 1 mg/ml aminocaproic acid and 2 mM PMSF) and phosphatase inhibitors (20 mM Na4P2O7 and 100 mM NaF). Immunoblots were performed as described previously (Sáez et al., 1997). Briefly, 50 μg (heart homogenates) or 150 μg (testis and HeLa cells) of protein were separated in 8% SDS-PAGE and electrotransfer. After incubation with anti-Cx43 polyclonal antibodies 71-0700 (1:500) and F100 (1:100) or monoclonal MAB3680 (1:500) blots were reacted with alkaline phosphatase-conjugated goat anti-rabbit IgG or anti-mouse IgG, respectively, and developed with substrate (BCIP/NBT tablets, Sigma). The reactivity of these antibodies was compared with that of a previously characterized anti-Cx43 antibody described elsewhere (Sáez et al., 1997). Cx40 and Cx43-transfected HeLa cells (Elfgang et al., 1995) were kindly provided by Dr. Klaus Willecke, Institute für Genetik, Universität Bonn, Germany.
Steroidal compounds were extracted from homogenized gonadal tissue from 10–15 individual mice of the same age, with diethyl ether. After solvent evaporation to dryness, extracts were resuspended in PBS and testosterone was determined in duplicate aliquots by a liquid phase radioimmunoassay (Sufi et al., 1986). Sensitivity of the assay was 6.8 pg/ml with a coefficient of variation of 3.6%. Data analysis was performed using a four parameter logistic transformation of the standard curve using the statistical package SigmaStat (Jandel Scientific, Sausalito, CA).
Specificity of Antibodies Against Cx43
Two polyclonal rabbit antibodies, 71-0700 and F100, and a monoclonal antibody, MAB3068, directed against unique amino-acid sequences of Cx43 were used. Figure 1 shows that by immunoblotting, antibody 71-0700 (A), MAB3068 (B) and antibody F100 (C) revealed bands of 40–45 kDa, in total homogenates from heart, adult testis, and in HeLa cells transfected with Cx43, but not in HeLa cells transfected with Cx40 or with the expression vector alone. These bands correspond to nonphosphorylated (NP) and several phosphorylated (P) forms of Cx43 (Sáez et al., 1997). The different banding patterns obtained with the various antibodies reflects the different degree of phosphorylation of Cx43 expressed in the studied cells and tissues. The specificity of antibodies to Cx43 was further studied by immunofluorescence (IF) in testicular sections at different ages. Figure 2 shows micrographs of sections from adult testis incubated with antibodies 71-0700 (A) and F100 (B). Both polyclonal antibodies gave a similar Cx43-labeling pattern in the interstitium as well as in the tubule. This distribution was also observed by staining with monoclonal antibody MAB3068 (Fig. 5). No labeling was detected in testicular sections incubated with the pre-immune serum F100 (Fig. 2C) or with FITC-secondary antibodies alone (Fig 2D). Furthermore, the reactivity of antiserum F100 was eliminated by adsorption with the antigenic peptide Cx43 (346–363) (Fig. 2E,F). None of the anti-Cx43 antibodies used in immunocytochemical studies labeled liver sections (not shown).
Expression of 3β-Hydroxy-5-Ene Steroid Dehydrogenase/5–4 Isomerase (3βHSD) During Postnatal Development Is Biphasic
The enzymatic complex 3βHSD is an essential component for the biosynthesis of all biologically active steroids, including androgens. In testicular sections of newborn animals, cells expressing 3βHSD were organized in round cell clumps (RCC) of large (>10 cells), medium (4–10) and small size (<4) at several intertubular areas (Fig. 3A). At 3 days postpartum (dpp), the size and number of 3βHSD-positive RCC decreased, particularly in the central part of the testis. In addition, elongated 3βHSD-positive cells were detected in some intertubular areas (Fig. 3B). These changes became more prominent at 7 dpp, when small cell clumps and mostly isolated elongated cells were found (Fig. 3C). By day 14, the number of intertubular areas that showed 3βHSD-positive cells increased (Fig. 3D). At 21 dpp, small RCC were again detected at most intertubular areas (Fig. 3E). Thereafter the size of 3βHSD positive RCC increased progressively at 26 (Fig. 3F) and 28 dpp (Fig. 3G), to reach a maximum at 35 (Fig. 3H) and 60 dpp (Fig. 3I). Most 3βHSD-positive cells exhibited lipid droplets under phase contrast microscopy (Fig. 3). The intensity of staining in 3βHSD cell clumps from different intertubular areas was heterogeneous within the same section (Fig. 3). Usually, the larger the cell clumps the more lipid droplets and more intense the 3βHSD reaction. Slight differences in intensity of the reactions were also found between cells within the same cluster. Cells in the clump that were more strongly labeled also exhibited more lipid droplets (not shown).
Biphasic Expression of Cx43 in the Interstitial Space Results From its Modulation in Leydig Cells
At birth, Cx43 was always detected in small, medium and large RCC, localized in the interstitial space (Fig. 4A,B, arrowheads). The size of Cx43-positive RCC and their spatial distribution were similar to that found for 3βHSD-positive cells (see above). Round cell aggregates of various sizes were unambiguously identified as LC in double-labeled sections stained with antibodies to Cx43 and 3βHSD. At birth, Cx43 IF-dots were found in the cytoplasm but mostly at cell membrane appositions (Fig. 5A). Interstitial cells surrounding 3βHSD-positive cells, very likely fibroblast and mesenchymal cells, were not labeled by Cx43 antibodies, whereas tubular cells were strongly labeled for Cx43 (Fig. 5A,C). From 3 dpp on, Cx43 labeling in the interstitium decreased progressively. Cx43 IF-dots were abundant in some cells, but absent in others from the same aggregate (not shown). At day 7, Cx43 was not found in most intertubular areas (Fig. 4C,D) and remained negative by 14 dpp (Fig. 4E,F). Double-labeling experiments showed that in addition to reduction in size of 3βHSD-positive cell clumps there was an important decrease in Cx43 expression at both ages (Fig. 5B,C). From day 21, interstitial Cx43-IF labeling again increased. Punctate IF dots were found at several intertubular areas (Fig. 4G,H arrowheads). The size of cell clumps positive for Cx43 continued to increase at 26 (Fig. 4I,J, arrowheads) and 28 dpp (Fig. 4K,L, arrowheads) to reach a maximum at 35 (Fig. 4M,N) and 60 dpp (Fig. 4O,P). In the adult animal, Cx43-IF-dots were localized mainly at cell membrane appositions (Fig. 5D). As the RCC size increased, lipid droplets and Cx43-IF dots were more abundant.
Developmental Changes in Intratesticular Testosterone Concentration
Figure 6 shows that intratesticular testosterone content also changed in a biphasic manner during postnatal development. The testicular testosterone content showed a marked reduction from 14 through 21 dpp when compared with newborn animals. At birth, the mean concentration of testosterone was 108 pg/mg of protein. This concentration subsequently decreased to 42 pg/mg of protein at 7 dpp and fell sharply to 3.7 and 4.2 pg/mg of protein at 14 and 21 dpp, respectively. From day 25 (20.1 pg/mg) onwards an increase in testosterone content was detected that became prominent at 35 (78.8 pg/mg) and 60 dpp (301.5 pg/mg). The testosterone content of samples collected on days 14, 21, 25, 26, was statistically significantly different from that at 0, 28, 35 and 60 dpp.
Connexin 43 Is Expressed in Cells of the Seminiferous Cords at Birth and Is Modulated During the Neonatal and Prepuberal Period
The seminiferous cord (SC) is formed by Sertoli and germinal cells. Mouse germ cell nuclear antigen, GNCA1, a common marker of the germ cell lineage, is known to be abundant in nuclei of prespermatogonia (gonocytes), spermatogonia and early spermatocytes (Enders and May, 1994). In mouse, spermatogenesis starts a few hours after birth (Vergoween et. al., 1993). We show that Cx43 is prominently expressed in SC cells from newborn mouse. Labeling of SC in the same section was homogeneous (Fig. 4A,B). In double-labeled sections Sertoli cells, identified by their positive staining with anti-Mullerian inhibitory hormone antibody, were found at the periphery of the SC (Fig. 7A). Abundant Cx43-IF dots were localized at cell membrane appositions between adjacent Sertoli cells as well as in their cytoplasmic projections that surround central gonocytes (Fig. 7A arrowhead). Gonocytes, identified with antibodies against GCNA1, were localized in the center of the SC (Fig. 7B, arrowhead). Between gonocytes Cx43 IF-dots were frequently found. Cx43 at this localization may result from its expression between gonocytes or between gonocytes and Sertoli cells.
During the first week of life, gonocytes migrate from the center to the periphery of the SC. At this latter region they proliferate (Vergoween et al., 1991) and by the end of the week most of them have differentiated to spermatogonia (Nebel et al., 1961). Sertoli cells continue their proliferation and change from a round to a droplet shape (Vergoween et al., 1991). Here we found that associated with these events, expression of Cx43 changed in SC cells. In double-labeled sections at 3 dpp, Cx43 increased in the center of the SC, localized in Sertoli cells that were not labeled with the anti-GCNA1 antibody (Fig. 7C). In the periphery, Cx43 was localized at cell borders between Sertoli cells and GCNA1-positive cells (Fig. 7C, arrowhead). At 7 dpp, Cx43 reactivity remained intense in the center of the SC and became almost undetectable at the periphery (Fig. 4C,D). In this latter region abundant and closely packed GCNA1-positive cells were found (Fig. 7D). Between adjacent GNCA1-positive cells, Cx43-IF labeling was localized occasionally (Fig. 7D arrow, and insert). In contrast, Cx43 IF-dots were frequently found between the luminal pole of GCNA1- positive cells and centrally localized Sertoli cell membranes (Fig. 7D arrowhead).
By 14 dpp, Sertoli cells stop proliferating and have differentiated to a polarized epithelium (Vergoween et al., 1991). Spermatogonia cells have differentiated up to their pachytene spermatocyte stage (Enders and May, 1994; Nebel et al., 1961). At this age, Cx43 expression changed again; labeling of SC was markedly heterogeneous (Fig. 4E,F). Three different patterns of Cx43 expression in the SC were found. In type one (T1), intense Cx43-IF labeling was localized in the center of the SC, and GNCA1-positive cells were localized in the periphery (Figs. 7E; 4E,F). In type two (T2), intense Cx43 reaction was localized in the middle of the SC, and GCNA1-positive cells were found in the center and at the periphery (Figs. 7E; 4E,F). In type three (T3), intense but more delineated Cx43-IF labeling was localized toward the basal region, and abundant GCNA1-positive cells were found in the center of the SC. In type II and type III SC, tiny Cx43 IF dots were also found between GCNA1-positive cells localized at the center (Fig. 7E; 4E,F). In the three different patterns of Cx43 expression in the SC, labeling was found at cell borders between negative and GNCA1 positive cells (Fig. 7, arrowheads).
From 21 dpp onwards, after the formation of the blood testis barrier, abundant delineated Cx43 was found at the basal third of Sertoli cells, within the area of localization of the junctional complex (Fig. 4G,H). From this time on, Cx43 intensity and distance toward the limiting border of the tubule varied between tubules from the same section (Fig. 4G–P). In the luminal compartment small Cx43 punctate staining was also found (Fig. 4G–P). Observations at these last dates are consistent with previous reports (Batias et al., 2000; Pelletier, 1996; Risley et al., 1992; Steger et al., 1999).
In these studies we have demonstrated that Cx43 expression in LC from mouse testis is regulated in an age- and functional-dependent manner from birth to adult life. In addition, it is clearly documented that Cx43 is expressed in Sertoli cells at birth and that it is also modulated during the neonatal and prepubertal period. Evidence suggesting that Cx43 forms gap junctions between Sertoli and germinal cells before and at the time of initiation of spermatogenesis is also provided.
Immunofluorescent staining of testis truly reflects expression of Cx43 because antibodies used for its detection were found to be highly specific. This was demonstrated by: a) similar labeling pattern in the interstitium and in the tubule at various developmental ages using three different antibodies directed against unique sequences of Cx43; b) polyclonal antibodies F100 and 71-0700 and monoclonal MAB3068 reacted with bands of 40–45 kDa in immunoblots of heart, adult testis and Cx43-transfected HeLa cells, but not in Cx40-transfectants. These bands correspond to phosphorylated and unphosphorylated forms of Cx43 (Sáez et al; 1997); c) no labeling was detected in liver sections with these antibodies; d) labeling of LC and tubular cells by antiserum F100 was eliminated by prior absorption with the immunogenic peptide; e) the pattern of Cx43 expression shown here by IF in adult animals is similar to that previously reported by others (see Introduction); and f) our description of the expression of Cx43 in Leydig, Sertoli and germinal cells in adult mouse testis is consistent with the recent localization of Cx43 gene transcript in these cell types by in situ RNA hybridization (Batias et al. 2000).
Biphasic fluctuation in interstitial Cx43 results from changes in Cx43 expression in LC because it was found to be temporally correlated with 3βHSD expression and intratesticular testosterone content along all postnatal development. Moreover, Cx43 was unambiguously localized in 3βHSD-positive cells by double labeling but not in other interstitial cell types that surround them. Luteinizing hormone may be the physiological trigger signal that controls Cx43 expression in LC during postnatal development. This possibility is suggested by the temporal correlation between fluctuations in testosterone content, 3βHSD and Cx43 expression found here, with the known biphasic changes in mouse LH serum levels during postnatal development (Berkowitz et al., 1979; Ogasawara et al., 1984). It is also suggested by the previously described modulation of intercellular communication and Cx43 expression induced by LH in cultured adult mouse Leydig cells (Pérez-Armendariz et al., 1996).
Results presented here suggest that expression of Cx43 may be participating in developmental processes required for adequate control of testosterone production and secretion. This is supported by the correlation between fluctuations in intratesticular testosterone content and Cx43 and 3βHSD expression and by the finding that modulation in Cx43 expression is an early event in Leydig cell developmental changes. Downregulation during neonatal and prepubertal life was found to occur in parallel with a decrease in 3βHSD expression, and Cx43 upregulation observed during puberty preceded both the main increase in 3βHSD-positive cells and testosterone release. One possible mechanism by which gap junctions may enhance testosterone release is by transfer of cAMP (Murray and Fletcher, 1984; Tsien and Weingart, 1974), induced by LH in some cells to LC expressing none or low numbers of LH receptor or steroidogenic enzymes levels. Transfer of cAMP may amplify the transcription of cAMP-dependent steroidogenic enzymes (Payne and Youngblood, 1995) as well as synchronize testosterone release between cells of the aggregate. Directed target disruption of Cx43 gene results in mice that die early after birth from heart failure (Reaume et al., 1995). Interestingly, knock-in mice in which the Cx40- or Cx32-coding region was substituted for Cx43 survive to adulthood but show failed spermatogenesis (Plum et al., 2000). Whether alterations in testosterone levels at this developmental stage contribute to the failure of differentiation of germinal cells in these mice remains to be determined.
This study provides evidence for the first time that in the seminiferous cord Cx43 is expressed between Sertoli and germinal cells before and during the first wave of spermatogenesis. Its expression in this site is indicated by Cx43-IF dots found in Sertoli cell cytoplasmic projections that surround germinal cells, as well as in cell borders between Sertoli and germinal cells at 0, 3, 7 and 14 days of age. It is also indicated by the occasional detection of Cx43 between adjacent GCNA1-positive cells. Thus, homotypic, heterotypic or heteromeric channels containing Cx43 (Bukauskas et al., 1995; Elenes et al., 1999; Elfgang et al., 1995, Oh et al., 1999) may communicate between these cell types. Cx43-mediated heterocellular contacts between Sertoli and germinal cells is consistent with recent findings of Cx43 gene transcript in germinal and Sertoli cells by in situ hybridization (Batias et al., 2000). It is also consistent with the description of dye transfer (Batias et al., 2000; Orth and Boehm, 1990) and gap junctions (Byers and Pelletier, 1991; McGinley et al., 1979) between Sertoli and germinal cells, and between adjacent germinal cells (Pelletier and Byers, 1992). Further studies using electron microscopic techniques will confirm whether Cx43-mediated heterocellular contacts exist between SC cells. Recently, failed maturation of remaining germinal cells has been found in knock-in Cx40 and Cx32 mice (Plum et al., 2000). Loss of Cx43 gap junctions between Sertoli and germinal cells may be a possible explanation for this.
Data presented here clearly show that Cx43 in Sertoli cells is expressed at birth and is modulated during the neonatal and prepuberal period. Gap junctions between neonatal Sertoli cells have also been documented (Nagano and Suzuki, 1976). Change of Cx43 spatial distribution in Sertoli cells from their apical region (3 dpp) to their basal domain (14 dpp) is correlated with the differentiation from a nonpolarized to a polarized epithelium (Vergoween et al., 1991). By the end of the second week post partum, the lumen of the tubule is formed (Flickinger, 1967) and gap junctions (Gilula et al., 1976; Nagano and Suzuki, 1976) and tight-junction protein ZO-1 (Byers et al., 1991) distribution changes from the apical to the basal pole of Sertoli cells. More recently, direct interaction (Giepmans and Moolenaar, 1998) and co-localization of the tight junction protein ZO-1 with Cx43 has been documented (Batias et al., 1999). Thus, it is likely that the heterogenous distribution of Cx43 between SC detected here at 14 dpp may reflect a redistribution of this connexin in Sertoli cells during the maturation of the junctional complex and formation of the hematotesticular barrier. These changes occur in parallel with the maturation and migration of germinal cells within the SC that may represent the first wave of spermatogenesis. Because Cx43 heterocellular contacts exist throughout early development, the possibility is raised that direct transfer of molecules from germinal to Sertoli cells mediated by this connexin may contribute to the regulation of Cx43 distribution at these stages in Sertoli cells.
We thank M. Roberto Chavira, Lourdes Boeck, and Gladys Garcés for their technical assistance. We thank Dr. F. Labrie, Dr. V. Luu-The, Dr. I. Mason, Dr. N. Josso and Dr. G.C. Enders for their gift of antibodies. We also thank Dr. O. Traub and Dr. E. Beyer for their advice in the generation of anti-connexin antibodies and Dr. K. Willecke for donation of HeLa cells transfected with Cx40 and Cx43. We also thank Dr. H. Rasgado-Flores and Dr. Robyn Hudson for critical reading of this manuscript.