|Home | About | Journals | Submit | Contact Us | Français|
In adult mammals such as rats, the blood-testis barrier (BTB) conferred by adjacent Sertoli cells in the seminiferous epithelium segregates post-meiotic germ cell development from the systemic circulation and is one of the tightest blood-tissue barriers. Yet it must “open” transiently at stage VIII of the epithelial cycle to accommodate the migration of preleptotene/leptotene spermatocytes. While this is a vital event of spermatogenesis, the mechanism(s) that regulates BTB dynamics is virtually unknown. Recent studies have suggested that transforming growth factor-β3 (TGF-β3) and tumor necrosis factor α (TNFα) secreted by Sertoli and germ cells into the microenvironment of the BTB are capable of inducing reversible BTB disruption in vivo, apparently by reducing the steady-state levels of occludin and zonula occludens-1 (ZO-1) at the BTB via the p38 mitogen activated protein (MAP) kinase signaling pathway. In this study, local administration of TGF-β3 (200 ng/testis) to the testis reversibly perturbed the BTB integrity in vivo. We next sought to delineate the mechanism by which these cytokines maintain the steady-state level of integral membrane proteins: occludin, junctional adhesion molecule-A (JAM-A) and N-cadherin at the BTB. Primary Sertoli cells cultured in vitro were shown to establish intact tight junctions and functional BTB within two days when assessed by transepithelial electrical resistance (TER) measurement across the cell epithelium. Sertoli cell integral membrane protein internalization at the BTB was assessed by biotinylation of cell surface proteins, to be followed by tracking the endocytosed/biotinylated proteins by using specific antibodies. Both TGF-β3 (3 ng/ml) and TNFα (10 ng/ml) were shown to significantly accelerate the kinetics of internalization of JAM-A, N-cadherin, and occludin versus controls. Treatment of cells with phenylarsine oxide (PAO) at 10 μM that blocks clathrin-mediated endocytosis was shown to inhibit the TGF-β3-induced protein internalization. This inhibition of TGF-β3-mediated protein endocytosis was further validated by silencing of clathrin. The specific effect of TGF-β3 on protein internalization was confirmed by RNAi using specific TGF-β receptor I (TβR1) siRNA duplexes. When TβR1 was knocked down, the TGF-β3-induced increase in the kinetics of JAM-A and occludin endocytosis was abolished, making them indistinguishable from controls, illustrating the specificity of the TGF-β3 effects on protein endocytosis. In summary, this report demonstrates for the first time that BTB dynamics are regulated by TGF-β3 and TNFα via an enhancement of protein endocytosis at the BTB.
Recent studies have shown that blood-testis barrier (BTB) dynamics during spermatogenesis are regulated, at least in part, by cytokines that determine the steady-state levels of integral membrane proteins at the BTB [for reviews, see (Lui and Cheng, 2007; Xia et al., 2005; Yan et al., 2007)]. For instance, it was shown that local administration of either TGF-β3 (Xia et al., 2006) or TNFα (Li et al., 2006) to testes of adult rats reversibly disrupted the BTB integrity when examined by dual-labeled immunofluorescence analysis using markers at the BTB, such as occludin, N-cadherin, JAM-A, and ZO-1. More important, local administration of TNFα to testes at concentration comparable to its endogenous level in the testis was shown to reversibly disrupt the BTB integrity. This was monitored by an in vivo functional assay by tracking the diffusion of a small fluorescent probe, such as fluorescein-5-isothiocyanate (FITC, Mr 389), from the systemic circulation to the adluminal compartment of the seminiferous epithelium following its administration at the jugular vein (Li et al., 2006). These results are consistent with recent observations in other epithelia, such as in small intestine and kidney, wherein cytokines are shown to be important regulators of tight junction (TJ) permeability barrier, such as in small intestine and kidney [for reviews, see (Walsh et al., 2000; Xia et al., 2005)]. These findings are important, since they illustrate that at the microenvironment of the BTB in the seminiferous epithelium, it is likely that cytokines, namely TNFα and TGF-β3 produced by Sertoli and/or germ cells (De et al., 1993; Skinner, 1993; Xia et al., 2006), contribute to the ‘restructuring’ and/or ‘opening’ of the BTB to facilitate the transit of preleptotene/leptotene spermatocytes across the BTB. This postulate is also supported by the observations that the expression of these cytokines was shown to be relatively high at or near the BTB during stage VIII of the epithelial cycle by immunohistochemistry (Lui et al., 2003a; Siu et al., 2003a; Xia et al., 2006). However, the mechanism by which cytokines induce a loss in the steady-state levels of integral membrane proteins at the BTB (e.g., occludin, JAM-A, N-cadherin) remains unknown.
Protein endocytosis plays a crucial role in regulating the steady-state levels of proteins at the cell junctions, which can be clathrin- or caveolin-dependent or via macropinocytosis [for reviews, see (Ivanov et al., 2005; Maxfield and McGraw, 2004)]. More importantly, it was postulated that the replacement of the apical ectoplasmic specialization [apical ES, a testis-specific actin-based atypical adherens junction (AJ) type] (Wong et al., 2008b) by apical tubulobulbar complex (apical TBC) to facilitate spermiation at stage VIII of the seminiferous epithelial cycle was mediated via protein internalization (Pelletier and Byers, 1992). Indeed, recent studies have supported this speculation wherein the adhesion domains of nectins 2 and 3 were found to be internalized as membrane vesicles near the TBC at spermiation (Guttman et al., 2004). In addition, recent studies have shown dynamin 2, a large GTPase known to be involved in protein endocytosis (McNiven et al., 2000; Sever et al., 2000) is structurally associated with integral membrane proteins at the BTB in rat testes (Lie et al., 2006). Furthermore, another isoform of dynamin, dynamin 3, was shown to be testis-specific (Kamitani et al., 2002). In light of these findings, we sought to examine if endocytosis is indeed taking place at the BTB of adult rat testes to regulate the steady-state levels of BTB- integral membrane proteins to facilitate the transit of preleptotene/leptotene spermatocytes during the epithelial cycle of spermatogenesis. This is the subject of this report.
Sprague-Dawley rats were purchased from Charles River Laboratories (Kingston, NY) and housed at the Rockefeller University Laboratory Animal Research Center with a 12 hr:12 hr light:dark cycle at 22 °C with access to rat chow and water ad libitum. The use of animals in this study was approved by the Rockefeller University Animal Care and Use Committee (Protocol Numbers: 03017 and 06018).
Primary Sertoli cell culture was prepared as previously described (Cheng et al., 1986; Xia et al., 2006). Freshly isolated Sertoli cells were cultured at 0.5 × 106 cells/cm2 on Matrigel (BD Biosciences)-coated dishes as described (Xia et al., 2006) in serum-free Ham’s F12 Nutrient Mixture and Dulbecco modified Eagle medium (F12/DMEM, 1:1, v/v, Sigma) containing HEPES (15 mM) and sodium bicarbonate (1.2 gm/L), supplemented with bovine insulin (10 μg/ml), human transferrin (5 μg/ml), epidermal growth factor (2.5 ng/ml), bacitracin (10 μg/ml), and gentamicin (20 μg/ml) as described (Mruk and Cheng, 1999). On the day of cell isolation, these cultures were designated as day 0. Matrigel was diluted 1:7 with F12/DMEM and dishes were coated 24 h prior to their use. These Sertoli cell cultures were incubated in a CO2 incubator at 35 °C in a humidified atmosphere with 5% air/95% CO2. About 36 hr thereafter, Sertoli cell cultures were subjected to a hypotonic treatment using 20 mM Tris, pH 7.4 at 22 °C for 2.5 min to lyse residual germ cells as described (Galdieri et al., 1981). Cultures were then washed twice in F12/DMEM. These Sertoli cells were contaminated with negligible germ cells when assessed by RT-PCR and/or immunoblotting using specific germ cell markers, such as c-kit receptor as described (Lee et al., 2004). Furthermore, by day 2 and thereafter these cultures were shown to form functional BTB when assessed by transepithelial electrical resistance (TER) measurement (Lui et al., 2001) across the Sertoli cell epithelium when cells were cultured on Matrigel-coated bicameral units. Additionally, functional BTB was detected as manifested by the presence of intact TJ and basal ES in these cultures on day 3 when examined by electron microscopy (Lee and Cheng, 2003; Siu et al., 2005). Indeed, these primary Sertoli cell cultures have been used by different laboratories including ours to study BTB function which mimic many of the functional and ultrastructural features of BTB in vivo (Byers et al., 1986; Chung et al., 2001; Janecki et al., 1991; Janecki et al., 1992). Media were replaced daily until these cultures were used on day 4 for the endocytosis assay to assess the effects of cytokines on the kinetics of endocytosis of integral membrane proteins at the BTB. Media collected on day 4 from selected cultures were used and served as Sertoli cell-conditioned media (SCCM).
GCCM was prepared from freshly isolated total germ cells using a nonenzymatic mechanical method as detailed elsewhere (Aravindan et al., 1996). In short, total germ cells without elongating/elongated spermatids isolated from adult rat testes were cultured in F12/DMEM supplemented with sodium pyruvate (2 mM) and sodium lactate (6 mM) as described (Aravindan et al., 1996) at 35 °C in a humidified atmosphere with 5% CO2/95% air (v/v) for 14–16 hr. Thereafter, media were collected, centrifuged at 800 g for 20 min to remove cellular debris and the supernatant was collected and concentrated using a Millipore YM-10 membrane in an Amicon Model 8050 ultrafiltration unit.
The in vivo BTB integrity assay was performed as earlier described (Li et al., 2006). In short, 200 ng recombinant TGF-β3 (R&D Systems, Minneapolis, MN) was administered locally to adult rat testis (~300 gm b.w.) on day 0 as described (Li et al., 2006). On day 2 and 14, rats (n=3 per treatment group for each time point) was anesthetized by ketamine HCl (60 mg/kg b.w., i.m.) with xylazine (10 mg/kg b.w., i.m.) as an analgesia. Rats without treatment/treated with vehicle (saline), or with CdCl2 for 3 days at 3 mg/kg b.w. via i.p. were served as negative and positive controls, respectively. A small incision of about 0.5–1 cm above the jugular vein was opened and 200 μl of FITC (1 mg/ml) in PBS was administered into the jugular vein using a 28-gauge needle. The incision was then stitched, and rats were allowed to recover. About 60–90 min thereafter, rats were euthanized by CO2 asphyxiation, and testes were immediately removed under aseptic conditions, and frozen in liquid nitrogen. Testes sections (about 8 μm) were obtained in a cryostat. All sections within an experimental set including controls and treatment groups were mounted with or without anti-fade reagent containing 4′,6′-diamidino-2-phenylindole (DAPI; Vector Laboratory, Burlingame, CA), and the green fluorescence for FITC was visualized in an Olympus BX40 fluorescent microscope. At least 100 tubules were randomly selected and acquired using an Olympus DP70 12.5 megapixel digital camera with the QCapture Software Suite (Version 2.56) (Quantitative Imaging Corp, Surrey, BC, Canada). Images were exported to TIFF format images. Distance of the FITC fluorescence that diffused away from the base of each seminiferous tubule (i.e., the relative location of the BTB near the basement membrane) (Df) in ~100 randomly selected tubules from two testes versus radius of the tubule (Rd) in treatment and control groups were computed. For oblique sections of seminiferous tubules, Rd was obtained by averaging the shortest and the longest distance from the basement membrane.
Endocytosis assay was performed essentially as earlier described with minor modifications (Le et al., 1999; Morimoto et al., 2005). Briefly, Sertoli cells cultured at 0.5 × 106 cells/cm2 on Matrigel-coated 6-well dishes for 4 days were washed twice with ice-cold PBS, incubated with 0.5 mg/ml Sulfo-NHS-SS-Biotin (Pierce) in PBS (10 mM sodium phosphate, 0.15M NaCl, pH 7.4 at 22 °C) containing 1 mM CaCl2 and 0.7 mM MgCl2 (PBS/CM) at 4 °C for 30 min to allow biotinylation of cell surface proteins. Excess Sulfo-NHS-SS-Biotin was quenched by 50 mM NH4Cl in PBS/CM at 4°C for 15 min. Then cells were washed twice with ice-cold PBS and incubated with F12/DMEM with (test) or without (control) cytokines (TGF-β3 at 3 ng/ml and TNFα at 10 ng/ml), or with concentrated GCCM at 50 μg/ml, at 35 °C for various time points in a humidified atmosphere with 95% air and 5% CO2 (v/v) to allow internalization of cell surface biotinylated proteins since endocytosis does not occur at 4 °C. At specified time points, cells were washed in cold PBS, incubated with a biotin stripping buffer [50 mM MESNA in 100 mM Tris/HCl (pH 8.6) containing 100 mM NaCl and 2.5 mM CaCl2] at 4 °C for 30 min to remove any remaining biotin or non-endocytosed biotinylated proteins on the cell surface, and quenched with a quenching buffer [5 mg/ml iodoacetamide in PBS/CM] at 4 °C for 15 min. Thereafter, cells were washed twice with ice-cold PBS and lysed in an IP lysis buffer [10 mM Tris, pH 7.4 at 22 °C, containing 0.15 M NaCl, 2 mM PMSF, 1 mM EGTA, 1% NP-40 (v/v), 1 mM sodium orthovanadate, 1 μg/ml leupeptin, 1 μg/ml aprotinin and 10% glycerol (v/v)]. Equal amount of cell lysates (about 400 μg protein) between samples in an experiment were incubated with NeutrAvidin beads (Pierce) to pull down biotinylated proteins, washed three times in PBS, and biotinylated proteins were extracted in SDS sample buffer (Xia and Cheng, 2005). Proteins were resolved by SDS-PAGE, to be followed by immunoblot analysis using corresponding specific antibodies (see Table 1) to assess the kinetics of internalization of different integral membrane proteins.
Pre-designed siRNA duplexes targeting TβRI (Genbank Accession Number: NM_012775; Cat. # 16708-48442), dynamin 2 (Genbank Accession Number: NM_013199; Cat. # 16708-197996) and dynamin 3 (Genbank Accession Number: NM_138538; Cat. # 16708-193600) were purchased from Ambion (Austin, TX) and their sequences are listed in Table 2. In short, Sertoli cells at 0.5 × 106 cells/cm2 were cultured for 2 days with negligible contamination of germ cells to allow the formation of an intact cell epithelium with functional TJ-barrier and the ultrastructures of BTB when examined by transepithelial electrical resistance (TER) and electron microscopy as described (Lui et al., 2003a; Siu et al., 2005). Thereafter, cells were transfected with specific siRNA duplex at 200 nM, using Oligofectamine (Invitrogen) for ~24 hr using protocols supplied by the vendor. Endocytosis assay was performed 2 days after transfection as described above and cultures were terminated at specified time point to assay changes in the kinetics of protein endocytosis. In selected experiments, cells that were transfected with specific siRNA duplex versus scrambled control siRNA duplex were harvested three days later for lysate preparation to confirm specific target protein knock down by immunoblot analysis.
Phenylarsine oxide (PAO) is an inhibitor of clathrin-dependent endocytic pathway (Ivanov et al., 2004). A recent report using specific inhibitors to clathrin (e.g., phenylarsine oxide, PAO) or caveolin (e.g., cholesterol oxidase, CO) have shown that protein endocytosis at the BTB using Sertoli cells cultured in vitro is likely mediated via the clathrin-dependent pathway (Yan et al., 2008). However, results derived from studies using specific inhibitors to study protein endocytosis should be cautiously interpreted (Ivanov, 2008); since PAO, besides blocking clathrin-mediated endocytosis, was shown to block macropinocytosis, phagocytosis (Frost et al., 1989; Massol et al., 1998), protein tyrosine phosphatase and Rho GTPase activity (Gerhard et al., 2003; Retta et al., 1996). PAO (Sigma-Aldrich) was used at 10 μM to examine its effects on TGF-β3-mediated enhanced endocytosis as described (Yan et al., 2008). Furthermore, to validate the earlier findings using inhibitors (Yan et al., 2008), we have re-examined the regulation of protein endocytosis at the BTB by cytokines (e.g., TGF-β3) using a more selective tool in which clathrin was knocked down by RNAi using specific clathrin siRNA vs. control scrambled duplexes. In short, Sertoli cells were isolated from 20-day-old rat testes (Day 0) as described above and plated at a density of 0.3 × 106 cells/cm2 on Matrigel-coated dishes where ultrastructures of BTB, such as TJ and basal ES, were detected by electron microscopy as described (Siu et al., 2005). Two days after isolation (Day 2), cells were transfected with 100 nM siRNA (ON-TARGET plus set of 4 duplexes LQ-090659-01; Dharmacon) against rat clathrin heavy chain (Genbank Accession Number: NM_019299). On-target plus siControl non-targeting pool (D-001810-10; Dharmacon) was used as the corresponding control. 10 μl of TransIT-TKO transfection reagent (Mirus Bio) was used to transfect a total volume of 2 ml reaction mixture per well in 6-well culture dishes. Two days after transfection (Day 4), endocytosis assay was performed as described above. Biotinylated proteins were pulled down by NeutrAvidin beads (Pierce), washed four times with IP lysis buffer and extracted in SDS-sample buffer. Cell lysate was also harvested to assess the steady-state protein level of clathrin heavy chain after silencing. Under our experimental conditions, a ~50–60% reduction in clathrin protein level was observed. In preliminary experiments, a cell density of 0.5 × 106 cells/cm2 was also used. However, at this higher cell density, Sertoli cells were required to be transfected twice in order to achieve acceptable knockdown when assessed by immunoblotting as previously described (Wong et al., 2008a), and endocytosis assay was performed 24 h after the second transfection (Day 8). Yet, endocytosis assays using Sertoli cells at higher density did not yield satisfactory result due to the diminished level of junction proteins on Day 8 of cells in culture (unpublished observations). Thus, the silencing condition was subsequently modified by using a lower cell density at 0.3 × 106 cells/cm2 vs. 0.5 × 106 cells/cm2, but BTB ultrastructures were visible by electron microscopy, and functional TJ-barrier was still formed, and Sertoli cells were transfected only once with the corresponding siRNA or control duplexes on Day 2, with the endocytosis assay performing on Day 4 to examine the effects on the kinetics of protein endocytosis in the presence or absence of TGF-β3.
Sertoli cells were cultured on coverslips at 0.5–1 × 105 cells/cm2 for dual-labeled immunofluorescence analysis. At specified time points following treatment versus controls, Sertoli cells were fixed in 4% paraformaldehyde (v/v) for 10 min, rinsed and treated with 10 mM glycine for 5 min to quench the aldehyde groups. Cells were permeabilized with 0.2% Triton X-100, rinsed and incubated with 10% normal goat serum in PBS for 15 min. For double-label immunofluorescence analysis, cells were incubated for 1 hr with the primary antibodies at appropriate dilutions (see Table 1), rinsed, and incubated for 1 hr with the relevant fluorescent secondary antibodies. Negative controls to assess staining specificity included incubation with normal IgG of the corresponding animal species of the primary antibodies or the omission of the primary antibodies. After extensive washing, coverslips were mounted on slides using Vectorshield mounting medium with DAPI (Vector Laboratory, Burlingame, CA). Images were examined and obtained by confocal microscopy and was performed at The Rockefeller University Bio-Imaging Resource Center with an inverted Zeiss LSM 510 Laser Scanning confocal microscope. Images were acquired using the Zeiss LSM 510 (v. 3.2) software and exported in TIFF format. Images were merged and analyzed using Adobe PhotoShop (Version 7.0).
Recombinant human TGF-β3 purchased from Calbiochem (Cat no. PF073) was resuspended and dissolved in 4 mM HCl containing 0.1% BSA (w/v). Just prior to its use, it was diluted in PBS to a concentration of 1 ng/μl. Adult rats (~300 gm b.w.) received 200 ng of recombinant TGF-β3 in a total volume of 200 μl per testis via intratesticular injection using a 28-gauge needle as described (Li et al., 2006) versus vehicle control with n = 2 rat per treatment or control group. Two days after TGF-β3 administration, rats were euthanized by CO2 asphyxiation, testes were removed and processed for electron microscopy as described (Li et al., 2006; Yan and Cheng, 2006). Electron micrographs were obtained using a JEOL 100CXII Electron Microscope (Peabody, CA) at 80 kV at the Rockefeller University Bio-Imaging Resource Center.
Immunoblotting and immunoprecipitation using the corresponding antibodies against different target proteins (see Table 1) were carried out as previously described (Xia and Cheng, 2005; Xia et al., 2006). Protein concentration in lysates was estimated by Coomassie blue-dye binding assay using BSA as a standard as described (Bradford, 1976). Proteins in GCCM were concentrated by ultrafiltration using an Amicon 8050 unit with YM-10 filters. Prior to their inclusion in the Sertoli cell cultures, GCCM was filtered via a 0.2-μm sterile filter unit and protein concentration was estimated and the desired concentration was appropriately adjusted by F12/DMEM under sterile conditions.
Each experiment was repeated at least three times using different batches of primary Sertoli cell cultures. Each time point in an experimental set contained triplicate cultures. Results of a treated sample group were compared to the corresponding control by ANOVA to be followed by Tukey’s Honest Significant Test or Dunnett’s test using the GB-STAT Statistical Analysis Software package (Version 7.0) (Dynamic Microsystems Inc., Silver Spring, MD).
As shown in Fig. 1, by local administration of recombinant human TGF-β3 (200 ng/testis) to adult rats (~300 gm b.w. at 100 days of age), the BTB integrity was shown to be disrupted by day 2 (Fig. 1A–B vs. D). This conclusion was reached since the BTB failed to restrict the FITC green fluorescence in the basal compartment of the seminiferous epithelium (see controls in Fig. 1A–B), resulting in the diffusion of FITC across the barrier. For instance, green fluorescence was detected in the seminiferous epithelium beyond the BTB in TGF-β3-treated rats, similar to rats treated with CdCl2 except that it was not that extensive (Fig. 1D vs. C, and A–B). Cadmium is an environmental toxicant known to disrupt BTB integrity irreversibly in rats (Hew et al., 1993; Setchell and Waites, 1970; Wong et al., 2004). As shown in Fig. 1E, the TGF-β3-induced disruptive effect on the BTB integrity is transient, since by 14 days after the treatment, the BTB was shown to restrict the diffusion of the FITC across the barrier, similar to control rats (Fig. 1E vs. A–B). To further validate the effects of TGF-β3 on the BTB following its local administration into the testis, electron microscopy was performed as shown in Fig. 2. The BTB in adult rat testes is created by TJ (see the ‘kisses’ denoted by the black arrows) at the Sertoli-Sertoli cell interface which is also present side-by-side with the basal ES (Fig. 2A). Basal ES is typified by the presence of actin filament bundles (see black arrowheads) sandwiched between the endoplasmic reticulum (ER) and the plasma membrane (note: the two apposing Sertoli cell plasma membranes are denoted by the two opposing arrowheads) (see Fig. 2A), which co-exist with TJ. However, in rats treated with TGF-β3 for 2 days, distinguishable ultrastructural damages were visible at the Sertoli-Sertoli cell interface (see white arrowheads in Fig. 2B), and both TJ and basal ES were no longer detactable at the BTB in this rat testis (Fig. 2B vs. 2A2A).
It is likely that at stage VIII of the seminiferous epithelial cycle of spermatogenesis, TGF-β3 is secreted by Sertoli and/or germ cells into the BTB microenvironment near the basement membrane of the seminiferous epithelium, contributing to its ‘restructuring’ (and/or ‘opening’) to allow the transit of preleptotene/leptotene spermatocytes across the barrier. Thus, it is important to determine if similar level of TGF-β3 can be reached in the testis endogenously. Fig. 3A–F summarizes the results of a study using a solid-phase based immunoblot assay to assess the relative levels of TGF-β3 and TNFα in SCCM, GCCM versus testes (Fig. 2A, C) by plotting the relative amount of corresponding cytokine in these media/lysates against the recombinant proteins. As shown in Fig. 3B, D and summarized in Fig. 3E, F, it was estimated that SCCM, GCCM and testis lysates (in about 300 μg total protein) contained 7.5±0.7, 5.8±0.4, and 2.5±0.5 ng of TGF-β3 (n = 3) versus 1.5±0.2, 4.5±0.6, and 1.9±0.3 ng of TNFα (n = 3), respectively. This was equivalent to about 1.2±0.3 and 0.91±0.13 μg (n = 3) TGF-β3 and TNFα/testis, respectively, assuming the testis weight of an adult rat is ~1.6 gm, which was used to prepare testis lysates using a testis:IP buffer ratio of 1:3 (w:v), and protein concentration was obtained by Bradford reagent (Bradford, 1976). Thus, these results illustrate that the in vivo effects of TGF-β3 on the BTB integrity as shown in Fig. 1D and Fig. 2 are physiologically relevant since this concentration of TGF-β3 or a combination of cytokines (e.g., TGF-β3 and TNFα) can possibly be achieved at the BTB microenvironment at the site between Sertoli cells and preleptotene/leptotene spermatocytes, which transiently disrupts the BTB to facilitate the transit of preleptotene spermatocytes.
Figure 4A summarizes the results of a representative set of experiments by monitoring the kinetics of endocytosis of several integral membrane proteins at the BTB: occludin, JAM-A, and N-cadherin. It was noted that the presence of TGF-β3 and TNFα enhanced the kinetics of endocytosis of biotinylated occludin, JAM-A and N-cadherin (Fig. 4A, B:a, b, c) from the Sertoli cell surface since significantly higher levels of these biotinylated integral membrane proteins were detected in the cytosol by 10 to 60 min (Fig. 4A, B). GCCM at 50 μg/ml was also effective in accelerating the internalization of biotinylated cell surface integral membrane proteins (Fig. 4A, B) possibly due to the presence of cytokines, in particular TNFα, in GCCM (see Fig. 3). However, the steady-state levels of the three BTB integral membrane proteins: occludin, JAM-A, and N-cadherin, in lysates from these cultures subjected to biotinylation and endocytosis assay with or without treatment with TGF-β3, TNF-α or GCCM vs. controls did not alter considerably (Fig. 4A, bottom four panels). These results were further verified by fluorescent microscopy as shown in Fig. 5A, B. For instance, in control Sertoli cell cultures without treatment with any cytokine, biotinylated proteins mostly localized at the cell surface (green fluorescence, FITC-streptavidin that bound to biotin), co-localizing with either occludin (red fluorescence) (Fig. 5A:a, b, c) or JAM-A (red fluorescence) (Fig. 5B:a, b, c) (see white arrowheads). By 30 min, cell surface biotinylated occludin (Fig. 5A:d, e, f) and JAM-A (Fig. 5B: d, e, f) were internalized (see white arrowheads); however, more occludin (Fig. 5A: h, i) and JAM-A (Fig. 5B: h, i) in the TGF-β3 treated Sertoli cell cultures were found to be internalized, moving away from the cell-cell interface (Fig. 5A: g, h, i and Fig. 5B: g, h, i). Furthermore, 3-dimensional projections of selected images (f & i) were reconstructed from a series of confocal images (z-stack) using the Imaris software (see f’ and i’ corresponding to f and i in Panels A and B), from which the protein internalization was better visualized. It is obvious that based on this analysis, TGF-β3 was capable of disrupting JAM-A- & occludin-based TJ fibrils via enhanced endocytosis, consistent with results shown in Fig. 4A, B.
To further validate the effects of TGF-β3 and to assess the role of dynamins in mediating protein endocytosis at the BTB, RNAi was used to knock down TβR1, dynamin 2 or dynamins 2 and 3. Treatment of Sertoli cells with either TβR1- or dynamin 2-specific siRNA duplex versus scrambled siRNA duplex (controls) was shown to specifically knockdown the steady-state level of the corresponding protein by ~50–60% without affecting the other protein, illustrating the specificity of this technique (Fig. 6A, B). Interestingly, partial knockdown of dynamin 2 in Sertoli cells by RNAi had no apparent effects on the endocytosis of JAM-A at the BTB (Fig. 6C, D). However, treatment of Sertoli cells with TβR1 siRNA duplex was shown to abolish the TGF-β3-induced acceleration of endocytosis of biotinylated JAM-A, but not TNFα-induced acceleration of protein endocytosis (Fig. 6E, F). Interestingly, while RNAi of dynamin 2 alone had no apparent effects on protein endocytosis (Fig. 6C, D), when both dynamins 2 and 3 were knocked down, protein internalization was disrupted significantly at the 60-min time point (Fig. 6G, H vs. C, D). Furthermore, silencing of dynamins 2 and 3 also disrupted the TGF-β3- and TNFα-induced acceleration of protein endocytosis at both 10- and 60-min (Fig. 6G, H).
Phenylarsine oxide (PAO) is a known inhibitor of clathrin-dependent pathway of protein endocytosis (Ivanov et al., 2004). The presence of PAO in the Sertoli cell cultures at 10 μM alone was shown to block the endogenous internalization of occludin, JAM-A and N-cadherin (Fig. 7A, B), illustrating the endocytosis of these proteins at the BTB is likely mediated via the clathrin-dependent pathway. The presence of PAO also abolished the TGF-β3-induced acceleration of protein endocytosis (Fig. 7A, B). However, PAO is known to block macropinocytosis, phagocytosis (Frost et al., 1989; Massol et al., 1998), protein tyrosine phosphatase and Rho GTPase activity (Gerhard et al., 2003; Retta et al., 1996). To further validate these data, a more specific approach by targeting clathrin was used. Specific siRNA duplex was used to knock down clathrin protein level by ~50–60% in primary Sertoli cells with established BTB as shown in Fig. 8A, B. Consistent with results obtained by using inhibitor such as PAO (see Fig. 7), the knockdown of clathrin by RNAi significantly blocked the accelerated endocytosis of JAM-A induced by TGF-β3 (Fig. 8C, D). While the use of RNAi to silence clathrin failed to block protein endocytosis as effective as an inhibitor (Fig. 8 vs. Fig. 7) since we only managed to knock-down ~60% of the clathrin steady-state protein level in these primary Sertoli cell cultures and some JAM-A continued to be endocytosed in the transfected Sertoli cells (Fig. 8C), it is obvious that the knockdown of clathrin blocked the TGF-β3-mediated acceleration of JAM-A endocytosis (Fig. 8C, D). Furthermore, the knockdown of clathrin by RNAi also considerably reduced BTB integral membrane protein (e.g., JAM-A) endocytosis (see Fig. 8C, top panel versus Fig. 4A, second panel) due to the remaining clathrin available to the Sertoli cells in the system. It is also noted that the steady-state protein level of JAM-A in the cell lysates from cultures subjected to clathrin knockdown with or without treatment with TGF-β3 did not alter significantly (Fig. 8C, lower panel, and Fig. 8E).
For more than three decades since the detailed morphological study illustrating the migration of preleptotene and leptotene spermatocytes across the BTB that takes place at stages VIII–IX of the seminiferous epithelial cycle of spermatogenesis in adult rats (Russell, 1977), the biochemical mechanism(s) that regulates this event, however, remains virtually unknown. Even though recent studies have shown that cytokines, such as TGF-β3 and TNFα, regulate the steady-state levels of integral membrane proteins at the BTB (e.g., occludin, ZO-1, claudins) (Hellani et al., 2000; Lui et al., 2001; Lui et al., 2003b; Siu et al., 2003a; Wong et al., 2004; Xia and Cheng, 2005), thereby determining the status of the BTB integrity plausibly via their effects on the transcriptional regulation of specific target genes such as claudins [for a review, see (Lui and Cheng, 2007)], the precise mechanism that is used by the testis involving cytokines to maintain the optimal integral membrane protein levels at the BTB remains obscure. Recent findings based on studies from different epithelia have shown that protein endocytosis is a novel mechanism utilized by a cell epithelium to rapidly alter cell junction dynamics, necessary to facilitate cell movement during embryogenesis, differentiation, and development (Ivanov et al., 2004; Le et al., 1999; Morimoto et al., 2005), it is of interest to determine if the transit of preleptotene/leptotene spermatocytes across the BTB employs similar mechanism involving cytokines, in particular Sertoli and germ cells are known to produce different cytokines in the seminiferous epithelium at relative high concentrations [for reviews, see (Cheng and Mruk, 2002; Mruk and Cheng, 2004; Skinner, 1993)].
In this report, it was shown that TGF-β3 when administered locally to the testis at 0.2 μg/testis, which is the concentration well within the endogenous range of TGF-β3 in normal testes, at 1.2±0.3 μg/testis when assessed by a solid-phase immunoblot-based assay, TGF-β3 was shown to reversibly disrupt the BTB integrity when assessed by a functional in vivo assay. These findings, coupled with the recent studies in which local administration of TNFα was also capable of inducing reversible BTB integrity disruption in vivo (Li et al., 2006), thus suggest that cytokines produced by Sertoli and germ cells locally [for reviews, see (Mruk and Cheng, 2004; Siu and Cheng, 2004; Skinner, 1993)] at the BTB microenvironment can regulate BTB dynamics possibly by regulating the steady-state levels and/or enhancing endocytosis of integral membrane proteins at the BTB (e.g., occludin, JAM-A and N-cadherin), causing a transient dissolution of the TJ-fibrils at the site, thus facilitating the transit of preleptotene/leptotene spermatocytes across the BTB. As shown in this report, the effects of TGF-β3 that enhance protein endocytosis appear to be specific since a partial knock-down of TβRI, the specific receptor for TGF-β, rendered the Sertoli cell cultures with functional BTB incapable of responding to TGF-β3 treatment to accelerate protein endocytosis. The schematic drawing depicted in Fig. 9 illustrating the hypothesis that cytokines (e.g., TGF-β3 and TNFα) contributed by Sertoli and germ cells to the microenvironment at BTB to facilitate spermatocytes in transit is further supported by the findings that TGF-β3 expression at the BTB is stage-specific and its localization at the site of the BTB in the seminiferous epithelium of adult rat testes is relatively high at stages VII–VIII of the epithelial cycle when examined by immunohistochemistry technique (Lui et al., 2003a; Xia et al., 2006). Furthermore, TβR1 was also detected at the BTB by immunohistochemistry at these stages (Xia et al., 2006). The present report thus strengthens the notion that as a preleptotene/leptotene spermatocyte from the basal compartment begins its transit crossing the BTB to enter the adluminal compartment to continue its further development, cytokines are released near its apical region to enhance endocytosis, thereby inducing transient restructuring of the BTB to facilitate cell movement as illustrated in Fig. 9. However, it remains to be determined if the TGF-β3-induced germ cell loss from the seminiferous epithelium as recently reported (Xia et al., 2006) is also mediated by accelerated endocytosis of proteins at the apical ES and/or desmosome-like junctions at the Sertoli-elongating/elongated spermatid and Sertoli-spermatocyte interface. Thus similar investigation should be expanded to the use of Sertoli-germ cell cocultures or similar systems in future studies.
It is of interest to note that by silencing dynamin 2 or both dynamins 2 and 3 via the use of specific dynamin duplexes, it failed to disrupt endocytosis of JAM-A at the BTB in the absence of cytokines. Yet the silencing of dynamins 2 and 3 can effectively block the TGF-β3- and TNFα-induced enhanced endocytosis of integral membrane proteins at the BTB. It seems that while these large GTPases are used by other epithelia to regulate protein endocytosis (McNiven et al., 2000), they are not involved in protein endocytosis under normal conditions. However, a surge of cytokines at the BTB microenvironment may have recruited or activated other adaptors (or kinases, phosphatases) which, in turn, render the physiological involvement of dynamins in this event to accelerate protein endocytosis. This postulate is supported by recent findings that a disruption of germ cell adhesion in the seminiferous epithelium is associated with a transient increase in the binding of dynamin 2 with β-catenin, pulling β-catenin away from the cadherins, thereby destabilizing Sertoli-germ cell adhesion (Lie et al., 2006). This possibility is also supported by recent findings that FAK and Src, both are non-receptor protein tyrosine kinases, are prominently present at the BTB as visualized by immunohistochemistry and fluorescent microscopy (Lee and Cheng, 2005; Siu et al., 2003b). These kinases can thus activate and recruit dynamins to the BTB site to accelerate protein endocytosis to facilitate the transit of preleptotene/leptotene spermatocytes.
In epithelia, protein endocytosis is mediated via either clathrin-dependent, caveolin-dependent, or clathrin and caveolin-independent (e.g., macropinocytosis) mechanism [for reviews, see (Ivanov et al., 2005; Maxfield and McGraw, 2004)]. A recent report using specific inhibitors to clathrin (e.g., PAO) or caveolin (e.g., cholesterol oxidase, CO) have shown that protein endocytosis at the BTB using Sertoli cells cultured in vitro is likely mediated via the clathrin-dependent pathway (Yan et al., 2008). However, many of these classical inhibitors employed to study protein endocytosis are not specific (Ivanov, 2008). For instance, PAO, besides blocking clathrin-mediated endocytosis, also blocked macropinocytosis and phagocytosis, protein tyrosine phosphatase and Rho GTPase activity (Frost et al., 1989; Gerhard et al., 2003; Massol et al., 1998; Retta et al., 1996). In this study, we have used a more specific approach by RNAi using specific siRNA duplexes to knock down clathrin versus scrambled control siRNA duplexes. Using primary Sertoli cell cultures with established BTB including the functional TJ-permeability barrier in vitro, we managed to knock down ~60% of the clathrin protein in these Sertoli cells. While Sertoli cells continued to endocytose BTB proteins (e.g., JAM-A) at a considerably level following the partial knockdown of clathrin, TGF-β3 was no longer capable of enhancing JAM-A endocytosis because of a lack of clathrin in the Sertoli cell epithelium. These findings thus support a recent report (Yan et al., 2008) and results reported herein using inhibitors, illustrating protein endocytosis, in particular TGF-β3-induced accelerated endocytosis, that occurs at the BTB is mediated by a clathrin-dependent mechanism.
Present studies in the field have shown that endocytosed proteins in epithelial cells can either (i) be recycled back to the cell surface via recycling endosomes or undergo de-ubiquitination, or (ii) be recruited to late endosome to be targeted to lysosomes or undergo ubiquitination and targeted to proteasomes for intracellular degradation [for reviews, see (Bright et al., 2005; Katzmann et al., 2002; Lui and Cheng, 2007; Maxfield and McGraw, 2004; Piper and Luzio, 2007). It remains to be determined if the TGF-β3- and TNFα-induced endocytosed proteins would be targeted to lysosomes or proteasomes for intracellular degradation, or the percentage of endocytosed proteins being targeted for intracellular degradation via lysosomes and/or proteasomes versus recycling are differentially regulated, thereby reducing the steady-state levels of integral membrane proteins at the BTB, destabilizing the barrier function at the BTB to facilitate preleptotene/leptotene spermatocyte migration during spermatogenesis. However, using Sertoli cells with established BTB, TGF-β2 and testosterone were shown to enhanced protein endocytosis at the BTB, such as occludin (Yan et al., 2008). But TGF-β2 apparently promoted endocytosed occludin to late endosomes for its intracellular degradation via an increase in association with Rab 9 (a late endosome marker); whereas testosterone stimulated protein recycling (Yan et al., 2008), possibly relocating integral membrane proteins at the BTB near the apical region of a preleptotene spermatocyte in transit to its basal region. This possibility is consistent with the recent findings that testosterone is important to BTB function (Meng et al., 2005; Wang et al., 2006). The present report not only confirms some of these earlier findings, it has unequivocally demonstrated that protein endocytosis at the BTB, such as JAM-A, is mediated via a clathrin-dependent mechanism. Also, besides TGF-β2, TNFα and TGF-β3 can also regulate BTB restructuring via protein endocytosis. In this context, it is of interest to note that the effects of TGF-β3 on the endocytosis of some BTB integral membrane proteins, such as JAM-A, are more potent than TGF-β2 (Yan et al., 2008). This is not entirely unprecedented, since all TGF-βs and their related family members (such as activins, inhibins, bone morphogenetic protein) exert their biological effects via interactions with one of the three TGF-β binding proteins (or receptors) but with different affinities, which, in turn, determines the different biological potencies of different ligands (i.e., TGF-βs or their members) [for reviews, see (Bovd et al., 1990; Massague, 2008)]. These findings also suggest that different cytokines may be expressed temporally and/or spatially at the BTB microenvironment during the epithelial cycle at the time preleptotene spermatocytes are in transit at the BTB by regulating the relocation of proteins at the Sertoli-Sertoli cell interface, such as from the apical region of a migrating spermatocyte to its basal region, via transcytosis. Nonetheless, this possibility must be vigorously investigated in future studies.
It is noted that in the endocytosis assays, there were variations in the recovery of endocytosed biotinylated proteins versus total biotinylated proteins between experiments. This is possibly due to the multi-step processing of samples in these assays: (i) lysates preparation, (ii) the pull-down of biotinylated total proteins by avidin-based beads, (iii) SDS-PAGE, to be followed by (iv) immunoblotting using corresponding antibodies to visualize a specific target protein in the cytosol over an experimental period to detect the endocytosed and biotinylated protein. Indeed, we had minimized intra-experimental variations by processing all samples within an experimental set simultaneously so that any loss of biotinylated proteins would occur uniformly across the samples in a given experiment. Yet, given the number of steps that were involved to process these samples biochemically, it is difficult to conclude that the amount of endocytosed junctional proteins (e.g., JAM-A) is small (such as those shown in Fig. 6E) because of the potential loss of samples during their processing. However, for data reported in Fig. 5B (see g–i versus a–f), these are the ‘real-time’ JAM-A at the cell-cell interface following TGF-β3 treatment at 30-min, but these results could not yield the time-dependent changes in endocytosis as those depicted in Fig. 6E nor be able to yield semi-quantitative information. But this technique is helpful to assess any changes in protein distribution (such as redistribution or mislocalization) in the Sertoli cells. Again, it is difficult to conclude that there was any loss of JAM-A at the cell-cell interface based on fluorescent staining by confocal microscopy except to conclude that more JAM-A was internalized following treatment with TGF-β3. These observations also expose the limitations of these two powerful and widely used techniques in the field, namely the biochemical-based endocytosis assay and the dual-labeled immunofluorescence analysis by confocal microscopy.
In summary, based on the results reported herein, we have provided a working model possibly used by the testis to regulate the transient ‘opening’ and ‘closing’ of the BTB during spermatogenesis involving cytokines and protein endocytosis as depicted in Figure 9. This model will now provide a basis for investigators in the field to design functional experiments to examine the intriguing regulation of junction dynamics at the BTB.
The authors thank Dr. Alison North at The Rockefeller University Bio-Imaging Resource Center for her helpful discussion on the use of confocal microscopy. We are also grateful for the excellent technical assistance of Ms. Eleana Sphicas at The Rockefeller University Bio-Imaging Resource Center for performing electron microscopy.
*This work was supported in part by grants from the National Institutes of Health (NICHD, U01 HD045908; R03 HD051512; U54 HD029990 Project 5 to CYC), and the CONRAD Program (CICCR, C1G 01-72).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.