Interleukin-1α is a regulator of the blood-testis barrier

Abstract

Throughout spermatogenesis, the Sertoli cell blood-testis barrier (BTB) is strictly regulated by cytokines, which mediate its timely restructuring, thereby allowing spermatocytes to enter the adluminal compartment of the seminiferous epithelium for development into spermatozoa. The aim herein was to investigate whether germ cells play a role in BTB restructuring via the action of interleukin-1α (IL-1α) since germ cells are known to control Sertoli cell production of this cytokine, and if yes, how these effects are mediated. When Sertoli cells were isolated from Sprague-Dawley rats and plated at high density, IL-1α (100 pg/ml) was shown to “open” the Sertoli cell barrier when its integrity was assessed by transepithelial electrical resistance measurements. Further investigation of Sertoli cells treated with IL-1α revealed striking changes in the cellular distribution of actin filaments when compared to untreated cells. These effects at the Sertoli cell barrier were mediated, in part, by epidermal growth factor receptor pathway substrate 8 (Eps8; an actin bundling and barbed-end capping protein) and actin-related protein 3 (Arp3; a component of the actin nucleation machinery). As important, an increase in the kinetics of occludin internalization but a decrease in its rate of degradation was noted following IL-1α treatment. These results indicate that IL-1α is a critical regulator of BTB dynamics.—Lie, P. P. Y., Cheng, C. Y., Mruk, D. D. Interleukin-1α is a regulator of the blood-testis barrier.

Throughout spermatogenesis in the mammalian testis, the blood-testis barrier (BTB), which is present between adjacent Sertoli cells, undergoes restructuring during stages VIII to XI of the seminiferous epithelial cycle to accommodate the entry of spermatocytes into the adluminal compartment of the seminiferous epithelium for their continued development (1–2). The current understanding is that spermatocytes cross the BTB while enclosed within an intermediate compartment that is sealed at both poles by several different types of Sertoli cell junctions, namely tight junctions (TJs), ectoplasmic specializations (ESs), and desmosome-like and gap junctions (3–4). After preleptotene spermatocytes have signaled to Sertoli cells (via a yet to be identified mechanism) that they require entry into the adluminal compartment, existing junctions situated above these spermatocytes have to disassemble. This event appears to be mediated first by the internalization of structural proteins, and second by the immediate trafficking of these proteins to the site below migrating preleptotene spermatocytes, where new junctions will assemble. As such, there is a brief moment during the seminiferous epithelial cycle, in which a migrating spermatocyte can be microscopically viewed as being trapped in between two barriers, the so-called intermediate compartment. This is somewhat analogous to a hospital isolation room: there are two doors, and both doors have to open eventually, but they cannot be opened at the same time. In this way, the integrity of the immunological barrier can be maintained during the passage of spermatocytes across the BTB, and this is critical for spermatogenesis. Restructuring of the BTB during spermatocyte movement is a complicated process that is coordinated in large part by Sertoli cell-derived cytokines, hormones, and other local factors, which regulate protein expression, localization, and turnover at this site (5–6). It is believed that preleptotene/leptotene spermatocytes also play an important role by producing cytokines, such as transforming growth factor (TGF)-β3 and tumor necrosis factor (TNF)α, which facilitate BTB restructuring (5–6). In this study, we investigate the role of interleukin-1α (IL-1α) in the restructuring of the Sertoli cell barrier with the aim of expanding our understanding of how spermatocytes cross this elusive but very important structure.

IL-1 is a proinflammatory cytokine that was initially described as a macrophage secretory factor and subsequently found to be comprised of IL-1α and IL-1β, two distinct proteins that share the IL-1 type I receptor (7–8). At present, the IL-1 family also includes a naturally occurring inhibitor known as IL-1 receptor antagonist (IL-1Ra), as well as a number of other recently discovered members (8–9). IL-1α is synthesized as a 31-kDa precursor protein that is cleaved into a 17-kDa mature protein by calpain, a cysteine protease (10–11). Unlike IL-1β, both precursor and mature IL-1α are biologically active. In the seminiferous epithelium of the adult testis, high levels of IL-1 bioactivity were detected during stages VIII to XI (12–13), coinciding with the release of spermatozoa at stage VIII and with the movement of spermatocytes across the BTB in vivo (1–2). While a low level of bioactive IL-1 was shown to be secreted into the spent medium by 20-d-old Sertoli cells (but not germ cells; ref. 14), its mechanism of secretion is not yet understood because IL-1 lacks a signal sequence. In addition, spermatocytes and round spermatids are known to produce both IL-1α mRNA and protein (15), as well as to regulate IL-1α synthesis by Sertoli cells. For instance, IL-1α expression was not detected in Sertoli cells when germ cells were depleted from the testis and spermatogenesis was halted by either exposure to radiation or treatment with busulfan (16). These intriguing results are in line with observations from our laboratory: using testis lysates for routine immunoblotting, IL-1α was undetectable after germ cells were depleted from the seminiferous epithelium by the contraceptive compound adjudin (unpublished results). Taken collectively, these data suggest that germ cells may be participating in the disassembly and assembly of the BTB by controlling the production of IL-1α by Sertoli cells. Moreover, in a recent study, we reported that intratesticular administration of recombinant IL-1α perturbed the F-actin network in Sertoli cells, which adversely affected the integrity of the BTB and resulted in the sloughing of germ cells (17). In this study, we expand our previous findings by using highly pure Sertoli cells as our in vitro model. These Sertoli cells were cultured at high density on a modified extracellular matrix, which allowed them to polarize and to assemble TJs, ESs, and desmosome-like and gap junctions, thereby mimicking the BTB in vivo. Herein, we report that the addition of IL-1α prompted disassembly of the BTB via a unique mechanism.

MATERIALS AND METHODS

Animals

The use of 20-d-old male Sprague-Dawley rats purchased from Charles River Laboratories (Kingston, MA, USA) was approved by The Rockefeller University (New York, NY, USA) Animal Use and Care Committee (protocols 06018 and 09016). Population Council laboratories are located on the campus of The Rockefeller University.

Sertoli cell cultures and their treatment with IL-1α

Sertoli cells isolated from 20-d-old rats, as described previously (18), were plated at cell densities that were predetermined as optimal for the different experiments described in this study. For the preparation of protein lysates, endocytosis assays, and protein degradation assays, Sertoli cells were plated at high density (0.5×10⁶ cells/cm²) on Matrigel (BD Biosciences, San Jose, CA, USA)-coated culture plates. For transepithelial electrical resistance (TER) measurements and confocal microscopy, Sertoli cells were plated at even higher densities (0.75–1.0×10⁶ cells/cm²) on Matrigel-coated Millicell HA cell culture inserts (Millipore, Billerica, MA, USA) and Corning Transwell polyester membrane inserts (Corning, Lowell, MA, USA), respectively. For epifluorescence microscopy, Sertoli cells were plated at low density (0.04×10⁶ cells/cm²) on Matrigel-coated glass coverslips but at a density that still permitted the assembly of functional TJs, basal ESs, and desmosomes (19, 20). Cells were cultured at 35°C for 3–4 d in F12/DMEM (Sigma-Aldrich, St. Louis, MO, USA) supplemented with gentamicin, bacitracin, and growth factors to allow the formation of a functional TJ permeability barrier (18, 21). Cultures were hypotonically treated 48 h after plating to remove residual germ cells, yielding Sertoli cells with ∼98% purity (22). On d 3 or 4 in vitro, highly pure Sertoli cells were treated with recombinant rat IL-1α at 100 pg/ml (R&D Systems, Minneapolis, MN, USA; 1:50,000 dilution of 5 μg stock dissolved in PBS, pH 7.4, containing 0.1% BSA, w/v, at 22°C) or vehicle control (1:50,000 dilution of PBS, pH 7.4, containing 0.1% BSA, w/v, at 22°C) diluted in F12/DMEM. Media containing IL-1α or vehicle were replaced daily thereafter. The concentration of IL-1α used in this study was within range of ED₅₀ values that reported different effects in cultured testicular cells (23–24).

Immunofluorescence analysis by epifluorescence and confocal microscopy

For confocal microscopy, Sertoli cells were plated at a considerably higher cell density to promote cell polarization. Prior to fixation, polyester membranes were excised from culture inserts. Sertoli cells cultured on polyester membranes (confocal microscopy) or glass coverslips (epifluorescence microscopy) were fixed in 4% paraformaldehyde (w/v) for 10 min, permeabilized in 0.1% Triton X-100 (v/v) for 4 min, and blocked with 1% BSA (w/v) for 30 min. For immunostaining, cells were incubated overnight with primary antibodies (1:100 dilution in PBS, pH 7.4, containing 1% BSA, w/v, at 22°C; Table 1), followed by a 30-min incubation with Alexa Fluor-conjugated secondary antibodies (Invitrogen, Carlsbad, CA, USA) at 1:100 dilution in PBS (pH 7.4) containing 1% BSA (w/v) at 22°C. For F-actin staining, cells were incubated with rhodamine-conjugated phalloidin (Invitrogen) at a 1:50 dilution together with the secondary antibody. All incubations were performed at room temperature. Cells were mounted in ProLong Gold antifade reagent with DAPI (Invitrogen).

Table 1.

Commercially obtained antibodies used in this study

Antibody	Host	Vendor	Catalog no.	Applications	IB dilution
Actin	Goat	Santa Cruz Biotechnology (Santa Cruz, CA, USA)	sc-1616	IB	1:300
Arp3	Mouse	Sigma-Aldrich (St. Louis, MO, USA)	A5979	IB, IF	1:3000
N-cadherin	Mouse	Invitrogen (Carlsbad, CA, USA)	33–3900	IF	N/A
N-cadherin	Rabbit	Santa Cruz Biotechnology	sc-7939	IB	1:200
CAR	Rabbit	Santa Cruz Biotechnology	sc-15405	IB	1:200
β-Catenin	Rabbit	Invitrogen	71–2700	IB, IF, IP	1:125
Eps8	Mouse	BD Biosciences (San Jose, CA, USA)	610143	IB, IF	1:5000
JAM-A	Rabbit	Invitrogen	36–1700	IB	1:125
Occludin	Rabbit	Invitrogen	71–1500	IB, IF	1:125
N-WASP	Rabbit	Santa Cruz Biotechnology	sc-20770	IB	1:200
ZO-1–(FITC)	Mouse	Invitrogen	33–9111	IF	N/A
ZO-1	Rabbit	Invitrogen	61–7300	IB	1:125

IB, immunoblotting; IF, immunofluorescence; IP, immunoprecipitation.

For epifluorescence microscopy, images were acquired with an Olympus BX61 fluorescence microscope (Olympus Imaging America, Center Valley, PA, USA) equipped with an Olympus DP70 12.5MPa digital camera and MicroSuite FIVE 1.224 software (Olympus Soft Imaging Solutions Corp., Munster, Germany). Tiff images were adjusted for brightness and contrast and merged in Photoshop 11.0 (Adobe Systems, San Jose, CA, USA). For confocal microscopy, image acquisition was performed with an inverted Zeiss LSM 510 laser scanning confocal microscope (Carl Zeiss Microimaging, Thornwood, NY, USA) and Zeiss LSM 510 software. Optical sections of ∼0.8 μm were collected at 0.25-μm intervals along the z axis to obtain a series of images (i.e., Z stack). Image deconvolution of Z stacks was performed with Huygens deconvolution software (Scientific Volume Imaging, Hilversum, Netherlands) to increase the signal-to-noise ratio. Z stacks were adjusted for brightness and contrast, and optical sections were reconstructed with ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA). Confocal microscopy was performed at The Rockefeller University Bio-Imaging Resource Center.

Endocytosis assays and protein degradation assays

Endocytosis assays for occludin and N-cadherin were performed as described previously with minor modifications (19). On d 4 in vitro, cell surface proteins were biotinylated with 0.8 mM sulfo-NHS-SS-biotin (Thermo Scientific, Rockford, IL, USA) in PBS-CM buffer (PBS, pH 7.4, containing 0.9 mM CaCl₂ and 0.33 mM MgCl₂, at 22°C) for 30 min at 4°C, and the reaction was quenched with 50 mM Tris in PBS-CM buffer for 15 min at 4°C. Endocytosis was initiated by incubating cells at 35°C with F12/DMEM containing IL-1α or vehicle. At different time points, remaining cell surface biotin was stripped with 2 washes of stripping buffer (50 mM MESNA in 100 mM Tris-HCl, pH 8.6, containing 100 mM NaCl and 2.5 mM CaCl₂, at 22°C) at 4°C for a total of 45 min, and the reaction was quenched with iodoacetamide (5 mg/ml) in PBS-CM buffer at 4°C for 15 min. Cell lysate was harvested in RIPA buffer (50 mM Tris, pH 7.4, containing 150 mM NaCl; 2 mM EGTA; 0.1% SDS, w/v; 1% Nonidet P-40, v/v; and protease and phosphatase inhibitor cocktails used at 1:100 dilutions; at 22°C). Protein degradation assays for occludin and N-cadherin were also performed. This assay followed the above described protocol except that biotin stripping and quenching steps were omitted, and cells were harvested in RIPA buffer immediately after being treated with F12/DMEM containing IL-1α or vehicle for different durations. The protein degradation assay did not discriminate between nondegraded proteins present on the cell surface or in the cytosol. Biotinylated proteins were isolated by incubating 180 or 80 μg protein (for endocytosis assays and protein degradation assays, respectively) with UltraLink Immobilized NeutrAvidin Plus (Thermo Scientific) overnight at 4°C, followed by immunoblotting using the corresponding antibodies (Table 1).

Statistical analysis

All experiments were performed ≥3 times using different batches of Sertoli cells. GB-STAT 7.0 software (Dynamic Microsystems Inc., Silver Spring, MD, USA) was used for statistical analyses. For experiments with multiple time points, 2-way ANOVA followed by Newman-Keuls test was used to compare the treatment with the control group at each time point.

General methods

TER across the Sertoli cell epithelium was quantified as described previously (21). Except for endocytosis assays and protein degradation assays, IP lysis buffer (50 mM Tris, pH 7.4, containing 150 mM NaCl; 2 mM EGTA; 10% glycerol, v/v; 1% Nonidet P-40, v/v; and protease and phosphatase inhibitor cocktails; at 22°C) was used to prepare cell lysates. Approximately 25 and 400 μg of protein was used for immunoblotting and immunoprecipitation, respectively, as described previously (25).

RESULTS

Sertoli cell TJ permeability barrier was perturbed as a result of disorganization of the actin network in vitro by IL-1α

Sertoli cells cultured at high density are known to assemble a functional TJ permeability barrier in vitro, as assessed by the TER technique (Fig. 1A), and this barrier mimics the BTB in vivo by the presence of ultrastructural features that define TJs, basal ESs, and desmosome-like junctions under electron microscopy (26–27). Using this in vitro Sertoli cell model, we now report findings that expand our earlier study on the regulation of the BTB by IL-1α in vivo (17). Treating Sertoli cells from d 3 onward (i.e., by d 3, a functional TJ permeability barrier had assembled) with recombinant rat IL-1α (17-kDa mature protein; 100 pg/ml or ∼6 pM) resulted in a decrease in TER by 12 h after treatment and in a leaky (i.e., disassembled) TJ barrier (i.e., ∼40 Ω·cm²) when compared to cells treated with vehicle (i.e., ∼60 Ω·cm²) (Fig. 1A). The dose used in this study is based on preliminary experiments which showed that higher doses of IL-1α failed to further decrease TER (data not shown). Also, doses <100 pg/ml were found to decrease TER dose dependently (data not shown). These results are in line with studies that reported IL-1α to function in the picomolar range when testicular cell cultures were used (23–24).

Figure 1.

IL-1α compromises the TJ permeability barrier as a result of disorganization of the actin network in Sertoli cells in vitro. A) Sertoli cells (1.0×10⁶ cells/cm²) having a functional TJ permeability barrier (assembled by d 3 in vitro) were treated with vehicle control (Veh) or IL-1α (100 pg/ml) onward of d 3, as indicated by arrows. D, day after treatment. TER was monitored both before and after vehicle or IL-1α treatment at specified time points. Data points represent means ±sd (n=3). Similar results were obtained in 5 independent experiments. B) Immunoblots investigating the steady-state levels of proteins important in the regulation of actin dynamics at the Sertoli cell barrier (i.e., BTB). Lysates from Sertoli cells (0.5×10⁶ cells/cm²) treated with vehicle or IL-1α for increasing periods of time were used for immunoblotting. Actin served as the loading control. C) Bar plot summarizing results from several independent experiments after normalizing each data point against its corresponding actin time point and then against its corresponding control at 0 h. Control in both experimental groups was arbitrarily set at 1. Bar plots are not shown for proteins that did not display significant changes. Data points represent means ± sd (n=4). D, E) Effects of IL-1α on the organization of the actin cytoskeleton and on the localization of actin-regulating proteins in Sertoli cells, as visualized by epifluorescence microscopy. F-actin (red; a, b, e, f) was costained with Eps8 (green; Dc–f) or Arp3 (green; Ec–f). Nuclei were visualized with DAPI (blue; a–d). Approximately 1000 cells in each experimental group in 3 independent experiments were analyzed; results are representative of these analyses. Scale bars = 25 μm. *P < 0.05, **P < 0.01 vs. vehicle; 2-way ANOVA followed by Newman-Keuls test.

Since an in vivo study reported that IL-1α compromised actin dynamics in the seminiferous epithelium (17), we investigated by immunoblotting and fluorescence microscopy the effects of IL-1α in vitro on F-actin and several actin-regulating proteins known to participate in BTB dynamics (Fig. 1B–E) (28–29). These included Eps8 (epidermal growth factor receptor pathway substrate 8; an actin bundling and barbed-end capping protein), Arp3 (actin-related protein 3; a component of the Arp2/3 protein complex which functions in actin nucleation) and N-WASP (neuronal Wiskott-Aldrich syndrome protein; an activator of the Arp2/3 complex). When F-actin in Sertoli cells was visualized by fluorescence microscopy 1 d after IL-1α treatment (Fig. 1D, E), actin filaments appeared disorganized and unevenly distributed (i.e., filaments had lost their unidirectional orientation), with thick clusters of filaments concentrating where three Sertoli cells met. After IL-1α treatment, F-actin within one cell appeared to extend into an adjacent cell, as opposed to vehicle treatment, in which F-actin was arranged in an orderly manner within the confines of each cell (Fig. 1D, E). From these IL-1α data, we conclude that the physiological end point of these disruptive changes involving F-actin and actin-regulating proteins was the decrease in TER (Fig. 1A). Moreover, the actin barbed-end capping protein Eps8 was also found to be mislocalized (Fig. 1D), but there was no change in its protein level (Fig. 1B). In vehicle-treated Sertoli cells, Eps8 localized primarily to the cell-cell interface with weak cytoplasmic staining, but IL-1α appeared to alter its cellular localization by forcing Eps8 into the nucleus. On the other hand, no apparent change in Arp3 localization was observed (Fig. 1E), except that its level increased significantly 2–3 d after IL-1α treatment when assessed by immunoblotting and epifluorescence microscopy (Fig. 1B, C, E). These results demonstrate that IL-1α facilitates barrier restructuring (i.e., “opening,” although in this context, this word is used cautiously because the BTB never really opens in vivo; see the introduction) in part via changes to the actin network.

BTB-constituent proteins were mislocalized in IL-1α-treated Sertoli cells

Given the adverse effects of IL-1α on the actin cytoskeleton, it was anticipated that actin-based BTB-constituent junctional protein complexes would also be affected. Thus, we examined first the colocalization of TJ proteins occludin and ZO-1, and second the colocalization of basal ES proteins N-cadherin and β-catenin by dual-labeled immunofluorescence staining (Fig. 2). Vehicle-treated Sertoli cells displayed strict colocalization of occludin with ZO-1 (Fig. 2A) and N-cadherin with β-catenin (Fig. 2B) at the cell-cell interface. Immunoreactive signals in control cells were discrete, and proteins were restricted to a finite area that corresponded to a functional Sertoli cell barrier. These results matched well with previously published reports in the field (19, 20). In contrast, IL-1α-treated Sertoli cells displayed BTB-constituent proteins, which were thickened at the cell-cell interface. Immunoreactive signals in treated cells were diffuse, and proteins failed to concentrate to a defined area, likely the result of a disrupted actin cytoskeleton. Although a moderate amount of colocalization between occludin and ZO-1 and between N-cadherin and β-catenin was noted in IL-1α-treated Sertoli cells, there were more areas in which colocalization was not obvious (Fig. 2), suggesting that IL-1α may have caused proteins that normally form a functional complex to disassociate.

Figure 2.

IL-1α alters the localization of BTB-constituent proteins to facilitate junction restructuring in Sertoli cells. Sertoli cells (0.04×10⁶ cells/cm²) were treated with vehicle control (Veh; a, b) or IL-1α (c, d) as described in Materials and Methods. D, day after treatment. Cells were coimmunostained with occludin (red; Aa–d) and ZO-1 (green; Ae–h) or N-cadherin (green; Ba–d) and β-catenin (red; Be–h). Corresponding merged images show sites of colocalization (yellow; i-l). Nuclei were visualized with DAPI (blue; a, c, e, g). Boxed areas in panels i and k are magnified in panels b, f, j and d, h, l, respectively. Brackets show a thickening in occludin/ZO-1 (Ai vs. k) and N-cadherin/β-catenin (Bi vs. k). Approximately 1000 cells in each experimental group in 5 independent experiments were analyzed; results are representative of these analyses. Scale bars = 25 μm (a, c, e, g, i, k); 10 μm (b, d, f, h, j, l).

At this point in our study, it was still unknown whether occludin, ZO-1, N-cadherin, and β-catenin concentrated to or moved away from the Sertoli cell barrier after IL-1α treatment. To address this question, confocal microscopy was performed (Fig. 3). In vehicle-treated Sertoli cells, the junctional complex (i.e., coexisting TJs, ESs, and desmosomes) was present between adjacent Sertoli cells, where occludin and ZO-1 (Fig. 3Ai), as well as N-cadherin and β-catenin (Fig. 3Bi), colocalized strongly. It is worth noting that diffuse immunoreactive signals corresponding to occludin, ZO-1, N-cadherin, and β-catenin were also observed at nonjunctional sites (i.e., apical and basal plasma membrane) in vehicle-treated Sertoli cells, possibly representing a protein pool with no discrete role in junction integrity, except perhaps as a standby reserve. On the other hand, IL-1α-treated cells did not possess extensive junctional complexes (i.e., note that the height of the junctional complex was greatly reduced in IL-1α-treated cells), albeit TJ proteins were still found to localize to the junctional complex (Fig. 3Aj, Bj). These results are in line with previously presented data (Fig. 1A) because IL-1α did not cause TER readings to decline completely to the basal level (Fig. 1A; see values at d 1), illustrating that a few remnant TJs had remained after treatment. Interestingly, virtually all N-cadherin and β-catenin had translocated to the apical plasma membrane, and there was little N-cadherin and β-catenin present between adjacent Sertoli cells (Fig. 3Bj). Also, occludin and ZO-1 were observed in the cytosol in IL-1α-treated cells, suggesting that these proteins may be internalized. Taken collectively, these confocal microscopy results substantiate the decrease in barrier function (Fig. 1A), as well as the disorganization of the actin network (Fig. 1D, E). More important, these findings illustrate that IL-1α can cause protein mislocalization at the Sertoli cell barrier, which is critical for BTB restructuring in vivo.

Figure 3.

IL-1α disrupts Sertoli cell junctions. A, B) Sertoli cells (0.75×10⁶ cells/cm²) were treated with vehicle control (Veh) or IL-1α, as described in Materials and Methods. D, day after treatment. Cells were coimmunostained with occludin (red; Aa, b) and ZO-1 (green; Ac, d) or N-cadherin (green; Ba, b) and β-catenin (red; Bc, d). Corresponding merged images show sites of colocalization (yellow; Ae–h, Be–h). Nuclei were visualized with DAPI (blue; e, f). a–h) Horizontal view of the Sertoli cell epithelium. Each column shows an identical optical slice from the x-y plane (i.e., parallel to the plane of cell attachment; see C), which is approximately halfway through the longitudinal dimension of Sertoli cells. i, j) Vertical view of the Sertoli cell epithelium. Each image is a reconstructed optical slice from the x-z plane (i.e., perpendicular to the plane of cell attachment; see D). Corresponding slice positions on the x-y plane are marked by dotted lines (g, h). Brackets in panels g, h correspond to relative position of brackets in panels i, j, respectively. Brackets also show a thickening in occludin/ZO-1 (Ag vs. h) and N-cadherin/β-catenin (Bg vs. h). Arrowheads indicate mislocalization of TJ and basal ES proteins (j). Approximately 200 cells in each experimental group in 3 independent experiments were analyzed; results are representative of these analyses. C) x-y section. D) x-z section. Scale bars = 8 μm.

IL-1α delayed protein degradation, thereby increasing the cellular levels of BTB-constituent proteins

By immunoblotting, we investigated whether there were any changes in the levels of TJ and ES proteins in Sertoli cells after treatment with IL-1α (Fig. 4A). Among the proteins that we examined, the levels of JAM-A (junctional adhesion molecule-A) and CAR (coxsackie and adenovirus receptor) were not affected by IL-1α treatment (Fig. 4A). However, the levels of occludin and ZO-1, as well as N-cadherin and β-catenin, all showed significant increases after treatment with IL-1α when compared to the vehicle control (Fig. 4). Except for β-catenin, which increased ∼1-fold by 11 h, proteins were up-regulated 2–3 d after IL-1α treatment. Interestingly, this increase in protein levels failed to rescue the leaky TJ permeability barrier brought about by IL-1α (Fig. 1A). To understand why IL-1α caused TJ and ES protein levels to increase in light of a disrupted Sertoli cell barrier (Fig. 1A), the kinetics of occludin (Fig. 5A, B) and N-cadherin (Fig. 5A, C) degradation were examined by protein degradation assays. In brief, Sertoli cell surface proteins were labeled by biotinylation, and cells were treated with IL-1α or vehicle. Nondegraded proteins (i.e., including those that remained on the cell surface, as well as those that had internalized after IL-1α treatment) were then isolated and detected by immunoblotting. After IL-1α treatment, a greater percentage of occludin and N-cadherin remained nondegraded in Sertoli cells as opposed to cells treated with vehicle, illustrating that IL-1α slowed the rate at which these proteins were degraded.

Figure 4.

IL-1α increases the steady-state levels of TJ and basal ES proteins in Sertoli cells in vitro. A) Lysates of Sertoli cells (0.5×10⁶ cells/cm²) treated with vehicle control (Veh) or IL-1α (100 pg/ml) for increasing periods of time were used for immunoblotting. Actin served as the loading control. D, day after treatment. B–E) Bar plots summarizing relative occludin (B), ZO-1 (C), N-cadherin (D), and β-catenin (E) results from several independent experiments after normalizing each data point against its corresponding actin time point and then against its corresponding control at 0 h. Control in both experimental groups was arbitrarily set at 1. Bar plots are not shown for proteins that did not display significant changes. Data points represent means ± sd (n=3–5). *P < 0.05, **P < 0.01 vs. vehicle; 2-way ANOVA followed by Newman-Keuls test.

Figure 5.

IL-1α delays the degradation of BTB integral membrane proteins, resulting in their accumulation in Sertoli cells. Protein degradation assays were performed as described in Materials and Methods. Sertoli cell surface proteins were labeled by biotinylation, and cells were treated with vehicle control (Veh) or IL-1α (100 pg/ml) for increasing periods of time. At termination, biotinylated proteins represented the amount of labeled protein that remained intact (i.e., nondegraded). A) Immunoblots investigating the amount of occludin and N-cadherin that remained nondegraded after IL-1α treatment. Actin served as the loading control. B, C) Graphs summarizing occludin (B) and N-cadherin (C) results from several independent experiments after normalizing each data point against its corresponding actin time point. Relative levels of nondegraded occludin and N-cadherin are expressed as a percentage of total level at 0 min. Data points represent means ± sd (n=3–5). *P < 0.05, **P < 0.01 vs. vehicle; 2-way ANOVA followed by Newman-Keuls test.

IL-1α accelerated the kinetics of occludin endocytosis in Sertoli cells

It was demonstrated that IL-1α slowed the rate at which occludin and N-cadherin were degraded (Fig. 5), causing these proteins to accumulate and to age within Sertoli cells (Fig. 4). It is possible that these aged proteins represented the pool of proteins that concentrated to nonjunctional sites, including the cytosol (Figs. 2 and 3). To address this question, we performed endocytosis assays using Sertoli cells treated with IL-1α or vehicle to assess the kinetics of protein internalization. IL-1α was shown to increase the rate of occludin internalization (i.e., 15–20 min after the addition of IL-1α into Sertoli cell cultures) when compared to the vehicle control (Fig. 6). Taken collectively, these results illustrate that IL-1α forced occludin into early endocytic vesicles, thereby rendering this protein nonfunctional at TJs, but that disorganization of the actin cytoskeleton prohibited the trafficking of occludin to late endosomes or the proteasome for degradation, resulting in its cellular accumulation.

Figure 6.

IL-1α facilitates the endocytosis of occludin in Sertoli cells. Endocytosis assays were performed as described in Materials and Methods. Sertoli cell surface proteins were labeled by biotinylation. Thereafter, endocytosis was initiated at 35°C in the presence of vehicle control (Veh) or IL-1α (100 pg/ml) for increasing periods of time. At termination, biotinylated proteins represented the amount of labeled protein that was endocytosed. A) Immunoblot investigating the amount of occludin that was internalized after IL-1α treatment. Actin served as the loading control. Controls included total biotinylated proteins without stripping as a positive control (total; this sample was diluted 1:4 for the isolation of biotinylated proteins, but no dilution was made for the actin immunoblot); no biotinylation as a negative control (no biotin); and stripping without endocytosis as a negative control (stripped). B) Graph summarizing results from several independent experiments after normalizing each data point against its corresponding actin time point. Relative level of endocytosed occludin at 5 min in vehicle was arbitrarily set as 1. Data points represent means ± sd (n=4–6). **P < 0.01 vs. vehicle; 2-way ANOVA followed by Newman-Keuls test.

DISCUSSION

In the present study, we investigated the role of IL-1α in the restructuring of the Sertoli cell barrier/BTB. We opted to study this cytokine because germ cells are known to control IL-1α production by Sertoli cells, and it is likely that germ cells (especially spermatocytes) regulate BTB restructuring either directly or indirectly. Herein, we show that IL-1α disrupted the Sertoli cell F-actin cytoskeleton in vitro, causing Arp3 to increase and Eps8 to mislocalize. After IL-1α treatment, the TJ protein occludin was also found to move away from the plasma membrane and to travel into the cytosol of Sertoli cells, where it accumulated, most likely as a nonfunctional protein. We arrive at this conclusion because the kinetics of occludin endocytosis and degradation were affected. The end point of these changes was the disassembly of the TJ permeability barrier in Sertoli cells. These observations, when taken together with the previously reported stage-specific expression pattern of IL-1α in the seminiferous epithelium, provide compelling evidence that IL-1α regulates BTB disassembly, in part, during the movement of spermatocytes from the basal to the adluminal compartment.

IL-1α is a regulator of the actin cytoskeleton whose integrity is critical for BTB function

It is well known that cytokines affect junction integrity in epithelia and endothelia, especially during processes such as inflammation, by affecting the contractility or the polymerization of actin filaments (30, 31). Herein, we report a novel mechanism used by IL-1α to disassemble the BTB, which affects the actin cytoskeleton and actin-regulating proteins Eps8 and Arp3. Eps8 is an actin barbed-end capping protein that can also function in actin bundling (32–33), whereas Arp3 is a component of the Arp2/3 protein complex that nucleates barbed ends on secondary actin filaments to form branches, such as those found in lamellipodia and invadopodia of migrating cells (34–35). Previous studies from our laboratory have demonstrated these proteins to have opposing functions during the movement of spermatocytes across the BTB. While Eps8 is needed to stabilize the BTB (28), the Arp2/3 protein complex is required for its restructuring (29). In this study, we demonstrate that IL-1α affects these proteins differentially and that these data seemingly support what occurs in vivo. At stage VIII of the seminiferous epithelial cycle, IL-1 bioactivity increases by ∼4-fold (12, 16) at which time Eps8 diminishes (28) and Arp3 peaks at the BTB, as visualized by immunofluorescence experiments (29). These findings illustrate that IL-1α destabilizes the actin cytoskeleton via Eps8 and Arp3, thereby disassembling the TJ permeability barrier and enabling spermatocytes to move upward toward the tubule lumen. Moreover, these changes are in agreement with earlier Eps8 RNAi experiments performed in Sertoli cells, which showed a disruption in actin filament organization and ZO-1 localization when compared to control cells (28). As such, the down-regulation in Eps8 at the BTB is mediated, at least in part, by germ cells that regulate IL-1α production by Sertoli cells.

IL-1α is unique in its ability to regulate BTB dynamics

Several cytokines, such as TGF-β3 and TNF-α, are known to regulate BTB dynamics via similar mechanisms that involve endocytosis and degradation of proteins (17, 36–41). Herein, we show that IL-1α also disassembles the TJ permeability barrier by internalizing occludin and sequestering it within the cytosol. Moreover, the kinetics of occludin and N-cadherin degradation were slowed after IL-1α treatment, resulting in their accumulation and increase within Sertoli cells, as determined by protein degradation assays and immunoblotting experiments. These data also support the up-regulation of ZO-1 and β-catenin, which bind to occludin and N-cadherin, respectively. Interestingly, none of these changes could reassemble the opened TJ permeability barrier, suggesting that these proteins may have been nonfunctional (i.e., aged) at the Sertoli cell barrier/BTB. We reason that disorganization of the actin cytoskeleton prohibited the trafficking of occludin to late endosomes or the proteasome for degradation. However, no increase in protein levels was observed in our previous in vivo study when recombinant IL-1α was administered intratesticularly (17). This discrepancy in results between the two studies may be due to the presence of germ cells in the testis, the rapid rate of IL-1α clearance in vivo, or the timing at which animals were sacrificed. Nevertheless, in both in vivo and in vitro studies, IL-1α affected the steady-state levels of BTB proteins differently than TGF-β3 and TNF-α. For instance, the latter cytokines were reported to down-regulate occludin, ZO-1, and N-cadherin (36–39), but IL-1α was shown to up-regulate their levels. Early morphological studies have demonstrated that movement of preleptotene spermatocytes across the BTB begins at stage VIII when the “new” BTB below spermatocytes is assembled, while the “old” BTB above spermatocytes is disassembled (1). Interestingly, TGF-β3 and TNF-α expression is restricted largely to the BTB during stages VII-VIII (36–37), suggesting that these cytokines may jump-start the degradation of BTB-constituent proteins prior to spermatocyte movement, whereas IL-1α functions later during the seminiferous epithelial cycle (12). At this point, it is not entirely known whether IL-1α can also facilitate the assembly of the new BTB below migrating preleptotene spermatocytes. Even though an up-regulation in BTB-constituent proteins failed to rescue the leaky TJ permeability barrier brought about by IL-1α, this possibility exists because our in vitro system did not include spermatocytes. Future studies should also investigate whether IL-1α cooperates with androgens in the restructuring of the BTB, since testosterone is known to mediate the recycling of internalized proteins back to the Sertoli cell surface (41). Moreover, since the IL-1α precursor is biologically active (42–43), studies should also investigate how IL-1α processing is regulated within the seminiferous epithelium and how the different forms of IL-1α contribute to the regulation of spermatogenesis.