Cdc42 is involved in NC1 peptide–regulated BTB dynamics through actin and microtubule cytoskeletal reorganization

Abstract

Noncollagenous domain 1 (NC1)-peptide is a biologically active peptide derived from the C-terminal region of collagen α3(IV) chain, a structural constituent protein at the basement membrane in the rat testis, likely via proteolytic cleavage of matrix metalloproteinase 9. Studies have shown that this NC1 peptide regulates testis function by inducing Sertoli cell blood-testis barrier (BTB) remodeling and is also capable of inducing elongate spermatid exfoliation through its disruptive effects on the organization of actin- and microtubule (MT)-based cytoskeletons at these cell adhesion sites. However, the underlying molecular mechanism remains unknown. NC1 peptide was found to exert its biologic effects through an activation of small GTPase cell division control protein 42 homolog (Cdc42) because cooverexpression of the dominant negative mutant of Cdc42 [namely, Cdc42-T17N (via a single mutation of amino acid residue 17 from the N terminus from Thr to Asn by site-directed mutagenesis, making it constitutively inactive)] and NC1 peptide was able to block the NC1 peptide–induced Sertoli cell tight junction–permeability barrier disruption. Their cooverexpression also blocked the NC1 peptide–induced misdistribution of BTB-associated proteins at the cell–cell interface and also disruptive cytoskeletal organization of F-actin and MTs through changes in spatial expression of the corresponding actin and MT regulatory proteins. Interestingly, NC1 peptide was also found to induce an up-regulation of phosphorylated (p)–ribosomal protein S6 (rpS6) (namely, p-rpS6-S235/S236) and a concomitant down-regulation of p–Akt1/2 (namely, p-Akt1-S473 and p-Akt2-S474), but these changes could not be blocked by overexpression of Cdc42-T17N. More importantly, NC1 peptide–induced Cdc42 activation was effectively blocked by treatment of Sertoli cell epithelium with a p-Akt1/2 activator SC79, which is also capable of blocking NC1 peptide–induced down-regulation of p-Akt1-S473 and p-Akt2/S474, but not p-rpS6-S235/S236 up-regulation. In summary, these findings illustrate that Cdc42 is working downstream of the mammalian target of rapamycin complex 1/rpS6/Akt1/2 signaling pathway to support NC1 peptide–mediated effects on Sertoli cell function in the testis using the rat as an animal model.—Su, W., Cheng, C. Y. Cdc42 is involved in NC1 peptide–regulated BTB dynamics through actin and microtubule cytoskeletal reorganization.

In rodent testes, the blood-testis barrier (BTB) undergoes extensive remodeling to facilitate the transport of preleptotene spermatocytes (differentiated from type B spermatogonia residing in the basal compartment of the seminiferous epithelium) across the barrier from late stage VII (when preleptotene spermatocytes first appear) through VIII–IX of the seminiferous epithelial cycle (1–3) through a mechanism that remains to be elucidated. Nonetheless, the immunologic barrier remains uncompromised during BTB remodeling so that cellular events pertinent to meiosis I/II and spermiogenesis can proceed in the specialized microenvironment of the adluminal compartment behind the BTB. It is envisioned that this is possible only if there are biomolecules produced locally in the seminiferous epithelium that coordinate the opening/diassembly of the old BTB above the preleptotene spermatocytes connected in clones via intercellular bridges under transport at the barrier while the new BTB behind these spermatocytes is being assembled (4, 5). Indeed, studies have shown that, using the rat testis as a study model, the seminiferous epithelium is producing several biologically active peptides to modulate BTB dynamics. For instance, it was shown that F5-peptide released from laminin-γ3 chain [a spermatid-specific apical ectoplasmic specialization (ES) adhesion protein (6–8)] at the apical ES, via the action of matrix metalloproteinase 2 (8), is capable of inducing BTB remodeling, making the barrier leaky (7, 9, 10), thereby supporting the transport of preleptotene spermatocytes across the BTB. Furthermore, F5-peptide, which induces BTB opening, is mediated through changes in the distribution and expression of signaling protein p-FAK-Tyr407 downstream (9). On the other hand, studies have shown that another biologically active 80-kDa fragment released at the C-terminal region of laminin-α2 chain, a constituent component of the basement membrane, containing the laminin globular domains 3, 4, and 5 (LG3/4/5, also known as the 80-kDa tail), designated LG3/4/5-peptide, is able to promote BTB function (11, 12), making it tighter. Unlike F5-peptide, LG3/4/5-peptide exerts its effects via the mammalian target of rapamycin complex 1 (mTORC1)/ribosomal protein S6 (rpS6)/protein kinase B (Akt)1/2 signaling pathway downstream (12). These findings illustrate the antagonistic effects of the F5- and LG3/4/5-peptide on the Sertoli cell BTB function, confirming the notion that the testis is capable of producing biomolecules to modulate BTB dynamics to support preleptotene spermatocyte transport at the barrier.

Interestingly, 2 recent studies using the rat testis as a study model have also demonstrated that the basement membrane releases a third biologically active peptide to modulate BTB dynamics known as the noncollagenous domain 1 (NC1) peptide (i.e., NC1 peptide) from collagen α3(IV) chain (13, 14), a major constituent component of the basement membrane (15, 16). NC1 domain is 28-kDa peptide containing ∼230 aa residues located at the C-terminal region of the collagen α3(IV) chain, behind the long collagenous domain of ∼1400 residues of Gly-Xaa-Yaa repeats (17, 18). Similar to F5-peptide, NC1-peptide also promotes BTB remodeling by making the barrier leaky reversibly when it is overexpressed (or through an inclusion of the purified recombinant NC1-peptide) in Sertoli cells cultured in vitro with an established tight junction (TJ) permeability barrier (13, 14). However, unlike the F5- and LG3/4/5-peptides, the signaling proteins or pathways downstream of NC1-peptide in the testis remains unknown. We sought to identify the signaling proteins and the pathways utilized by NC1-peptide to regulate BTB dynamics because this information, if known, will be crucial to design functional experiments to provide mechanistic insights on the concerted effects of these 3 peptides; namely, the F5-, NC1-, and LG3/4/5-peptides to regulate the opening and closing of the BTB during the transport of preleptotene spermatocytes at the BTB in the rat testis.

MATERIALS AND METHODS

Animals and ethics statement

Sprague-Dawley male pups in groups of 10 at 16–18 d of age were obtained from Charles River Laboratories (Wilimington, MA, USA). Each of the 10 male pups were accompanied with a foster mother per cage, and they were housed at the Rockefeller University Comparative Bioscience Center (New York, NY, USA) in a temperature-controlled environment of 20 ± 1°C with a 12-h light/dark cycle and had free access to standard raw chow and water. Rats were kept in accordance with the applicable portions of the Animal Welfare Act and the guidelines in the Guide for the Care and Use of Laboratory Animals [National Institutes of Health (NIH), Bethesda, MD, USA]. Ten pups at 20 d of age were used for isolation of primary Sertoli cells for all the experiments reported herein. The use of rats for all our reported experiment was approved by the Rockefeller University Institutional Animal Care and Use Committee (15-780H and 18-043H). Rats, including pups and foster mothers, were euthanized by CO₂ asphyxiation using slow (20–30%/min) displacement of chamber air from compressed carbon dioxide via a CO₂ regulator connected to a euthanasia chamber approved by the Rockefeller University Laboratory Safety and Environmental Health. The number of animals used in our studies were the minimum required to obtain sufficient data for meaningful statistical analysis.

Antibodies

Antibodies used for various experiments were purchased commercially, which had been characterized earlier from our laboratory and listed in Table 1, including their working dilutions for different applications and the Resource Identification Initiative (RRID) numbers.

TABLE 1.

Antibodies used in different experiments in this study

Antibody (RRID)	Host species	Vendor	Catalog no.	Working dilution
				IB	IF
Collagen NC1-peptide (AB_2800519)	Rabbit	C.Y.C. laboratorya		1:200
Cdc42 (AB_398244)	Mouse	BD Biosciences (San Jose, CA, USA)	610929	1:1000
Cdc42 (AB_1961757)	Rabbit	NewEast Biosciences	21010	1:500
RhoA (AB_1961797)	Rabbit	NewEast Biosciences	21009	1:500
Actin (AB_630836)	Goat	Santa Cruz Biotechnology (Dallas, TX, USA)	sc-1616	1:300
Active Cdc42-GTP (AB_1961759)	Mouse	NewEast Biosciences	26905	IP
Active RhoA-GTP (AB_1961799)	Mouse	NewEast Biosciences	26904	IP
Occludin (AB_2533977)	Rabbit	Thermo Fisher Scientific	71-1500	1:300	1:100
ZO-1 (AB_2533983)	Rabbit	Thermo Fisher Scientific	61-7300	1:300	1:100
CAR (AB_2087557)	Rabbit	Santa Cruz Biotechnology	sc-15405	1:200	1:50
N-cadherin (AB_2313779)	Mouse	Thermo Fisher Scientific	33-3900		1:100
N-cadherin (AB_647794)	Rabbit	Santa Cruz Biotechnology	sc-7939	1:200
β-Catenin (AB_138792)	Mouse	Thermo Fisher Scientific	138400	1:300	1:100
p38 MAPK (AB_330713)	Rabbit	Cell Signaling Technology (Danvers, MA, USA)	9212	1:1000
p-p38 MAPK (AB_331641)	Rabbit	Cell Signaling Technology	9211	1:1000
ERK1/2 (AB_390779)	Rabbit	Cell Signaling Technology	4695	1:1000
p-ERK1/2 (AB_2315112)	Rabbit	Cell Signaling Technology	4370	1:1000
Akt1/2 (AB_329827)	Rabbit	Cell Signaling Technology	9272	1:1000
p-Akt1 T308 (AB_2255933)	Rabbit	Cell Signaling Technology	2965	1:1000
p-Akt1 S473 (AB_2315049)	Rabbit	Cell Signaling Technology	4060	1:1000
p-Akt2 S474 (AB_2630347)	Rabbit	Cell Signaling Technology	8599	1:1000
rpS6 (AB_331355)	Rabbit	Cell Signaling Technology	2217	1:1000
p-rpS6 S235/236 (AB_916156)	Rabbit	Cell Signaling Technology	4858	1:1000
Arp3 (AB_476749)	Mouse	MilliporeSigma	A5979	1:3000	1:50
Eps8 (AB_397544)	Mouse	Thermo Fisher Scientific	610143	1:5000	1:50
Palladin (AB_2158782)	Rabbit	Proteintech (Rosemont, IL, USA)	15492-1-AP	1:1000	1:100
GAPDH (AB_2107448)	Mouse	Abcam (Cambridge, MA, USA)	ab8245	1:1000
MARK4 (AB_2140610)	Rabbit	Proteintech	20174-1-AP	1:2000
EB1 (AB_397891)	Mouse	BD Biosciences	610534	1:1000	1:200
MAP-1A (RRID:AB_649150)	goat	Santa Cruz Biotechnology	sc8969	1:200
α-Tubulin (AB_2241126)	Mouse	Abcam	ab7291	1:1000	1:200
β-Tubulin (AB_2210370)	Rabbit	Abcam	ab6046	1:1000
Detyrosinated α-tubulin (AB_869990)	Rabbit	Abcam	ab48389	1:1000	1:200
Goat IgG-HRP (AB_634811)	Bovine	Santa Cruz Biotechnology	sc-2350	1:3000
Rabbit IgG-HRP (AB_634837)	Bovine	Santa Cruz Biotechnology	sc-2370	1:3000
Mouse IgG-HRP (AB_634824)	Bovine	Santa Cruz Biotechnology	sc-2371	1:3000
Rabbit IgG-Alexa Fluor 488 (AB_2576217)	Goat	Thermo Fisher Scientific	A-11034		1:200
Mouse IgG-Alexa Fluor 488 (AB_2534088)	Goat	Thermo Fisher Scientific	A-11029		1:200

^aWong et al. (14). HRP, horseradish peroxidase.

Primary Sertoli cell cultures

Sertoli cells were isolated from 20-d-old rat testes as previously detailed by Mruk et al. (19). In brief, freshly isolated Sertoli cells were plated on Matrigel (Thermo Fisher Scientific, Waltham, MA, USA) dishes in F12/DMEM (MilliporeSigma, Burlington, MA, USA) supplemented with gentamicin (20 µg/ml), human transferrin (5 µg/ml), bovine insulin (10 µg/ml), epidermal growth factor (EGF) (2.5 ng/ml), and bacitracin (5 µg/ml). For different experiments, Sertoli cells were plated at: 1) 1.2 × 10⁶/cm² on bicameral units (Millicell-HA mixed cellulose esters–based cell culture inserts, 0.45 µm, 12-mm diameter, effective surface area at 0.6 cm²; with units placed in 24-well dishes; MilliporeSigma) with 0.5 ml F12/DMEM (containing supplements) each in the apical and basal compartment for transepithelial electrical resistance (TER) assay; 2) 0.03 × 10⁶/cm² on Matrigel-coated coverslips for immunofluorescence study (to be placed in 12-well dishes containing 2-ml F12/DMEM/well); or 3) 0.4 × 10⁶/cm² on Matrigel-coated 6-well dishes (containing 5 ml F12/DMEM/well) for all other experiments including immunoblot analysis (IB), cell division control protein 42 homolog (Cdc42), or RhoA pull-down assays that quantified the activated forms of Cdc42 or RhoA, actin polymerization kinetics or bundling assays, and microtubule (MT) polymerization and kinetics assays. These Sertoli cells cultured in vitro established a functional TJ permeability barrier that mimicked the Sertoli cell BTB in vivo within 2 d, which has been widely used by investigators in the field as a model system to study Sertoli cell BTB biology (20–25).

Assessment of Sertoli cell TJ permeability barrier function

The Sertoli cell TJ permeability barrier function was assessed by quantifying TER across the Sertoli cell epithelium on a daily basis wherein cells were cultured on Matrigel-coated bicameral units as previously described in refs. 19 and 26. Each experiment contained triplicate or quadruple bicameral units for each treatment vs. control groups. Reported data were representative from a single experiment of n = 3 independent experiments using different batches of Sertoli cells, which yielded similar results.

Transfection of Sertoli cells with plasmid DNA for protein overexpression and treatment with Akt1/2 activator SC79

For overexpression experiments, cDNA encoding NC1 peptide (13, 14) or dominant negative mutant of Cdc42 [i.e., Cdc42-T17N, by replacing amino acid residue Thr17, from the N terminus, with Asn17 by site-directed mutagenesis as previously described by Wong et al. (27)] was cloned into mammalian cell expression vector pCI-neo (Promega, Madison, WI, USA). Cultured primary Sertoli cells were transfected with plasmid DNA at 0.8 μg/10⁶ cells on day 2.5 using LipoJet (SignaGen Laboratories, Rockville, MD, USA) transfection reagent for 24 h essentially as previously described by Chen et al. (28). For cotransfection of 2 genes for their overexpression (namely, NC1 peptide and Cdc42-T17N mutant), a total of 1.0 μg DNA/10⁶ cells (0.5 μg DNA of each target gene/10⁶ cells) was used (based on findings of pilot experiments) with LipoJet reagent for 24 h. Thereafter, Sertoli cells were rinsed and cultured in fresh F12/DMEM for an additional 24 or 48 h for either IB or Cdc42 activation assay/biochemical cytoskeletal assays, respectively. For TER measurement that quantified the integrity of the TJ permeability barrier function across the Sertoli cell epithelium cultured on bicameral units, TER was recorded daily until d 7. For experiments using the Akt1/2 activator SC79 [2-amino-6-chloro-α-cyano-3-(ethoxycarbonyl)-4H-1-benzopyran-4-acetic acid ethyl ester, Mr 364.78, a cell membrane–permeable compound that interacts with Akt plecktrin homology domain phosphatidylinositol-(3,4,5)-trisphosphate (PIP3) binding pocket, making it more susceptible to phosphorylation by upstream kinases to rescue Akt1/2 activity] (29), SC79 was dissolved in DMSO at a working dilution of 25 μg/μl, and a final concentration of 2 μg/μl (5.5 µM) in F12/DMEM was used to treat Sertoli cells for 30 min before and after transfection, respectively, with the same amount of DMSO used in control groups as previously described by Gao et al. (30).

Cdc42 GTPase and RhoA GTPase activity analysis

Activated Cdc42 (i.e., Cdc42-GTP) vs. activated RhoA (i.e., RhoA-GTP) in Sertoli cell lysates were estimated using specific activation assay kits for Cdc42 vs. RhoA obtained from NewEast Biosciences (King of Prussia, PA, USA). This was an antibody-based pull-down assay using the corresponding specific anti-Cdc42-GTP or anti-RhoA-GTP antibody to pull down the activated Cdc42 or RhoA, and the level of Cdc42 or RhoA was then quantified by immunoblotting using corresponding anti-Cdc42 or anti-RhoA antibody (Table 1) according to the manufacturer’s protocol. In brief, primary Sertoli cells cultured at 0.4 × 10⁶/cm² from treatment vs. control groups were lysed and collected in 1× lysis/assay buffer [including protease inhibitors: PMSF (1 mM), leupeptin (10 μg/ml), and aprotinin (10 μg/ml)]. After sonication and centrifugation (12,000 g) at 4°C for 1 h, supernatants were collected on ice for immediate use. For positive or negative controls, after addition of EDTA (20 mM), non- or slowly hydrolyzable G-protein-activating analog of GTP (GTPγS) (100 µM) or 1 mM guanosine diphosphate (GDP; 1 mM) from the Cdc42 or RhoA Activation Assay Kit was added to Sertoli cell protein lysate, respectively, and incubated at 30°C for 30 min with agitation, to be followed by adding MgCl₂ (1 M) to a final concentration of 60 mM on ice. In order to remove any possible nonspecific protein-protein interactions with IgG, a preclearing step was included wherein 0.8–1.0 mg (in 0.5–1.0 ml) protein lysate for each sample was incubated with 2 μg normal mouse IgG for 1 h before incubating with 10 μl Protein A/G agarose beads for another 1 h. Thus, nonspecific interaction proteins were then removed by centrifugation (1000 g) at 4°C and supernatants were collected in and sequentially added with 2 μg configuration-specific antiactive Cdc42 (or antiactive RhoA) monoclonal antibody and 20 μl resuspended Protein A/G agarose beads. Sample tubes were then incubated at 4°C for 1 h with gentle agitation. Thereafter, immunocomplexes that bound beads were pelleted by centrifugation (1000 g) and washed with 0.5 ml 1× lysis/assay buffer (3 times) before being resuspended in 60 μl SDS sample buffer (containing reducing reagent 2-ME) for immunoblotting using an anti-Cdc42 pAb (Table 1).

MT spin-down assay

This assay was used to estimate the relative level of polymerized tubules (i.e., MTs) vs. nonpolymerized (i.e., free tubulins) in Sertoli cell lysates in treatment vs. control groups by using the MT spin-down assay kits (BK038; Cytoskeleton, Denver, CO, USA) according to the manufacture’s protocols. In short, cultured Sertoli cells were collected in a specific regimen and homogenized with a 25-gauge needle using 35°C prewarmed lysis/MT stabilization buffer (100 mM piperazine-N, N′-bis (PIPES), containing 30% glyercol, 0.1% 2-ME, 0.1% Tween-20, 0.1% Nonidet P-40, 0.1% Triton X-100, 1 mM EGTA, 5 mM MgCl₂, pH 6.9, at 22°C) freshly added with 0.2 mM GTP and 1 mM ATP. For positive or negative controls, paclitaxel (Taxol, 30 μM; known to induce MT stabilizing) or CaCl₂ (4 mM; known to induce MT depolymerization), respectively, was added to normal Sertoli cells (prior to their termination) before being subjected to homogenization. Samples were immediately centrifuged at 2000 g for 5 min at 35°C to remove cellular debris before centrifugation at 100,000 g to precipitate MTs (i.e., polymerized tubulins) vs. free tubulin monomers retained in the supernatant. The collected supernatant and the pellet (resuspended in 2 mM CaCl₂) were used for immunoblotting using an anti–β-tubulin antibody (Table 1).

Tubulin polymerization kinetics assay

Sertoli cell lysates in treatment vs. control groups were used to estimate the kinetics of MT polymerization from tubulins using in vitro tubulin polymerization assay kits (BK011P) according to the manufacturer protocols. In brief, Sertoli cells at specified time points as indicated in the regimen were homogenized in the lysis/MT stabilization buffer with a 25-gauge needle, to be followed by centrifugation at 20,800 g for 1 h at 4°C to remove cellular debris. Then 5 μl supernatant containing 10 or 20 μg protein of each sample between treatment and control groups, or 5 μl paclitaxel (Taxol, 3 μM; served positive control), or 5 μl CaCl₂ (0.5 mM; served as negative control) was incubated with 50 μl tubulin reaction mix in a 96-well black polystyrene microplate with flat (black)-bottom dishes. Polymerization was allowed to take place, which was estimated by an increase in fluorescence intensity in each well due to the incorporation of a fluorescent reporter into MTs during polymerization. The fluorescence kinetics were monitored in a FilterMax F5 Multi-Mode Microplate Reader via fluorescence detection from the top, and data were acquired by Multi-Mode Analysis Software 3.4 at 37°C using the following settings: measurement type, kinetics, 100 cycles, 20-s intervals; excitation wavelength, 360 nm; emission wavelength, 430 nm. Linear regression analysis was then performed at the intervals of exponential increase in fluorescence intensity to estimate the tubulin polymerization rate by comparing treatment and control groups.

F-actin spin-down assay and polymerization assays

The assays to monitor the relative levels of polymerized actin filaments in Sertoli cell lysates (i.e., bundles of actin filaments) vs. nonpolymerized actins (i.e., unbunbled actin filaments or free globular actins) and the rate (i.e., kinetics) of actin polymerization between treatment and control groups were performed as previously described in refs. 31 and 32.

Immunofluorescence analysis, MT, and F-actin staining

Immunofluorescence microscopy (IF) and F-actin staining were performed as previously described in refs. 26 and 30. MTs were stained using a specific antibody against α-tubulin (Table 1), which together with β-tubulin create the α-/β-tubulin oligomers, which are the building blocks of MTs. In brief, Sertoli cells cultured on coverslips were fixed in either 4% paraformaldehyde (in PBS, 10 mM sodium phosphate, 0.15 M NaCl, pH 7.4, at 22°C) or ice-cold methanol for 10 min, permeabilized in 0.1% Triton X-100, and blocked in goat serum (10%) or bovine serum albumin (5%) (in PBS). Cells were than used for incubation with the corresponding primary antibodies and the corresponding secondary antibodies (Table 1). Cell nuclei were visualized by staining with DAPI. Slides were mounted in Prolong Gold Antifade reagent (Thermo Fisher Scientific). For F-actin staining, cells were incubated with Alexa Fluor 488-phalloidin (green fluorescence) (Thermo Fisher Scientific). Fluorescence images were visualized using a Nikon Eclipse 90i Fluorescence Microscope system equipped with both Nikon Ds-Qi1Mc and DsFi1 digital cameras. Images were acquired using Nikon NIS Elements AR 3.2 software package (Nikon, Tokyo, Japan) and saved in TIFF format. All images shown were untouched original captured images, and image overlaps were performed using Adobe Photoshop CS6 (Adobe, San Jose, CA, USA). All images were representative findings from a typical experiment from n = 3 independent experiments, which yielded similar results. All coverslips from a single experiment including both treatment and control groups were processed simultaneously to avoid interexperimental variations.

Sertoli cell lysate preparation, protein assay, and IB

Sertoli cell cultures at specific time points were lysed in immunoprecipitation (IP)/lysis buffer [50 mM Tris, containing 0.15 M NaCl, 1% Nonidet P-40, 1 mM EGTA, 10% glycerol, pH 7.4 at 22°C, freshly supplemented with protease inhibitor mixture (MilliporeSigma) and phosphatase inhibitor cocktail II (MilliporeSigma)] to obtain cell lysates as previously described in refs. 26 and 30, which were then used for IB to assess changes in the steady-state levels of different target proteins (Table 1) and corresponding biochemical assays. Protein estimation was performed using Bio-Rad DC protein assay kits (Bio-Rad, Hercules, CA, USA) with bovine serum albumin as a standard, and protein concentration was obtained by spectrophotometry using a Bio-Rad Model 680 Plate Reader. IB was performed using 40 µg protein per sample as described by ECL (33) using an ImageQuant LAS 4000 Mini (GE Healthcare, Waukesha, WI, USA) Imaging System and ImageQuant Software package (v.1.3). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and β-actin served as protein loading controls. Intensities of IB protein bands were quantified by ImageJ 1.45s software (NIH; http://rsbweb.nih.gov/ij). All samples within an experimental group were processed simultaneously to avoid interexperimental variations. Each sample had triplicate culture wells, dishes, and coverslips in both treatment and control groups from n = 3 independent experiments.

Statistical analysis

Each experiment including treatment and control groups had triplicate cultures from n = 3 independent experiments, which yielded similar results. For bar graphs, each data point is a mean ± sd of n = 3 experiments. Data were subjected to either Student’s t test (for 2-group comparisons) or 2-way ANOVA (for multigroup comparisons) using the repeated-measures model followed by Dunnett’s test to compare changes between treatment vs. control groups by assessing within-experimental effects, which were the focus of the analysis. Statistical analysis was performed using the GB-STAT statistical analysis software package (v. 7.0; Dynamic Microsystems, Silver Spring, MD, USA). Values of P < 0.05 were taken as statistically significant.

RESULTS

NC1 peptide–induced Sertoli cell TJ barrier disruption is mediated through an activation of Cdc42

The regimen used for the study that examined the involvement of Cdc42, an Rho GTPase, in NC1 peptide–induced BTB disruption is shown in Fig. 1A. In brief, Sertoli cells isolated from 20-d-old rat testes and cultured in F12/DMEM in vitro was used on d 2.5 for transfection for overexpression of NC1 peptide (13, 14). Overexpression of NC1 peptide in Sertoli cells following transfection was confirmed by IB as shown in Fig. 1B. However, the steady-state level of either Cdc42 or RhoA, 2 members of the Rho GTPase family, earlier shown to modulate F-actin (34, 35), was not induced (Fig. 1B). Interestingly, when the steady-state levels of the activated form of RhoA and Cdc42 were assessed, activated Cdc42 was considerably induced, but not RhoA, following overexpression of NC1 peptide (Fig. 1C, see also composite data on the bar graph). The induction of activated Cdc42 level by NC1 peptide was shown to be specific, as noted in the next 2 experiments. First, overexpression of a dominant negative mutant of Cdc42 [namely, Cdc42-T17N by mutating amino acid residue Thr17 to Asn17 to make it constitutively inactive (27)] in Sertoli cells (or following its cotransfection with pCI-neo/NC1 peptide) was confirmed (Fig. 1D); and its cotransfection with pCI-neo/NC1 peptide did not perturb the expression of either NC1 peptide or Cdc42 (Fig. 1D). Second, overexpression of the dominant negative mutant of Cdc42 (pCI-neo/Cdc42-T17N) with NC1 peptide was able to eliminate the NC1 peptide–induced activated Cdc42 expression (Fig. 1E, see also composite data in the bar graph in lower panel), illustrating the involvement of Cdc42 in NC1 peptide–mediated effects in Sertoli cell function. This notion was next supported by data shown in Fig. 1F, wherein NC1 peptide–induced Sertoli cell TJ permeability barrier disruption following NC1 peptide overexpression in Sertoli cells was blocked by cotransfection of Sertoli cells with pCI-neo/Cdc42-T17N mutant and pCI-neo/NC1 peptide.

Figure 1.

NC1-peptide–induced Sertoli cell TJ barrier disruption is mediated through Cdc42, which can be blocked by overexpression of Cdc42-T17N, a dominant negative mutant of Cdc42, in Sertoli cell epithelium. A) Regimens used for this study with Sertoli cells cultured in vitro at 0.4 × 10⁶ cells/cm² that formed an intact epithelium (see Materials and Methods) for different experiments in this report. B) A study by IB that confirmed overexpression of NC1 peptide following its overexpression in Sertoli cells without affecting the steady-state levels of Cdc42 or RhoA. β-Actin served as the protein loading control. Composite data of IB were shown on the right panel, wherein each bar is the mean ± sd of n = 3 experiments. Each experiment had triplicate cultures. **P < 0.01 by ANOVA. C) A pull-down assay was used to assess an activation of Cdc42 (i.e., GTP-bound Cdc42), but not RhoA, following overexpression of NC1 peptide in Sertoli cells with the corresponding negative (−ve) and positive (+ve) controls. N.s., not significantly different. Composite data of n = 3 experiments were shown on the right panel. **P < 0.01 by ANOVA. D) Overexpression of Cdc42-T17N mutant (a dominant negative mutant of Cdc42 by mutating residue Thr17 to Asn17) was confirmed by IB; however, its overexpression alone or cooverexpression with NC1 peptide did not alter the steady-state level of either NC1 peptide or Cdc42. The composite data of n = 3 independent experiments of this study were shown on the right panel. E) Overexpression of Cdc42-T17N mutant was able to eliminate the NC1 peptide–induced Cdc42 activation by quantifying the steady-state level of activated Cdc42 (i.e., Cdc432-GTP) by pull-down assays as noted by IB, including both −ve and +ve controls, with the composite data of n = 3 experiments in the lower panel. **P < 0.01 by ANOVA. F) A study by TER that quantified the Sertoli cell TJ permeability barrier function wherein overexpression of Cdc42-T17N mutant abolished the NC1 peptide–induced Sertoli cell TJ barrier disruption from a representative experiment with n = 4 bicameral units. Three additional experiments using different batches of Sertoli cells yielded similar results. **P < 0.01 by ANOVA by comparing to the other 3 groups. Uncropped blots corresponding to blots in B–E are shown in Supplemental Fig. S1. Ctrl, control.

NC1 peptide–induced Sertoli cell TJ disruption involves the mTORC1/rpS6/Akt1/2 signaling pathway, independent of Cdc42-mediated changes in the distribution of TJ- and basal ES proteins at the Sertoli cell BTB

We next delineated the signaling proteins/pathways that worked in concert with Cdc42 to mediate the effects of NC1 peptide by perturbing Sertoli cell BTB function (shown in Fig. 2). Using the regimen shown in Fig. 1A, wherein Sertoli cells were transfected with either NC1 peptide or Cdc42-T17N mutant alone or cotransfected with NC1 peptide and Cdc42-T17N mutant for IB (Fig. 2A), overexpression of NC1 peptide or Cdc42-T17N mutant vs. NC1 peptide + Cdc42-T17N mutant in different groups was confirmed following overexpression of the corresponding plasmid DNA (see also composite data of IB in bar graphs in the lower panel of Fig. 2A). The expression of BTB-associated adhesion protein complexes of the TJ [e.g., occludin/zonula occludens-1 (ZO-1), coxsackie and adenovirus receptor (CAR)/ZO-1] and basal ES (e.g., N-cadherin/β-catenin), as well as members of the MAPKs (e.g., p38, ERK1/2, and their activated/phosphorylated forms), were unaffected by either treatment (Fig. 2A). Interestingly, a considerable up-regulation of phosphorylated (p)-rpS6-235/S236 but not total rpS6 [note: rpS6 is the downstream signaling protein of mTORC1 and a phosphorylatable protein tranlation regulator (36)] and a down-regulation of p-Akt1-S473 and p-Akt2-S474 (also a downstream signaling protein of the mTORC1/rpS6 signaling complex) (31, 37, but not p-Akt1-T308 (nor total Akt1/2/), were noted (Fig. 2A, see also composite data in bar graph in the lower panel) following overexpression of NC1 peptide. In this context, it is of interest to note that these changes, (namely, an up-regulation of p-rpS6 and a down-regulation of p-Akt1/2) were earlier shown to be involved in perturbing the Sertoli cell TJ barrier function in vitro (31, 37) and in vivo, as well as other spermatogenic function (e.g., germ cell adhesion, spermatid polarity) in vivo including the use of genetic model (38, 39). However, the use of Cdc42-T17N mutant failed to interfere with the NC1 peptide–induced p-rpS6-S235/S236 up-regulation or the p-Akt1-S473 and p-Akt2-S474 down-regulation (Fig. 2A), suggesting that Cdc42 was working downstream of the mTORC1/rpS6/Akt1/2 signaling pathway (or independent of the mTORC1/rpS6/Akt1/2 pathway) because the mTORC1/rpS6/Akt1/2 signaling pathway was activated by NC1 peptide, independent of the Cdc42-mediated pathway. Additionally, NC1 peptide, which induced disruptive changes on the distribution of TJ (e.g., occludin, CAR, ZO-1) and basal ES (e.g., N-cadherin, β-catenin) proteins at the Sertoli cell–cell interface, as noted by IF, were being blocked by overexpression of the Cdc42-T17N mutant, whereas Cdc42-T17N mutant overexpression per se had no notable effects on the distribution of these BTB-associated proteins (Fig. 2B).

Figure 2.

NC1 peptide–induced Sertoli cell TJ barrier disruption involves the mTORC1/rpS6/Akt1/2 signaling pathway, independent of the Cdc42-mediated changes in protein distribution at the Sertoli cell–cell interface. A) Results of IB analysis using the regimen noted in Fig. 1A from the 4 different groups (G), confirming overexpression of NC1 peptide vs. Cdc42-T17N mutant alone and NC1 peptide + Cdc42-T17N mutant (cotransfection). The treatment groups did not affect the steady-state levels of TJ and basal ES proteins or members of the MAPKs; namely, p38, ERK1/2, and their activated/phosphorylated forms. However, an activation of p-rpS6-S235/S236 (the activated form of rpS6), but not the total rpS6, was noted following NC1 peptide overexpression but not Cdc42-T17N mutant, concomitant with a down-regulation of p-Akt1-S473 and p-Akt2-S474 but not total Akts or p-Akt1-T308. However, overexpression of Cdc42-T17N mutant failed to interfere with the NC1 peptide–induced up-regulation of p-rpS6 or down-regulation of p-Akt1/2. Notable changes in the expression of signaling proteins following NC1 peptide overexpression with or without Cdc42-T17N cooverexpression as noted by IB were boxed in red. See also composite down in the bar graphs in lower panel, with each bar representing the mean ± sd of n = 3 experiments. Uncropped blots corresponding to blots are shown in Supplemental Fig. S2. **P < 0.01 by ANOVA. B) A study by IF showed that NC1 peptide–induced misdistribution of TJ (e.g., occludin, CAR, ZO-1) and basal ES (e.g., N-cadherin, β-catenin) proteins, but overexpression of Cdc42-T17N alone had no effects. However, Cdc42-T17N was capable of blocking the NC1 peptide–induced protein misdistribution at the Sertoli cell–cell interface. Scale bar, 40 µm, which applies to all other micrographs.

NC1 peptide–induced disruptive changes in actin cytoskeletal organization in Sertoli cells are blocked by overexpression of Cdc42-T17N mutant

Because both TJ and basal ES adhesion protein utilize F-actin for their attachment, and NC1 peptide was recently shown to perturb BTB and spermatogenic function through changes in actin-based cytoskeletal organization, we next examined whether cotransfection of Sertoli cells with pCI-neo/Cdc42-T17N mutant and pCI-neo/NC1 peptide could also block the NC1 peptide–induced changes in cytoskeletal organization. Using the regimen shown in Fig. 1A, we examined any changes in the expression of the actin regulatory proteins by IB following different treatment (Fig. 3A). Overexpression of either NC1 peptide, Cdc42-T17N mutant or NC1 peptide + Cdc42-T17N had no apparent effects on the expression of actin-related protein (Arp) 3 [which together with Arp2 creates the Arp2/3 complex known to induce branched actin polymerization, converting bundled actin filaments into a branched network (40), earlier found in the testis (41)] (Fig. 3A). However, overexpression of NC1 peptide down-regulated the expression of epidermal growth factor receptor pathway substrate 8 (Eps8), an actin barbed-end capping and bundling protein in the testis (42, 43), but not paladin, an actin cross-linking/bundling protein in the testis (44–46), which was blocked by cotransfection of Sertoli cells with Cdc42-T17N mutant and NC1 peptide (Fig. 3A; see also composite bar graph data shown on the right panel). We next examined the organization of actin filaments across the Sertoli cell cytosol as noted in Fig. 3B. In control cells transfected with empty vector (pCI-neo/Ctrl), actin filaments stretched across the entire cell cytosol, similar to cells transfected with Cdc42-T17N mutant alone (Fig. 3B). However, overexpression of NC1 peptide was found to perturb the Sertoli cell TJ barrier function (Fig. 1F) as well as the distribution of TJ and basal ES proteins at the Sertoli cell–cell interface (Fig. 2B), also capable of inducing gross cytoskeletal organization of actin filaments (Fig. 3B). For instance, actin filaments no longer bundled tightly and stretched across the entire cell cytosol, as noted in both the control and pCI-neo/Cdc42-T17N mutant groups, following NC1 peptide overexpression (Fig. 3B). These changes thus contributed to a disruption of the Sertoli cell TJ permeability barrier function as noted in Fig. 1F. These changes in actin filament bundles across the Sertoli cells perhaps were due to a down-regulation of the actin barbed-end capping and bundling protein Eps8 (Fig. 3A; see also composite data on the right panel vs. Fig. 3B). Using a biochemical-based assay that monitored the relative ability of cell lysate to polymerize actin filaments into bundles, it was noted that overexpression of pCI-neo/NC1 peptide in Sertoli cells considerably down-regulated the ability of cells to polymerize actin filaments into bundles, but cotransfections with Cdc42-T17N and NC1 peptide was able to block the disruptive effects of NC1 peptide on actin filament bundling activity (Fig. 3C). This finding is also consistent with results of another biochemical-based assay by monitoring the kinetics of actin polymerization using Sertoli cell lysates, wherein overexpression of NC1 peptide was able to down-regulate the kinetics of actin polymerization activity in Sertoli cell lysates, but this NC1 peptide–induced disruptive effect was also blocked by Cdc42-T17N mutant following cotransfections of this mutant with NC1 peptide (Fig. 3D).

Figure 3.

NC1 peptide–induced Sertoli cell TJ barrier disruption involves Cdc42-mediated disruptive changes in F-actin organization, actin microfilament bundling, and polymerization kinetics, which can be blocked by overexpression of Cdc42-T17N mutant. A) Regimen noted in Fig. 1A was used for this study by IB. Overexpression of NC1 peptide was found to induce a down-regulation of Eps8 expression (boxed in red) but had no apparent effect on the branched actin nucleation protein Arp3 nor the actin bundling protein palladin. Although Cdc42-T17N overexpression alone had no effect on the steady-state level of Eps8, its cooverexpression with NC1 peptide was capable of blocking the NC1 peptide–induced down-regulation of Eps8. Composite data of bar graphs were shown on the right panel, wherein each bar represents the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. B) IF analysis of F-actin (green fluorescence) in Sertoli cells wherein actin filaments stretched across the Sertoli cell cytosol in the control and Cdc42-T17N group, whereas NC1 peptide induced considerably disorganization of actin filaments in which they no longer appropriately bundled and stretched across the cell cytosol as noted in control cells. However, overexpression of Cdc42-T17N blocked the NC1 peptide–induced disorganization of actin filaments. Consistent with findings noted in A, there was a considerable down-regulation and misdistribution of actin barbed-end capping and bundling protein Eps8 in Sertoli cells following NC1 peptide overexpression, which were corrected by cotransfection of Sertoli with Cdc42-T17N. Scale bar, 40 µm, which applies to other micrographs. C) A study by biochemical assay to assess the actin filament bundling activity in Sertoli cells wherein NC1 peptide considerably perturbed actin bundling activity which was blocked by cotransfection with Cdc42-T17N mutant, with β-actin and GAPDH serving as the protein loading controls. The bar graph below is the composite data of n = 3 experiments. Uncropped blots corresponding to blots in A and C are shown in Supplemental Fig. S3. D) A study that assessed the actin polymerization kinetics as detailed in Materials and Methods. Indeed, NC1 peptide considerably perturbed actin polymerization (left panel) and its kinetics (middle panel), which was assessed during the first 20 min of the assay (i.e., exponential phase) and noted in the composite data (right panel), wherein each bar represents the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. +ve, positive; –ve, negative; Ctrl, control; n.s., not significantly different.

NC1 peptide–induced disruptive changes in MT cytoskeletal organization in Sertoli cells can be blocked by overexpression of Cdc42-T17N mutant

In this context, it is of interest to investigate whether Cdc42 is also involved in NC1 peptide–mediated changes in MT dynamics. It is noted that MT affinity-regulating kinase 4 (MARK4) is a Ser/Thr protein kinase found in the rat testis (47) that is known to phosphorylate MT-associated proteins (MAPs, such as MAP-1a, which promotes MT stabilization by binding onto tubulins), which leads to MAP detachment from MTs, thereby causing MT catastrophe (48, 49). On the other hand, end-binding protein 1 (EB1) [an MT plus-end tracking protein (+TIP)] is known to stabilize MTs in mammalian cells (50), including Sertoli cells in the testis (51). Moreover, detyrosinated α-tubulin formed by removing C-terminal Tyr from α-tubulin by exposing Glu at the newly formed C terminus, which is known to render MT less dynamic by conferring MT stability (52). Using the regiment shown in Fig. 1A, overexpression of NC1 peptide induced down-regulation on the steady-state level of EB1 and detyrosinated α-tubulin but not MARK4 and MAP-1a (Fig. 4A, see also composite data summarized in bar graphs on the right panels). However, the NC1 peptide–induced down-regulation of EB1 and detyrosinated α-tubulin was blocked by Cdc42-T17N mutant (Fig. 4A), supporting the notion that NC1 peptide is working in concert with Cdc42 to exert its regulatory effect on MT organization. These changes, including a down-regulation of EB1 and detryosinated α-tubulin induced by NC1 peptide, thus contributed to MT destabilization as noted in Fig. 4B. In control cells (pCI-neo/Ctrl) or cells transfected with pCI-neo/Cdc432-T17N mutant alone, MTs (visualized by α-tubulin staining, because α- and β-tubulin form the α-/β-tubulin dimers, which are the building blocks of MTs) stretched across the Sertoli cell cytosol to support cell shape and cell function (Fig. 4B). Also, detyrosinated α-tubulin and EB1, which are known to promote MT stabilization, also stretched out across the cell cytosol (Fig. 4B). However, overexpression of NC1 peptide in Sertoli cells caused truncation of MTs wherein MTs wrapped around the cell periphery instead of stretching out across the cell cytosol (Fig. 4B). Meanwhile, signals of both detyrosinated α-tubulin and EB1 became considerably weakened after NC1 peptide overexpression and also retracted from cell cytosol and moved close to the cell nuclei (Fig. 4B). These changes, noted in Fig. 4B, thus contributed to Sertoli cell TJ barrier disruption as noted in Fig. 1F. Cotransfection of Sertoli cells with Cdc42-T17N mutant and NC1 peptide was found to block the NC1 peptide–induced MT cytoskeletal disorganization as noted in the last column in Fig. 4B, possibly mediated by proper distribution of detyrosinated and EB1 (and also their proper expression, see Fig. 4A). Using a biochemical assay to monitor the relative level of polymerized MTs in Sertoli cell lysates following various treatment vs. control cells (pCI-neo/Ctrl), it was noted that the overexpression of NC1 peptide in Sertoli cells indeed perturb MT polymerization, which was blocked by cotransfection of Sertoli cells with Cdc42-T17N mutant (Fig. 4C, see also the composite data shown in the bar graph in the lower panel). We next used another biochemical assay that monitored the kinetics of MT polymerization as noted in Fig. 4D. Consistent with the MT spin-down assay that assessed the relative level of polymerized MTs, NC1 peptide was indeed found to perturb the kinetics of MT polymerization considerably, as noted in the bar graph on the right panel, when the polymerization rate was assessed in real time and compared between control and treatment groups (Fig. 4D). However, overexpression of Cdc42-T17N mutant was able to block the NC1 peptide–induced disruption on MT polymerization kinetics (Fig. 4D).

Figure 4.

NC1 peptide perturbs MT cytoskeletal organization, which can be blocked by Cdc42-T17N mutant. In this study, we sought to determine whether the NC1 peptide–induced BTB dysfunction through its effects on cytoskeletal organization is a Cdc42-dependent event. A) Using the regimen in Fig. 1A for IB analysis, NC1 peptide was found to down-regulate the expression of EB1 and detyrosinated α-tubulin, but not MARK4 and MAP-1a, which were blocked by Cdc42-T17N mutant (see IB, boxed in red). Composite data are shown in the histograms, with each bar representing the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. B) NC1 peptide was also found to perturb the organization of MTs (visualized by staining with α-tubulin, which together with β-tubulin create the α/β-tubulin dimers, which are the building blocks of MTs), detyrosinated α-tubulin [the stabilized form of MTs (73)], and EB1 [known to promote MT stabilization (50)], considerably different from the control group (and overexpression of Cdc42-T17N mutant had no effects on their distribution). However, these disruptive changes were corrected by cotransfecting cells with NC1 peptide and Cdc42-T17N mutant. Representative findings of an experiment and a total of n = 3 experiments yielded similar results. Scale bar, 40 µm, which applies to other micrographs. C) Results of a study to assess the ability of the Sertoli cell lysates to polymerize MTs including negative (CaCl₂, 2 mM) and positive (Taxol, 20 µM) controls known to induce MT depolymerization and confer MT stabilization, respectively. NC1 peptide was found to considerably perturb Sertoli cell MT polymerization (but not Cdc42-T17N mutant alone), which was blocked by Cdc42-T17N mutant. See also composite histogram data below, with each bar representing the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. Uncropped blots corresponding to blots in A and C are shown in Supplemental Fig. S4. D) Results of the MT polymerization assay to assess an increase in polymerization rate (over time) in the corresponding control and treatment groups vs. the negative and positive control groups (left panel). Histogram on the right panel illustrates the relative polymerization rate based on analysis of the kinetics data during the 20 min of the polymerization assay (left panel) with the pCI-neo/control (Ctrl) group arbitrarily set at 1 (see Materials and Methods). Each bar is the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. +ve, positive; –ve, negative; n.s., not significantly different.

NC1 peptide–induced disruptive changes in TJ and basal ES protein distribution at the Sertoli cell–cell interface can be blocked by SC79, an activator of Akt1/2

We next examined the involvement of rpS6 and Akt1/2 in NC1 peptide–mediated changes in protein distribution at the Sertoli cell–cell interface that in turn supported the barrier function by using SC79, a specific activator of Akt1/2 (29). We sought to examine the physiologic relationship between Cdc42 and the mTORC1/rpS6/Akt1/2 signaling pathway to support the NC1 peptide–mediated effects on BTB dynamics. Using the regimen shown in Fig. 5A, it was found that NC1 peptide was able to induce an activation of Cdc42 (i.e., Cdc42-GTP) by using a Cdc42-specific pull-down assay (Fig. 5B), consistent with data shown in Fig. 1C. Interestingly, SC79 was capable of blocking NC1 peptide–induced Cdc42 activation (Fig. 5B), illustrating p-Akt1/2 is working in concert with Cdc42 to modulate the effects of NC1 peptide at the Sertoli cell BTB. This notion was further supported by findings shown in Fig. 5C, wherein NC1 peptide was found to induce disruptive distribution of TJ (e.g., occludin, CAR, and ZO-1) and basal ES (e.g., N-cadherin, β-catenin) proteins at the Sertoli cell–cell interface such that these proteins no longer tightly localized at the site to support the BTB; instead, they were internalized into the cell cytosol (Fig. 5C). However, treatment of cells with SC79 was capable of retaining these proteins at the cell–cell interface, even with NC1 peptide overexpressed in these Sertoli cells (Fig. 5C). Interestingly, a study by IB showed that the use of SC79 was also capable of blocking the NC1 peptide–induced down-regulation of p-Akt1-S473 and p-Akt2-S474 expression (Fig. 5D), but it had no effect on the NC1 peptide–induced up-regulation of p-rpS6-S235/S236 (Fig. 5D). Collectively, these findings suggest that Cdc42 was working downstream of the mTORC1/rpS6/Akt1/2 signaling pathway such that an activation of p-rpS6 (i.e., up-regulation of p-rpS6-S235/S236) by NC1 peptide that associated with an inactivation of p-Akt1/2 (i.e., down-regulation of p-Akt1-S473 and p-Akt2-S474) can be blocked by using a chemical-based Akt1/2 activator SC79 (i.e., by up-regulating p-Akt1-S473 and p-Akt2-S474) without affecting p-rpS6-S235/S236 expression upstream (Fig. 5C, D). This conclusion was supported by findings in which the use of SC79 that yielded the phenotypes noted in Fig. 5C were similar to the use of Cdc42-T17N mutant as shown in Fig. 2B.

Figure 5.

NC1 peptide that induces Sertoli cell BTB function is mediated through Cdc42/p-Akt1/2 pathway but involves independent action of the mTORC1/p-rpS6 signaling complex. This study sought to further examine whether the p-rpS6-S235/S236 up-regulation and the p-Akt1/2 down-regulation induced by NC1 peptide are working in concert or independently by using SC79 (29), a specific Akt1/2 activator. A) Regimen used for different analyses in this study with n = 3 experiment wherein each experiment had triplicate dishes. B) In this representative pull-down assay that assessed the steady-state level of activated Cdc42 (i.e., GTP-bound Cdc42), NC1 peptide was capable of inducing activated Cdc42 level. Although SC79 itself had no effects on the activated Cdc42 level, it was able to block the NC1 peptide–induced Cdc42 activation, consistent with the findings shown in Fig. 2 regarding the involvement of p-Akt1/2 signaling protein downstream of Cdc42 during NC1 peptide–induced Sertoli cell BTB disruption. Positive (+ve) and negative (−ve) controls shown herein confirmed the validity of this assay. Bar graph on the right panel is the composite data, wherein each bar is the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. C) In controls (pCI-neo/Ctrl), TJ (e.g., occludin, CAR, ZO-1) (green fluorescence) and basal ES (e.g., N-cadherin, β-catenin) (green fluorescence) proteins are tightly localized at the Sertoli cell–cell interface to confer Sertoli TJ barrier function. However, overexpression of NC1 peptide induced redistribution of these proteins, causing their internalization via endocytosis. Although SC79 per se had no effects on the distribution of these proteins, SC79 was able to block the NC1 peptide–induced redistribution of these BTB-associated proteins, making them similar to control cells, confirming the notion that the p-Akt1/2 signaling protein is downstream of NC1 peptide–induced Cdc42 activation. Representative findings of an experiment from n = 3 experiments, which yielded similar results. Scale bar, 40 µm, which applies to other micrographs. D) IB analysis of samples from the 3 treatment groups vs. the control (pCI-neo/Ctrl) group using about 40 µg protein per lane. NC1 peptide overexpression was confirmed, and the steady-state levels of total Cdc42 and BTB-associated protein were not perturbed in treatment groups. However, NC1 peptide induced an up-regulation of p-rpS6-S235/S236 (but not total rpS6) and a down-regulation of p-Akt1-S473 and p-Akt2-S474 (but not p-Akt1-T308 or total Akt1/2), similar to data shown in Fig. 2. However, unlike Cdc42-T17N mutant, SC79 was able to block NC1 peptide–induced Akt1/2 down-regulation but not p-rpS6-S235/S236 up-regulation, illustrating that p-Akt1/2 signaling protein and p-rpS6 were working independent to one another downstream of NC1 peptide–mediated effects on Sertoli cell function. Notable changes in the expression of signaling proteins following NC1 peptide overexpression with or without treatment of SC79 as noted by IB were boxed in red. β-Actin and GAPDH served as protein loading controls. Bar graphs in the bottom panel are composite data of the IB shown in the top panel, with each bar representing the mean ± sd of n = 3 experiments. **P < 0.01 by ANOVA. Uncropped blots corresponding to blots in B and D shown in Supplemental Fig. S5. G, group; n.s., not significantly different.

NC1 peptide–induced disruptive changes on the organization of actin- and MT-based cytoskeletons can be blocked by the Akt1/2 activator SC79

We next confirmed the notion summarized above regarding the NC1-mTORC1/rpS6/Akt1/2-Cdc42 regulatory pathway. Using the regimen shown in Fig. 5A, NC1 peptide was found to down-regulate the expression of actin barbed-end capping and bundling protein Eps8 but not actin-cross-linking/bundling protein palladin or branched actin polymerization protein Arp3 (Fig. 6A), consistent with the data shown in Fig. 3A. It was noted that SC79, an Akt1/2 activator, was capable of blocking the NC1 peptide–induced Eps8 down-regulation (Fig. 6A). More importantly, SC79, which blocked the down-regulation of p-Akt1/2, as noted in Fig. 5D, was also able to rescue Sertoli cells from the NC1 peptide–induced actin cytoskeletal disorganization and capable of restoring proper distribution of Eps8 at the Sertoli cell–cell interface, which was perturbed by NC1 peptide (Fig. 6B). Furthermore, NC1 peptide was found to down-regulate EB1 and detyrosinated α-tubulin but had no effects on MARK4 or MAP-1a (Fig. 6A), consistent with data shown in Fig. 4A. Importantly, treatment of Sertoli cells with SC79 was capable of blocking the NC1 peptide–mediated down-regulation of EB1 and detyrosinated α-tubulin (Fig. 6A). Furthermore, these effects of SC79 on MT binding/regulatory protein expression noted by IB and shown in Fig. 6A were consistent with findings when the organization of MTs across the Sertoli cell cytosol was examined (Fig. 6B). It was shown that the NC1 peptide–induced MT cytoskeletal disorganization was also blocked by cotreatment of Sertoli cells with SC79 and overexpression of NC1 peptide, including distribution of detyrosinated α-tubulin and EB1 (Fig. 6B). Based on the findings, Fig. 7 summarizes the cascade of signaling events mediated by NC1 peptide that modulates Sertoli cell BTB function through changes in the organization of actin- and MT-based cytoskeletons.

Figure 6.

NC1 peptide–induced Sertoli cell cytoskeletal disorganization is mediated through the Cdc42/p-Akt1/2 pathway. We next sought to examine whether the use of SC79 to activate p-Akt1/2 signaling protein that was found to restore the distribution of BTB-associated proteins at the Sertoli cell–cell interface was also capable of restoring actin- and MT-based cytoskeletal organization through corrective expression of actin and MT regulatory proteins. The regimen shown in Fig. 5A was used for this study. A) Representative findings of an IB experiment is shown, and the composite data are summarized in histograms shown in the lower panel, with each bar representing the mean ± sd of n = 3 experiments. N.s., not significantly different. **P < 0.01 by ANOVA. IB data boxed in red represent proteins that were down-regulated following overexpression of NC1 peptide, but treatment of cells was able to restore the NC1 peptide–mediated down-regulation. Uncropped blots corresponding to blots are shown in Supplemental Fig. S6. B) Representative IF results from n = 3 experiments, which yielded similar results. The NC1 peptide–induced F-actin disorganization through changes in distribution (and down-regulation) of actin barbed-end capping and bundling protein Eps8 was correctively restored by SC79, an activator of p-Akt1/2. Similarly, SC79 was also capable of restoring the cytoskeletal disorganization of MTs by restoring the corrective distribution and also expression of detyrosinated α-tubulin [known to confer MT stabilization (73)] and EB1 [a +TIP also known to confer MT stabilization (50)]. Scale bar, 40 µm, which applies to other micrographs.

Figure 7.

Schematic drawing that illustrates the signaling pathway involving mTORC1/rpS6/Akt1/2 and Cdc42 utilized by NC1 peptide to induce Sertoli cell BTB remodeling in the testis.

DISCUSSION

Basement membrane in the testis is a modified form of extracellular matrix (ECM), and its major constituent proteins are laminins (e.g., laminin α2 chains) and type IV collagen chains (15, 16, 53). Type IV collagen is a triple helical structure constituted by 3 α chains, which create the building blocks (monomers) to form the collagen network via dimerization and self-association to create the collagen suprastructure (18, 54). There are 6 genetically distinct α chains of α1 to α6, and α1 (IV) to α4 (IV) chains are mostly found in rodent testes, with collagen α3(IV) being the most predominant collagen chains in the basement membrane of rat testes (55–58). NC1 domain is an ∼230-aa residue fragment of 28 kDa derived from the C-terminal region of collagen α3(IV) chain in the basement membrane, likely generated via the action of matrix metalloproteinase 9 through an activation and up-regulation of TNF-α in the rat testis (14, 58). The NC1 fragment derived from collagen IV chains, such as tumstatin and canstatin, depending on the collagen types and α chain type, has been shown to regulate cell adhesion, proliferation, angiogenesis, and apoptosis in different cell types via its interactions with cell surface receptors, such as integrins (59), in different epithelia (60–63). Studies have also shown that during tumorgenesis, the lack of oxygen that leads to hypoxia among tumor cells induces the secretion of cytokines, which in turn activate surrounding stromal ECM to produce proteases and angiogenic factors. This ECM breakdown induces the release of ECM fragments such as fragments from type IV and XIX collagen chains, which are matrikines (or matricryptins) capable of acting as angiogenesis inhibitors to restrict tumor progression (64), which include NC1 peptide from collagen α2 (IV) chain known as canstatin (an antitumor peptide) (65). Earlier studies have shown that the NC1 peptide derived from collagen α3 (IV) chain in the testis is a potent biologically active peptide that induces Sertoli cell BTB remodeling using primary Sertoli cells cultured in vitro and also testes in vivo (13, 14) through its disruptive effects on the F-actin and MT cytoskeletal networks at the BTB site; namely, the organization of actin filaments and MTs in Sertoli cells at the basal ES/TJ that constitute the BTB (13). These earlier studies were performed using either purified recombinant NC1 peptide or through its overexpression in primary Sertoli cells in vitro or testes in vivo with an NC1 peptide cDNA clone into mammalian expression vector pCI-neo. Additionally, NC1 peptide is also capable of perturbing actin filaments and MTs at the apical ES (the testis-specific adherens junction at the Sertoli-spermatid interface) following its overexpression in the testis, leading to germ cell exfoliation (13). For instance, the organization of F-actin and MT networks across the seminiferous epithelium was grossly disrupted in adult rat testes following overexpression of NC1 peptide (13), consistent with the earlier concept that constituent proteins in the ECM (e.g., collagen IV chains) are crucial to support cytoskeletal filament assembly to modulate cell spreading, adhesion, and function (66, 67).

In this report, we have provided mechanistic insights regarding the molecular mechanism by which NC1 peptide perturbs Sertoli cell TJ permeability barrier function. Overexpression of NC1 peptide, which induces Sertoli cell TJ barrier disruption, is likely mediated through an initial but specific activation of Cdc42, but not RhoA, because the use of a Cdc42T17N dominant negative mutant (i.e., a constitutively inactive mutant of Cdc42) for cotransfection was shown to block the disruptive effects of NC1 peptide on F-actin organization by conferring proper distribution of actin and MT regulatory proteins in Sertoli cell epithelium, which in turn, restored the organization of F-actin and MT networks across the Sertoli cell cytosol. As noted in our findings, overexpression of NC1 peptide in Sertoli cell epithelium induces disorganization of actin filaments (such as branching of actin filaments and defragmentation instead of a linear network of actin filaments appearing bundled in the basal ES) across the Sertoli cell cytosol due to the activation of Cdc42, which is known to activate the neural Wiskott–Aldrich Syndrome protein (N-WASP)-Arp2/3 complex to cause branched actin nucleation (68, 69). Furthermore, it is noted that the NC1 peptide–induced Sertoli cell TJ barrier disruption is accompanied by an up-regulation of p-rpS6-S235/S236 and a down-regulation of p-Akt1-S473 and p-Akt2-S474, which is consistent with findings of earlier studies (31, 37). Studies have shown that rpS6 is the downstream signaling protein of mTORC1 [created by the binding of mTOR to regulatory associated protein of mTOR) (70)], which creates the mTORC1/rpS6 signaling complex (36, 70), and the activation of the mTORC1/rpS6 complex were earlier shown to perturb the Sertoli BTB function by making the barrier leaky through disruptive changes in the organization of F-actin across the Sertoli cell epithelium (31, 37, 71). These observations are also consistent with studies using genetic models by specific inactivation of mTOR in mice wherein the loss of mTOR in Sertoli cells led to extensive seminiferous epithelial damage with tubules devoid of germ cells and infertility and accompanied with a surge in p-rpS6 (i.e., phosphorylated and activated rpS6) level (39). Furthermore, specific deletion of regulatory associated protein of mTOR (Raptor) in mice, thereby inactivating mTORC1, also led to male infertility with extensive disruption of cytoskeletal organization of F-actin, MTs, and vimentin across the seminiferous epithelium (72). Additionally, overexpression of the p-rpS6 mutant [i.e., p-rpS6-S235E/S236E and p-rpS6-S240E/S244E (a quadruple phosphomimetic, i.e., constitutively active, mutant)] in Sertoli cells in vitro or in testes in vivo was also found to induce BTB disruption and was accompanied with extensive disorganization of the cytoskeletons (37, 38). Interestingly, this p-rpS6 activation is accompanied by a down-regulation of p-Akt1-S473 and p-Akt2-S474, which is likely the downstream signaling protein of the mTORC1/rpS6 signaling complex, and is consistent with earlier reports (31, 37).

In this report, we have demonstrated unequivocally that NC1 peptide induces not only an activation of Cdc42 but also an up-regulation of p-rpS6-S235/S236 and a down-regulation of p-Akt1-S473 and p-Akt2-S474. These changes in signaling protein expression are being accompanied by a disruption of the Sertoli cell TJ permeability barrier function mediated by mislocalization of TJ and basal ES proteins at the Sertoli cell–cell interface, which are caused by NC1 peptide–induced cytoskeletal disorganization across the Sertoli cell cytosol. The involvement of Cdc42 in NC1 peptide–mediated Sertoli cell TJ barrier disruption is confirmed by the observation that the use of a Cdc42-T17N mutant was capable of blocking the NC1 peptide–mediated Sertoli cell TJ barrier disruption due to misdistribution of proteins at the Sertoli cell barrier sites, resulting from cytoskeletal disorganization. Cdc42-T17N mutant was also able to restore cytoskeletal organization of F-actin and MTs across Sertoli cells, conferring proper distribution of TJ and basal ES proteins at the cell–cell interface. More importantly, the use of an Akt1/2 activator, SC79, was able to have similar phenotypic effects as of Cdc42-T17N mutant on the Sertoli cell cytoskeletal organization and BTB-associated protein distribution at the cell–cell interface, which was also capable of restoring the NC1 peptide–induced p-Akt1/2 down-regulation. SC79 had no effects on the NC1 peptide–induced p-rpS6-S235/S236 up-regulation because it is a specific chemical activator of p-Akt1/2 by causing conformational changes in Akt, thereby enhancing Akt phosphorylation and activation (29). Collectively, these findings support the notion that Cdc42 is the downstream regulator of the mTORC1/rpS6/Akt1/2 signaling pathway, which in turn modulates the action of NC1 peptide to induce changes in the organization of the F-actin and MT cytoskeletons to support BTB function during the epithelial cycle of spermatogenesis.

Glossary

Arp

actin-related protein

BTB

blood-testis barrier

CAR

coxsackie and adenovirus receptor

Cdc42

cell division control protein 42 homolog

EB1

end-binding protein 1

ECM

extracellular matrix

Eps8

epidermal growth factor receptor pathway substrate 8

ES

ectoplasmic specialization

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GDP

guanosine diphosphate

IB

immunoblot analysis

IF

immunofluorescence microscopy

IP

immunoprecipitation

LG

laminin globular domain

MAP

MT-associated protein

MARK4

MT affinity-regulating kinase 4

MT

microtubule

mTORC1

mammalian target of rapamycin complex 1

NC1

noncollagenous domain 1

rpS6

ribosomal protein S6

RRID

Resource Identification Initiative

TER

transepithelial electrical resistance

+TIP

microtubule plus–end tracking protein

TJ

tight junction

ZO-1

zonula occludens-1

AUTHOR CONTRIBUTIONS

C. Y. Cheng conceived the study; W. Su and C. Y. Cheng designed research; W. Su and C. Y. Cheng performed research; W. Su and C. Y. Cheng contributed new reagents/analytic tools; W. Su and C. Y. Cheng performed data analysis; W. Su and C. Y. Cheng prepared all figures; W. Su and C. Y. Cheng wrote the paper; and all authors read and approved the final manuscript.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.