Connexin 43 reboots meiosis and reseals blood-testis barrier following toxicant-mediated aspermatogenesis and barrier disruption

Abstract

Earlier studies have shown that rats treated with an acute dose of 1-(2,4-dichlorobenzyl)-1H-indazole-3-carbohydrazide (adjudin, a male contraceptive under development) causes permanent infertility due to irreversible blood-testis barrier (BTB) disruption even though the population of undifferentiated spermatogonia remains similar to normal rat testes, because spermatogonia fail to differentiate into spermatocytes to enter meiosis. Since other studies have illustrated the significance of connexin 43 (Cx43)-based gap junction in maintaining the homeostasis of BTB in the rat testis and the phenotypes of Sertoli cell-conditional Cx43 knockout mice share many of the similarities of the adjudin-treated rats, we sought to examine if overexpression of Cx43 in these adjudin-treated rats would reseal the disrupted BTB and reinitiate spermatogenesis. A full-length Cx43 cloned into mammalian expression vector pCI-neo was used to transfect testes of adjudin-treated rats versus empty vector. It was found that overexpression of Cx43 indeed resealed the Sertoli cell tight junction–permeability barrier based on a functional in vivo assay in tubules displaying signs of meiosis as noted by the presence of round spermatids. Thus, these findings suggest that overexpression of Cx43 reinitiated spermatogenesis at least through the steps of meiosis to generate round spermatids in testes of rats treated with an acute dose of adjudin that led to aspermatogenesis. It was also noted that the round spermatids underwent eventual degeneration with the formation of multinucleated cells following Cx43 overexpression due to the failure of spermiogenesis because no elongating/elongated spermatids were detected in any of the tubules examined. The mechanism by which overexpression of Cx43 reboots meiosis and rescues BTB function was also examined. In summary, overexpression of Cx43 in the testis with aspermatogenesis reboots meiosis and reseals toxicant-induced BTB disruption, even though it fails to support round spermatids to enter spermiogenesis.—Li, N., Mruk, D. D., Mok, K.-W., Li, M. W. M., Wong, C. K. C., Lee, W. M., Han, D., Silvestrini, B., Cheng, C. Y. Connexin 43 reboots meiosis and reseals blood-testis barrier following toxicant-mediated aspermatogenesis and barrier disruption.

Studies using different animal models have illustrated the importance of the blood–testis barrier (BTB) and spermatogenesis. For instance, a delay of BTB assembly by treating neonatal rats with diethylstilbestrol (a synthetic nonsteroidal estrogen) also delays meiosis in which pachytene spermatocytes undergo degeneration instead of differentiating into zygotene and diplotene spermatocytes to enter meiosis (1), illustrating the importance of a functional BTB to spermatogenesis. Treatment of adult rats with cadmium (2) or glycerol (3), which are known to disrupt tight junction (TJ) fibrils at the BTB, also leads to spermatogenesis failure because these rats are sterile due to a permanent BTB damage. Consistent with these findings, treatment of adult rats with an acute high-dose of 1-(2,4-dichlorobenzyl)-1H-indazole-3-carbohydrazide (adjudin), a male contraceptive under investigation (4), was shown to induce irreversible BTB disruption (5). Interestingly, in these animals, the population of undifferentiated spermatogonia and spermatogonial stem cells remains largely unaffected, yet these spermatogonia fail to differentiate into spermatocytes due to meiosis arrest because no round spermatids are detected in these rats, unlike rats treated with a low dose of adjudin when the BTB is only transiently disrupted and reinitiation of meiosis closely matches the timeline on the resumption of a functional BTB (5). In this context, it is of interest to note that the BTB, unlike other blood-tissue barriers that are conferred almost exclusively by endothelial TJ barrier, is constituted by both the TJ and gap junction (GJ) that coexist with a testis-specific anchoring junction known as ectoplasmic specialization (ES) (6–8). Studies have shown that GJ plays a prominent role in defining the BTB function (9, 10). For instance, GJ is crucial to provide the needed crosstalks to maintain the homeostasis of BTB so that different junction types can be coordinated to confer the integrity of the immunologic barrier (11). Thus, it is not surprisingly that the Sertoli cell-specific deletion of connexin 43 (Cx43) in mice leads to spermatogenesis failure as a result of meiotic arrest in which spermatogonia fail to differentiate into spermatocytes (12). The BTB in these Sertoli cell-specific Cx43 knockout (KO) mice also displays significant defects, such as no GJ can be found at the Sertoli cell-cell interface, mislocalization of zonula occludens 1 (ZO-1) and N-cadherin at the BTB, and a loss of GJ communication function (13). Collectively, these findings illustrate the likely involvement of GJ in BTB function and its significant role in spermatogenesis. To better understand the role of GJ in maintaining the BTB function and spermatogenesis, we expand our earlier study using the acute adjudin animal model to investigate if an overexpression of Cx43 in these rats would rescue spermatogenesis in particular meiosis and if the overexpression of Cx43 in these rats would reseal the disrupted BTB. If it did, we also sought to understand the underlying molecular mechanism(s) by which Cx43 regulates BTB function.

MATERIALS AND METHODS

Animals

Sprague-Dawley rats (300–325 g body weight) were purchased from Charles River Laboratories (Kingston, NY, USA) and maintained at the Rockefeller University Comparative Bioscience Center. The use of the animals in this study was approved by the Rockefeller University Institutional Animal Care and Use Committee with Protocol Numbers 12-506 and 15-780-H, and the use of recombinant DNA (e.g., Cx43 full-length cDNA clone and other relevant plasmid DNAs) and siRNA duplexes for in vitro and in vivo experiments reported herein was approved by the Rockefeller University Institutional Biosafety Committee with Approval Number 2-15-04-007.

Cloning of Cx43 into pCI-neo mammalian expression vector

The full-length coding sequence of rat Cx43 (GenBank Accession Number BC081842.1), was amplified from rat Sertoli cell cDNAs using AccuPrime Pfx DNA polymerase (Invitrogen, Carlsbad, CA, USA), and was cloned into the pCI-neo mammalian expression vector (Promega, Madison, WI, USA) by nested PCR using 2 sets of primer pairs (Table 1) including the corresponding EcoRI and XbaI restriction enzyme recognition sites as described (14, 15). This full-length cDNA clone of 1149 bp was confirmed by direct nucleotide sequence analysis at Genewiz. pCI-neo empty vector and Cx43/pCI-neo plasmid DNA was purified using the EndoFree Plasmid Mega Kit (Qiagen, Germantown, MD, USA) to remove possible contamination of endotoxins prior to their use for transfection.

TABLE 1.

Primers used for cloning and qPCR experiments in this report

Target gene	GenBank accession number	Primer sequence, 5′–3′				Expected size (bp)
		Sense	Nucleotide position	Antisense	Nucleotide position
Connexin 43 (full-length cDNA)	BC081842.1	TAAGTGAAAGAGAGGTGCCC	169–188	CTCCTCCATAATCGACAGCT	1358–1377	1209
Connexin 43 (full-length clone)	BC081842.1	ATGAATTCGCCAACATGGGTGACTGGAGTGCCTTG	193–213	ATTCTAGATTAAATCTCCAGGTCATCAGGCCGAGGCCTG	1311–1341	1149
Connexin 43 (qPCR)	BC081842.1	GCTGAGAACCTACATCATCAG	649–669	TGAGAGGAAGCAGTCTACC	781–799	151
GAPDH	NM_017008.4	GCTGGTCATCAACGGGAAAC	267–286	GGTGAAGACGCCAGTAGAC	360–378	112

Nested PCR was used to obtain the full-length Cx43 coding sequence with appropriate restriction sites for its cloning into pCI-neo mammalian expression vector (Promega). The first primer pair of Cx43 was used to amplify the full-length Cx43 sequence from Sertoli cell total cDNAs by PCR using AccuPrime Pfx DNA polymerase (Invitrogen). The second primer pair then was used to introduce an EcoRI (underlined, sense primer) and an XbaI (underlined, antisense primer) restriction enzyme recognition site, respectively, to this full-length Cx43 sequence for its directional cloning into the pCI-neo plasmid. The Kozak sequence GCCAAC (italics), which plays an important role in the initiation of eukaryotic mRNA translation process, was also added before the initiation codon (ATG). This clone was confirmed by direct nucleotide sequencing at Genewiz (South Plainfield, NJ, USA). The third and fourth primer pairs are for q-PCR experiments to quantify the steady-state mRNA levels of Cx43 and GAPDH in samples, respectively, with GAPDH serving as the qPCR loading control.

Treatment regimen—administration of adjudin and transfection of full-length Cx43 cDNA cloned into pCI-neo mammalian expression vector in vivo

An adjudin suspension was prepared at 40 mg/ml in Milli-Q water containing 0.05% (wt/vol) methylcellulose (Sigma-Aldrich, St. Louis, MO, USA) as described (14). Adult rats (∼320 g body weight) received a single dose of adjudin at 125 mg/kg body weight by oral gavage at time 0 as illustrated in Fig. 1A. Twelve weeks following adjudin treatment when virtually all seminiferous tubules were devoid of all germ cell types in the seminiferous epithelium except spermatogonia, each testis was transfected with either Cx43 (pCI-neo/Cx43, containing the full-length Cx43 cDNA) or empty vector (pCI-neo alone) at 12, 12.5, and 13 wk via an intratesticular injection as described elsewhere (15) using a transfection mixture consisting of 15 µg pCI-neo or Cx43/pCI-neo plasmid DNA in 150 µl TransIT-EE hydrodynamic Delivery Solution (Mirus Bio, Madison, WI, USA) via a 29 gauge insulin syringe. Rats were terminated at 14, 15, and 18 wk, euthanized by CO₂ asphyxiation using slow (20–30%/min) displacement of chamber air with compressed carbon dioxide in a euthanasia chamber approved by Rockefeller University Laboratory Safety. Thereafter, testes were immediately removed, either fixed in Bouin’s fixative or frozen in liquid nitrogen, and both the body and the testis weights were recorded. For histologic analysis by hematoxylin and eosin staining using paraffin sections, testes were fixed in Bouin’s fixative (Polysciences, Warrington, PA, USA), and the paraffin sections were processed at the Laboratory of Comparative Pathology (LCP, Memorial Sloan-Kettering Cancer Center, New York, NY, USA). For immunoblot, quantitative PCR (qPCR), and dual-labeled immunofluorescence analysis, testes were snap-frozen in liquid nitrogen and kept at −80°C until use.

Figure 1.

Overexpression of Cx43 reboots meiosis in rats with aspermatogenesis caused by an acute dose of adjudin. A) Treatment regimen. Adult rats (~320 g body weight) were treated with a single acute dose of adjudin (125 mg/kg body weight) by oral gavage at time 0 known to induce aspermatogenesis (5). For overexpression (O/E) of Cx43, rats were transfected thrice with a transfection performed on wk (W) 12, 12.5 and 13, using plasmid DNA containing a full-length Cx43 cDNA cloned into the mammalian expression vector pCI-neo (pCI-neo/Cx43) vs. empty vector (pCI-neo, control). Thereafter, rats in each group (n = 3 or 4 rats per time point) were terminated on 14, 15, and 18 wk. A testis from each rat (n = 3 or 4 rats) was used for histologic analysis and the other was snap-frozen in liquid nitrogen for immunoblotting or immunofluorescence analysis. B) Hematoxylin and eosin staining of paraffin sections of testes. Rats from adjudin alone group were terminated at time 0 (normal), and 12 and 18 wk. Normal spermatogenesis was seen in normal rats at time 0. Virtually, all the tubules examined were devoid of germ cells at 12 and 18 wk after adjudin treatment (i.e., adjudin alone) except for spermatogonia. Rats treated with adjudin and transfected with empty vector (adjudin + pCI-neo) terminated at 14, 15, and 18 wk displayed similar phenotypes as of the adjudin alone group. However, rats from the pCI-neo/Cx43 group in which Cx43 was overexpressed in the testis, resumption of spermatogenesis was detected in some tubules at 14, 15, and 18 wk because round spermatids were found in many tubules, illustrating meiosis has been rebooted. However, round spermatids failed to develop into elongating and elongated spermatids via spermiogenesis; instead, they formed multinucleated cells (see blue arrowheads) in some tubules, destined to undergo degeneration. At 18 wk, meiosis was clearly visible in some tubules as shown herein (see green boxed area, illustrating meiotic germ cells at anaphase). A micrograph shown herein is an enlarged image of the boxed area (in red) in the inset at lower magnification. Scale bar, 60 µm in micrograph boxed in red; 160 µm in inset in black. Images in the lower 2 panels are also magnified in green or blue boxed area and shown in the corresponding inset. Scale bar, 30 µm in green or blue inset.

Assessing Cx43 overexpression efficiency in the testis in vivo

The efficiency of Cx43 overexpression in the rat testes in vivo was estimated by 3 approaches. First, a DsRed2 cDNA construct was cloned using a primer pair specific to DsRed 2 with the pIRES2-DsRed2 plasmid DNA (Clontech, Mountain View, CA, USA) served as a template by PCR as detailed elsewhere (15). This cDNA was then ligated to the MluI/XbaI site of the pCI-neo vector to obtain the DsRed2-pCI-neo vector (15). About 10 µg of DsRed2-pCI-neo plasmid DNA and the same amount of pCI-neo/Cx43 DNA (a total of 20 µg DNA) suspended in 150 µl TansIT-EE delivery solution were administered to the right testis of an adult rat with n = 3 rats on d 0, 3, and 5 (i.e., a total of 3 transfections) for their cotransfections, and rats were euthanized on d 8 (i.e., 3 d after the last transfection), with the left testis of the same rat received pCI-neo plasmid DNA alone to serve as a control. Tubules (about 100 tubules per testis with n = 3 rats) with red fluorescence (DsRed2) in the seminiferous epithelium were randomly selected and scored as tubules with successful transfection and overexpression. Second, lysates (60 µg protein) from the testis transfected with pCI-neo/Cx43 vs. pCI-neo (Ctrl) from rats treated with adjudin at 14, 15, and 18 wk (see Fig. 1A for the regimen) were used for immunoblotting to assess the steady-state Cx43 protein level to confirm overexpression of Cx43 using a specific anti-Cx43 antibody (Table 2). Third, RNAs isolated from these rats at 14, 15, and 18 wk pretreated with DNase to eliminate genomic DNA contamination were used for q-PCR to assess the Cx43 steady-state mRNA level to confirm successful overexpression of Cx43 using primer pair specific to Cx43 (Table 1).

TABLE 2.

Antibodies used for different experiments in this report

Antibody	Host species	Vendor	Catalog no.	Working dilution
				IB	IF
Actin	Goat	Santa Cruz Biotechnology (Santa Cruz, CA, USA)	sc-1616	1:300
Connexin 43	Rabbit	Cell Signaling Technology (Danvers, MA, USA)	3512s	1:1000	1:100
Eps8	Mouse	BD Biosciences (San Jose, CA, USA)	610144		1:50
Formin 1	Mouse	Abcam (Cambridge, MA, USA)	ab68058		1:50
γ-Catenin	Mouse	BD Biosciences	610254		1:100
N-Cadherin	Mouse	Invitrogen (Carlsbad, CA, USA)	33-3900		1:100
Occludin	Rabbit	Invitrogen	71-1500		1:50
ZO-1	Rabbit	Invitrogen	61-7300		1:50
p-FAK-Tyr407	Rabbit	Invitrogen	44650G		1:100
Rabbit IgG-Alexa Fluor 488	Goat	Invitrogen	A11034		1:250
Mouse IgG-Alexa Fluor 488	Goat	Invitrogen	A11029		1:250
Mouse IgG-Alexa Fluor 555	Goat	Invitrogen	A21424		1:250
Goat IgG-HRP	Bovine	Santa Cruz Biotechnology	sc-2350	1:3000
Rabbit IgG-HRP	Bovine	Santa Cruz Biotechnology	sc-2370	1:3000

IB, immunoblot; IF, immunofluorescence microscopy. All the primary antibodies used herein cross-reacted with the corresponding target proteins in rats as noted by the manufacturers and confirmed in our pilot/preliminary experiments.

Immunoblot analysis and protein estimation

Testis lysate was obtained by suspending samples of testes in immunoprecipitation lysis buffer [50 mM Tris, containing 0.15 M NaCl, 1% Nonidet P-40 (vol/vol), 1 mM EGTA, 2 mM N-ethylmaleimide, 10% glycerol (vol/vol), pH 7.4 at 22°C, freshly supplemented with protease inhibitor mixture (Sigma-Aldrich) and phosphatase inhibitor cocktail II and III (Sigma-Aldrich)] using a tissue/buffer ratio of 1:5, to be followed by a brief sonication of ∼8 s using a Cole-Parmer Ultrasonic Processor (Model CPX130PB, Cole-Parmer, Chicago, IL, USA), whereas the sample in microfuge tube was placed in ice. Thereafter, samples were centrifuged at 15,000 g for 15 min to obtain the clear supernatant as the protein lysate. Protein concentration was estimated by Bio-Rad D_c protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA). Immunoblot analysis used 100 µg total protein/lane for detecting Cx43. Chemiluminescence was performed using an ECL reagent prepared in our laboratory as described previously (16) and a FujiFilm LAS-4000 Mini Luminescent Image Analyzer (FujiFilm Incorporated, Stamford, CT, USA) for image acquisition. β-Actin served as a protein loading control. All samples within an experimental group were processed simultaneously to avoid interexperimental variations.

q-PCR

Total RNA from tissues was obtained by TRIzol (Invitrogen, Eugene, OR, USA). Genomic DNA was removed from the RNA samples by incubating mRNA with DNase I at room temperature for 15 min, and it was followed by a reverse transcriptional reaction by using oligo-dT as the primer. For qPCR, cDNA were mixed with a SYBR Green PCR master mix (Applied Biosystems, Foster City, CA, USA) and corresponding primer pair (Table 1), glyceraldehyde 3-phosphate dehydrogenase (GAPDH), served as an internal reference. The reaction was carried out by using an ABI 7900HT Fast Real-Time PCR system at the Genomic Resource Center (The Rockefeller University, New York, NY, USA). Data were analyzed by ABI SDS 2.3 software package (Applied Biosystems).

Dual-labeled immunofluorescence analysis

Frozen sections (7 µm in thickness) were obtained by cryostat at −22°C, which were then fixed in 4% paraformaldehyde in PBS (wt/vol) for 10 min and permeabilized with Triton X-100 (0.1%, vol/vol) in PBS for 10 min. Sections were blocked by 4% bovine serum albumin (wt/vol) in PBS, to be followed by an overnight incubation with primary antibodies (Table 2). Thereafter, sections were incubated with corresponding Alexa Fluor-conjugated secondary antibodies at a dilution of 1:250 for 30 min. Sections were finally mounted with Prolong Gold anti-fade reagent with DAPI (Life Technologies, Grand Island, NY, USA). To assess staining specificity, negative controls included the substitution of the primary antibody with IgG corresponding to the animal species used to obtain the primary antibody (Table 2) and also substitution of the secondary antibody with normal goal IgG. For F-actin staining, testis sections were incubated with rhodamine-phalloidin (Life Technologies) at a dilution of 1:100 for 30 min. Fluorescence images were acquired using an Olympus BX61 fluorescence microscope with a built-in Olympus DP70 12.5-MPx digital camera and analyzed using the Olympus MicroSuite Five software (version 1.224, Olympus Soft Imaging Solution Corp, Tokyo, Japan). Image overlay and analysis were performed using PhotoShop in Adobe Creative Suite Design Premium software package (version 3.0, Adobe Systems, San Jose, CA, USA).

Assessment of spermatogenesis status

Resumption of meiosis in rats following overexpression of Cx43 in rats treated with an acute dose of adjudin (adjudin + pCI-neo/Cx43) that induced aspermatogenesis vs. adjudin-treated rats without overexpression of Cx43 (Adjudin alone) or overexpressed with empty vector without Cx43 (adjudin + pCI-neo) was monitored using paraffin sections following hematoxylin and eosin staining. In brief, the cross-section of a tubule containing >10 round spermatids in the seminiferous epithelium was scored as a tubule with meiosis. At least 100 tubules were randomly scored from each rat testis, and a total of 3 rats were examined at 14, 15, and 18 wk.

Image analysis

To obtain semiquantitative data for comparing changes in the distribution of proteins at the BTB, such as F-actin, basal ES-associated proteins (e.g., N-cadherin, γ-catenin), TJ-associated proteins (e.g., occludin, ZO-1), and actin microfilament regulatory proteins [e.g., formin 1, Eps8, epidermal growth factor receptor pathway substrate 8 (Eps8)], the distribution of each target protein near the basement membrane (annotated by white dashed line as shown in corresponding micrographs) was measured and annotated by yellow bracket in control group without Cx43 overexpression (pCI-neo) vs. white bracket in Cx43 overexpression group (pCI-neo/Cx43). Four readings were taken from the cross section of a seminiferous tubule at 4 opposite ends of a tubule to obtain the average for each tubule. Approximately 50 tubules per testis having meiosis as confirmed by the presence of round spermatids in the seminiferous epithelium in pCI-neo/Cx43 group was scored vs. pCI-neo control group following adjudin treatment with n = 3 rats (i.e., a total of 150 randomly selected tubules were scored) including normal testes. The distribution of a target protein in normal testes was arbitrarily set at 1 against which statistical comparison was performed. For experiments in which the intensity of fluorescence signal for a target protein, such as p-focal adhesion kinase (FAK)-Tyr⁴⁰⁷, was quantified for comparison between treatment and control groups, the software ImageJ 1.45 obtained from http://rsbweb.nih.gov/ij (National Institutes of Health, Bethesda, MD, USA) was used. About 50 tubules per rat testis of n = 3 rats with a total of 150 randomly selected tubules having round spermatids in the lumen were examined and scored, illustrating resumption of meiosis in adjudin + pCI-neo/Cx43 were compared with adjudin + pCI-neo (empty vector) vs. normal testes control groups.

BTB integrity assay

The BTB integrity in rats overexpressed with Cx43 vs. empty vector (control) previously treated with adjudin was assessed using an in vivo assay as described earlier (17). This assay is based on the ability of an intact BTB to block the permeation of a small fluorescence tag from the basal compartment to the apical compartment of the seminiferous epithelium. In brief, adult rats (n = 3, ∼320 g body weight) treated with a single dose of adjudin by oral gavage using the treatment regimen shown in Fig. 1A were transfected with pCI-neo empty vector vs. pCI-neo/Cx43 vector thrice. At 15 wk, rats were subjected to anesthesia using ketamine HCl (∼60 mg/kg body weight)/xylazine (10 mg/kg body weight) (Sigma-Aldrich) intramuscularly. A small incision of ∼0.5 cm was made using a scissors over the skin of jugular vein under aseptic conditions to expose the vein. Thereafter, 200 µl of FITC-inulin (M_r 4.6 kDa, Sigma-Aldrich) dissolved in PBS at 5 mg/ml was administered into the jugular vein with a 28 gauge needle. The wound was closed by a 9 mm Autoclip wound clip (Becton Dickinson, Franklin Lakes, NJ, USA) and rats were kept on a 35°C heating pad for 30 min before they were euthanized by CO₂ asphyxiation. Testes were removed immediately and snap-frozen in liquid nitrogen. Frozen sections (10 µm in thickness) were mounted with Prolong Gold antifade reagent with DAPI (Life Technologies). The distance traveled by the FITC-inulin from the basement membrane (i.e., relative location of the BTB near the basement membrane) to the apical compartment (D_Signal) in a tubule vs. the radius of the tubule (D_Radius) was recorded in ∼150 randomly selected tubules from n = 3 rats (i.e., ∼ 50 randomly selected tubules per rat) per treatment and control group to assess BTB integrity. For oblique sections of tubules, D_Radius was obtained by averaging the shortest and the longest distance from the basement membrane.

Statistical analysis

Data from treated animals were compared to the corresponding controls by ANOVA to be followed by Tukey’s honest significant test or Dunnett’s test using the GB-STAT Statistical Analysis software package (version 7.0; Dynamic Microsystems Incorporated, Silver Spring, MD, USA). All experiments shown herein were results of a representative experiment; however, all experiments were repeated at least 3 times, and both the control and treatment groups were conducted in a single experimental session to avoid interexperimental variations. The number of animals used for both the control and treatment groups at each time point was 3 or 4 rats as noted in the regimen summarized in Fig. 1A.

RESULTS

Overexpression of Cx43 in the testis reboots meiosis following adjudin-induced aspermatogenesis

Using a regimen designed based on pilot experiments and shown in Fig. 1A, a single acute dose of adjudin at 125 mg/kg body weight by oral gavage caused sterility in adult rats in which virtually all the tubules were devoid of germ cells except for the population of spermatogonia, which remained relatively unaffected (top panel in Fig. 1B), consistent with an earlier report (5). A full-length Cx43 cDNA was first obtained by nested PCR using Sertoli cell cDNAs transcribed from total RNAs and specific primer pairs (Table 1), and the identity of this clone was confirmed by direct nucleotide sequencing and cloned into a mammalian expression vector of pCI-neo at the EcoRI and XbaI sites (Table 1). Overexpression of Cx43 in the testis in vivo using pCI-neo as the expression vector at 12, 12.5, and 13 wk at a transfection efficacy of ∼30% as earlier reported (15, 18) and rats (n = 4 at each time point) terminated by week 14, 15, and 18 vs. rats transfected with empty pCI-neo vector (n = 3 rats) (Fig. 1A) was found to induce resumption of meiosis (Fig. 1B). Transfection efficiency was further estimated by using a DsRed2-pCI-neo cDNA construct in which a red fluorescence DsRed2 protein tag was cloned into pCI-neo, which was cotransfected with pCI-neo/Cx43 into the testis of n = 3 normal rats, and red fluorescence in ∼300 randomly selected cross sections of tubules (∼100 tubules per testis with n = 3 rats) were scored; representative data are shown in Fig. 2A. Approximately 30% of the scored tubules were found to be positively transfected. A tubule was considered to be successfully transfected when >15 red fluorescence aggregates were found within the seminiferous epithelium of a tubule as shown in Fig. 2A. Fluorescence (DsRed2) was associated mostly with the cytosol of Sertoli cells and some early spermatocytes but not postmeiotic spermatids in particular elongating/elongated spermatids, which are highly differentiated and metabolically quiescent germ cells (Fig. 2A). Most importantly, it was noted that round spermatids were detected in a significant number of tubules in testes overexpressed with Cx43 vs. adjudin-treated rats overexpressed with empty vector alone or rats treated with adjudin alone (Figs. 1B and 2B) of 3 rats. However, elongating and elongated spermatids remained undetected in Cx43 overexpressed testes (Fig. 1B), illustrating spermiogenesis remained arrested. The specificity of the anti-Cx43 antibody (Table 2) was assessed by immunoblotting in which 3 Cx43 immunoreactive bands, the result of differential phosphorylation that led to this heterogeneity as earlier reported (19), were identified designated p0, p1, and p2 with an apparent M_r of 39, 41, and 43 kDa (Fig. 2C), consistent with findings of earlier reports (19, 20). Furthermore, adjudin caused a considerable down-regulation on the steady-state protein level of Cx43 by 14 wk post-treatment based on immunoblot analysis as shown in Fig. 2C.

Figure 2.

Assessment of transfection efficiency for overexpression of Cx43 in the testis that restores meiosis following treatment of rats with an acute dose of adjudin that led to meiotic arrest. A) Transfection of the rat testes in vivo with pCI-neo empty vector (left panel) vs. DsRed2-pCI-neo cotransfected with pCI-neo/Cx43 (right panel), illustrating successful transfection of the testis with plasmid DNA in which DsRed2 (red fluorescence, annotated by yellow arrowheads) in testes cotransfected with DsRed2-pCI-neo and pCI-neo/Cx43 vector. The transfection efficiency was estimated to be ∼30% by scoring about 300 randomly selected cross sections of tubules with n = 3 rats (i.e., ∼100 tubules per rat). Scale bar, 30 µm. B) To assess the status of spermatogenesis in the testis following overexpression of Cx43 from rat testes with meiotic arrest after treatment with an acute dose of adjudin (125 mg/kg body weight, by oral gavage), cross section of a tubule containing at least 10 round spermatids was scored as a tubule with meiosis as noted in Fig. 1B. At least 100 tubules were randomly scored from each rat testis by 14, 15, and 18 wk (W), and a total of 3 rats was examined at each of these points in both Cx43 overexpression group (adjudin + pCI-neo/Cx43) vs. empty vector alone (adjudin + pCI-neo) control group. *P < 0.05; **P < 0.01. C) The specificity of the anti-Cx43 antibody was assessed by immunoblotting, showing 3 immunoreactive Cx43 bands of 39, 41, and 43 kDa, illustrating the heterogeneity of Cx43 due to differential phosphorylation of Cx43 as earlier reported (19). The 2 samples on the left represent control (normal, adult rats) testes and the 2 samples on the right represent samples from rats treated with a single dose of adjudin (125 mg/kg body weight by oral gavage) for 14 wk (n = 3 rats; i.e., transfected with pCI-neo empty vector alone, see Fig. 1A for regimen) from a representative experiment.

Overexpression of Cx43 in the testis leads to proper and normal localization of Cx43 at the BTB

We next examined the efficacy of Cx43 overexpression in testes of rats treated with an acute dose of adjudin via oral gavage. It was noted that in rats transfected with the pCI-neo vector containing the full-length Cx43 cDNA for overexpression of Cx43 vs. empty vector (control), the body weight was relatively similar but a mild but statistically insignificant increase in testis weight was detected (Fig. 3A). A study by immunoblotting (Fig. 3B) and q-PCR (Fig. 3C, Table 1) confirmed overexpression of Cx43 in rat testes transfected with pCI-neo containing the full-length Cx43 cDNA. In normal rat testes, Cx43 was localized prominently at the BTB (Fig. 3D), consistent with earlier reports (9, 21). However, rats treated with adjudin but overexpressed with empty vector (control), the expression of Cx43 was considerably weakened in the seminiferous epithelium vs. normal testes and virtually no Cx43 was detected at the BTB (annotated by white dash line, see 2 left panels in Fig. 3E vs. Fig. 3D). Overexpression of Cx43 in the testis from adjudin-treated rats was found to stimulate meiosis because round spermatids and spermatocytes were detected in the seminiferous epithelium, and in these rat testes, Cx43 became properly organized, making them similar to the normal rat testes (see 2 right panels in Fig. 3E vs. Fig. 3D and 2 left panels in Fig. 3E). The fluorescence staining of Cx43 shown herein and other target proteins in this report was specific because negative control (see yellow boxed image in Fig. 3D) showed no green fluorescence signals of Cx43.

Figure 3.

Overexpression of Cx43 in rat testes with aspermatogenesis following an acute dose of adjudin treatment restores the expression and localization of Cx43 at the BTB in tubules displaying signs of meiosis as illustrated by the presence of round spermatids. A) Overexpression of Cx43 in testes of adjudin (an acute dose) treated rats (pCI-neo/Cx43) vs. empty vector alone (pCI-neo) also treated with adjudin had no apparent effects on body weight on 14, 15 and 18 wk (W) even though the testis weight mildly increased, but not statistically significant, in rats of the pCI-neo/Cx43 group. B, C) Overexpression of Cx43 in the rats of pCI-neo/Cx43 group was confirmed by immunoblotting (B) and q-PCR (C), illustrating Cx43 was indeed overexpressed in testes from rats treated with adjudin (125 mg/kg body weight, by oral gavage). ^*P < 0.05. D) Immunofluorescence analysis of normal testes illustrating Cx43 (green fluorescence) was prominently expressed in the basal compartment, near the basement membrane (annotated by a dashed white line), consistent with its localization at the BTB. The yellow boxed image is a negative control (-ve Ctrl) in which the primary Cx43 antibody was substituted by normal rabbit IgG that showed no fluorescence staining of Cx43, illustrating the specificity of staining shown herein and also in (E). Scale bar, 80 µm, and inset shown below in green, which is the enlarged image of the green boxed area above, 20 µm, which apply to the other micrograph of the same panel; inset in yellow boxed image, 250 µm. E) By 14, 15, and 18 wk, the expression of Cx43 remained down-regulated in testes transfected with empty vector (adjudin + pCI-neo) and virtually no Cx43 was detected at the BTB (see the 2 left panels vs. control testes shown in D). However, overexpression of Cx43 in adjudin-treated testes (adjudin + pCI-neo/Cx43) displayed an up-regulation of Cx43, and Cx43 was found to localize above the basement membrane (annotated by a dashed white line) consistent with its localization at the BTB in tubules displaying signs of meiosis as supported by the presence of round spermatids in the seminiferous epithelium. Scale bar, 150 µm, and inset, 50 µm, which is the enlarged image of the boxed area; these scale bars apply to all other micrographs in (E).

Overexpression of Cx43 leads to reorganization of F-actin and basal ES proteins at the BTB in tubules of testes with resumption of meiosis following adjudin-induced aspermatogenesis

We speculated that the proper localization of Cx43 in tubules with signs of meiosis that followed Cx43 overexpression in adjudin-induced aspermatogenesis in the testis was the result of a reorganization of F-actin network; we thus assessed the organization of F-actin vs. basal ES proteins N-cadherin and γ-catenin, which used actin microfilaments for attachment. As noted in Fig. 4A, overexpression of Cx43 in testes of adjudin-treated rats that had signs of meiosis as illustrated by the presence of round spermatids displayed proper localization of F-actin at the BTB located right above the basement membrane (annotated by dashed white line; right panel in Fig. 4A), making the F-actin network at the BTB analogous to normal rat testes (right vs. left panel in Fig. 4A). These findings are consistent with the distribution of the 2 basal ES proteins N-cadherin and γ-catenin at the BTB in tubules with signs of meiosis, making them similar to normal rat testes (right vs. left panel in Fig. 4A). Changes in the organization and/or distribution of F-actin, N-cadherin, and γ-catenin in tubules following overexpression of Cx43 vs. empty vector control group and normal testes were semiquantitatively assessed by measuring the relative distribution of these proteins at the BTB and shown in Fig. 4B. These findings also support the notion that the BTB might have been resealed in these tubules.

Figure 4.

Overexpression of Cx43 in adjudin-treated rats with aspermatogenesis restores F-actin organization and proper localization of basal ES proteins N-cadherin and γ-catenin at the BTB. A) In normal rat testes (left panel), F-actin (red fluorescence, top panel), N-cadherin (green fluorescence, middle panel), and γ-catenin (green fluorescence, bottom panel) were tightly localized (see white brackets) at the BTB (annotated by white arrowheads), adjacent to the basement membrane (annotated by a dashed white line) in the seminiferous epithelium. Following treatment of rats with adjudin, the epithelium became devoid of all germ cells except for spermatogonia in the basal compartment by 14, 15, and 18 wk (W), and F-actin, N-cadherin and γ-catenin were no longer tightly localized at the BTB in pCI-neo (empty vector) control rats; instead, they were diffusely localized at the site (see yellow brackets). However, overexpression of Cx43 in these rat testes (pCI-neo/Cx43) restored F-actin organization (see white brackets) in tubules with signs of meiosis. Cx43 overexpression also restored localization of basal ES proteins N-cadherin and γ-catenin at the BTB in tubules (see white brackets) displaying signs of meiosis as noted by the presence of round spermatids because these basal ES proteins became tightly localized at the BTB, analogous to the normal testis. Box areas are magnified and shown in the inset of the corresponding micrograph. Yellow boxed images shown for normal testes were negative control (-ve Ctrl) in which the primary antibody was substituted by the corresponding mouse IgG that yielded no green fluorescence staining, illustrating the specificity of staining shown herein. Scale bar, 80 µm; inset, in red or in white, which is the enlarged image of the corresponding red (in normal testis) or white (adjudin-treated testis) boxed area in the micrograph at lower magnification, 30 µm; and yellow boxed –ve Ctrl, 160 µm, which apply to corresponding micrographs. These results are representative findings of an experiment that was repeated using n = 3 rats and yielded similar results. B) Image analysis that semiquantitatively summarized findings shown in (A) regarding the relative distribution of F-actin, N-cadherin, and γ-catenin at the BTB by comparing the extent of fluorescence signaling in control testes transfected with empty vector (pCI-neo) (see yellow bracket in A) vs. rat testes with overexpression of Cx43 (pCI-neo/Cx43) (see white bracket in A) on 14, 15, and 18 wk following treatment with adjudin vs. normal testes (control). Analysis was performed by quantifying the distribution of a target protein (e.g., F-actin, N-cadherin, γ-catenin) at the BTB near the basement membrane (annotated by a dashed white line) as annotated by the white vs. the yellow bracket in corresponding animal groups. About 50 tubules per rat testis of n = 3 rats with a total of 150 randomly selected tubules were scored. For rats in the Cx43 overexpression (pCI-neo/Cx43) group, tubules having round spermatids in the seminiferous epithelium, illustrating resumption of meiosis, were selected for comparison. For each tubule, 4 readings were taken at opposite ends of the cross-section, and their average was then used for comparison. The distribution of each target protein in normal control testes was arbitrarily set at 1 against which statistical analysis and comparison were performed. **P < 0.01.

Overexpression of Cx43 leads to reorganization of TJ proteins at the BTB in tubules of testes with resumption of meiosis following adjudin-induced aspermatogenesis

We then assessed whether overexpression of Cx43 induced proper localization of TJ-associated proteins occludin and ZO-1 at the BTB in tubules with signs of meiosis as indicated by the presence of round spermatids in adjudin-treated rats because these proteins also utilized actin microfilaments for attachment. Consistent with findings of the basal ES proteins shown in Fig. 4, Cx43 overexpression also led to proper organization of occludin and ZO-1 at the BTB, located right above the basement membrane (annotated by dashed white line), making them similar to their distribution at the BTB of normal rat testes (right panel vs. left panel in Fig. 5A). These observations are also supported when the distribution of these proteins were semiquantitatively analyzed and shown in Fig. 5B. These findings also support the notion that the BTB might have been resealed in tubules with successful Cx43 overexpression in rat testes with aspermatogenesis induced by an acute dose of adjudin.

Figure 5.

Overexpression of Cx43 in adjudin-treated rats with aspermatogenesis restores organization of TJ-associated proteins occludin and ZO-1 at the BTB. A) In normal rat testes, TJ proteins occludin (green fluorescence, top panel) and ZO-1 (green fluorescence, bottom panel) were tightly localized (see white bracket) at the BTB (annotated by white arrowheads), adjacent to the basement membrane (annotated by a dashed white line) in the basal compartment. Following treatment of rats with an acute dose of adjudin (125 mg/kg body weight) by oral gavage and transfected with pCI-neo vector alone (pCI-neo, empty vector), the seminiferous epithelium became devoid of all germ cells except for spermatogonia in the basal compartment by 14, 15, and 18 wk (W). Occludin and ZO-1 were no longer tightly localized at the BTB as of normal rat testes; instead, they were diffusely localized at the site (see yellow brackets). Also, in some tubules, ZO-1 was also down-regulated as illustrated herein. However, following overexpression of Cx43 in these rat testes (pCI-neo/Cx43), the localization of occludin, and the expression and localization of ZO-1 were restored at the BTB in tubules displaying signs of meiosis as noted by the presence of round spermatids. Boxed areas are magnified in insets. Yellow boxed images shown for normal testes were negative control (-ve Ctrl) in which the primary antibody was substituted by the IgG of the corresponding animal species that yielded no green fluorescence staining, illustrating the specificity of staining shown herein. Scale bar, 80 µm; inset in red or in white, which is the enlarged image of the corresponding red (in normal testis) or white (adjudin-treated testis) boxed area in the micrograph at lower magnification, 30 µm; yellow boxed –ve Ctrl, 160 µm, which applies to corresponding micrographs. These results are representative findings of an experiment that was repeated using n = 3 rats and yielded similar results. B) Image analysis that semiquantitatively summarized findings shown in (A) regarding the relative distribution of occludin and ZO-1 at the BTB was performed as described in Fig. 4. **P < 0.01.

Overexpression of Cx43 that confers proper organization of F-actin network at the BTB in adjudin-treated testes is mediated through corrective spatiotemporal expression of actin regulatory proteins formin 1 and Eps8

We then examined the molecular basis by which overexpression of Cx43 recused the proper organization of F-actin network at the BTB to structurally support basal ES and TJ adhesion protein complexes of N-cadherin-γ-catenin (Fig. 5) and occludin-ZO-1 (Fig. 6). We used 2 actin regulatory proteins as markers in this study: formin 1, an actin nucleation protein known to promote the formation of long stretches of microfilaments necessary to assemble into actin filament bundles at the basal ES (22, 23), and an actin barbed end capping and bundling protein known to confer actin microfilaments their bundled configuration at the ES (24). It was noted that both formin 1 and Eps8 became properly expressed at the BTB (annotated by white arrowheads) near the basement membrane (annotated by a dashed line), similar to normal rat testes, in tubules of testes following overexpression of Cx43, displaying signs of meiosis as indicated by the presence of round spermatids (Fig. 6A). These data were also supported by findings when the distribution of these proteins were analyzed semiquantitatively and shown in Fig. 6B. The presence of these actin nucleation and bundling proteins thus support the proper organization of actin microfilaments at the ES, which may in turn be used to reseal the disrupted barrier to confer the BTB function.

Figure 6.

Overexpression of Cx43 in adjudin-treated rats with aspermatogenesis restores the spatiotemporal expression of formin 1 and Eps8 in the seminiferous epithelium. A) In normal rat testes, actin nucleation protein formin 1 (green fluorescence, top panel) that promotes the formation of long stretches of actin microfilaments necessary to maintain basal ES/BTB integrity and also actin microfilament barbed end capping and bundling protein Eps8 (green fluorescence, bottom panel) are both localized at the BTB (annotated by white arrowheads) adjacent to the basement membrane (annotated by a dashed white line) in the epithelium. Following treatment of rats with adjudin and transfected with pCI-neo vector alone (control), the epithelium became devoid of all germ cells except for spermatogonia in the basal compartment by 14, 15, and 18 wk (W). Formin 1 and Eps8 were diffusely localized (annotated by yellow brackets) and considerably down-regulated at the BTB in rats transfected with pCI-neo empty vector after adjudin treatment vs. normal testes. However, overexpression of Cx43 in pCI-neo/Cx43 rats caused robust expression of formin 1 by 14, 15 and 18 wk, tightly localized at the BTB (annotated by white brackets), likely being used to restore the actin microfilament bundles via nucleation to form actin microfilaments; Eps8 was also tightly localized at the BTB (annotated by white brackets), plausibly being used to assemble actin microfilaments generated by formin 1 into actin filament bundles to support basal ES/BTB function. Yellow boxed images shown for normal testes were negative control (-ve Ctrl) in which the primary antibody was substituted by the IgG of the corresponding animal species that yielded no green or red fluorescence staining, illustrating the specificity of staining for either formin 1 or Eps8 shown herein. Box areas are images magnified and shown in insets. Scale bar, 80 µm; inset in red or in white, which is the enlarged image of the corresponding red (in normal testis) or white (adjudin-treated testis) boxed area in the micrograph at lower magnification, 30 µm; yellow boxed -ve Ctrl, 160 µm, which apply to corresponding micrographs. These results are representative findings of an experiment that was repeated using n = 3 rats and yielded similar results. B) Image analysis that semiquantitatively summarized findings shown in (A) regarding the relative distribution of formin 1 and Eps8 at the BTB was performed as described in Fig. 4. **P < 0.01.

Overexpression of Cx43 confers proper expression of BTB regulatory signaling protein p-FAK-Tyr⁴⁰⁷ at the BTB

Studies by using phosphomimetic (i.e., constitutively active) vs. nonphosphorylatable (i.e., constitutively inactive) mutants of p-FAK-Tyr⁴⁰⁷ have illustrated that p-FAK-Tyr⁴⁰⁷ is a crucial regulator of ES function at the BTB in the testis, which is likely the result of its effects on Arp2/3-neuronal Wiskott-Aldrich Syndrome protein intrinsic activity (25) because the Apr2/3 complex is a known branched actin nucleation-inducing protein (26, 27). In the normal testis, p-FAK-Tyr⁴⁰⁷ was prominently expressed at the BTB (Fig. 7A), consistent with an earlier report (25). However, treatment of rats with an acute dose of adjudin down-regulated p-FAK-Tyr⁴⁰⁷ expression considerably, but overexpression of Cx43 in these adjudin-treated rats was found to correct the spatiotemporal expression of p-FAK-Tyr⁴⁰⁷, making the expression and the distribution of this regulatory protein analogous to the normal control testes in tubules displaying signs of meiosis due to the presence of round spermatids in the seminiferous epithelium (Fig. 7A). This conclusion was supported by semiquantitatively analyzed data by quantifying the fluorescence intensity of p-FAK-Tyr⁴⁰⁷ in these tubules vs. the corresponding control groups (Fig. 7B).

Figure 7.

Overexpression of Cx43 in adjudin-treated rats with aspermatogenesis restores the spatiotemporal expression of regulatory protein p-FAK-Tyr⁴⁰⁷ at the BTB. A) In normal testes, the known BTB regulatory protein kinase, p-FAK-Tyr⁴⁰⁷ (25), was localized at the BTB (annotated by white arrowheads), near the basement membrane (annotated by a dashed white line). Following adjudin treatment, the expression of p-FAK-Tyr⁴⁰⁷ was considerably down-regulated. However, overexpression of Cx43 that restored meiosis as illustrated by the presence of round spermatids in the epithelium was found to have the spatiotemporal expression of p-FAK-Tyr⁴⁰⁷ similar to the normal testis. Yellow boxed image shown for normal testis was negative control (-ve Ctrl) in which the primary antibody was substituted by normal rabbit IgG that yielded no green fluorescence staining, illustrating the specificity of staining shown herein. Scale bar, 80 µm; inset in red or in white, which is the enlarged image of the corresponding red (in normal testis) or white (adjudin-treated testis) boxed area in the micrograph at lower magnification, 30 µm; yellow boxed –ve Ctrl, 160 µm, which apply to corresponding micrographs. These results are representative findings of an experiment which was repeated using n = 3 rats and yielded similar results. B) Image analysis that semiquantitatively summarized findings shown in (A) regarding the relative fluorescence intensity of p-FAK-Tyr⁴⁰⁷ at the BTB by using ImageJ software package. About 50 tubules per rat testis of n = 3 rats (i.e., a total of 150 tubules) were randomly selected and scored. For rats in pCI-neo/Cx43 group, tubules having round spermatids in the seminiferous epithelium illustrating resumption of meiosis were randomly selected and scored. For each tubule, 4 readings were taken such as those shown in insets at opposite ends of a tubule and their average was used for comparison. Fluorescence intensity of p-FAK-Tyr⁴⁰⁷ in normal testes was arbitrarily set at 1 against which statistical analysis and comparison were performed. **P < 0.01.

Overexpression of Cx43 re-establishes the disrupted TJ-permeability barrier by resealing the BTB in testes following adjudin-induced aspermatogenesis

Based on findings shown in Figs. 4–7 that the disrupted BTB induced by an acute dose of adjudin might have been resealed, at least in part, following overexpression of Cx43 in the testis with signs of meiosis as indicated by the presence of round spermatids, we next examined the BTB integrity using a functional assay in vivo. In brief, rats were treated with adjudin and received transfections for overexpression of Cx43 vs. empty vector (control) using the treatment regimen shown in Fig. 1A, and rats on wk 15 (15W) were used for the BTB integrity assay with n = 3 rats for each group including normal and the 2 treatment groups. In normal testes, the functional BTB blocked the entry of fluorescence tag (inulin-FITC, molecular mass at ∼5 kDa) into the adluminal compartment in which the basement membrane adjacent to the BTB was annotated by a dashed white line (Fig. 8A). Following an acute dose of adjudin, fluorescence tag was detected well into the adluminal compartment in >80% of the tubules scored (Fig. 8A, B). However, overexpression of Cx43 in the testis was found to reseal the disrupted BTB in many of the tubules examined because a considerable number of tubules displayed signs of functional BTB by blocking the entry of fluorescence tag into the adluminal compartment (Fig. 8A, B). It was noted that the number of tubules in which the BTB displayed signs of being resealed was ∼25% (Fig. 8B), consistent with the transfection efficiency of ∼30% shown in Fig. 2A. These findings thus support the notion that overexpression of Cx43 rescued the disrupted BTB, thereby allowing or facilitating the resumption of spermatogenesis, which in turn produced round spermatids via meiosis. However, round spermatids failed to enter spermiogenesis to produce elongating and elongated spermatids. In short, overexpression of Cx43 had rescued and resealed the adjudin disrupted BTB, at least in part, to reboot and sustain meiosis, with the production of round spermatids.

Figure 8.

Overexpression of Cx43 rescues the disrupted BTB mediated by an acute dose of adjudin through resealing of the TJ-permeability barrier and reboots meiosis. A) BTB integrity was performed at 15 wk. Functional BTB blocked the diffusion of inulin-FITC (M_r, ∼5 kDa; green fluorescence) from the basement membrane (annotated by dashed white lines) by traversing the basal compartment (annotated by red brackets), so that green fluorescence was not detected in the adluminal compartment in the epithelium of tubules in normal control testes (left panel) but limited to the basal compartment, with DAPI-stained cell nuclei shown in the bottom panel. In rats treated with adjudin and transfected with only pCI-neo empty vector by 15 wk, the BTB was grossly disrupted; green fluorescence was detected well inside the adluminal compartment (red brackets) as noted in middle panel. Overexpression of Cx43 in adjudin-treated testes transfected with pCI-neo vector containing the full-length Cx43 cDNA (adjudin + pCI-neo/Cx43) resealed the BTB in tubules that had signs of meiosis as indicated by the presence of round spermatids (annotated by orange arrowheads), which prevented the entry of green fluorescence tag into the adluminal compartment (see the small red bracket in fluorescence images in the right panel). Scale bar, 60 µm on the left micrograph in each panel, which applies to the remaining micrographs of the same panel. B) Data from the BTB integrity assay were semiquantitatively analyzed by measuring the distance of fluorescence tag traveled from the basement membrane (annotated by dashed white lines shown in A) in each seminiferous tubule (D_Signal) vs. the radius of the tubule (D_Radius). For oblique sections, the radius of the tubule was obtained by averaging the longest and shortest distance from the basement membrane. Each bar is a mean ± sd of 90 tubules that were randomly selected and scored from testes of 3 rats. *P < 0.05; **P < 0.01.

DISCUSSION

As noted in our earlier report that adult rats treated with an acute dose of adjudin at 125 or 250 mg/kg body weight by oral gavage failed to reinitiate spermatogenesis, which was also associated with a permanently disrupted BTB (5). This observation is consistent with earlier studies reporting that toxicant-induced irreversible BTB disruption, such as cadmium and glycerol, is also associated with sterility in male rats (2, 3, 28, 29). Furthermore, a delay of BTB assembly by 4 wk when neonatal rats were treated with diethylstilbestrol, a synthetic nonsteroidal estrogen and an endocrine disruptor, also delayed spermiation by the same amount of time in which spermatocytes failed to enter meiosis to develop into haploid spermatids but were found to undergo degeneration (1). Furthermore, the onset of meiosis closely follows the assembly of a functional BTB in the mammalian testes (30–33), including both rodents and humans. Collectively, these findings illustrate the significance of the BTB to spermatogenesis in particular meiosis. Preliminary studies conducted at the time in which adult rats were treated with an acute dose of adjudin had shown a considerable down-regulation and mislocalization of Cx43 in the seminiferous epithelium in particular at the BTB, alerting us to consider if Cx43 was playing any role in impeding these rats to reinitiate spermatogenesis or meiosis and also to the resealing of the disrupted BTB. This consideration was also based on 2 reports that the Cx43-based GJ is necessary to maintain the Sertoli cell TJ permeability barrier function (21), in particular the reassembly of the TJ during BTB remodeling based on a study using a Ca²⁺-switch model to assess the involvement of Cx43 in Sertoli cell TJ reassembly (11). Emerging evidence has illustrated the importance of GJ to coordinate cellular events that take place at the BTB to maintain its homeostasis and to coordinate other cellular functions necessary to maintain the epithelial cycle of spermatogenesis (9, 10, 34). Additionally, studies have shown that although global Cx43 KO in mice leads to neonatal lethality due to cardiac malformation (35), Sertoli cell-specific KO of Cx43 leads to an interesting phenotype (12) that partially resembles the acute dose adjudin-induced aspermatogenesis (4, 5). For instance, the Sertoli cell-specific Cx43 KO mice had reduced testis weight but no apparent effects on the male reproductive tract and the associated organs such as prostate and seminal vesicles (12). Also, spermatogonia in these mice failed to differentiate into spermatocytes to enter meiosis to develop into haploid spermatids so that these Sertoli cell-specific Cx43 KO mice are infertile (12). Furthermore, the localization of BTB-associated proteins in these mice, such as TJ-associated protein ZO-1 and basal ES-associated protein N-cadherin, was found to be perturbed because these proteins no longer localized properly to the BTB near the basement membrane (13). But there are also some dissimilarities between the 2 models. First, the population of spermatogonia in adjudin-treated rats remains relatively unaltered (5), whereas Sertoli cell-specific deletion of Cx43 leads to a reduced spermatogonia population (12). Second, the population of Sertoli cells is not grossly affected in adjudin-treated rats (4), whereas Sertoli cells remains mitotically active in Sertoli cell-specific Cx43 KO mice (36), and aggregates of undifferentiated Sertoli cells are being shed from the epithelium and found in tubule lumen in these mice (12) but not in adjudin-treated rat testes (4). In this context, it is of interest to note that in germ cell-specific Cx43 KO mice, these mice remain fertile without the noted phenotypes seen in the Sertoli cell-specific Cx43 KO mice (37). The findings using germ cell-specific Cx43 KO mice support the notion that the deleted Cx43 function in germ cells can be superseded by other connexins, but not when Cx43 was deleted in Sertoli cells, because the testis is known to express at least 10 different connexins in rodents, such as Cx33 and Cx26 (10, 38). Consistent with the 2 earlier reports that GJ is crucial to maintain BTB function (11, 21), it was shown herein that the adjudin-induced BTB disruption can be rescued and resealed in tubules displaying signs of Cx43 overexpression based on the use of a functional assay, which is also supported by proper redistribution of TJ- and basal ES-associated proteins at the BTB in tubules that had signs of meiosis such as the presence of round spermatids. The resealing of BTB is likely mediated by Cx43-induced F-actin reorganization, possibly involving p-FAK-Tyr⁴⁰⁷ as noted herein, which is an important regulator of actin nucleation at the basal ES of the BTB (25). These findings are also in agreement with recent reports that Cx43 and other GJ proteins play a crucial role on the integrity and function of other tissue barriers such as the blood–brain barrier (39) and epidermal barrier (40). Furthermore, the percentage of tubules that had their disrupted BTB resealed was estimated to be ∼25% as shown in Fig. 8, which is consistent with the transfection efficiency of ∼30% noted in findings summarized in Fig. 2A, supporting the notion that overexpression of Cx43 in testes with adjudin-induced aspermatogenesis can reboot meiosis to generate round spermatids.

A striking observation in this study is the functional relationship between Cx43 and the actin-based cytoskeletal function. For instance, testes with aspermatogenesis induced by an acute dose adjudin treatment displayed improper organization of F-actin at the BTB. This thus fails to support the localization of both TJ- (e.g., occludin-ZO-1) and basal ES-associated (e.g., N-cadherin-γ-catenin) adhesion protein complexes to confer BTB integrity. However, overexpression of Cx43 in these testes was found to induce reorganization of F-actin at the BTB, conferring proper localization of both TJ and basal ES adhesion protein complexes to the BTB. These changes are modulated by corrective spatiotemporal expression of actin regulatory proteins at the ES, namely formin 1 and Eps8, making the adjudin-treated testis analogous to the expression pattern detected in normal rat testes. These findings thus suggest that there is a functional crosstalk between Cx43-based GJ and actin-based cytoskeleton in the testis during spermatogenesis. Recent studies have supported such a functional relationship between GJ and cytoskeletal function (41, 42). For instance, Cx43 is involved in B lymphocyte motility and transendothelial cell migration by regulating actin-based cytoskeletal organization based on studies using Cx43-specific shRNA (small hairpin RNA) for Cx43 knockdown (41). A knockdown of Cx43 in B lymphocytes (e.g., WEHI 231 cells, a B-lymphoma cell line) disrupted the organization of actin microfilaments in these cells, leading a failure of transendothelial cell migration across bEND.3 and SVEC4-10 cells (both are endothelial cell lines) using an in vitro cell migration assay (41). Furthermore, pannexin 1, another GJ protein found in mammalian cells, is also known to be involved in remodeling of actin cytoskeleton in C6 cells (a rat glioma cell line) because C6 cells with pannexin1 aggregates displayed robust F-actin microfilament networks, but treatment of these cells with either carbenoxolone or probenecid, which specifically and directly blocked pannexin 1 channels, also disrupted the actin microfilaments (43). Because our findings have further demonstrated that this re-establishment of F-actin network at the BTB is mediated by actin regulatory proteins formin 1 and Eps8, it is likely that the re-established Cx43-based GJ channels provide the necessary chemical signaling function to confer the needed spatiotemporal expression of formin 1 and Eps8.

In this context, it is of interest to note that when adjudin was administered to adult rats at 50 mg/kg body weight, the effective dose that induces reversible infertility in rats, a time-dependent recovery of spermatogenesis via reinitiation of spermatogenesis from the remaining spermatogonia in the tubules was noted (4, 5). However, utilizing a single acute dose at 125 or 250 mg/kg body weight, spermatogenesis failed to repopulate the lost germ cells in the tubules, likely the result of a damaged BTB (5), unless Cx43 was overexpressed in the testis to re-establish the disrupted BTB to sustain meiosis I/II as noted herein in tubules wherein the BTB was resealed. We speculate that this resumption of spermatogenesis through meiosis I/II induced by Cx43 overexpression is possibly the result of the needed intercellular communications mediated by Cx43-based GJ, perhaps involving miRNAs (microRNAs) being transported between Sertoli cells and the residual undifferentiated spermatogonia because adjudin induced a down-regulation and mislocalization of Cx43 in the testis as noted herein. This speculation is supported by 2 lines of evidence. First, germ cells contain a large number of miRNAs, which can be used to regulate Sertoli cell function, including initiation of meiosis and spermatogenesis (44–46). Second, GJ is known to be involved in transporting miRNAs to regulate various cellular functions (47–51). However, it is obvious that overexpression of Cx43 alone fails to re-establish the entire events of spermatogenesis, in particular spermiogenesis, because multinucleated spermatids are detected in some tubules, which illustrate the onset of germ cell apoptosis, necrosis, and/or Sertoli cell injury (4, 52–55), and these cells would eventually be phagocytosed by Sertoli cells for their disposal.

In summary, we have demonstrated that overexpression of Cx43 in the testis can reseal the disrupted BTB following adjudin treatment by conferring proper localization of TJ and basal ES proteins to the BTB microenvironment. This also repairs spermatogenesis, at least through meiosis. These findings are of interest because they suggest that toxicant-induced BTB dysfunction and male infertility including humans induced by other toxicants, such as cadmium, phthalates, and PFOS (perfluorooctanesulfonate), can possibly be therapeutically managed by overexpressing Cx43 because these toxicants are recently also shown to exert their effects via changes in the actin-based cytoskeleton at the Sertoli cell BTB both in rodents and humans (56–58).

Glossary

BTB

blood-testis barrier

Cx43

connexin 43

Eps8

epidermal growth factor receptor pathway substrate 8

ES

ectoplasmic specialization

FAK

focal adhesion kinase

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GJ

gap junction

KO

knockout

miRNA

microRNA

qPCR

quantitative PCR

shRNA

small hairpin RNA

TJ

tight junction

ZO-1

zonula occludens 1