Connexin 43 a check-point component of cell proliferation implicated in a wide range of human testis diseases

Abstract

Gap junction channels link cytoplasms of adjacent cells. Connexins, their constitutive proteins, are essential in cell homeostasis and are implicated in numerous physiological processes. Spermatogenesis is a sophisticated model of germ cell proliferation, differentiation, survival, and apoptosis, in which a connexin isotype, connexin 43, plays a crucial role as evidenced by genomic approaches based on gene deletion. The balance between cell proliferation/differentiation/apoptosis is a prerequisite for maintaining levels of spermatozoa essential for fertility and for limiting anarchic cell proliferation, a major risk of testis tumor. The present review highlights the emerging role of connexins in testis pathogenesis, focusing specifically on two intimately interconnected human testicular diseases (azoospermia with impaired spermatogenesis and testicular germ cell tumors), whose incidence increased during the last decades. This work proposes connexin 43 as a potential cancer diagnostic and prognostic marker, as well as a promising therapeutic target for testicular diseases.

Introduction

Disruption of testis development and of spermatogenesis, consequently to neonatal exposure to endocrine disrupters, is now widely recognized [1]. Identification of the mechanisms that drive these features is important because fetal exposure is held responsible for the increasing incidence of male infertility and testicular cancer (reviewed in [2, 3]). Spermatogenesis is a highly regulated process of germ cell proliferation, differentiation, survival, and apoptosis, starting from gonocytes to spermatocytes and giving rise to spermatids, the future spermatozoa, which occurs in the testis within the seminiferous epithelium (Fig. 1). During spermatogenesis, the spermatogonia (2n) undergo mitosis, followed by a cellular transformation from type B spermatogonia into spermatocytes, which enter meiosis to form spermatids (1n) and finally develop into spermatozoa via spermiogenesis. Spermatogenesis and spermiogenesis occur within the seminiferous epithelium and their normal progression depends on cell–cell interactions between the germ cells and Sertoli cells, which have been described as the “nurse cells”, capable of creating a proper environment for the development of germ cells and to provide essential nutriments and growth factors. This local control may be partly mediated through direct intercellular communication channels made up by gap junctions and a family of constituted proteins, the connexins (Cx) [4].

Fig. 1.

Schematization of testis section illustrating the seminiferous tubules composed by Sertoli cells and germ cells at different stages of development and the interstitium with Leydig cells

Based on our recent findings, we hypothesized that testicular Cx dysregulation could partly participate in the etiopathology of human infertility and testicular cancer [5, 6]. A large number of experimental studies suggest that targeting Cx may be a novel approach either to inhibit tumor cell growth directly or to restore normal cell growth, indicating that the therapeutic potential of connexins is undeniable [7]. This review focuses on the regulation and the function of Cx43, the predominant testicular gap junction protein, in human male infertility and testicular germ cell cancer with emphasis on their therapeutic potentials.

Human male infertility

Unexplained male infertility occurs in 30–40 % of men with abnormal semen parameters [8]. Obstructive azoospermia with normal spermatogenesis and nonobstructive azoospermia with distinct degrees of abnormal spermatogenesis represent the most cases of male infertility. The origins of spermatogenic defects in infertile patients are multifactorial. Endocrine disruption of testis development during neonatal period, due to environmental pollution, genetic, and epigenetic factors, is the most potential explanation evoked for the unexplained male infertility [2, 8, 9]. This defect can lead to testicular dysgenesis, male infertility, and to an increased risk of testicular malignancy, as recently suggested [2]. However, the underlying molecular mechanisms that drive these testicular pathologies are mostly unknown, and remain to be investigated [10].

Within the seminiferous epithelium, cell–cell interaction between the somatic cells, Sertoli cells, and the different germ cells are essential in the local control of spermatogenesis and are mediated through paracrine pathways and direct intercellular communication channels made up by Cx [4]. By using transgenic animals, it has been well established that one of these Cx, Cx43, the most predominant Cx in the testis, is a prerequisite for normal spermatogenesis [11–16] and it has been hypothesized that testicular Cx dysregulations could partly participate in the etiopathology of human male infertility [5, 6].

Gap junction channels, connexins, and associated human diseases

Gap junctions are intercellular plasma membrane channels that create electric and metabolic coupling from cell to cell in a wide variety of tissues. Small molecules (<1 kDa), metabolic precursors, nutrients, and signaling factors such as inositol 1,4,5-tri-phosphate, cAMP, cGMP, glutathione, and ions can be transferred directly from one cell to adjacent cells through gap junction channels whereas larger molecules as proteins, polysaccharides and complex lipids are retained within the cytoplasm (reviewed in [17]). Gap junctions are formed by the docking of two hemi-channels, termed connexons, present at opposing plasma membranes of adjacent cells, resulting from the oligomerization of six molecules of Cx (Fig. 2). To date, at least 21 members of this multigene family have been cloned in mammals [18]. Current nomenclature is based on their molecular weights. This process of molecular signal exchange through gap junctions, called gap junction intercellular communication (GJIC), has been involved in several cellular processes including control of cell growth, intercellular synchronization, and co-operation and hormone responsiveness. Cx can be regulated at different levels including transcription, mRNA processing, protein synthesis, post-translational modification (phosphorylation), assembly in the Golgi apparatus, trafficking to the plasma membrane, connexon docking, and gating of the channel (reviewed in [19]).

Fig. 2.

Schematic architecture of gap junction plaque, gap junction channel, gap junction hemi-channel or connexon and connexin. Each intercellular channel or gap junction provides an axial channel that interconnects the cytoplasms of two adjacent cells (cell 1 and cell 2) and allow the direct exchange of molecules with a relative molecular mass up to 1 kDa. Hemi-channels or connexons are formed of six connexin subunits that exhibit four transmembranous domains (M1–M4), two extracellular loops (E1 and E2), and three cytoplasmic domains including one intracellular loop (IL) and cytoplasmic NH₂- and COOH-termini

Disruption of GJIC could be due to different processes such as mutations of Cx, altered expression of the Cx genes, or impaired trafficking of the protein to the plasma membrane [20]. Genetic approaches have uncovered a still growing number of mutations in Cx genes related to human diseases, including peripheral and central neuropathies, deafness, skin disease, cataracts, and cardiovascular dysfunctions, suggesting that Cx are crucial for a large number of physiological processes [21, 22] (see Table 1). In addition, it has been reported that a loss of gap junctions also appears as one of the primary events of uncontrolled cell proliferation, giving rise to tumor development [23]. Indeed, impaired GJIC and Cx expression have been reported in many human tumors, in nearly all malignant cell lines and after exposure to tumor promoters and oncogene expression (Table 1) [24, 25]. Cx have been considered as tumor-suppressor genes since Cx transfection into transformed cells restores in most cases normal cell growth, leading to the proposal that pharmacological Cx up-regulation may have therapeutic implications in cancer [26].

Table 1.

Connexin-associated human diseases and cancers (adapted from [21–25])

Human diseases	Connexins
X-linked Charcot–Marie–Tooth disease (CMTX)	Cx32
Atrial fibrillation	Cx40
Viscero-atrial heterotaxia	Cx43
Cataract	Cx46, Cx50
Oculodentodigital dysplasia (ODDD)	Cx43
Clouston syndrome	Cx30
Hearing loss (non-syndromic or associated with skin	Cx26, Cx30, Cx31
Keratitis ichthyosis deafness syndrome	Cx26, Cx30
Erythrokeratodermia variabilis	Cx30.3, Cx31
Pelizaeus–Merzbacher-like	Cx46.6, Cx47
Vohwinkel syndrome	Cx26

Human cancers	Connexins
Head and neck carcinoma	Cx31.1
Nasopharyngeal carcinoma	Cx43, Cx45
Meningioma and hemangiopericytoma	Cx26, Cx43
Human larynx carcinoma	Cx26, Cx30, Cx32, Cx40
Liver tumor	Cx32, Cx43
Colon cancer	Cx26, Cx43
Esophageal cancer	Cx26, Cx32, Cx43
Breast cancer	Cx26, Cx43
Mesothelioma tumor	Cx37, Cx43, Cx45
Glioblastoma	Cx43
Lung cancer	Cx43
Adrenocortical tumors	Cx43
Renal cell cancer	Cx32, Cx43
Cervical carcinoma	Cx43
Ovarian carcinoma	Cx43
Endometrial carcinoma	Cx43
Prostate tumor	Cx26, Cx32, Cx43
Bladder cancer	Cx26
Thyroid carcinoma	Cx26, Cx32, Cx43
Testis cancer	Cx26, Cx40, Cx43
Hepatocarcinogenesis	Cx26, Cx32
Skin tumor	Cx26, Cx30

Gap junction, connexin 43, and testicular diseases

Connexin and testis

Transcripts for at least 11 different Cx (Cx26, Cx30.2, Cx31, Cx31.1, Cx32, Cx33, Cx37, Cx40, Cx43, Cx46, and Cx50) have been detected in the rodent testis [27] and many of them are also present in the male genital tract within prostate, seminal vesicles, epididymis, and corpus cavernosum (reviewed in [28]). In the testis, Cx43, also named GJA1 (gap junction protein alpha 1), is present in Leydig cells and in seminiferous epithelium between Sertoli cells and between Sertoli cell and germ cells [29].

In the human testis, Cx40 transcripts were identified [30], Cx26 protein has been detected within seminiferous tubules [31] and Cx31.9 was found expressed in testicular vascular smooth muscle [32, 33]. The Cx43, which appears to be the most abundant Cx within the human testis, is ubiquitously expressed within most cell types [29]. Cx43 mRNA and protein are present in Leydig cells and within the seminiferous tubules between Sertoli cells and spermatogonia or primary spermatocytes [34–37].

To characterize the role of these Cx isoforms into the reproductive function, genomic deletion of these proteins was achieved. If knock-out (KO) animals for Cx31, Cx32, Cx40, Cx46, and Cx50 show normal fertility [38], Cx43-deficient mice die at birth [39] but exhibit a 50 % depletion in primordial germ cells [11]. To overcome this perinatal lethality, testes from Cx43-null mutant fetuses were grafted under the kidney capsules of adult males [13], or Cx43 gene was substituted by the coding sequences of Cx26, Cx32, or Cx40 [12, 16]. In both cases, the reproductive function was strongly impaired with abnormal proliferation and differentiation of germ cells or spermatogenesis arrest at the level of spermatogonia, leading to a Sertoli-cell-only (SCO) phenotype. Even if these studies supported the unique role of Cx43 in spermatogenesis, the contribution of Cx43 deletion in other tissues was nevertheless not verified. Thus, in order to avoid both perinatal lethality and pleiotropic effects on other developmental tissues, authors generated a conditional Cx43 KO mouse using the Cre/loxP recombination system, which lacks the Cx43 gene solely in Sertoli cells by crossing two transgenic mouse lines, AMH-Cre mice and Cx43-floxed LacZ mice [14, 15]. Adult male transgenic animals showed normal testis descent and development of the genital tract. However, the size of the testes was reduced and most seminiferous tubules were devoid of germ cells, with abnormal proliferation and intermediate phenotype expression of Sertoli cells [40]. If these conditional KO experiments confirmed the key function of Cx43 in spermatogenesis, they did not, however, give more information on the role of this Cx within each testis cells.

Since fetuses from Cx43 KO mice exhibit a marked depletion in primordial germ cells (PGCs), it is likely that Cx43 participates in the control of PGC survival and/or migration [11]. Previous studies demonstrated that Cx43 is essential for PGC motility in 8.5- and 11.5-day-old embryos and cell survival in older embryos, with abnormal p53 activation playing a crucial role in the apoptotic loss of PGCs at the latter stages in the Cx43 KO mouse embryos [41]. Within Sertoli cells, the detection of Cx43 expression has been found to correlate with cell differentiation [35, 42]. Then, it has been hypothesized that Cx43 could actively control Sertoli cell proliferation and differentiation during neonatal period, as suggested by the observations in Sertoli cell conditional Cx43 KO mice [14, 15]. Recently, we reinforced this possibility by developing a Cx inhibitory mimetic peptide strategy, which reveals that Cx43 between Sertoli cells control Sertoli cell proliferation whereas Cx43 between Sertoli cells and spermatogonia regulate germ cell survival rather than germ cell proliferation [43].

During the time course of testis development, it has been postulated that Cx43 gap junctions, present in the perinatal testis, control Sertoli cell differentiation and through this way maintain the number of germ cells (gonocytes or primitive spermatogonia). Afterwards, Cx43 could control germ cell differentiation and appears essential for meiotic progression of spermatocytes [44, 45]. Sertoli cells would ensure metabolic and signaling coupling to germ cells through Cx43 gap junction channels and allow synchronization of male germ cell proliferation, differentiation, and survival [43]. Consistently, microarray analysis of pre-pubertal mice in which Cx43 was specifically depleted in Sertoli cells revealed that an important amount of germ cell-specific genes, essential for mitotic and meiotic progression of spermatogenesis, were down-regulated. Moreover, other genes controlling transcription, metabolism, cell migration, and cytoskeleton organization were also altered [46]. Recent studies suggested that Cx43 could also indirectly drive the spermatogenic process by controlling tight junction proteins, which form the blood–testis barrier (BTB) [47, 48]. The involvement of such a regulation is of interest since it is well established that the BTB plays a crucial role in spermatogenesis by forming an intricate proteinaceous network, which segregates and protects the post-meiotic germ cell from unwanted biomolecules or environmental toxicants.

Connexin expression in testis with impaired spermatogenesis

Although endocrine disruption of testis development, during the neonatal period, is the most potential explanation evoked for both the increased incidence of male hypofertility and testicular malignancy, the molecular mechanisms at the origin of these testicular pathologies are unknown. To date, there is strong evidence suggesting that most cases of this testicular dysgenesis syndrome are due to environmental factors (reviewed in [2]). We recently hypothesized that testicular Cx43 dysregulations could participate in the etiopathology of human male infertility and testis cancer since Cx43 is (1) essential for spermatogenesis, (2) altered during tumoral process, (3) a specific target for testicular Cx (reviewed in [5, 6, 25]).

There are few data reported in the literature that inform on the status of testicular gap junctions in the pathological testis of patients with abnormal spermatogenesis. Previous freeze-fracture studies have not reported any variation in the number of gap junctional particles in the seminiferous tubules of azoospermic and oligozoospermic patients [49]. However, no information was given on the origin and degree of the spermatogenesis defect. Inversely, gap junction-like cell membrane specializations are very rare in hypo- or azoospermic patients [50] and are affected in SCO seminiferous tubules [51]. This alteration is accompanied by a decrease in immunoreactive Cx43 [34]. Moreover, recent studies from our laboratory demonstrated that the Cx43 mRNA amount decrease in the testis of rat, with impaired spermatogenesis following experimental cryptorchidism, is concomitant with a disappearance of Cx43 protein in the seminal epithelium, whereas Cx43 is still present in the interstitial tissue [35]. Similar observations were reported in other species [52]. In addition to cryptorchidism, inflammation of the male reproductive tract, resulting from interstitial infiltration of T lymphocytes, is an important etiological factor of infertility [53]. A recent study reported that the germ-cell loss observed in rats with experimental autoimmune orchitis is associated with early alterations of Cx43 cell–cell gap junction communication [54]. In azoospermic patients with severe spermatogenesis failure, we reported that the disturbed expression of Cx43 mRNA might be related to a defect in the maturation of Sertoli cells [35]. Thus, the possibility that Sertoli cell Cx43 impairment could be a sign of undifferentiated Sertoli cell functionality has been hypothesized. More precisely, other authors found a significant positive correlation between the histological score and intensity of the testicular Cx43 expression, when the spermatogenesis defect was evaluated in oligozoospermic men by Johnsen score [36]. While Cx43 staining is quite similar in human testes with hypospermatogenesis and spermatogenic arrest at the level of round spermatids or spermatocytes compared to healthy normal adult testis, seminiferous tubules with spermatogenic arrest at the level of spermatogonia and Sertoli-cell-only syndrome were completely immunonegative [34]. From these data, we suggested that the communication between Sertoli cells and germ cells through Cx43 may be important for the progression of spermatogenesis specifically at some precise stages of germ cell development. For unknown reasons, sterility induced by epididymal and vasal ligation was able to affect Cx43 expression in the testis of the experimental rat model, suggesting that Cx43 is a key molecule in the reproductive process [55].

Although these findings demonstrate that testicular Cx43 expression is reduced in mutant rodents, with testicular defects, and in hypofertile men, the existence of a direct relationship between disrupted Cx and spermatogenesis arrest has not been clearly established. Consequently, it is still questionable if the altered Cx expression is a consequence of impaired testicular function rather than the cause of the testicular pathology (reviewed in [28]). Since experimental studies have demonstrated that Cx43 is essential in the initiation and maintenance of spermatogenesis, as revealed by the analysis of mice lacking Cx43 through germ line knock-out of Gja1 [11–16], the possibility that Cx43 drives and controls spermatogenesis has been suggested. To test this hypothesis, phenotype analysis of Cx43 gene mutation and consequence on testicular function have to be analyzed. Over the last decades, the role of Cx signaling in human physiology has been highlighted by the discovery of numerous Cx gene mutations and Cx protein alteration, in a wide range of tissues. These Cx defects cause a large variety of severe human pathologies, such as myelin-related neuropathy diseases, skin diseases, hearing loss, congenital cataract, or more complex syndromes, such as cardiovascular dysfunctions and the oculodentodigital (ODDD) dysplasia (reviewed in [21, 22]) (Table 1). However, no effective and consequent testis dysfunction was examined and reported in men for most Cx gene mutations (reviewed in [56]). Since Cx43 is the predominant Cx in the human testis, the consequence of Cx43 mutation needs to be analyzed. To date, there is only one report in which human genital tract abnormalities (hypospadias, undescended testis) have been reported in male patients and infertility in a female [57]. Experimentally, a Cx43 mutant from a mouse model for ODDD, with a Cx43 dominant loss of function, exhibits impaired oogenesis due to granulosa cells dysfunction [58]. This interesting result is in agreement with the existence of a direct relationship between altered Cx43 expression and reproductive defects, since granulosa cells and Sertoli cells play similar supporting functions for germ cells during oogenesis and spermatogenesis, respectively. However, it must remain cautious because in the last data cited phenotypic descriptions of ODDD male patients have not concerned reproductive function. Thus, since this feature has not been investigated in the clinical protocols, these patients have not been reported to be infertile. Recently, the exploration of Gja1^Jrt/+ mice, which carry a dominant mutation that causes an amino acid substitution (G60S) and mimic the phenotype of ODDD, revealed a loss of germ cells in some seminiferous tubules and reduced sperm count and sperm velocity parameters, supporting, for the first time, the possibility of subfertility in ODDD human males [59]. Thus, it is possible that, as in mutant mice, many ODDD patients could present signs of hypofertility, but this defect is not sufficiently severe to lead to infertility. This hypothesis is in agreement with our previous data demonstrating a relationship between the decrease in testicular Cx43 and the severity of spermatogenic defect in men [35]. Further studies must be developed in male subjects harboring Cx43 mutations with ODDD syndrome and accurate analysis of reproductive function and sperm parameters (number, viability, velocity) must be performed.

Testis cancer and connexin 43

Competent differentiating Sertoli cells are required for spermatogenesis, spermiogenesis, and the production of normal spermatozoa. Therefore, dysfunction of Sertoli cells leads to hypospermatogenesis, which may lead to hypofertility or complete infertility [60]. During neonatal development, Sertoli cell dysfunction is associated with an arrest of gonocyte production at an early stage of development. These undifferentiated primitive germ cells are thought to be the forerunners to carcinoma-in situ (CIS) testis syndrome that, in turn, can develop into testicular germ cell cancer (reviewed in [61]). Interestingly, it has been proposed that testis tumor progression may be correlated with a disrupted Sertoli cell function resulting from an abnormal and undifferentiated status of the somatic cells (reviewed in [62]). According to this hypothesis, epidemiologic studies demonstrated an increased incidence of testicular cancer in infertile men who exhibit abnormal semen analysis [63–67]. In addition to genetic abnormalities due to SRY mutations, male infertility and testicular cancer have been associated with defects in DNA repair genes, tumor suppressor gene mutations, and epimutations [68]. Several genes that could be involved in the pathogenesis of testicular germ cell tumors have been identified [69, 70]. In addition to these genes, Cx43 could be a good candidate in the etiopathogeny of testis diseases. In agreement with this hypothesis, Cx gene expression is abnormally down-regulated in most tumoral tissues (listed in Table 1) and cell lines [23–25]. Moreover, restoration of normal phenotype in transformed cells, by transfection of exogenous Cxs, gave rise to the concept that the gap junction proteins can act as tumor suppressor genes (reviewed in [26, 71]). In the human testis, the two gap junction proteins Cx43 and Cx26 were found to be associated with neoplasic progression of CIS. Indeed, down-regulation of Cx43 expression was found in the pathological testes of patients with CIS, Sertoli cell tumor, seminoma, and non-seminomatous germ cell tumors, as compared to healthy testis [30, 31, 72, 73]. It is noteworthy that the down-regulation of Cx43 gene expression during tumoral progression from CIS to seminoma is concomitantly associated with an up-regulation of another Cx, Cx26 [31]. Cx26 overexpression and cytoplasmic accumulation has also been reported in numerous carcinomas (pancreas, head and neck, breast, colon, prostate), in keratinocyte-derived skin tumors and in human papillary thyroid and follicular thyroid cancers, suggesting a non-tissue-specific role of this Cx in the tumoral process (reviewed in [23]). Thus, it is possible that the compensatory mechanism that occurs between Cx43 and Cx26 was unable to overtake the function of Cx43, since Cx43 exerts a unique role in the testis as mentioned above.

In addition to reduced mRNA and protein expression, delocalization of the membrane protein within the cytoplasm has been reported in many human cancers such as prostate, colorectal, gastric adenocarcinomas, breast, and hepatocellular carcinoma [74–78]. In neoplasic cells originating from a seminoma cell line, we reported that Cx43 protein was aberrantly trafficked with accumulation of the protein within the cell cytoplasms [72]. Overexpression of wild-type Cx43 by transfection of a Cx43 vector not only restored gap junctional intercellular communication but also blocked abnormal proliferation of these cells. In agreement with this previous study, aberrant cytoplasmic Cx43 accumulation has also been demonstrated in pure human testicular seminoma tissues but the precise intracytoplasmic compartment was not clearly evidenced [79]. In Leydig cell tumors, the presence of Cx43 was detected within early endosomes, suggesting an excessive internalization of the gap junction protein [80]. Although the role of this intracytoplasmic presence of the gap junction protein is questionable, this aberrant localization in tumoral tissue has been suggested to stimulate tumor processing [74, 81] and has been observed in many cell types exposed to carcinogens (reviewed in [25]). This result is in agreement with previous findings demonstrating that Cx itself could control cell growth independently of cell–cell communication [82–84].

Consequently from these data, several causes for the invasive potential of CIS cells could be advanced. First, it may be speculated that the independence of these pre-invasive tumor cells to proliferate anarchically could result from impaired control of germ cells proliferation by mature Sertoli cells through functional Cx43 gap junctions (Fig. 3). This could happen if developing Sertoli cells failed to differentiate due to Cxs disruptor [5] and/or mature Sertoli cells dedifferentiate [31, 85]. Indeed, disruption of Sertoli cell control on germ cell proliferation is in agreement with our previous experimental observations demonstrating that (1) the transfer of signaling molecules through gap junctions occurs unidirectionally from Sertoli to germ cells [86], (2) such an independence may also occur in physiological conditions during specific stages of spermatogenesis between Sertoli cells and some germ cells, such as spermatogonia [87]. Second, it is also possible that the undifferentiated and preinvasive tumor germ cells, which escape Sertoli cell control, could accumulate Cx43 protein within their cytoplasms. This aberrant intracytoplasmic accumulation could further favor human seminoma development through GJIC-independent mechanisms. Such dual effect of Cx has been previously suggested in the breast where Cx first acts as a tumor suppressor in the primary tumors and then as an enhancer of tumor progression (reviewed in [88]). This possibility is also supported by recent findings demonstrating that a loss of intercellular gap junctions and a gain of intracytoplasmic Cx26 and Cx32 can play an important role in the formation of gastric and colorectal adenocarcinomas [77, 78]. Interestingly, cytoplasmic Cx32 exerts stimulatory effects on some steps of hepatocellular carcinoma progression, such as invasion and metastasis, when the cells have acquired a malignant phenotype [81]. The molecular basis of such a GJIC-independent role of Cxs is still unknown. Nevertheless, a recent study evidenced that cytosolic Cx43 regulates the dynamic of microtubule network required for cell migration as well as cellular polarity [89]. Thus, it could be postulated that such drastic effects on cytoskeleton and polarity could participate in the epithelial–mesenchymal transition initiating cancer development and metastasis.

Fig. 3.

Schematic illustration of human testicular diseases associated with impairment of Cx43. In physiological situations, Sertoli cell Cx43 independently or dependently of gap junction channels participates in the control of germ cell proliferation, differentiation, survival, and apoptosis required for spermatogenesis. In physiopathological conditions, for example under the influence of genetic, epigenetic and/or environmental factors, a decreased or complete rupture of the control of germ cells by Sertoli cells through Cx43 can conduce either to hypospermatogenesis, which first leads to man hypofertility, or to abnormal tumor cell proliferation, which finally leads to testicular germ cell tumors

Altogether, these data suggest that the down-regulation of Cx43 and the shift in its localization, from the cell membrane to the cytoplasm, could be early events of tumor progression associated with uncontrolled cell proliferation. These two events could serve as indicators of carcinogenesis and, as such, as additional neoplasic biomarkers during tumor germ cell progression. In the testis, the Cx43 down-regulation and delocalization [31, 80], concomitant with Cx26 up-regulation, have been postulated to be early signs of carcinogenesis [73]. The recent demonstration of a similar shift of expression and delocalization in some areas of the contralateral testis biopsies of patients with non-seminomatous testicular germ cells, although no CIS was evidenced, strongly supports this assumption and could be indicative of malignant dormant germ cell tumors [90].

Clinical and therapeutic applications of connexin 43 in testicular diseases

Cx43 as diagnostic biomarker of testicular diseases

The incidence of two diseases related to germ cell proliferation, hypospermatogenesis and testicular cancer, has been reported to continuously increase over the last 3–5 decades in men of different industrialized countries [91, 92]. In addition, epidemiological studies argued for an increased risk of testicular germ cell cancer among infertile men [66, 93–95]. Consequently, the early recognition of the precursor lesion before CIS stage in the non-obstructive azoospermic testes and the determination of imaging and/or serum tumors markers in the failing human testes are challenging issues. Indeed, tumor marker identification could be useful in the prognostic diagnosis and staging of the disease, for monitoring the therapeutic response and for analyzing tumor recurrence. Several markers of testis cancers have been evaluated in serum, semen, and tissue samples [96]. Regarding testicular dysgenesis syndrome, recent studies proposed the Hiwi protein and chromosome 12 aneuploidy, DNA mismatch repair, and Y-chromosome instability as possible connections between male infertility and testicular germ cell tumor [95]. The present review suggests that alterations of Cx expression, particularly its aberrant localization within testicular cells, could be a prognostic indicator for uncontrolled cell proliferation and consequently of tumor progression in the infertile testis. Analysis of Cx43 disruption may be particularly useful for screening young men at high risk of testicular cancer (hypofertility, bilateral cryptorchidism, contralateral testicular cancer).

Characterization of noninvasive prognostic biomarkers is another strategy that could be developed to facilitate the management of infertile patients and to evaluate the potential of cancer progression. However, previous studies have reported that analysis of the levels of AFP, hCG, PLAP, and LDH, the most routinely used markers, do not appear as good predictors of testicular cancer in this population of men at high risk [97, 98] and few of them exhibited high diagnostic specificity and sensitivity [99]. Recent studies have demonstrated that the production patterns in body fluids of microRNAs (miRNAs) is highly correlated with various pathologies and could be used as a novel class of noninvasive biomarkers to diagnose and monitor human cancers [100, 101]. These RNA-based elements are small non-coding RNAs with a broad range of functions essentially in gene translation regulation. Recently, altered profiles of miRNAs have been reported in the seminal plasma of infertile men [102] and in human testis samples with CIS [103]. Interestingly, circulating specific miRNAs and miRNAs coding for Cx have been postulated to be biomarkers or mediators of cardiovascular diseases [104, 105]. Thus, further research is needed to determine if alterations of circulating levels of miRNAs targeting human Cx43 mRNA, in the serum and also in the seminal plasma, could be associated with male hypofertility and the increased risk of testicular germ cell tumor development.

Cx43 as a potential therapeutic target

Based on the experimental data that support a tumor suppressor role for Cx, (reviewed in [26, 71]), the gap junction proteins have been viewed as potential therapeutic targets for inhibiting tumor development. Thus, it is likely that the restoration of Cx expression could be an attractive strategy to restore male fertility and inhibit tumor germ cell development. Various strategies, including the use of therapeutic agents, Cx gene therapy or both approaches, have been developed [106].

Combination of Cx-dependent tumor-suppressive effect and chemotherapeutic agents, which has been used for clinical tumor treatment, could be effective in enhancing the sensitivity of the tumor to cytotoxic treatment. Previous studies demonstrated that overexpression of Cx43 increases cell susceptibility to several common chemotherapeutic agents, including etoposide, paclitaxel (Taxol), and doxorubicin in a gap junction communication-dependent or -independent manner [24, 25, 106, 107]. The enhanced cell toxicity can result from intercellular diffusion of toxic/apoptotic signals mediated through gap junction channels and the degree of this toxic “bystander” effect generally correlates with the level of GJIC (reviewed in [108]). The possibility that Cx43 sensitizes cells to drug-initiated cytotoxicity, possibly through hemichannel-mediated effects on intracellular oxidative status, has also been hypothesized [109]. In addition, recent findings demonstrated that, in response to cisplatin, a widely used chemotherapeutic agent to treat testis cancer, gap junctions propagate opposite in vitro effects: protective in normal cells and toxic among cancer testicular cells [110]. However, little data are related to pathological testis in in vivo situations. This is worrying since epidemiological studies argued, during the last decades, for both an increased incidence of testis cancer, which is the most common malignant disease occurring in young adult men, and for an increased risk of testicular germ cell tumor development among the infertile male population [66, 93–95].

From our point of view, in the research of therapeutic physiological or chemical effectors that can target human testis diseases, a particular interest has been given to the granulocyte colony-stimulating factor (G-CSF). G-CSF is a member of the hematopoietic growth factor family, which controls the proliferation, differentiation, and survival of hematopoietic progenitor cells [111]. This cytokine, which has recently been described as a protector of spermatogenesis [112], has successfully been used in combination with chemotherapeutic agents in the treatment of testis cancer. Testicular germ cell tumor is the most frequent cancer occurring in young men and originates from transformed gonocytes or undifferentiated spermatogonia, which respectively derived from fetal germ cells and adult germ stem cells. Seminoma is the most frequent (50–70 %) testicular germ cell tumor. Interestingly, G-CSF, which increases Cx43 expression and redistribution of Cx43 at the plasma membrane, has recently been used as a therapeutic tool for heart disease [113, 114]. Whether such a treatment could be useful for restoring Cx43 expression in the pathological testes, stimulating germ cell progression in the testis of man with hypospermatogenesis or blocking abnormal germ cell proliferation in the tumoral testis, deserves to be investigated.

In addition, since in most cases of testicular tumors, Cx43 have been detected mainly present within the cytoplasm and absent at the plasma membrane, the hypothesis to stimulate the trafficking of Cx43 towards the membrane or to inhibit the endocytosis of membranous Cx43, has been postulated to correct this defect. However, to date, a better understanding of the molecular interactions that occur between Cxs and their protein partners, during the full trafficking of the gap junction proteins, is of primary importance (reviewed in [115]). Thereafter, a future clinical challenge will be to identify new pharmaceutical agents, capable of restoring normal Cx43 trafficking, localization, and functionality of gap junction channels, and consequently to prevent or reverse the tumoral processing in the pathological testis.

Another strategy to increase Cx expression in the pathological testis is Cx gene transfer. In the testis, germ cells are appropriate target cells for generating transgenic animals whereas the somatic cells, Sertoli and Leydig cells are more appropriate target cells for clinical applications. This molecular approach has been successfully developed in laboratory animals for correcting infertility resulting from Sertoli cell defect [116]. Thus, Cx gene transfer has been proposed as a revolutionary advance for reproductive treatment in patients with male infertility and for testicular cancers [117]. There is growing evidence that germ cell tumors, which represent 95 % of all testicular cancers, derive from germ cells that exhibit at a precise time of development abnormal gene programming, and that testicular tumoral progression is strongly reinforced by the absence or altered communication between Sertoli cells and germ cells. However, due to the complexity of germ cell tumor pathogenesis, the multitude of potential candidate male fertility-associated genes [118, 119] and the potential side-effects caused by the integration of the transgene in the germ line, there is to date rare data on gene therapy for germ cell tumors. Such genetic approach has been, however, successfully developed for REIC/Dkk-3 gene, which is down-regulated in seminoma and which can be used as a gene-therapeutic agent against testicular cancer [120]. In the absence of a general strategy that focuses on restoration of specific germ cell-deficient proteins, which may be missing or mutated in the diseased individual, therapeutically Cx43 increase through Cx transfer gene therapy could be another useful alternative for controlling tumor progression. Indeed, overexpression of Cx43, in pathological Sertoli cells unable to control their own proliferation and differentiation and those of germ cells, could reverse hypospermatogenesis in the azoospermic testis and prevent tumor progression. In keeping with this hypothesis, Cx43 gene transfer has been successfully validated in animal models [121, 122] and attempted in preclinical human studies for diseases, such as loss of hearing function [123] and atrial arrhythmias [124].

Another clinical approach for enhancing the reduced Cx43 levels found in the pathologic testis is to control miRNAs. As discussed above, their involvement in the neoplasic development of germ cell tumors has been well documented (reviewed in [100, 101, 125]). In addition, alteration of miRNAs associated with male infertility and testis tumor has also been reported [102, 103]. Thus, viewed from this angle, the discovery of potential regulatory miRNAs targeting human Cx43 is of interest for developing a new form of gene therapy. Cx43 has been shown to be targeted by two miRNAs miR-1 and miR-206 that are able to down-regulate Cx43 expression [126, 127]. In agreement with these findings, a recent clinical study reported that the misregulation of Cx43, resulting from altered miR-1 processing, could be responsible of the human cardiac dysfunctions in patients with myotonic dystrophy [128]. Although miR-206 was previously viewed as a muscle-specific miRNA, another study demonstrated its presence in osteoblasts and its capacity to inhibit their differentiation [129]. In the testis, it is unknown if Cx43, a target of miR-1 and of miR-206, could be controlled by these two miRNAs and if their disturbing expression could be associated with testicular diseases. In this context, it is worth noting that the expression of miR-1 is altered in testicular tissues of patients with non-obstructive azoospermia [130]. Thus, modulation of miRNA either by using anti-sense oligonucleotides or by transfection or infection could provide a new potential therapeutic window to control Cx43 expression and subsequently testis infertility and neoplasia. Interestingly, it has been suggested that down-regulation of another microRNA-383 is associated with male infertility and is able to promote testicular germ cell tumors [131].

Conclusive remarks

Increased incidence of male infertility and testicular cancer is a general feature of the last decades. In the testis, the gap junction protein Cx43, which is present between the somatic Sertoli cells and between Sertoli and germ cells, is essential for normal spermatogenesis and appears to be generally impaired in testicular diseases. In seminiferous tubules, Sertoli cells exert crucial functions: they initiate, promote, and maintain spermatogenesis and Cx43 plays an essential role in these process as recently evidenced by the use of transgenic Sertoli cell conditional Cx43 knock-out mice [14, 15]. As proposed for other human pathologies such as heart diseases [132, 133] and erectile dysfunction [134], the present review raises hope that the restoration of Cx expression and/or gap junction intercellular communication between Sertoli cells and germ cells might be a significant tool to correct impaired spermatogenesis and to prevent possible associated testicular germ cell tumors in the hypofertile man population. Transcriptional and post-transcriptional processings of mRNA could successfully reinitialize Cx43 expression in most cancer cells [106], either by overexpression of the lacking or altered Cx or by modulating the miRNAs that control Cx43 mRNA. If miR-1 and miR-206 are the two best-explored miRNA regulators of Cx expression, the potential role of other identified miRNAs, such as miR-145 and miR-218 [135, 136], or still unidentified miRNAs, which also target Cx43, the crucial predominant Cx for spermatogenesis, remains to be explored. If this strategy for enhancing Cx43 levels by gene-transfer technique is clinically challenging in many organs, such an approach raises, however, practical difficulties that still seem to be far from being resolved for testicular diseases. Moreover, the other testicular Cxs and corresponding miRNAs remain to be examined in these pathologies to determine their therapeutic potentials. Indeed, controlling the expression of key Cxs, which are able to regulate other Cx isoforms and/or proteins of the BTB, could be a powerful therapeutic approach. Gene transfer into the testis also requires careful consideration, particularly for biosafety and ethical concerns, since the transgene does not pass to the germ line if used for gene therapy in the testis [137]. This strategy must have clinical potential for treatment of specific types of infertility and testicular cancer associated with a deficiency of Cx43 in the somatic Sertoli cells and must use technologies that specifically target the impaired somatic cells. To date, adenoviral vectors appear to be the most appropriate biotechnology tools for future clinical applications, since there is consequent information from clinical trials concerning other diseases [138]. In conclusion, therapeutic opportunities offered by gap junctional intercellular communication mediated by Cx43 in the testis are still widely open. New findings regarding the regulation of both Cx43 and Cx43 partners provide original tools to support this approach in hypofertile patients with impaired spermatogenesis and/or testicular cancer.

Abbreviations

AFP

Alpha-fetoprotein

AMH

Anti-Müllerian hormone

BTB

Blood–testis barrier

cAMP

Cyclic adenosine monophosphate

cGMP

Cyclic guanosine monophosphate

CIS

Carcinoma in situ

Cx

Connexin

G-CSF

Granulocyte colony-stimulating factor

GJA1

Gap junction protein alpha 1

GJIC

Gap junction intercellular communication

hCG

Human chorionic gonadotropin

KO

Knock-out

LDH

Lactate dehydrogenase

miRNA

MicroRNA

ODDD

Oculodentodigital

PGC

Primordial germ cell

PLAP

Placental alkaline phosphatase

SCO

Sertoli-cell-only