Minireview: Regulation of Gap Junction Dynamics by Nuclear Hormone Receptors and Their Ligands

Abstract

Gap junctions are plasma membrane channels comprising connexin proteins that mediate intercellular permeability and communication. The presence, composition, and function of gap junctions can be regulated by diverse sets of physiological signals. Evidence from many hormone-responsive tissues has shown that connexin expression, modification, stability, and localization can be targeted by nuclear hormone receptors and their ligands through both transcriptional and nontranscriptional mechanisms. The focus of this review is to discuss molecular, cellular, and physiological studies that directly link receptor- and ligand-triggered signaling pathways to the regulation of gap junction dynamics.

Cell-cell interactions mediated by intercellular junctions are indispensable for a wide range of complex cellular and physiological processes, and selective disruptions in intercellular communication and contact can trigger tissue dysfunction and the onset of a variety of physiological disorders. The four major intercellular junctions can be functionally categorized into communicating junctions (gap junctions), anchoring junctions (adherens junctions and desmosomes), and occluding junctions (tight junctions). In recent years, the identity, organization, and function of the transmembrane and intracellular membrane-associated proteins that form junctional complexes have been intensely examined (1, 2). All of the intercellular junctions have highly dynamic structures that can be coordinately regulated in response to diverse sets of extracellular, intracellular, and metabolic signals (3, 4). Steroids and other small molecule ligands that act through nuclear hormone receptors (5, 6, 7) have emerged as an important class of regulators of intercellular junctional complexes. Cellular exposure to these receptor ligands can efficiently coordinate the function and accessibility of junctional components and control the assembly, disassembly, and maintenance of intercellular junctions. Nuclear hormone receptors have been shown to regulate the dynamics of cell-cell interactions through transcriptional signaling and through nontranscriptional membrane effects. Depending on the physiological context, receptor-dependent pathways can target the expression, modification, stability, function, and/or localization of specific structural and/or regulatory components of junctional complexes.

Control of Intercellular Communication through Gap Junctions

Gap junctions allow for direct communication and functional coordination between cells through electrical and metabolic coupling of adjacent cells. These junctions allow passage of a variety of inorganic ions and small water-soluble molecules between cells, including signaling molecules such as cAMP and inostiol-3-phosphate, amino acids, nucleotides, vitamins, and sugars. The intercellular movement of these types of molecules can coordinate either an enhanced or attenuated receptor-dependent signaling in neighboring cells within tissues. Gap junctions comprise connexin protein complexes that form channels called connexons. Six connexin proteins assemble to form a connexon that holds adjacent cells together at a distance of 2–4 nM apart. Each connexin protein has four transmembrane domains that form the pore, with both the amino and carboxy termini facing into the cytoplasmic environment, and two extracellular loops that tether connexin proteins together and play a role in cell-cell recognition. There are 21 connexin gene products (denoted as Cx with the molecular mass of the corresponding connexin protein) that have been identified in different tissues (8). The expressed connexins include isoforms of several of the connexin genes, which are categorized into α, β, and γ forms based on gene homology, structure, and sequence motifs. The composition of the connexin proteins within each connexon dictates the permeability of each gap junction channel (8), and the aberrant expression of connexin proteins or expression of mutated connexin gene products is associated with many pathologies including cancer (9).

Gap junctions can dynamically fluctuate between open and closed states in response to various cellular or extracellular stimuli. The connexins themselves, as well as the associated regulatory components, have been shown to be direct or indirect targets of nuclear hormone receptor signaling. The studies analyzing the nuclear hormone receptor control of connexin gene expression and/or control of connexin gene promoter activity were mostly accomplished using the mouse, rat, or human genes (8, 10). Consistent with these studies, the promoters of each connexin gene contain several intriguing consensus binding elements for members of the nuclear hormone receptor gene family or for associated transcription factors that can functionally interact with these receptors. As pointed out in the following discussion, these promoter sites can potentially account for the receptor-dependent regulation of connexin gene expression in many of the tested systems. Also, in a few systems, evidence suggests that steroids can act through nontranscriptional membrane pathways to regulate gap junctions. Table 1 summarizes the key information on nuclear hormone ligand regulation of connexins that will be discussed in detail in the remainder of this review.

Table 1.

Summary of connexin regulation by nuclear hormone receptor ligands

Connexin	Nuclear hormone receptor ligand	Effects on connexin gene products	Cell/tissue type
Cx26	Estrogens progesterone	↑ (mRNA and protein)	Reproductive tissue
		↓ (Protein)
	Progesterone	↑ (Protein)	Ectopic endometrium
	Glucocorticoids	↑ (mRNA and protein)	Hepatocytes
	Retinoic acid	↑ (mRNA and protein)	Epidermis
	Carotenoids	↑ (Protein)	Embryonic fibroblasts
Cx32	Estrogens progesterone	↑ (Protein)	Reproductive tissue
		↓ (Protein)
	Estrogens	↑ (mRNA)	Brain
	Androgens	Stabilized localization at PM	Prostate
	Glucocorticoids	↑ (mRNA and protein)	Hepatocytes
Cx36	Estrogens progesterone	↑ (mRNA)	Brain
		↓ (mRNA)
	Glucocorticoids	↑ (mRNA)	Pancreas
Cx37	Progesterone	↓ (mRNA)	Reproductive tissue
Cx40	T3	↑ (mRNA)	Cardiac atria
	Absence of RARα	↓ (Protein)	Spermatocytes
Cx43	Estrogens progesterone	↑ (mRNA and protein)	Reproductive tissue (contains ERE)
		↓ (mRNA and protein)
	Estrogens progesterone	↑ Localization at PM	Reproductive tissue
		↓ Localization at PM
	Estrogens	↑ (mRNA)	Brain, cardiac tissue
	Androgens	↑ (Protein)	Testis
	Glucocorticoids	↓ (mRNA)	Pancreas
	Glucocorticoids	Phosphorylation via MAPK	Tissue-specific effects
	Low levels of aldosterone	↑ (mRNA and protein)	Ventricular myocytes
	T3	↑ (Protein)	Cardiomyocytes
	T3	↑ (mRNA and protein)	Sertoli cells
	Vitamin D	Induce differentiation	Bone marrow
	RAR	Stimulate promoter activity	Embryonal carcinoma cells (contain RARE)
	Retinoic acid	↑ (mRNA and protein)	Epidermis, embryonic carcinoma cells, canine lens epithelial cells
	Absence of RARβ	↓ (mRNA)	Testis
	Carotenoids	↑ (mRNA)	Fibroblasts
	Carotenoids	↑ (mRNA and protein)	Mammary epithelial cells

ERE, Estrogen response element; PM, plasma membrane; RARE, retinoic acid response element.

Estrogens and Progesterone

Numerous cellular and physiological studies have implicated both estrogen receptor (ER) subtypes (ERα and ERβ) and the progesterone receptor (PR) as critical regulators of gap junctional intercellular communication in reproductive tissue, as well as in a variety of other tissues including the heart, brain, and liver. The biological significance of estrogen and progesterone regulation of gap junctions is linked to the profound roles of these steroid hormones on mammary, ovarian, and endometrial tissue proliferation, development, and differentiation. Estrogens and progestins can have complementary or opposing actions on the dynamics of gap junction permeability, as well as on the synthesis and trafficking of connexin gene products that depend upon on the physiological context and tissue phenotype.

Opposing actions of estrogens and progesterone on myometrial gap junctions

Reproductive tissues are highly steroid responsive in which estrogen is generally associated with an enhancement of gap junctions, and progesterone is associated with the attenuation of gap junction-mediated communication. Characterization of myometrial reproductive tissue in rodents has provided useful insights into the importance of the Cx43 connexin gene as a target of the opposing actions of estrogens and progesterone (11). Physiological studies show that myometrial Cx43 gene transcript levels are low in nonpregnant rats, and then rise after 10 d of gestation, reaching a maximum during labor in a process that is associated with the formation of functional gap junctions (12). In both nonpregnant and pregnant contexts, generally estrogen stimulates and progesterone inhibits Cx43 expression. The steady state increase in myometrial Cx43 transcript levels is strongly positively correlated with a higher ratio of plasma estrogen to progesterone (12, 13). Estradiol treatment of nonpregnant immature rats, or ovariectomized mature rats, results in an increase in Cx43 mRNA levels and induction of gap junctions whereas coinjection of progesterone with estradiol prevents the induction of Cx43 transcripts (13). When pregnant rats are given estrogen, an induction of Cx43 expression and gap junctions is followed by ablation of the fetus, indicating that the stimulated metabolic and electrical coupling of uterine muscle cells that increases during labor is estrogen responsive and concomitant with development of gap junctions (13).

In rodents, exposure to exogenous progesterone in vivo abolished the increase in Cx43 transcripts in myometrial tissue normally observed in late pregnancy under physiological conditions in which estradiol levels are high. Conversely, treatment of rats with a progesterone receptor antagonist on d 15 of gestation, when Cx43 mRNA levels are relatively low, results in a relatively rapid rise in Cx43 transcript levels that is maximal after 23–38 h, which is then followed by a preterm delivery several days later (14). Furthermore, treatment of rats with the PR antagonist RU486 induces an increase in myometrial gap junctions that is associated with muscle contractility during labor (15). The level of Cx43 transcripts and protein, as well as Cx43-containing gap junctions, is stimulated with the onset of labor in myometrial tissue and the development of synchronized uterine contractions, which follows a significant increase in the estrogen to progesterone ratio (16). In a complementary study, deletion of the Cx43 gene was shown to impair the estrogen-dependent development of new blood vessel within the stromal compartment, which results in the arrest of embryo growth, termination of pregnancy, and aberrant differentiation of uterine stromal cells (17).

Targeting of the Cx43 connexin by estrogens and progesterone through receptor-dependent responses

A complete understanding of the molecular mechanism by which ER and PR regulate gap junctions in reproductive tissue remains elusive. However, evidence has emerged that these steroid receptors can regulate expression of the connexin genes through receptor-dependent responses in processes that can be highly influenced by other cellular factors. During the preimplantation phase of early pregnancy, estrogens induce expression of Cx26 in the uterine epithelium and Cx43 in the uterine stroma by ER-mediated responses, whereas during implantation and decidualization, expression of both connexins is maintained by ER-independent embryonic signals (18). It is worth noting that, in contrast to myometrial cells, during early pregnancy, progesterone does not suppress estrogen-induced Cx43 gene expression in differentiating uterine stromal cells (17, 18).

An analysis of ERα-knockout and ERβ-knockout mice showed that, in the endometrium, estrogen-induced Cx26 gene expression is dependent on functional ERα (18), although this response requires newly synthesized transcription factors (19). The Cx43 gene promoter contains several half-palindromic estrogen response elements and several activator protein 1 (AP-1) DNA sites, and when linked to reporter plasmids, estrogens stimulate Cx43 gene promoter activity in transfected HeLa cells (20) and in rodent myometrial cells (21). Before the stimulation of Cx43 connexin gene expression, estrogen treatment of rats strongly induces c-Fos and c-Jun levels in vivo (21, 22). Evidence also suggests that protein kinase C signaling plays a role in the up-regulation of Cx43 gene expression through one of the AP-1 elements in the Cx43 promoter (23). These observations suggest that the ER control of Cx43 transcription in the myometrial tissue could result from a combination of direct receptor interactions with the Cx43 promoter and complementary parallel responses that involve the protein kinase C activation of c-Jun and the stimulated expression of components of the AP-1 transcription complex.

In addition to the potential for direct interactions of nuclear hormone receptors with connexin gene promoters, several studies have shown that steroid-regulated cofactors for nuclear hormone receptors play a role in the estradiol or progesterone control of connexin gene expression. Ini, a nuclear zinc-finger protein identified by screening an estrogen-induced rat myometrial expression library, has been shown to bind to the proximal Cx43 promoter and enhance the ERα dependent expression of Cx43 promoter activity. Ini was shown to selectively stimulate the transcriptional activator function-1 (AF-1) region of ERα independent of the AF-2 region of the receptor (24). Estrogen strongly up-regulates Ini transcripts in the myometrium, suggesting that this AF-1-specific coactivator of ERα plays a key role in enhancing the estrogen-stimulated expression of Cx43 in a physiologically appropriate context, such as during labor. PR represses gene expression associated with the increased contractile activity of the myometrium involved in the initiation of labor. As part of this process, PR represses Cx43 gene transcription through a mechanism that is dependent on the presence of an AP-1 site in the Cx43 promoter (25) and on the recruitment of the p54nrb transcriptional corepressor of PR to the Cx gene promoter. Mutation of the AP-1 site abolished PR-mediated repression of Cx43 expression and decreased the recruitment of both PR and p54nrb to the Cx43 promoter, whereas knockdown of p54nrb prevented the PR repression of Cx43 repression. It was proposed that the decreased expression of p54nrb at the onset of labor derepresses the PR-mediated inhibition of Cx43 expression and thereby contributes to the initiation of labor (25).

In addition to the steroid-dependent effects on expression of Cx43 gene, one study has shown that estrogen stimulates and progesterone attenuates trafficking of Cx43 protein from the Golgi and assembly into functional gap junctions at the plasma membrane in a process that is required for effective myometrial contractions during labor (26). Conceivably, any one of many protein trafficking and membrane localization factors or their upstream regulators that control Cx43 compartmentalization could be under steroid control. Further studies will hopefully elucidate the mechanism of this intriguing steroid response.

Estrogen regulation of Cx43 expression in myometrial tissue can display opposite effects that depend on the presence of the two ER subtypes, ERα and ERβ. Generally, ERα activity is associated with the stimulated expression of Cx43 gene products and functional gap junctions whereas the presence of ERβ is correlated with the loss of gap junctions and Cx43 expression. In immature rats, which express high levels of ERα in myometrial tissue, estrogen strongly induces gap junctions in a process that requires de novo protein synthesis and can be inhibited by the tamoxifen, a selective ER modulator that antagonizes ERα in this tissue (27). ERβ expression is increased in human term myometrial tissue, and during gestation this ER subtype appears to inhibit Cx43 expression. This effect is due, in part, to the interaction of ERβ with the AP-1 transcription complex in a manner that blocks transcriptional activity. During labor, in which the presence of functional gap junctions helps synchronize the myometrium, estrogen induces Cx43 expression during a time frame in which ERβ levels are almost undetectable and ERα levels are relatively high (28).

Estrogen and progesterone control of Cx26, Cx32, and Cx37 connexin expression and function

Although the receptor-mediated mechanisms have not been characterized, estrogen and progesterone have been shown to regulate expression of the Cx26, Cx32, and Cx37 connexins individually or in coordination with Cx43 in reproductive tissues. For example, the progesterone control of the reproductive cycle is associated with dynamic changes in gap junctions and expression of gap junction proteins in the endometrial glandular epithelium. Cx26 and Cx32 are weakly expressed in the proliferative phase, increase during ovulation and the secretory phase, and then decrease during the late-secretory phase (29). In another study, Cx37 transcript and protein expression in granulosa cells were dependent on the follicular stage of luteal development of sheep ovaries, with an inverse correlation observed between Cx37 gene product expression and progesterone levels during the estrous cycle (30).

The steroid-regulated expression of connexin genes show distinct patterns in endometriotic tissue as compared with the normal uterine endometrium. In ectopic endometrium of human patients, in which tissue that normally lines the uterus is found outside of the uterus, expression of Cx43 is correlated with high serum levels of 17β-estradiol, and expression of Cx26 is correlated with high progesterone concentrations after progestin treatment (31). In endometrial cancers, expression of PR correlated negatively with the expression Cx26, which in turn is correlated positively with ERα expression. The Cx26 and Cx43 connexins are detected in the cytoplasm in approximately 70% of the studied endometrial cancers rather than as part of the gap junction complex, indicating that loss of functional gap junctions is due to combinations of altered expression and aberrant localization of the connexins (32). Tissue microarrays from invasive breast carcinomas revealed that expression of ERα may be linked to the expression of Cx43 and Cx32 whereas expression of PR correlated positively with Cx43 levels. However, expression of these connexins was not linked to patient outcomes and is not considered indicators of breast cancer prognosis (33). It is tempting to consider that differences in expression and function of tissue-specific transcriptional regulators associated with the connexin promoters as well as steroid responsive gene products that can modulate connexin expression, stability, and intracellular transport may play a critical role in the cellular control of gap junctions.

Estrogen and progesterone regulation of connexins in the brain

Estrogens and progesterone regulate gap junctions within the brain in a region-specific manner that may be important in regulating reproductive function through GnRH. In the rat preoptic area, a region containing GnRH cell bodies, treatment with estrogen, progesterone, or a combination of the two steroids increased Cx43 protein levels in female rats (34). In contrast, treatment with estrogen and progesterone decreased Cx43 levels in male rats. In both female and male rats, combinations of estrogen and progesterone down-regulated Cx43 levels in the hypothalamus, a region that contains GnRH nerve endings (34). These results indicate that steroid hormones differentially regulate Cx43 expression in specific regions of the brain in a sexually dimorphic manner and that gap junctional communication may contribute to the control of the estrous cycle and sexual behavior in female rodents. The preoptic area of the brain is important in sex-specific induction of GnRH surge and may explain the differential regulation of Cx43 expression in male vs. female rats. In the suprachiasmatic nucleus, estrogen increased expression of Cx36 mRNA, and this increase is inhibited by progesterone. In contrast, neither estrogen nor progesterone has an effect on the levels of Cx36 mRNA in the cerebral cortex (35). Estrogen has also been shown to increase the mRNA levels of Cx32 in the suprachiasmatic nucleus and of Cx43 in the cerebral cortex of female rats (36). Consistent with the enhanced connexin expression observed in the brain, estrogen has a beneficial effect on arrhythmias by attenuating the reduction in gap junction proteins, such as Cx43, that is observed after infraction (37). Mechanistic explanations for these intriguing observations have not been delineated.

Androgens

Depending on the system, androgen signaling can be associated with either the stimulation or disruption of gap junctions through either transcriptional or nontranscriptional pathways. Testosterone and dihydrotestosterone inhibit gap junction-mediated intercellular communication in human bladder carcinoma cell lines (38). In contrast, an in vivo experiment using male rats showed that treatment with testosterone and LH releasing hormone accelerated the appearance of gap junctions during development (39). The level of Cx43 mRNA and protein expression can be differentially regulated in porcine ovaries, which depends on the timing of androgen or antiandrogen treatment and on phase of follicular development (40). For example, treatment with flutamide, a potent antiandrogen, during either the neonatal or prenatal developmental stage, down-regulated Cx43 gene expression in preantral follicles. In contrast, treatment with this antiandrogen stimulated Cx43 expression in antral follicles during neonatal steroid treatment but had no effect with prenatal treatment (40). Antiandrogens have also been shown to reduce Cx43 expression with a concurrent disruption of spermatogenesis in porcine testis (41). Consistent with this observation, a decrease in Cx43 expression correlated with the disruption of androgen responsiveness due to a decrease in androgen receptor expression in boar epididymis (42).

Other studies suggest that androgens can regulate gap junctions through membrane effects. In cultured rat Sertoli cells, treatment with natural and synthetic androgens rapidly induced an increase in intracellular calcium levels within 20–30 sec of steroid treatment that was coupled to the stimulation of gap junctions in the absence of de novo protein synthesis (43). When Sertoli and cardiac cells are pretreated with the antiandrogen cyproterone acetate, which blocks nuclear androgen receptor activity, there is no effect on the ability of androgens to stimulate gap junctions (44). It was therefore suggested that androgens can trigger this response either through some sort of membrane receptor-like molecule or perhaps by interacting with the membrane in a way that alters the conformation of gap junction channels that alter their functional state.

In human prostate cancer cells, androgen treatment results in the stabilized localization of Cx32 protein at the plasma membrane, sequestering this connexin away from the early secretory compartment where the protein can be degraded (45). The plasticity of this process is likely due in part to the relatively short half-lives of many of the connexin proteins. An important future direction will be to identify and functionally characterize androgen-regulated target genes and signaling cascades that direct the localization and stabilization of connexin proteins.

Glucocorticoids and Mineralocorticoids

Both transcriptional and membrane actions of glucocorticoid hormones have been shown to control gap junctions in different cell and tissue systems. In rat hepatocytes, the synthetic glucocorticoid dexamethasone induces transcripts for both the Cx26 and Cx32 genes under conditions in which this steroid stimulates gap junctional communication (46). Treatment with dexamethasone, but not with other steroid hormones such as estradiol, progesterone, testosterone, or aldosterone, enhanced Cx32 and Cx26 expression in rat hepatoma cells in a dose-dependent manner that is consistent with a glucocorticoid receptor (GR)-dependent response (47). Tissue-specific differences in the glucocorticoid control of connexin gene expression have been observed in several different systems. For example, in dexamethasone-induced insulin-resistant rats, Cx36 transcript expression was augmented whereas expression of Cx43 gene products was diminished in pancreatic islet cells (48). Relatively little is known about mineralocorticoid control of gap junction protein expression. One study showed that 24-h treatment of cultured rat ventricular myocytes with 10 nm aldosterone induced expression of Cx43 transcripts and protein under conditions in which the conduction velocity was increased (49). Higher concentrations (1 μm) of aldosterone inhibited Cx43 protein levels, and pretreatment with the mineralocorticoid receptor (MR) antagonist eplerenone prevented the mineralocorticoid effects on Cx43 expression. The transcriptional mechanism by which the GR and MR potentially regulates connexin gene expression remains relatively uncharacterized. Based on the presence of consensus sequences for transcription factor-binding sites within the human Cx26, Cx32, and Cx43 gene promoters, it is tempting to consider that in some cell types, connexin genes may be targets of both GR and MR through transcription factors such as CCAAT enhancer binding protein or Sp1 that directly interact with these steroid receptors.

One recent study has shown evidence for transcription-independent plasma membrane actions of GR that regulated gap junction communication and proliferation of neural progenitor cells (50). A nongenomic pathway was proposed in which lipid raft-associated GR triggers the MAPK-dependent phosphorylation of Cx43 protein in a process that requires caveolin-1 and the c-Src tyrosine kinase. Cx43 protein phosphorylation at specific sites can enhance or attenuate gap junction communication in a tissue-specific manner (51). Thus, combinations of steroid receptor-mediated transcription-dependent and transcription-independent cascades are likely to account for the effects of glucocorticoids on gap junction levels, function, and connexin composition.

Thyroid Hormones

Several studies provided evidence that the Cx43 connexin gene is a biologically important target of thyroid hormones that is associated with the control of gap junctional communication involved in myocardial electrical and metabolic synchronizations within heart muscle. For example, treatment of cultured neonatal cardiomyocytes with T₃ up-regulated expression of Cx43 protein and accelerated gap junction formation (52). Expression of a dominant-negative form of the thyroid hormone receptor in mouse cardiomyocytes functionally established that thyroid hormones up-regulate expression of Cx40 transcripts in the cardiac atria through a receptor-dependent process (53). The effects of thyroid hormone on Cx43 expression are not limited to heart muscle cells. One study documented that thyroid hormones regulate Sertoli cell proliferation and differentiation by increasing the levels of Cx43 gene products in a time- and dose-dependent manner (54), which mediates maturation of Cx43-containing gap junctions in rat Sertoli cells (55). Furthermore, T₃ regulates expression of Cx43 in brook trout testis, and this response was proposed to be is carried out by T₃ targeting different sites on the Cx43 promoter (56). Although the promoter of the Cx43 connection gene has consensus sites for transcription factors that have been shown to interact with thyroid hormone receptors (57), no studies to date have directly established the thyroid hormone receptor-mediated mechanism by which T₃ regulates expression of this connexin.

Vitamin D

Vitamin D₃ plays important roles in the regulation of cell proliferation and cell-cell interactions through gap junctions. In patients with renal cell carcinoma, serum levels of vitamin D are lower than those in healthy individuals, and the presence of vitamin D is correlated with enhanced gap junction intercellular communication (58). In murine fibroblast cells, vitamin D₃ induces cell-cell communication via gap junction at concentration of 0.01 μm and 1.0 μm, whereas higher concentrations of vitamin D₃ have a suppressive effect on gap junctions that can be reversed with retinoic acid (59). The underlying mechanism for this intriguing observation has not been defined. Interesting findings have also emerged relating the vitamin D₃ control of gap junctions and bone differentiation. For example, in mouse bone marrow cultures, Cx43 expression is detected in the marrow stromal cells and in mature osteoclasts, and when these cells are treated with carbenoxolone, a gap junction inhibitor, prostaglandin E2, and vitamin D₃ are inhibited from stimulating osteoclast differentiation (60).

Retinoic Acid and Carotenoids

Vitamin A can be found in two forms: retinols and carotenes. In several different cell systems, Cx43 gene expression is regulated by retinoic acid, a vitamin A metabolite, and/or certain carotenoids, which are naturally occurring tetraterpenoid organic pigments found in plants and certain animals that include vitamin A precursors (61). Given the extensive information on nuclear hormone receptors associated with cellular responsiveness to retinoids and carotenoids (62), only a few studies have characterized the direct involvement of receptors for specific retinoid and carotenoid compounds in controlling connexin gene transcription (10). The retinoic acid regulation of Cx43 expression was observed in retinoic acid receptor (RAR) expressing F9 mouse embryonal carcinoma cells in which a retinoic acid response region of the Cx43 promoter was found near the transcriptional start site. DNA binding and mutagenic assays showed that an Sp1/Sp3 GC box was critical for the retinoic acid response. No retinoic acid response elements were found, suggesting that the RAR could potentially interact with Sp1 and/or Sp3 to stimulate Cx43 promoter activity (63). Consistent with these mechanistic studies, topical retinoic acid treatment of all-trans-retinoic acid to human epidermis selectively increased the transcript and protein levels for both the Cx26 and Cx43 connexins in vivo, with no effects on expression of Cx32, Cx37, and Cx40 (64).

Treatment of F9 mouse embryonal carcinoma cells with the Ro 41–5253 RAR antagonist suppressed the retinoid induction of Cx43 promoter-luciferase reporter plasmids, but did not suppress Cx43 promoter activity by the non-provitamin A carotenoids astaxanthin or lycopene (63). All-trans-retinoic acid, which acts through the RAR, increased Cx43 transcripts and protein and enhanced gap junction intercellular communication in primary canine lens epithelial cells in manner consistent with a receptor-dependent process (65). In RARα-deficient mice, a markedly reduced expression of Cx40 was observed in aberrant pachytene spermatocytes. Bioinformatics analysis implicated this connexin as a direct RARα responsive gene target because of the presence of a potential retinoic acid response element in its promoter (66). Retinoid X receptor-β is highly expressed in the testis, and retinoid X receptor-β-deficient mice display decreased levels of testis Cx43 transcripts along with abnormal spermatogenesis due to altered Sertoli cell function (67). In contrast to these studies, transgenic mice that express a constitutively active form of RAR have dilated cardiomyopathy along with down-regulated expression of cardiac muscle Cx43, N-cadherin, and β-catenin (68). These results implicate tissue-specific roles for RAR via its interactions with other coactivators, corepressors, and/or other transcription factors that control Cx43 gene transcription.

Carotenoids have been to shown to induce gap junction communication and regulate expression of Cx43 transcripts in murine fibroblasts (69), human fetal skin fibroblasts, and human breast cancer cell lines (70). For example, treatment of fibroblasts with 0.1 μm lycopene, a carotenoid found in tomatoes, and a slightly higher amount of acyl-retinoic acid (1.0 μm) stimulated gap junctions (70). It was proposed that this response was due to the interaction between lycopene and the RAR that subsequently leads to the synthesis of Cx43 mRNA. In several types of tumorigenic human breast cancer cells and in nontumorigenic human mammary epithelial cells, 48 h exposure to 10 μm lycopene coordinately enhanced expression of protein and transcript levels for RARα and Cx43 (71), suggesting the potential importance of RARα-mediated transcriptional control of this connexin gene. The effects of carotenoids are not limited to controlling Cx43 expression. A transgenic mouse line carrying the crtB gene encoding phytoene synthase produces an endogenous phytoene, which is a carotenoid, and displays a significant increase in Cx26 levels in embryonic fibroblasts compared with cells from wild-type animals (72).

Perspectives

An extensive set of cellular and in vivo studies have demonstrated that steroids and other ligands of nuclear hormone receptors can acutely regulate gap junction dynamics and intercellular communication in a physiologically appropriate manner. The cellular pathways under direct receptor control can be quite diverse. As summarized in Table 1, the expression, stability, modification, localization, and function of the connexin genes can be targets of transcriptional signaling of nuclear hormone receptors. Furthermore, in a few systems some evidence has emerged for the regulation of gap junctions by membrane actions of ligands for this family of receptors. Conceivably, such exquisite regulatory cascades facilitate critical transitions in the composition and function of gap junctions and intercellular communication. Individual components and downstream targets within the receptor-activated cellular pathways represent intriguing pharmaceutical targets for the control of pathologies associated with alterations in gap junction-mediated intercellular communication. To date, only relatively limited molecular information has emerged that can account for the nuclear hormone receptor control of gap junction dynamics within the context of a large body of cellular and physiological studies. An exciting future direction will be to delineate the precise mechanistic links by which nuclear hormone receptor-mediated transcriptional signaling and certain ligand-activated nontranscriptional effects can target connexin promoter activity, connection protein stability and localization, and the formation of functional gap junctions in a tissue-specific manner.

Search Strategies

For this review, PubMed was searched using either gap junction or connexin gene expression in combination with each of the following terms: steroids, steroid receptors, nuclear receptors, thyroid hormone, retinoic acid, vitamin D, androgen, estrogen, progesterone, glucocorticoid, intercellular communication, cell-cell adhesion, cell-cell communication, cell-cell interaction, or cell signaling pathways.

NURSA Molecule Pages^†:

• Nuclear Receptors: ER-α | ER-β | PR | GR | MR | RAR-α | RXR-β;

• Ligands: 17β-estradiol | Progesterone | Dihydrotestosterone | Testosterone | Dexamethasone | Aldosterone | Thyroid hormone | 1α25-dihydroxyvitamin D₃.