Minireview: Steroid/Nuclear Receptor-Regulated Dynamics of Occluding and Anchoring Junctions

Abstract

A diverse set of physiological signals control intercellular interactions by regulating the structure and function of occluding junctions (tight junctions) and anchoring junctions (adherens junctions and desmosomes). These plasma membrane junctions are comprised of multiprotein complexes of transmembrane and cytoplasmic peripheral plasma membrane proteins. Evidence from many hormone-responsive tissues has shown that expression, modification, molecular interactions, stability, and localization of junctional complex-associated proteins can be targeted by nuclear hormone receptors and their ligands through transcriptional and nontranscriptional mechanisms. The focus of this minireview is to discuss molecular, cellular, and physiological studies that directly link nuclear receptor- and ligand-triggered signaling pathways to the regulation of occluding and anchoring junction dynamics.

Cell-cell interactions are indispensable for a wide range of complex cellular and physiological processes, and selective disruptions in intercellular communication can trigger tissue dysfunction and the onset of a variety of physiological disorders. Intercellular junctions control the nature and efficacy of cell-cell interactions. In recent years, the structure, function, and composition of junctional complexes have been intensely examined in many mammalian cell systems (1). The four major intercellular junctions are comprised of multiprotein complexes at the plasma membrane that include both intracellular peripheral membrane proteins and intercellular transmembrane protein components. The intracellular junctions can be functionally categorized into occluding junctions (tight junctions), anchoring junctions (adherens junctions and desmosomes), and communicating junctions (gap junctions) (2). Each of these junctions has 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 hormone ligands, which act through nuclear receptors (5–7), have emerged as an important class of in vivo regulators that can efficiently control the assembly, disassembly, function, and maintenance of intercellular junctions. Depending on the physiological and cellular contexts, the dynamics of cell-cell interactions can be controlled by receptor-dependent transcriptional signaling and in some systems through ligand-induced nontranscriptional membrane effects. This minireview will discuss the evidence supporting how nuclear receptors and their ligands target the expression, modification, stability, function, and/or localization of specific structural and/or accessory components of tight junctions and adherens junctions.

Control of intercellular communication through anchoring and occluding junctions

Cell-cell interactions mediated by occluding and anchoring junctions play critical roles in many physiologically important processes such as cell polarity, intercellular permeability, cellular migration, and attachment (8, 9). With the exception of some tissue types, the tight junction, or zonula occludens, is the most apical structure of the junctional complex and forms a continuous seal around the lateral circumference of adjacent cells to regulate the selective diffusion of solutes on the basis of size and charge through a paracellular pathway (10, 11). In addition to this gate function, the tight junction serves as a cellular fence by restricting the lateral diffusion of lipids and membrane proteins between the apical and basolateral regions. Thus, tight junctions assist in the maintenance of cell surface polarity by physically defining the border between the compositionally and functionally distinct apical and basolateral membrane domains (12, 13). The adherens junction (or zonula adherens or belt desmosome) is situated immediately basal to the tight junction and is responsible for intercellular adhesion between neighboring cells (14). Both intercellular junctions have been proposed to associate with the perijunctional actin cytoskeleton through multiprotein complexes (13, 14), suggesting that the cytoskeleton, tight junctions, and adherens junctions form an integrated functional unit that regulates the architecture of epithelial cells. The nature, specific combination, and organization of membrane components in occluding and anchoring junctions control the functional dynamics of each junction, such as the strength of intercellular adhesion and restriction of intercellular permeability as well as mediating the intracellular signaling from the apical junctional complex.

Tight junctions

Depending on the tissue origin and cell phenotype, transmembrane proteins detected within tight junctions can be occludins, claudins, or junctional adhesion molecules (JAMs). In highly organized tight junctions, the transmembrane protein components from adjacent cells form a branching network of sealing strands in which the extracellular domains directly contact each other. In different cell and tissue types, occludins or members of the claudin gene family can regulate paracellular permeability determined by intercellular flux of small molecule tracers. Tight junction protein complexes include several types of intracellular peripheral membrane proteins that play critical roles in tight junction dynamics, cell signaling processes, and anchor the sealing strands to the actin of the cytoskeleton. The zonula occluden (ZO) proteins, most notably ZO-1 and its related family members ZO-2 and ZO-3, are tethered to the cytoplasmic tails of occludins or claudins and serve as scaffolding proteins for other tight junctional proteins and the actin cytoskeleton (15–17). Other junctional complex proteins and cell signaling molecules such as an atypical protein kinase C (PKC), a G protein subunit, and the Ras target afadin/AF-6, which binds to ZO-1, are recruited to the tight junction protein complex (3, 4). Table 1 summarizes the information on nuclear hormones ligand regulation of tight junctional protein components that will be discussed in detail.

Table 1.

Summary of Occluding Junction Proteins Regulated by Nuclear Hormone Receptor Ligands

Occluding Junction Protein	Nuclear Hormone Receptor Ligand	Effects on Occluding Junction Gene Products	Cell/Tissue Type	References
Occludin	Estrogen	↑ (mRNA)	Brain	(22)
	Estrogen	↑ (mRNA and protein)	Intestinal epithelium	(23)
	Estrogen	↓ (mRNA and protein)	Endothelial cells	(24)
	Progesterone	↑ (protein)	Reproductive tissue	(38)
	Androgen	↑ (protein, junctional localization)	Sertoli cells	(45)
	Glucocorticoids	↑ (junctional localization)	Endothelial cells	(57)
	Glucocorticoids	↑ (protein, GR promoter binding)	Brain microvascular cells	(60)
	Glucocorticoids	↑ (mRNA and protein)	Retinal endothelial cells	(66, 67)
	Glucocorticoids	↑ (protein, junctional localization)	Rodent mammary epithelial cells	(71, 74, 75)
	Vitamin D	↓ (Junctional localization)	Colon cancer cells	(87)
	Retinoic acid	↑ (protein, mRNA)	Murine embryonic carcinoma cells	(92)
	Retinoic acid	↑ (mRNA)	Canine kidney cells	(93)
	Retinoic acid	↑ (junctional localization)	Hepatocytes	(94)
	All-trans retinoic acid	↑ (protein, mRNA)	Myopia retinal pigment epithelium	(101)
Claudin-1	Progesterone	↑ (protein)	Cervix epithelium	(37)
	Androgens	↑ (mRNA, junctional localization)	Sertoli cells	(45)
	Retinoic acid	↑ (mRNA and protein)	Canine kidney cells	(93)
	All-trans retinoic acid	↑ (protein, mRNA)	Myopia retinal pigment epithelium	(101)
Claudin-2	Progesterone	↓ (mRNA)	Cervix epithelium	(37)
	Vitamin D	↑ (protein)	Enterocytes	(85)
	Retinoic acid	↑ (protein)	Keratinocytes, colorectal adenocarcinoma cells	(96, 97)
	PPAR	↓ (mRNA)	Urothelium	(105)
Claudin-3	PPAR	↑ (protein, mRNA)	Urothelium	(105)
Claudin-4	Progesterone	↑ (protein)	Reproductive tissue	(38)
	Androgen	↑ (protein)	Sertoli cells	(49)
	PPAR	↑ (protein, mRNA, junctional localization)	Urothelium	(105)
Claudin-5	Estrogen	↑ (mRNA and protein, ER promoter binding)	Endothelial cells	(25, 26)
	Glucocorticoids	↑ (protein)	Brain	(59)
	Glucocorticoids	↑ (mRNA and protein)	Retina endothelial cells	(62, 66, 67)
	Retinoic acid	↑ (protein, junctional localization)	Urothelium	(105)
Claudin-6	Estrogen	↑ (protein, mRNA)	Breast cancer cells	(28)
	Retinoic acid	↑ (protein, mRNA)	Murine embryonic carcinoma cells	(92)
Claudin-7	Retinoic acid	↑ (mRNA)	Murine embryonic carcinoma cells	(92)
Claudin-8	Androgen	↑ (protein)	Prostate gland	(49)
Claudin-11	Androgen	↑ (protein, mRNA, junctional localization)	Sertoli cells	(45)
Claudin-12	Vitamin D	↑ (protein)	Enterocytes	(85)
JAM-A	Estrogen	↑ (mRNA and protein)	Intestinal epithelium	(23)
ZO-1	Androgen	↑ (protein)	Sertoli cells	(48)
	Glucocorticoids	↑ (junctional localization)	Endothelial cells	(57)
	Glucocorticoids	↑ (mRNA and protein)	Retinas	(62)
	Glucocorticoids	↑ (protein, junctional localization)	Rodent mammary epithelial cells	(71, 72, 77)
	Glucocorticoids	↑ (junctional localization)	Human endometrial cancer cells	(83)
	Vitamin D	↑ (protein, junctional localization)	Colorectal cancer cells	(84)
	Retinoic acid	↑ (protein, mRNA)	Murine embryonic carcinoma cells	(92)
	Retinoic acid	↑ (protein, mRNA)	Canine kidney cells	(93)
	Retinoic acid	↑ (interaction with occludin)	Hepatocytes	(94)

Adherens junctions

In epithelial tissue, this type of anchoring junction can form an intercellular adhesion belt (zonula adherens) or an attachment point to the extracellular matrix (adhesion plaques). The engagement of cell-cell contacts at the adherens junction is mediated by intercellular adhesion between extracellular domains of single-pass transmembrane cadherin molecules expressed in adjacent cells. This process triggers the formation of an intracellular protein complex at the inner surface of the plasma membrane that includes cadherin, α-catenin, β-catenin, plakoglobin, and the actin cytoskeleton. The conserved cytoplasmic tails of the cadherins interact with intracellular catenins, which serves to help link the plasma membrane junctional complex to the actin cytoskeleton (18, 19). Formation of adherens junctions is considered a prerequisite to the formation of tight junctions. Activation of cell-cell adhesion can trigger the recruitment and organization of tight junction components to the points of intercellular contact, and in a complementary manner, disruption of cadherin-based adhesion has been shown to prevent tight junction formation (20, 21). Information on nuclear hormone ligand-mediated regulation of adherens junctions discussed in this minireview is summarized in Table 2.

Table 2.

Summary of Anchoring Junction Proteins Regulated by Nuclear Hormone Receptor Ligands

Anchoring Junction Protein	Nuclear Hormone Receptor Ligand	Effects on Anchoring Junction Gene Products	Cell/Tissue Type	References
E-cadherin	Estrogen	↓ (junctional localization)	MCF-7 breast cancer cells	(27)
	Estrogen	Altered subcellular localization	MCF-7 breast cancer cells	(30)
	Androgen	↑ (protein)	LNCaP prostate cancer cells	(50)
	Androgen	↓ (mRNA), AR promoter interactions	Nonmetastatic breast cancer cells	(55)
	Glucocorticoids	↑ (protein, altered localization)	Brain microvascular cells	(64)
	Vitamin D	↑ (protein, junctional localization)	Colorectal cancer cells, keratinocytes	(84, 90)
	Retinoic acid	↑ (Interaction with β-catenin, junctional localization)	Hepatocytes	(94)
	All-trans retinoic acid	↑ (junctional localization)	Renal carcinoma cells	(95)
	Retinoic acid	↑ (protein)	Breast cancer cells	(100)
	All trans retinoic acid	↑ (protein, mRNA)	Myopia pigment epithelium	(101)
Cadherin-6	Progesterone	↓ (protein)	Endometrial cancer cells	(41)
N-cadherin	Estrogen	↓ (protein, mRNA)	Somatolactotropic GH3 cells	(32)
	Androgen	↑ (protein)	Sertoli cells	(48)
	Vitamin D	↓ (protein)	Osteoblast-like cells	(89)
α-catenin	Estrogen (membrane effect)	↓ (junctional localization)	Endothelial cells	(33)
	Progesterone	↑ (mRNA)	Endometrial cancer cells	(42)
β-catenin	Estrogen	↓ (protein, mRNA)	Somatolactotropic GH3 cells	(32)
	Estrogen	↑ (protein)	Prostate cancer cells	(34)
		↑ (interaction with E-cadherin	Colorectal cancer cells	(36)
	Androgen	↑ (protein)	Sertoli cells	(48)
	Glucocorticoids	↑ (protein, mRNA, junctional localization)	Rat mammary epithelial tumor cells	(74, 81)
	Glucocorticoids	↑ (protein stability, junctional localization)	Rat mammary epithelial tumor cells	(82)
	Thyroid hormone	↑ (interaction with cadherins)	Osteoblast-like cells	(91)
	Retinoic acid	↑ (interaction with E-cadherin	Hepatocytes	(94)
	All trans retinoic acid	↑ (junctional localization)	Renal carcinoma cells	(95)
Vinculin	Estrogen	↓ (junctional localization)	Breast cancer cells	(31)
	Thyroid hormone	↓ (junctional localization)	Osteoblast-like cells	(91)
Plakoglobin	Estrogen	↓ (junctional localization)	Breast cancer cells	(31)
	Glucocorticoids	↑ (junctional localization)	Endothelial cells	(57)

Estrogens

Estrogens have been implicated as critical hormone regulators of tight junctions and adherens junctions in a variety of tissue types, and in many of the examined cell systems the effects of 17β-estradiol are associated with and likely caused by the altered expression of specific tight junction or adherens junction components. In most instances, specific roles for the two estrogen receptor (ER) isotypes, ER-α and ER-β, have not been directly evaluated or distinguished, although the current evidence suggests that ER-dependent transcriptional responses that can significantly differ in a tissue-dependent context. Tight junctions play an important role in maintaining the blood-brain barrier, and in the female brain tissue of ovariectomized mice, estrogen up-regulates transcript expression of the occludin transmembrane tight junction protein (22). Generally consistent with this observation, in the intestinal epithelium, 17β-estradiol reinforces the epithelial barrier by stimulating an ER-β dependent increase in expression of occludin and JAM-A transcripts and protein (23). Estrogens can induce a concentration and time dependent biphasic effect on tight junctions in human endothelial cells (24). At a concentration between the nanomolar and micromolar range, estrogen treatment decreased the level of occludin and increased paracellular permeability. However, at concentrations in the picomolar range, estrogen stimulated an increase in the expression of occludin with a concomitant decrease in paracellular permeability.

A biphasic down-regulation of transepithelial electrical resistance (TER), which is a measure of tight junction functionality, was also observed over a 4-day treatment with estrogen. Although the protein and the mRNA levels of occludin are lower after exposure to estrogen, the levels of claudin-1, claudin-5, or ZO-1 tight junction protein gene products remain unaltered (24). In contrast, in murine endothelial cells derived from the brain and the heart, estrogens induce expression of claudin-5 protein and transcripts and enhance vascular structural integrity (25). A recent follow-up study using promoter constructs, mutagenesis, and chromatin immunoprecipitation provided the first in-depth evidence that ER-α and ER-β induce murine claudin-5 gene transcription by direct binding to both a nonconsensus estrogen response element and a specificity protein-1 site in the claudin-5 promoter (26). This response is driven by cooperative interactions between estrogen receptors and the specificity protein-1 transcription factor (26). Taken together, current evidence shows that the estrogen regulation of tight junction integrity can vary significantly in a tissue- and cell-dependent manner, and an important future direction will be to understand the precise role of nuclear ERs in most of these systems.

Even with cell systems derived from the same tissue type, such human mammary tissue and breast cancer cells, estrogens can display different regulatory effects on junctional complexes. In MCF-7 cells, a highly estrogen-responsive human breast cancer cell line representative of a luminal A subtype, estrogens stimulate expression of the Bcl-2 antiapoptotic protein that leads to a decrease in cell surface expression of E-cadherin and disruption of junctional complexes (27). Culturing MCF-7 breast cancer cells in the absence of estrogen down-regulates bcl-2 expression and up-regulates membrane localization of junctional proteins (27). In contrast, other studies have shown that estrogens have positive effects on expression of junctional components in breast cancer cells. For example, in MCF-7 cells, treatment with either 17β-estradiol or the propylpyrazole triol ER-α selective agonist strongly up-regulates protein and mRNA expression of claudin-6 (28). Exposure to the ER antagonist ICI 182,780 blocks the ER agonist induction of claudin-6 expression, which implicates a receptor-dependent estrogen control of claudin-6 gene transcription. Although ER-α interactions with the CLDN6 (claudin-6 gene) promoter have not been experimentally evaluated, sequence analysis reveals two consensus ER DNA binding elements in the CLDN6 promoter, suggesting the possibility of this tight junction protein being a direct ER-α transcriptional target. It is also interesting to note that phytoestrogens have been shown to abolish the stimulatory effects of 17β-estradiol on cell proliferation in human mammary epithelial cells by altering the distribution of tight junctional proteins such as ZO-1 and thereby disrupting the formation of hormone-dependent malignancies (29).

Tumor growth and metastasis are highly dependent on the dynamics of functional intercellular junctional complexes, and several studies implicate estrogens in regulating cytoskeletal components that in turn control the organization of adherens junctions. Treatment of MCF-7 breast cancer cells with 17β-estradiol induces the rearrangement of F-actin cytoskeleton and formation of lamellipodial structures with few focal contacts, whereas exposure to ER antagonists in the presence of estrogens suppresses development of lamellipodial structures (30). The adherens junction component E-cadherin is found within lamellipodial structures only after 17β-estradiol treatment, indicating that adherens junctions are under estrogen control (30). 17β-Estradiol triggers a decrease in vinculin- and talin-rich cell-matrix adhesion plaques and a reduction in the amount of vinculin that localizes within MCF-7 cell-cell contacts as well as causes a loss in the localization of F-actin and plakoglobin (31). These observations suggest that estrogens contribute to the ability of breast cancer cells to proliferate in multiple cell layers due in part to the disruption of cell-cell adhesion. Consistent with this concept, estrogen treatment of rat somatolactotropic GH3 cells induces a nonadherent phenotype that is reversed upon treatment with antiestrogens (32). Estrogen treatment reduces, and exposure to the ICI 182,780 ER antagonist increases, N-cadherin and β-catenin protein and transcript levels with corresponding effects on cell adhesion and proliferation. Also, exposure of cells to N-cadherin antibodies blocks the ability of antiestrogens to induce the formation of adherent aggregate of cells (32), which implicates the importance of this adherens junction protein in mediating ER-regulated cell aggregation.

Estrogen signaling through plasma membrane-associated processes has been shown to play an important role in the integrity of cellular junctions in certain tissue types. For example, 17β-estradiol interactions with the plasma membrane, presumably through some sort of membrane-associated receptor, can transiently disrupt uterine microvascular endothelial cell adherens junctions (33). Treatment with either membrane permeable 17β-estradiol or with the non-cell membrane-permeable 17β-estradiol-linked BSA increased monolayer permeability of endothelial cell monolayers by the disruption of adherens junctions, which was triggered by the detachment of adherens junction proteins from the cytoskeleton, resulting in the redistribution of α-catenin from the plasma membrane. Cotreatment of cells with estrogens and the 4-amino-5-(4-chlorophenyl)-7-(t-butyl)pyrazolo[3,4-d] pyrimidine Src-family tyrosine kinase inhibitor prevented the steroid-induced phosphorylation and redistribution of adherens junction proteins (33), implicating a critical role for Src activity in this estrogen-mediated membrane response.

Consistent with a potential role of cell signaling cascades in regulating the junction effects of estrogen-related compounds, in both androgen-dependent and -independent prostate cancer cells, 2-methoxyestrodiol, which is derived from estrogen and has anticancer activities, was shown to up-regulate β-catenin levels through a cellular process that required MAPK kinase (MEK)-ERK2 signaling (34). Although the precise mechanism is not well understood, the cytostatic effect of 2-methoxyestrodiol was proposed to be mediated by the inhibition of β-catenin translocation into the nucleus with a corresponding enhancement of membrane associated β-catenin and increase in cell-cell adhesion (34).

The steroid receptor coactivator-3 (SRC-3), a member of a family of nuclear receptor coactivators that regulates the efficacy of receptor-mediated gene transcription, is expressed at high levels in many estrogen-responsive human reproductive cancer cell types, such as ovarian and breast cancers. In ovarian cancer cells, inhibition of SRC-3 activity caused a reduction in cell spreading and migration with altered intercellular localization of the focal adhesion kinase, which is highly involved in regulating cell-cell interaction (35). In a mouse model of adenomatous polyposis coli-associated colorectal cancer, ovariectomy increases intestinal adenomas, and estrogen replacement therapy of the animals reduces the intestinal adenomas to baseline with a concomitant up-regulation of ER-β levels (36). When these mice are treated with estrogen or phytoestrogens, a more efficient association between β-catenin and E-cadherin is observed, which is proposed to enhance enterocyte migration and intercellular adhesion (36). An important future area of investigation will be to uncover the ER-β and/or SRC-3 coactivator controlled target genes involved in directly regulating adherens junction dynamics.

Progesterone

In reproductive tissues, progesterone can regulate both adherens junctions and tight junctions, although relatively little is known about receptor-dependent mechanisms. The cervix epithelium has several functions that rely on junctional complexes, including the maintenance of fluid balance, protection from environmental stimuli, and paracellular transport of solute via tight junctions. Progesterone production is associated with the negative regulation of the permeability barrier during cervical ripening (37). During this time frame, expression of claudin-1 and claudin-2 tight junction protein gene transcripts are temporally regulated in opposite directions, with claudin-1 mRNA expression increasing and claudin-2 expression declining (37). Furthermore, claudin-1 localizes to the cell periphery at the end of pregnancy. In cervical ripening-deficient mice, elevated local levels of progesterone correlated with aberrant expression of claudin-1 and claudin-2. These mice also fail to form claudin-1-enriched tight junctions and have abnormal expression of genes involved in epithelial terminal differentiation (37). Progesterone was shown to control the expression of claudin-4 and occludin during blastocyst implantation in humans. The levels of both tight junctions proteins increase at day 1 of pregnancy and eventually become concentrated to the tight junctions at the time of implantation (38). Progesterone also influences the development, polarity and junctional complex organization of the blastocyst by attenuating Wnt signaling and the β-catenin control of tight junction and adherens junction dynamics (39). Within mammary epithelium, progesterone regulates tight junction dynamics so that the tissue becomes impermeable to transepithelial transport of molecules during lactation (40), and an important future direction will be to identify the progesterone regulated genes and characterize the precise receptor-dependent transcriptional mechanism operating in this system.

Progesterone acts through its cognate progesterone receptor (PR) of which there are two isotypes, PRA and PRB. In the normal uterine epithelium, the progesterone inhibition of cell proliferation and endometrial carcinogenesis is correlated with the disruption in expression of either one or both of these receptor isoforms. Hec50co cells, a poorly differentiated endometrial cancer cell line, express a small amount of PRA but lack ER and PRB. In these cells, progesterone inhibits the expression of the cellular adhesion molecules fibronectin, integrins, and cadherin-6, which modulate endometrial cancer cell invasiveness and metastatic potential (41). In cells transfected with exogenous PRB, progesterone induced a greater down-regulation of expression of each of these adhesion molecules, which is consistent with the loss of adherens junctions (41).

Estrogen and progesterone

Estrogen and progesterone have been shown to act antagonistically in certain normal and transformed female reproductive tract cell systems to control cellular differentiation and proliferation. The adherens junction protein E-cadherin plays a role in the detachment of cancer cells, which can regulate cancer cell invasion and metastasis. In Ishikawa cells, a well-differentiated human endometrial cancer cell line, progesterone reverses the estrogen suppressed cell-cell aggregation activity and inhibition of transcript levels of E-cadherin, α-catenin and β-catenin (42). This observation suggests that progestins induce the differentiation of cultured endometrial cancer cells in part through the up-regulation of adherens junction proteins. Consistent with the observations, in human umbilical cord derived endothelial cells, 17β-estradiol decreases and progesterone enhances endothelial cell barrier properties (43).

Androgens

Androgen receptor (AR) signaling regulates tight junction dynamics between Sertoli cells (44), which is important for formation of the blood-testis barrier and spermatogenesis. Evidence from several studies suggests that members of the claudin gene family of tight junction proteins are likely AR transcriptional targets. Using a quantitative RT-PCR analysis, T was shown to induce claudin-11 expression in immature mouse Sertoli cells (45). Claudin-3 is an androgen-inducible gene, and in mice that lack Sertoli cell-specific AR, expression of claudin-3 is significantly reduced, which in turn increases the permeability of the blood-testis barrier and compromises testicular immune privileges (46). Sertoli cells form functional tight junctions as measured by TER 3 days after treatment with T or dihydrotestosterone (DHT). Both androgens stimulate expression of claudin-1 transcripts as well as induce localization of claudin-11 and occludin to intercellular contacts (45, 47).

Exposure of Sertoli cells to heat causes a decrease in AR expression as well as a concomitant drop in levels of two adherens junction proteins, N-cadherin and β-catenin, as well as the ZO-1 tight junction protein (48). Treatment with the AR antagonist flutamide mimics these heat-induced changes, whereas the expression of exogenous AR strongly up-regulated the expression of N-cadherin, β-catenin, and ZO-1 (48). This observation strongly suggests that disruption in AR expression caused by heat treatment in Sertoli cells likely triggers the heat-induced changes in junctional proteins (48). In benign prostate glands, low serum levels of T leads to a decrease in gene expression of claudin-4 and claudin-8 that results in defective tight junctions. This effect is reversed upon T supplementation (49). Conceivably, the AR regulated expression of claudin gene family members may result from a combination of primary (direct) and secondary (indirect) transcriptional effects of AR, which may depend on the tissue type, environmental state, and the influence of other hormonal regulators.

In prostate cancer cells, a reduction in the level of the adherens junction protein E-cadherin is associated with low tumor grade and poor prognosis. An early study demonstrated that treatment of LNCaP cells, a hormone-sensitive human prostate cancer cell line, with estrogen and androgen increases expression of E-cadherin protein that was not reversed by the addition of the corresponding receptor antagonists (50). β-Catenin plays a role in prostate development and tumorigenesis through its effects on E-cadherin-mediated cell adhesion and the Wnt-dependent signal transduction pathway. In prostate cancer cells, β-catenin was shown to directly interact with the AR but not with other steroid receptors such as ER-α, PR-β, and the glucocorticoid receptor (GR) (51). β-Catenin augments the ligand-dependent activity of AR by interacting with the ligand-binding domain of AR and its own armadillo repeats in the N terminus. When E-cadherin is expressed in E-cadherin-negative prostate cancer cells, β-catenin relocalizes to the cell periphery at the site of junctional complexes with a concomitant reduction in AR-mediated transcription (51). This observation suggests that the loss of the adherens junction protein E-cadherin can lead to an increase in the amount of β-catenin protein, which in turn contributes to a more malignant phenotype that increases AR activity during prostate cancer progression.

After expression of exogenous wild-type AR in androgen-insensitive human prostatic adenocarcinoma cells, treatment with DHT induced the translocation of AR to the nucleus with a concomitant disruption in cell-cell adhesion. This response was accompanied by a reduction in the level of E-cadherin expression that correlated with an inhibition of proliferation and increase in apoptosis (52). T-activated AR signaling can modify the actin cytoskeleton through the control of signal transduction components, which in turn alters the interactions and function of junctional proteins. For example, androgen treatment of LNCaP prostate cancer cells induces the association of phosphorylated focal adhesion kinase with phosphatidylinositol 3-kinase (PI3K) and activation of Cdc42/Rac1, which results in the organization of the actin cytoskeleton (53). Adherens junctions play an important role in spermatogenesis. By disrupting adherens junction formation through the suppression of intratesticular T levels in rats, it has been shown that adherens junction dynamics are regulated by the activation of kinases such as PI3K and c-Src (54). This effect is correlated with the disrupted interaction between E-cadherin and β-catenin along with a concomitant reduction in c-Src-dependent tyrosine phosphorylation (54).

In nonmetastatic breast cancer cells, DHT-bound AR represses E-cadherin gene transcript expression in the context of inducing changes in cell morphology and promoting tumor metastasis (55). AR and histone deacetylase-1 synergistically down-regulate E-cadherin gene expression, and importantly, activated AR mediates this transcriptional response by its direct binding to the E-cadherin promoter both in vitro and in vivo (55). This study directly established that E-cadherin is a transcriptional target gene of AR. Furthermore, when nonmetastatic breast cancer cells are transplanted into mice, DHT treatment stimulates tumor development due to the loss of E-cadherin. This observation is consistent with clinical data from breast cancer patients with invasive ductal carcinomas showing that high levels of nuclear AR are associated with significantly reduced E-cadherin expression (55).

Glucocorticoids

A role for glucocorticoids in controlling tight junction dynamics through regulated changes in tight junction protein expression and/or function has been well documented in a variety of tissue types including mammary, brain, liver, blood, lung, and intestine. Although many studies suggest that GR-stimulated or -inhibited gene transcription accounts for the effects on tight junctions, a unifying mechanism of action and common GR-regulated target genes has not been identified, in part due to tissue and cell dependent differences in glucocorticoid responses. In a few systems, emerging evidence implicates occludin and potentially claudin-5 as primary glucocorticoid-responsive tight junction genes, whereas other studies show that glucocorticoids regulate expression of cell signaling molecules, such as the serum- and glucocorticoid-inducible protein kinase (Sgk-1), which in turn alter the function and/or assembly of junctional complexes (see later discussion in this section).

In a relatively early study, freeze-fracture/thin section approaches using H4-IIE hepatoma cells showed that treatment with the synthetic glucocorticoid dexamethasone (Dex) was sufficient to stimulate the formation of tight junctions (56). Tight junctions are important in endothelial cell function because a tightening of endothelial cell-cell contacts is a critical process at the end of angiogenesis, which allows the controlled transfer of solute between the blood stream and solid tumors. Hydrocortisone and dibutytyl cAMP induce a permeability barrier in endothelial cells that stimulates junctional localization of adherens junction protein plakoglobin and tight junctions proteins, ZO-1, ZO-2, and occludin (57). Interestingly, this study also showed that coculturing of human umbilical cord vein endothelial cells and murine smooth muscle-like 10T1/2 cells induces the tightening of endothelial cell-to-cell contacts and increases junctional localization of plakoglobin, ZO-1, ZO-2, and occludin. In human corneal epithelial cells, Dex protects the hypoxia-induced loss of the tight junction protein ZO-1 from the cell periphery by reversing the down-regulation of ZO-1 expression (58). In these cells, there were no observed effects on expression or localization of the tight junction protein occludin or of the adherens junction proteins E-cadherin and β-catenin. Similarly, glucocorticoids increase the barrier properties of the blood-nerve barrier by selectively up-regulating expression of claudin-5 (59). The human CLDN5 (claudin-5) gene promoter contains a consensus glucocorticoid response element (GRE)-like sequence; however, in contrast to the analysis of ER interactions with the murine claudin-5 gene promoter (26), the direct interactions of the GR with this putative GRE has not been evaluated. Therefore, an important future direction in this field that is needed to establish the mechanism by which glucocorticoids stimulate expression of claudin-5 in different tissue types.

One of the important roles for the tight junction protein occludin is being an essential element of the blood-brain barrier, and glucocorticoid treatment of the cEND brain microvascular cell line up-regulates expression of occludin through a GR-dependent process (60). Chromatin immunoprecipitation and site-directed mutagenesis of the occluding promoter demonstrated that GR transactivation of occludin gene expression is mediated by a single imperfect GRE (5′-ACATGTGTTTACAAAT-3′) within the occludin promoter that surprisingly contains a 4-bp instead of a 3-bp spacer between two highly degenerate half-sites (60). The intracellular barrier function conferred by tight junctions is critical to functioning of intestinal epithelial cells, and defective intestinal epithelial tight junction barrier contributes to inflammation of Crohn's disease. Treatment with either Dex or prednisolone prevents the TNF-α induced an increase in the permeability of Caco-2 intestinal epithelial cells by inhibiting the TNF-α-induced increase in myosin light-chain kinase (MLCK) protein expression (61). The prednisolone bound GR was shown to directly bind to a GRE within the MLCK promoter, which blocks the TNF-α-mediated activation of the MLCK promoter (61).

TNF-α levels are elevated in the vitreous of diabetic patients and in the retinas of diabetic rats, which correlated with an increase in vascular permeability that occurs through a decrease in protein and mRNA levels of ZO-1 and claudin-5 along with disruption in localization of these proteins from the cell periphery. Treatment of bovine retinal endothelial cells or rat retinas with Dex reverses the TNF effects on tight junction protein expression through GR transactivation and nuclear factor-κB transrepression (62). Glucocorticoids serve as the most effective antiinflammatory drugs in the treatment of asthma. Treatment of human airway epithelial cell monolayers with Dex increases tight junction formation that is reversed upon knockdown of epidermal growth factor (EGF) receptor (63). It was further demonstrated that Dex increases EGF receptor phosphorylation without changes in total receptor expression levels, suggesting a role for EGF receptor signaling in maintaining the GR effects on tight junction dynamics (63). The precise GR regulated pathway that accounts for this observation has not been uncovered.

Glucocorticoids can have a range of effects on various tissue types that involve changes in adherens junction organization and function. Using tissue surface tensiotometry, a technique that measures cohesivity of tissue under physiological conditions, Dex was shown to decrease invasiveness and increases aggregate cohesivity in brain tumors (64). The decrease in invasiveness correlated with an increase in aggregate cohesivity but not with expression or function of N-cadherin (64). In contrast, Dex treatment of cEND brain microvascular cells increases E-cadherin protein levels and stimulates the relocalization of E-cadherin protein to the cytoskeleton (65). This study also showed that Dex induced the formation of cobblestone cellular morphology and reinforcement of adherens junctions that is concomitant with an increase in E-cadherin attachment to the actin cytoskeleton. Treatment of primary retinal endothelial cells with glucocorticoids stimulated expression of occludin and claudin-5 transcripts and protein that is reversed upon treatment with the GR antagonist RU486. Consistent with observations in cEND brain microvascular cells (60), a noncanonical GRE in the occludin promoter was shown to be responsible for the glucocorticoid-induced response in the retinal endothelial cells (66). Analogous promoter analysis of the claudin-5 gene awaits further studies. Furthermore, small interfering RNA knockdown experiments demonstrate that the p54/NONO transcription factor in the retinal endothelial cells is necessary for the glucocorticoid stimulation of occludin and claudin-5 gene expression and for the barrier inducing effects of this steroid (67). In bronchial airway epithelial cells, treatment with Dex reverses the asthma-associated disrupting effects on E-cadherin expression that is caused by damage to junctional complexes. This effect is dependent on the down-regulation of phosphorylated ERK1/2 levels after Dex treatment (68).

The molecular and cellular mechanisms by which glucocorticoids regulate tight junction and adherens junction dynamics have been intensely studied in rodent mammary epithelial cells. These studies provide strong evidence for the involvement of steroid regulated cell signaling events in the control of interacellular junctional complexes. With one exception (see later discussion), the primary GR-regulated signal transduction genes have not been defined. In the 31EG4 nontumorigenic mouse mammary epithelial cell line, indicative of the formation of functional tight junctions, Dex dose-dependent increased TER and decreased apical to basolateral paracellular permeability that correlated with glucocorticoid receptor occupancy and required de novo protein synthesis and extracellular calcium (69). Under these conditions, the production or localization of actin, ZO-1, and E-cadherin was not altered. In a follow-up study, an inhibition of phosphatase activity reversed the Dex-mediated stimulation of TER, increased tight junction permeability, and caused cellular redistribution of ZO-1 (70). These results implicate a critical role of cellular phosphorylation/dephosphorylation signaling cascades in the ability of glucocorticoids to stimulate tight junction function. In 31EG4 mammary epithelial cells, TGF-β and Dex were shown to antagonistically regulate the formation of tight junctions (71). Exposure of 31EG4 cells to TGF-β in the presence of up to 24 hours of pretreatment with Dex inhibits the steroid induced increase in TER and disrupts the Dex-regulated cellular staining pattern for ZO-1 and F-actin at the cell periphery without altering the level of expressed ZO-1 protein. In contrast, 48 hours of pretreatment with Dex stabilizes the cellular tight junction organization in a manner that is resistant to the disruptive effects of TGF-β (71).

Using the highly glucocorticoid-responsive Con8 rat mammary epithelial tumor cell line, Dex was shown to induce a significant increase in TER and decreases paracellular transport across the monolayer in a GR-dependent process (72). After glucocorticoid treatment, the tight junction protein ZO-1 is localized to the cell periphery, indicating its participation in the formation of the apical junction complex. In this same study, constitutive expression of TGF-α was shown to reverse the glucocorticoid-induced formation of tight junctions, disrupt the steroid regulated localization of ZO-1 at the cell periphery, and prevent the Dex-mediated antiproliferative response (72). These results demonstrate that the proliferation and differentiation of rat mammary epithelial cancer cells may involve opposing signals between glucocorticoids and TGF-α receptor signaling. In a follow-up study, Dex was shown to induce the polarity of Con8 cells with distinct apical and basal sides (73). When TGF-α is administered to the basolateral but not the apical plasma membrane compartment, the glucocorticoid-induced increase in TER is disrupted and ZO-1 localization becomes highly disorganized. Intriguingly, TGF-α induces its proliferative response equally on both sides of the cell monolayer, suggesting the glucocorticoid induced apical junction complex forms a cellular architecture that selectively enables polarized responses to growth factor signaling (73).

The Con8 mammary epithelial tumor cells were used to establish that different sets of glucocorticoid-regulated cell signaling components play critical roles in the GR control of tight junctions. One study showed that Dex strongly down-regulates the expression of the actin-bundling protein fascin and that the expression of constitutively active fascin prevented the Dex-induced increase in TER and relocalization of β-catenin and occludin to the cell periphery (74). Interestingly, expression of either the actin binding domain or the β-catenin binding domain of fascin act in a dominant negative fashion to prevent the inhibitory effects of wild-type fascin on the steroid induced formation of tight junctions. Growth factor receptor signaling appears to disrupt this response because the constitutive, basolateral production of TGF-α ablates the Dex-induced down-regulation of fascin and prevents both apical junction remodeling and the increase in TER (75). Furthermore, expression of dominant-negative RasN17 or treatment with a PI3K inhibitor, which selectively inhibits growth factor receptor signaling cascades, attenuated the TGF-α-mediated disruption of TER and apical junction remodeling and regulation of fascin protein levels (75). Consistent with a direct role for these cell signaling components in the glucocorticoid regulation of tight junction dynamics, Ras and the p85 regulatory subunit of PI3K were shown to be recruited to the regions of cell-cell contact upon Dex treatment. Expression of dominant-negative RasN17 or treatment with PI3K inhibitors attenuated the Dex-induced TER without effecting remodeling of the apical junction complex (76), which suggest that downstream components of growth factor signaling pathways act at one of the later steps in the GR-regulated formation of functional tight junctions.

Associated with the glucocorticoid regulation of the apical junction complex, Dex rapidly stimulates expression of the inhibitor of differentiation-1 (Id-1) protein, a serum-inducible transcriptional repressor (77). Expression of exogenous Id-1 facilitates the Dex-induced TER without effecting reorganization of ZO-1, β-catenin, and F-actin. Knockdown of Id-1 production prevents both Dex-induced increase in TER and reorganization of apical junction proteins, which demonstrates that Id-1 is required for this process (77). Dex was also shown to down-regulate the RhoA small GTPase before stimulating TER of the Con8 cell monolayer. In functional tests for this pathway, expression of exogenous active RhoA was shown to disrupt the Dex-induced tight junction effects, whereas the expression of dominant-negative RhoA mimics the Dex-induced junctional response in the absence of steroid (78). Consistent with these observations, expression of exogenous Rnd3/RhoE, a GTPase deficient Rho family member and RhoA antagonist, induces tight junction sealing and relocalization of β-catenin and ZO-1 to the cell periphery in the absence of glucocorticoids (79). In a follow-up study, glucocorticoids were shown to regulate the expression and/or activity of the RhoA downstream effectors ROCK1 and ROCK2 in the rodent mammary epithelial tumor cells (80). Dex treatment of Con8 cells increases ROCK2 protein levels and activity and simultaneously down-regulates ROCK-1 kinase activity without affecting its protein levels. When these cells are treated with the Y-27632 ROCK inhibitor, the glucocorticoid induced and the Rnd3-induced stimulation in tight junction sealing are ablated (80).

The adherens junction protein β-catenin plays an important role in the glucocorticoid regulation of both adherens and tight junction formation. Treatment of Con8 rat mammary epithelial tumor cells with Dex results in the up-regulation of β-catenin protein and transcript levels and the down-regulation of the phosphorylated form of β-catenin (81). Indirect immunofluorescence revealed that β-catenin localizes to the cell periphery upon Dex treatment and that these effects are not due to inhibition of cell proliferation. Expression of dominant-negative RasN17 prevents the Dex-induced up-regulation of β-catenin protein expression but did not alter the β-catenin relocalization at the cell membrane. Consistent with the generally opposing actions of glucocorticoids and TGF-α signaling on cell-cell interactions in mammary epithelial cells, the constitutive production of TGF-α prevented the Dex-induced relocalization of β-catenin without altering its protein levels (81).

In one of the more decisive studies that characterized steroid induced cell signaling pathways involved in the glucocorticoid regulation of apical junction formation, the GR-mediated stimulation of Sgk expression through a GRE in the Sgk promoter was shown to trigger the Dex stimulation of tight junction dynamics and apical junction formation in Con8 mammary epithelial tumor cells (82). Sgk is the only identified primary GR responsive signal transduction gene that has been established to play a role in apical junction dynamics. Glucocorticoid-induced Sgk activates a signaling cascade that results in the stabilization and membrane localization of the nonphosphorylated form of β-catenin. In the absence of glucocorticoids, β-catenin is phosphorylated by glycogen-synthase kinase-3 (GSK-3) and phosphorylated β-catenin is either transported into the nucleus or is degraded in the cytoplasm. In either scenario, β-catenin is inaccessible to the apical junction complex. In the presence of glucocorticoids, the Dex-induced expression of Sgk phosphorylates GSK-3, which activates GSK-3 and thereby signals the phosphorylation and ubiquitin-mediated degradation of GSK-3. The loss of GSK-3 prevents phosphorylation of β-catenin, allowing a stable form of β-catenin to participate in adherens junction assembly and subsequently tight junction sealing (82).

The effects of glucocorticoids on junctional complex dynamics were investigated in Ishikawa cells (83), a highly steroid responsive human endometrium cancer cell line. This study showed an intriguing functional connection between GR signaling and that of the HER2 member of the EGF receptor gene family. Treatment with Dex up-regulates the HER2 and prevents arecoline, the causative agent of precancerous oral lesions, from down-regulating ZO-1 protein expression and disrupting ZO-1 localization to the cell periphery (83). Furthermore, in Ishikawa cells expressing exogenous HER2, Dex treatment induces the organization of the overall junctional complex. Our recent results show that glucocorticoids stimulate a bimolecular interaction between ZO-1 and occludin protein that is mutually required for both tight junction proteins to localize to the cell periphery. Furthermore, this GR-dependent response requires AMP-activated protein-kinase activity, the loss of c-Src signaling, and the expression of Kruppel-like factor-9 transcript (Kapadia, B. J., and G. L. Firestone, unpublished results).

Vitamin D

A role of vitamin D has been implicated in the regulation of tight junctions within the intestinal epithelium. The active form of vitamin D, 1α,25-dihydroxyvitamin D3 [1,25(OH)2D3] is known to promote intestinal calcium absorption across the intestinal epithelium monolayers via both paracellular and transcellular pathways. In Caco-2 cells, a human epithelial colorectal adenocarcinoma cell line, treatment with 1,25(OH)2D3 induced the expression of ZO-1 and E-cadherin proteins within junctions and increased the TER (84), which is a functional measure of tight junction sealing. Small interfering RNA knockdown of the vitamin D receptor strongly reduces the expression of tight junctional proteins and inhibits the monolayer TER. This observation demonstrates that the vitamin D receptor plays a role in the homeostasis of the mucosal barrier by preserving the integrity of junctional complexes (84). In the small and large intestines of female rats, the vitamin D receptor and membrane-associated rapid response steroid binding receptors are expressed at high levels. Acute vitamin D treatment stimulates the solvent drag-induced calcium transport, which is abolished by exposure to pharmacological inhibitors of PI3K, PKC, or MEK (85). A longer 3-day treatment with vitamin D increases transcellular active duodenal calcium transport. There is no detectable widening of tight junctions; however, vitamin D decreases the TER and increases the sodium and chloride permeability through the paracellular pathway (85). This study suggests that PI3K, PKC, and MEK signaling pathways are required for vitamin D to enhance solvent drag-induced calcium transport by changing the charge-selectivity of the duodenal epithelium (85).

In enterocytes, 1,25(OH)2D3 up-regulates the expression of claudin-2 and claudin-12 tight junction proteins through a vitamin D receptor-dependent process. Claudin-2 and claudin-12 form paracellular calcium channels in the intestinal epithelium, which likely provides a mechanistic perspective on how vitamin D maintains calcium homeostasis (86). In human colon cancer cells, treatment with vitamin D hinders the localization of occludin at the plasma membrane in a process that requires the inactivation of RhoA, a member of the Ras superfamily of small GTP-binding proteins (87).

Vitamin D3 is an important regulator of bone metabolism, and cadherin-mediated cell-cell adhesion is a critical for the development and survival of osteogenic cells. Using microarray gene analysis of human cancer cells, the gene targets of the active form of vitamin D3, 1,25(OH)2D3, include those involved in cellular adhesion such as E-cadherin and α-catenin (88). In murine osteoblast-like cells, vitamin D3 strongly suppresses protein levels of N-cadherin and pancadherin, which affects apical junction dynamics and cell-cell contact (89). Vitamin D3 has also been shown to stimulate the assembly of adherens junctions and desmosomes in cultured human keratinocytes (90). Upon treatment with vitamin D3, E-cadherin, P-cadherin, α-catenin, and vinculin relocate to the cell periphery, indicating the presence of functional adherens junctions. Exposure to pharmacological inhibitors of PKC activity inhibits the vitamin D3-stimulated formation of adherens junctions that is reversed upon activation of PKC. This observation implicates a role for PKC signaling in the regulation of vitamin D3 induced adherens junctions (90). An important future direction will be to identify the primary vitamin D3 regulated genes that control apical junction dynamics and cell-cell interactions.

Thyroid hormones

Relatively little is known about the thyroid hormone control of the apical junctions. Thyroid hormones are produced in the thyroid gland by follicular cells that are highly polarized with a distinct apical and basolateral membrane. Expression of the tight junctional proteins occludin and claudin have been shown to change during the maturation procession of proximal tubules in response to increases in the serum levels of T₃ (91). Based on coimmunoprecipitation results, T₃ treatment of murine osteoblast-like cells was shown to increase the association of pancadherin with β-catenin. In addition, exposure to T₃ causes vinculin to localize to the junctional complex (91). These results, in combination with previously discussed studies (89), indicate that T₃ and vitamin D3 have opposite influences on the functional competence of cell-cell contacts and adherens junction formation.

Retinoic acid and carotenoids

Retinoids function through two members of the nuclear receptor superfamily of ligand-dependent receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs). This class of nuclear receptor ligands induces differentiation of columnar epithelial cells and prevents metaplasia into stratified squamous epithelial cells that are void of tight junctions, suggesting a role for these ligands in regulating the structure and function of tight junctions in columnar epithelial cells. Indeed, treatment of F9 murine embryonic carcinoma cells with retinoic acid generated highly organized tight junctions and induced the transcript expression of the ZO-1, occludin, claudin-6, and claudin-7 tight junction protein genes through RXR-RAR heterodimers (92). Further evidence suggests that in response to retinoic acid, RARα stimulates barrier function of epithelial cells via tight junction formation due to the induced expression of occludin, claudin-1, claudin-4, and ZO-1 (93), although the transcriptional mechanism of this response has not yet been characterized.

In hepatocytes, retinoic acid treatment induces an interaction between β-catenin and E-cadherin and stimulates the interaction between ZO-1 and occludin through a mechanism that involves a reduction in tyrosine-phosphorylation levels of β-catenin and ZO-1 (94). These retinoic acid-induced protein complexes are localized to the cell periphery. A similar observation was made in metastatic human renal carcinoma cells, in which treatment with all-transretinoic acid induces the localization of E-cadherin and β-catenin to the cell periphery in a process that is dependent on the dephosphorylation of tyrosine residues in β-catenin (95). In human keratinocytes, all-transretinoic acid increases the expression of claudin-2, implicating its role in epidermal differentiation (96). Retinoids have been shown to stimulation the differentiation of human epithelial colorectal adenocarcinoma cells that corresponds with an increase in the expression of claudin-2 and the enhancement of TER (97). Consistent with the differentiation effects of retinoids, treatment of Sertoli cells with retinoid acid initiated tight junction formation, whereas the expression of a dominant-negative form of RARα was shown to disrupt the formation of a blood-testis-barrier due in part to the down-regulation of occludin gene expression (98). Although suggestive of occludin being a primary gene target of RAR-α, the molecular mechanisms underlying these interesting observations have not been elucidated.

One of the components of intercalated disks between myocytes is the adherens junction. Retinoic acid plays an important role in cardiomyogenesis, and these early cardiogenic effects are dependent on the presence of N-cadherin (99). In SKBR3 cells, a HER2-positive human breast cancer cell line, treatment with retinoic acid leads to differentiation and reduction in cell proliferation. This response is mediated through the retinoic acid-induced expression of cadherin levels (100). A recent study showed that all-trans retinoic acid promotes epithelial barrier function of myopia retinal pigment epithelium by rapidly stimulating expression of the E-cadherin adherens junction component as well as of the occludin and claudin-1 tight junction proteins (101). Taken together, these studies suggest that the retinoic acid control of adherens junction components through RAR/RXR-mediated transcriptional effects is integral to the differentiation and antiproliferative effects of this nuclear receptor ligand.

Peroxisome proliferator-activated receptors

Peroxisome proliferator-activated receptors (PPAR) are a part of the nuclear hormone receptor superfamily of ligand-activated transcription factors. PPAR-RXR heterodimers regulate transcription by binding to PPAR response elements in the promoter of target genes. Ligands for PPARs include free fatty acids and signaling molecules derived from the oxidation of 20-carbon essential fatty acids (102). PPARs have been implicated in the regulation of tight junctions in several cell systems. As with other nuclear receptors, the effect of PPARs on junctional complexes is also tissue and cell type specific. In untreated ileum tissue, ZO-1 localization is disorganized and the expression of occludin and β-catenin is relatively low. Exposure to endogenous PPAR-alpha ligands reduces tight junction permeability in the ileum tissue (103). In contract, alkylphospholipids, which are also ligands for PPARs, can modulate cancer cell invasion through changes in adherens junction protein complexes. For example, in human colorectal cancer cells, exposure to alkylphospholipids causes a decrease in TER and changes the solubility of ZO-1 and occludin tight junction proteins (104).

PPARs are involved in the regulation of tight junctional proteins in the urothelium, in which tight junctions maintain urothelial barrier functions. In the differentiated urothelium, claudin-3, claudin-4, claudin-5, claudin-7, ZO-1, and occludin proteins are expressed, whereas in a proliferating urothelium, claudin-1 is highly expressed. When PPAR is activated, the expression of claudin-3 mRNA and protein is altered, depending on the differentiated state of the urothelium (105). Additionally, after PPAR activation, claudin-2 mRNA is down-regulated, whereas claudin-4 and claudin-5 protein levels are up-regulated, and the tight junction proteins relocalize to the cell periphery. This observation suggests that PPAR is associated with stimulating tight junction organization and inducing tight junction protein expression that leads to a more differentiated urothelium (105).

Tight junctions regulate the integrity of the blood-brain barrier, which among other functions, plays a role in HIV trafficking into the brain and the development of central nervous system (CNS) complications associated with HIV. It has been shown that expression of exogenous PPAR-α or PPAR-γ at high levels attenuates HIV-mediated deregulation of tight junction proteins by down-regulating proteosomal activity and matrix metalloproteinases. Thus, the reestablishment of tight junctions by PPAR signaling may confer a protective role against an HIV-induced disruption of brain endothelial cells (106).

Summary and perspectives

Intercellular junctional complexes have highly dynamic structures whose permeability and adhesive properties can be acutely regulated by nuclear receptor-dependent transcriptional signaling and in some systems by the membrane actions of ligands for this family of receptors. Emerging molecular, cellular, and in vivo evidence has uncovered a diverse set of nuclear receptor-directed targets, such as specific tight junction and adherens junction protein genes, signal transduction molecules, transcription factors, and other regulatory molecules that have the potential to control the expression, modification, stability, function, and/or localization of specific structural and/or accessory components of junctional complexes (summarized in Tables 1 and 2). Such regulatory pathways could conceivably facilitate critical transitions in plasma membrane composition and function that control cell-cell interactions and cellular responsiveness to environmental cues in a physiologically appropriate manner. Within the context of a large body of cellular and physiological studies, in only a few systems have the precise mechanisms been evaluated by which nuclear hormone receptor mediated transcriptional signaling and certain ligand activated nontranscriptional effects control the assembly, disassembly, and maintenance of intercellular junctions. In many cases, even in which changes in expression of specific junctional complex proteins have been observed, primary (direct) or secondary (indirect) effects of nuclear receptor signaling have not been clearly delineated.

An exciting future direction will be to define the direct cellular targets of nuclear receptors and their ligands that control the dynamics of intercellular junctions in a tissue-specific manner. Another intriguing forthcoming experimental path will be to identify individual components of nuclear receptor-activated cellular cascades as pharmaceutical targets for the potential development of therapeutic strategies to control aberrant physiological and disease states associated with apparent alterations in intercellular permeability and adhesion. For example, in glucocorticoid induced glaucoma, the junctional protein ZO-1 has been proposed to be a potential target for developing new modalities for glaucoma therapy (107). Furthermore, the complex phenotype of mice carrying a null mutation in the occludin gene (108) suggests that this junctional component, and other components, represents potentially exciting therapeutic targets for a range of complex physiological disorders such as those affecting the endothelial cells that form the blood brain barrier (109) and in patients with intestinal permeability defects (110, 111).