Tight Junction-Based Epithelial Microenvironment and Cell Proliferation

Home , Claudin, Occludin, Tight junction

Oncogene (2008) 27, 6930–6938 & 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $32.00 www.nature.com/onc REVIEW Tight junction-based epithelial microenvironment and cell proliferation

S Tsukita1, Y Yamazaki, T Katsuno, A Tamura and S Tsukita2

Laboratory of Biological Science, Graduate School of Frontier Biosciences/Graduate School of Medicine, Osaka University, Suita, Osaka, Japan

Belt-like tight junctions (TJs), referred to as zonula bodies of multicellular organisms. The sheet-like divi- occludens, have long been regarded as a specialized sions allow diffusion barrier formation and selective differentiation of epithelial cell membranes. They are permeation of substances and ions (permselectivity), required for cell adhesion and paracellular barrier func- both excluding unnecessary or toxic molecules and tions, and are now thought to be partly involved in fence including necessary components to maintain home- functions and in cell polarization. Recently, the molecular ostasis. The combination of the belt-like adherens bases of TJs have gradually been unveiled. TJs are junctions (AJs) and tight junctions (TJs) have important constructed by TJ strands, whose basic frameworks are functions in the formation of epithelial cell sheets and composed of integral membrane proteins with four also in the formation of paracellular permselective transmembrane domains, designated claudins. The claudin barriers through their functions as septa (Mitic and family is supposedly composed of at least 24 members in Anderson, 1998; Tsukita and Furuse, 1999; Hartsock mice and humans. Other types of integral membrane and Nelson, 2008). Paracellular barrier functions with proteins with four transmembrane domains, namely occlu- permselectivity are unique to the epithelial cell system din and tricellulin, as well as the single transmembrane and regulate the internal homeostasis of ions and solutes proteins, JAMs (junctional adhesion molecules)and CAR in the body. As the proteinaceous components of TJs, (coxsackie and adenovirus receptor), are associated with such as claudins, occludin, tricellulin, JAMs and CAR, TJ strands, and the high-level organization of TJ strands is have been characterized (Furuse et al., 1993; Liu et al., likely to be established by membrane-anchored scaffolding 2000; Cohen et al., 2001; Ikenouchi et al., 2005), proteins, such as ZO-1/2. Recent functional analyses of evidence has accumulated that TJs are the main claudins in cell cultures and in mice have suggested that components of paracellular permselective barriers (Tsu- claudin-based TJs may have pivotal functions in the kita et al., 2001; Van Itallie and Anderson, 2006a; regulation of the epithelial microenvironment, which is Angelow et al., 2008). Recently, the epithelial-specific critical for various biological functions such as control of TJ-based permselective barrier system has been sug- cell proliferation. These represent the dawn of ‘Barriology’ gested to be involved in the regulation of cell prolifera- (defined by Shoichiro Tsukita as the science of barriers in tion, possibly by regulating the microenvironment multicellular organisms). Taken together with recent around epithelial cells (Tamura et al., 2008). This notion reports regarding changes in claudin expression levels, has attracted particular attention because over 90% of understanding the regulation of the TJ-based microenvir- cancers are of epithelial origin. This review will present onment system will provide new insights into the regula- some interesting points regarding the TJ-based epithelial tion of polarization in the respect of epithelial microenvironment and cell proliferation. microenvironment system and new viewpoints for devel- oping anticancer strategies. Oncogene (2008) 27, 6930–6938; doi:10.1038/onc.2008.344 Ultrastructural identification of TJ: implications for Keywords: tight junction; cell adhesion; paracellular their regulatory roles in the epithelial microenvironment barrier; paracellular permeability; permselectivity; epithelial cell More than half a century ago, local modification of the apical side of epithelial cells was recognized at the light and electron microscopic levels to affect cell-cell adhesion and paracellular barrier functions. Detailed Introduction electron microscopic analyses in 1963 revealed the morphological bases for this local modification and Epithelial cell sheets cover the internal and external defined them as junctional complexes composed of TJs, surfaces of organs and form compartments within the AJs and desmosomes (Farquhar and Palade, 1963). Among these junctional complexes, TJs are usually Correspondence: Dr S Tsukita, Laboratory of Biological Science, located closest to the lumen of the epithelia. Character- Graduate School of Frontier Biosciences/Graduate School of istically, the TJ-associated membranes of adjoining cells Medicine, Osaka University, Suita, Osaka 565-0871, Japan. E-mail: [email protected] come together, obliterating the intercellular space by 1Sachiko Tsukita in full name. fusion of their outer leaflets at the thin-section electron 2Shoichiro Tsukita in full name. microscopic level (the so-called kissing points of plasma TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6931 the epithelial physiology, paracellular permeability/ transport constitutes a substantial part of the total permeability/transport through epithelial and endothelial cell sheets, the ratios of which vary relative to transcellular transport, depending on different cell types (Van Itallie et al., 2003). Thus, the TJ physiology should be an important contributor to the epithelial and endothelial physiology. As an example, it is well known that paracellular ionic permeability dominates the transcellular ionic permeability in the small intestine (Frizzell and Schultz, 1972; Okada et al., 1977), and has an important function in the intestinal physiology, such as in the absorption of nutrients (Hopfer, 1977; Turk et al., 1991). Similarly, re-absorption of ions through paracellular pathways in the kidney is very important for regulating the ionic homeostasis of the body (Simon et al., 1999). Non-ionic solutes are thought to permeate through paracellular as well as transcellular pathways (Watson et al., 2001) and the molecular mechanisms for the permeabilities for ions and solutes appear to vary depending on the cell type, with the characteristics of the conductance-flux paradox (Yap et al., 1998; Van Itallie and Anderson, 2006a). It appears that the solute flux of Figure 1 Electron microscopic images of tight junctions (TJs), the low-capacity/size-independent pathway, the high- adherens junctions (AJs) and desmosomes. TJs are located at the capacity/size-restrictive pathway and ion-based electri- most apical parts of the plasma membranes of intestinal epithelial cells, whereas AJs and desmosomes are localized in the more basal cal permeability together determine the paracellular parts of the lateral membranes, as revealed by thin-section electron permeability. The proportions of their different con- microscopic images (a and c). Note the kissing points of TJs (c). TJs tributions to the overall permeability define the para- are represented by TJ strands at the apical parts of the plasma cellular permeability of individual tissue types. In any membranes in freeze-fracture replica images (b and d). Micrographs courtesy of H.Sasaki (b and c) (see references, Tamura et al., 2008 case, the molecular bases for TJ strands should be in and Tsukita et al., 2001). Bars, 100 nm. accord with the physiological aspects of the TJ concerned. membranes) (Figure 1). Consistent with such ultrastructural aspects, experiments with mass external tracers have suggested that TJs act as barriers to Molecular bases for TJ diffusion of the tracer by sealing off the lumen from the intercellular spaces. Thus, TJs are considered to have Although it was most likely that the intramembranous critical functions in regulating the microenvironment or particles in the freeze-fracture replica samples reflected circumstances around epithelial cells. the integral membrane proteins, the real nature of the TJ Another key ultrastructural aspect of TJs was strands, lipids and/or proteins remained a mystery until characterized by freeze-fracture replica electron micro- the proteinaceous components constituting the TJ scopy (Staehelin, 1973). Intramembranous TJ strands or strands were identified (Kachar and Reese, 1982; Pinto fibrils appear to underlie the lines of fusion of the outer da Silva and Kachar, 1982; Verkleij, 1984). Hence, when leaflets of the adjoining cell membranes at the epithelial a novel integral membrane protein with four transmem- apical sites. Thus, TJ strands have been suggested to be brane domains of B60 kDa, termed occludin (Furuse responsible for the intercellular sealing of epithelial/ et al., 1993), was identified as a component of TJ strands endothelial cell sheets (Figure 1). from hepatic junctional fractions (Tsukita and Tsukita, 1989), it was regarded as the Holy Grail in this field (Gumbiner, 1993). However, knockout mice for occludin still possessed TJ strands, suggesting that occludin TJ as a permselective barrier: implications for the was not necessarily required for TJ strand formation TJ-based epithelial microenvironment system (Saitou et al., 2000) (Figure 2). It was subsequently revealed that the molecular backbones of TJ strands Further evidence for the barrier functions of TJs was were likely to be constituted by the relatively large gene shown by electrophysiological approaches as well as by family of claudins, which have four transmembrane tracer barrier assays at the electron microscopic level. domains of B23 kDa (Tsukita and Furuse, 2000). It is Detailed epithelial electrophysiological analyses sug- generally accepted that the claudin family consists of at gested the existence of paracellular aqueous pores or least 24 members in mice and humans (Tsukita et al., channels within TJs (Diamond, 1977; Claude, 1978; 2001; Van Itallie and Anderson, 2006a), with the fish Powell, 1981; Reuss, 1992; Gumbiner, 1993). Based on Takifugu having 56 claudin genes, together suggesting

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6932

Figure 2 Schematic drawings of epithelial tight junctions (TJs) and adherens junctions (AJs) and TJ membrane proteins. (a) Belt-like TJs and AJs. In epithelial cells, TJs and AJs are arranged in a belt-like fashion to form paracellular permselective barriers at the apical part of the lateral membranes. (b) TJ membrane proteins. TJ membrane proteins are comprised of claudins, which constitute the basic framework of TJ strands, occludin, tricellulin, JAMs and CAR.

very highly sophisticated roles of TJ-based microenvir- et al., 2002). Occludin and JAM-1 were shown to be onments in multicellular organisms (Loh et al., 2004). associated with ZO-1 through both non-PDZ and PDZ The intercellular and intracellular interactions bet- domains (Itoh et al., 2001; Schmidt et al., 2004). Recent ween the multigene claudin family members have not studies on ZO-1/2/3-deficient cultured epithelial cells been examined in detail. Although the homotypic and/ have revealed unequivocally that ZO-1/2 determines or heterotypic interactions of claudin-1, -2, -3, -5, -6, where and when the claudins are polymerized to form -11, -14 and -19 in forming TJ strands have been TJ strands with barrier functions (Umeda et al., 2006). examined in claudin-free L fibroblastic cells, more These cells lacking in ZO-1/2/3, are normal in their basic detailed studies are required to fully understand the apico-basal polarity, but are abnormal in the fine critical molecular bases of TJ strands (Furuse et al., distribution of polarity proteins, such as PAR3, 1999; Morita et al., 2002; Krause et al., 2008). Among suggesting that tight junctions might be involved in the the family members, claudin-11 and -19 each appear to fine-tuning of cell polarity, although they are probably form TJ strands by themselves, based on both knockout not involved in the initiation of apico-basal polarity. mouse studies and cell culture studies (Gow et al., 1999; The recent findings that ZO-1- or ZO-2-deficient mice Miyamoto et al., 2005). showed the embryonic lethal phenotype, respectively, Recently, another novel integral membrane protein further suggest their essential roles in development with four transmembrane domains of B70 kDa was (Katsuno et al., 2008; Xu et al., 2008). identified and designated tricellulin (Ikenouchi et al., Although ZO-1/2 spatio-temporally regulate the 2005). Tricellulin is enriched in the tricellular TJs at formation of functional TJs, the specific characteristics positions where three cells meet, and it is assumed to of the TJ-dependent paracellular permselective barrier play roles in the formation of bicellular and tricellular functions are assumed to be essentially determined by TJs. In addition to the four transmembrane integral the extracellular domains of the TJ-constituting integral membrane proteins, single transmembrane proteins, membrane proteins expressed in that particular cell, designated JAMs and CAR (Liu et al., 2000; Cohen although the possible preferences of the scaffolding et al., 2001), were also identified as TJ strand- proteins for particular TJ strand components are constitutive proteins (Figure 2). Neither tricellulin and unclear. occludin, which both contain MARVEL (MAL and related proteins for vesicle trafficking and membrane link) domains (Sa´ nchez-Pulido et al., 2002), nor JAMs Mysterious molecular mechanisms of TJ physiology and CAR cannot form TJ strands by themselves, and implied by cellular and molecular studies are thus assumed to play modulating rather than basic roles in TJ organization. The molecular characteristics of TJs basically define The above-described TJ strand-constituting integral their barrier functions. However, the detailed molecular membrane proteins are highly organized by peripheral mechanisms underlying the variable barrier functions membrane-associated cytoskeletal scaffolding proteins. and the variety of permselectivities of individual cell The key connecting points for this issue are provided by types remain to be clarified at the cellular and molecular the PDZ-binding motifs of claudins, which bind to levels, as well as in knockout mice. The most essential PDZ-containing membrane scaffolding proteins such as question is how claudins and other TJ components form ZO-1/2/3 and MUPP1 (Itoh et al., 1999; Hamazaki paracellular barriers, and also how they may form the,

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6933

Figure 3 Schematic drawing of claudin molecules. The acidic hydrophilic, basic hydrophilic, uncharged hydrophilic and nonpolar hydrophobic amino acid residues are shown in distinct colors. The functional aspects are shown below each molecular model. T, functional effects of transfection of each claudin into cultured epithelial cells; M, MDCK cells; LLC, LLC-PK1 cells; D, human diseases caused by mutations in the corresponding claudin gene; KO, phenotypes that appear upon knockout of the corresponding claudin gene in mice; TER, transepithelial resistance; Extra, extracellular space; PM, plasma membrane; Cyto, cytoplasm. as yet, hypothetical pores/channels that have permsele- Yap et al., 1998). Second, transfection or knockdown of ctivity for ions and non-ionic solutes. Furthermore, it is individual claudins or their combinations in cultured possible that the dynamic nature of TJ strands caused by epithelial cells led us to speculate that claudin-2 and -15 polymerization of claudins with other TJ components contribute to cation channels/pores (Furuse et al., 2001; contributes to the paracellular permeation (Sasaki et al., Van Itallie et al., 2003), whereas claudin-4, -7 and -10a 2003; Matsuda et al., 2004). The in vivo dynamism of TJs contribute to anion channels/pores or cation barriers in epithelial cells is still a contentious issue. However, if (Hou et al., 2006; Van Itallie et al., 2006b) (Figure 3). such a dynamism does occur in vivo, it could represent a The plus or minus charges of the first loops of claudin paracellular permeation mechanism, that is not asso- molecules thus appear to have functions in paracellular ciated with permselectivity. These points may corre- epithelial permeation (Colegio et al., 2003). Third, the spond to the long-standing puzzle in the field, paracellular passage of nonionic solutes is size-depen- designated the conductance-flux paradox, thus contri- dent with an estimated diameter of 4 A˚ at most, buting to paracellular permeation in a characteristic although this is not the case for permselectivity (Watson fashion (Yap et al., 1998; Van Itallie and Anderson, et al., 2001; Van Itallie et al., 2008). In these paracellular 2006a). permeabilities, ions and non-ionic solutes are thought to At the present time, several aspects of the molecular pass across the TJ strands through channels/pores, mechanisms of TJ physiology have been implied by although we do not have any direct evidence for the cellular and molecular biology studies, as described actual existence of paracellular channels/pores of TJ below, although the detailed molecular mechanisms still strands, nor do we have any evidence for whether the remain unclear. First, ZO-1/2-deficient cultured epithe- putative paracellular channel/pores are any more than lial cells, which form epithelial cell sheets but lack TJs simple gaps or mismatches in the polymerization of (Umeda et al., 2006), exhibited no paracellular barrier claudins or related TJ proteins in TJ strands. These functions, but did exhibit the normal epithelial cell sheet issues are really challenging for TJ physiology. Fourth, formation (mediated by cell-cell adhesion through the in contrast, paracellular permeability without belt-like arrangement of AJs). Thus, the proteinaceous permselectivity is assumed to depend on the dynamic TJ strands are dominantly responsible for the para- features of TJ strands. Specifically, fluorescent live-cell cellular barrier functions, consistent with the dominant imaging has revealed dynamic aspects of TJ strands in L contribution of TJs, compared with the minor contribu- cells (Sasaki et al., 2003), thereby leading to the proposal tion of the lateral intercellular space, to paracellular of TJ dynamism in epithelial cells, although such an idea resistance that was revealed by physiological measure- remains to be solved in future studies and knockout ments (Kottra et al., 1993a; Kottra and Fro¨ mter, 1993b; mouse studies.

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6934 Mysterious molecular mechanisms of TJ physiology How segregation of the apical and lateral implied by gene mutation studies and knockout microenvironments by permselective TJs is related mouse studies to cell proliferation

Similar mysterious issues to those implied by cellular The apical and lateral microenvironments of epithelial and molecular studies have also been deduced cells are segregated by TJ-dependent paracellular from human mutation studies of claudins and knockout barriers (Tsukita et al., 2001; Van Itallie et al., 2006a). mouse studies of claudins and occludin (Figure 3). This segregation provides one answer to the question of Hypomagnesemia was reported to result from mutations how ligand–receptor interactions are controlled in in paracellin (claudin-16) (Simon et al., 1999). Similarly, epithelia (Figure 4). In this respect, it was reported that hypomagnesemia and hypercalcinuria were reported airway epithelia constitutively produced both a ligand, to be induced by mutations in claudin-19 as well as the growth factor heregulin and its receptors, erbB2, in claudin-16 (Konrad et al., 2006). Knockout mice erbB3 and erbB4 (Vermeer et al., 2003). However, in for claudin-1 and -5 exhibit severe newborn lethal cases where TJs were formed, the ligand and receptors phenotypes because of defective water loss from their were segregated and the receptor activation was skin and defects in their blood–brain barrier, respec- silenced. This model predicts that increased paracellular tively (Morita et al., 1999a; Furuse et al., 2002; Nitta permeability as well as complete epithelial disruption et al., 2003). Knockout mice for claudin-11 and -14 were would allow the ligand to access its receptors (Vermeer associated with deafness arising from paracellular et al., 2003). This system has been recognized as a repair barrier-based permselective defects in the inner lymph system for damage to epithelial layers (Chao et al., system (Ben-Yosef et al., 2003; Kitajiri et al., 2004; 2003). Wattenhofer et al., 2005). It is noteworthy that The ionic conditions around epithelial cells could be deficiency of claudin-11 and -19 lead to defects in TJ another factor involved in regulating cell proliferation. formation and defective electrical sealing of glial myelin In this respect, a working model is presented, suggesting sheath cells in the central and peripheral nervous that a change in the intracellular pH regulation could systems, respectively (Gow et al., 1999; Morita et al., alter cell proliferation, as observed in the colonic 1999b; Miyamoto et al., 2005). Taken together with hyperplasia exhibited by knockout mice for the anion other immunostaining data, it is likely that claudin-11 and -19 each form single-species claudin-based barriers without other TJ-constitutive transmembrane proteins. A recent claudin-15-knockout mouse study has provided new in vivo insights for the functions of claudins (Tamura et al., 2008). Claudin-15-deficiency leads to megaintestine and decreases in intestinal epithelial paracellular ion permeability. However, in the claudin-15-deficient megaintestine phenotype, no sign of tumorigenesis was detected, indicating that the normal cell proliferation was simply enhanced by the significantly increased number of transit-amplifying intestinal cells. No evidence was detected that claudin- 15 constituted a platform of any signaling cascades to lead to cell proliferation at TJs. Although clarification of the possibility that claudins provide cues to regulate cell proliferation is still pending, we would like to present the hypothesis that the megaintestine phenotype may be based on defects in TJ-based paracellular permselective barrier functions induced by the deficiency of claudin-15. These findings are consistent with the data obtained in cell culture analyses indicating that Figure 4 Images of the segregation of the apical and lateral claudin-15 increased permeability for sodium ions in microenvironments created by the permselective barrier arising anion-selective epithelial cell sheets (Van Itallie et al., from the extracellular domains of tight junction (TJ) membrane 2003). Taken together, these studies provide possible proteins. (Right half) Receptor-ligand segregation. When the general views of the functions of the paracellular epithelia constitutively produce both a ligand (GF) and its receptors (GFR), the ligand and receptors are segregated and the permselective barriers for cell proliferation, involving receptor activation is silenced when TJs are formed ( Â ’TJ). the role of TJ-based epithelial microenvironment in cell Epithelial disruption would allow the ligand to access its receptors. proliferation. This notion is consistent with the case of (Left half) Ionic conditional segregation. The paracellular permse- claudin-like genes in Drosophila such as Megatrachea lective barrier has an important function in regulating the ionic microenvironment within the lateral one. The ionic balance between and Sinuous, whose mutations reportedly cause ab- the apical and lateral microenvironments should be regulated by normalities in the epithelial tube size control (Behr et al., the selective ionic permeation through TJs (J’TJ) (ZO, zonula 2003; Wu et al., 2004). occludens; belt-like TJ. ZA, zonula adherens; belt-like AJ).

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6935 exchanger AE1/2 (anion exchanger 1/2) (Toivola et al., Table 1 Changes in expression of claudins in cancer cells 2004) and NHE3 (Schultheis et al., 1998). By extending Claudin 1 Breast Down b this line of interpretations, the apical and lateral Cervical Down a microenvironments of ions may be potential contribu- Colorectal Up a tory causes of hyperproliferation, possibly because of Epidermis Up a intercellular pH regulation or possibly due to electric Esophagus Up a Down a cues (Song et al., 2002). Hence, based on the idea that Gastric Up a paracellular permselective barriers have critical func- Hepatocyte Down a tions in regulating the ionic characteristics of the lateral Prostate Up a microenvironment, the ionic balance between the apical Claudin 2 Cervical Down a and lateral microenvironments should be regulated by Colorectal Up a TJs (Figure 4). In addition it is possible that the TJ- based fine-tuning of cell polarization might contribute to Claudin 3 Breast Up a establish ionic microenvironments through the control Colorectal Up a of the lateral distribution of channels/transporters. Esophagus Up a Gastric Up a Although the direct evidence of the critical functions Ovary Up c of the apical and lateral microenvironments in cell Pancreas Up a proliferation has not been accumulated, the megaintes- Prostate Up a tine phenotype is fully suggestive of such an implication. Uterus Up a Some kinds of bacterial infection and inflammation Claudin 4 Biliary tract Up a are closely related to epithelial cell proliferation. When a Bladder Up a bacterial infection, such as H. pylori, is established in Down a epithelia, TJs are disrupted and disappear, leading to Cervical Down a invasion of the bacterial components and/or toxin from Colorectal Up a Down a the apical to the lateral sides of epithelial cells to induce Epidermis Up a inflammation and cell proliferation of epithelial cells Down a (Amieva et al., 2003). In addition, when some inflam- Esophagus Up a mation-inducing substances penetrate through the para- Mesothelioma Down a Gastric Up b cellular TJ-based barriers, inflammation could take Down a place with epithelial cell proliferation. These inflamma- Ovary Up c tion/cell proliferation processes contribute to form the Pancreas Up b bases for tumorigenesis (Macarthur et al., 2004). Prostate Up a Recently, it has been reported that the expression levels Uterus Up a of various types of claudins are significantly upregulated Claudin 5 Pancreas Up a or downregulated in various types of cancers, support- Prostate Down a ing the existence of a relationship between the TJ-based barrier functions and cell proliferation (Kominsky, Claudin 7 Breast Down a Cervical Down a 2006; Gonza´ lez-Mariscal et al., 2007). Colorectal Up a esophagus Up a Down a Gastric Up a Changes in the expression levels of claudins induced Head and neck Down a Ovary Up a by carcinogenesis: implications for direct and/or indirect Uterus Down a roles of claudins in cell proliferation Claudin 10 Liver Up a Recently, it has been reported that the expression levels of many types of claudins are altered in cancer cells Claudin 16 Breast Down a compared with normal cells (Table 1). It appears that Claudin 18 Pancreas Up a the same type of claudin can be be up or downregulated Gastric Down a depending on the type of cancer, and the extent of the Claudin 23 Gastric Down a up or downregulation of the claudin expression levels Occludin Breast Down a vary depending on the cancer concerned (Kominsky, Cervical Down a 2006; Gonza´ lez-Mariscal et al., 2007). For example, Colon Down a claudin-1 and -7 reportedly decreased to almost Epidermis Up a undetectable levels in some cancers from detectable Down a levels in control subjects (Lioni et al., 2007). On the Gastrointestinal Down a Prostate Down a other hand, the expression levels of claudin-1, -2, -3 and Lung Down a -7 were reported to show more than 10-fold differences, and even more than 100-fold differences in some cases, Number of reported references (1997B)1B2 ¼ a, 3B4 ¼ b, 5B ¼ c. in more than 10 types of cancers (Pan et al., 2007). However, the direct or indirect molecular mechanisms

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6936 underlying the changes in the expression levels of claudins are the binding partners of Clostridium perfrin- claudins are still unclear. As a simple example, in the gens enterotoxin (CPE), commonly associated with case of the epithelial–mesenchymal transition, the over- C. perfringens type A, which causes food poisoning all expression levels of claudins were decreased (Ike- effects and is capable of inducing cytolysis of mamma- nouchi et al., 2003). In the presence of HGF and EGF, lian cells in as little as 5 min (Katahira et al., 1997). The some claudins were reportedly upregulated whereas binding affinities of CPE to claudins have been others were downregulated (Singh and Harris, 2004). estimated as claudin-4>-6>-7>-3>-14>-8 with no Furthermore, even when the expression levels did not affinities for claudin-1, -2, -5 and -10 (Fujita et al., change, it was reported that the localizations of claudins 2000). Thus, the cancer cells with increased expression were altered (Lioni et al., 2007). In addition, phosphor- levels of CPE-binding type claudins may be more ylation of claudins was reported to decrease TJ susceptible to the cytolytic effects of CPE. Therefore, functions (D’Souza et al., 2007; Tatum et al., 2007). controlled administration of CPE to cancer cells On the other hand, it is noteworthy that mutations may represent a possible strategy for cancer therapy causing tumorigenesis have not been reported in any of (Kominsky, 2006; Gonza´ lez-Mariscal et al., 2007). the claudin genes, although the possibility of cancer- When the C-terminal domain of CPE, without the dependent mutations of claudins has not been excluded. cytotoxic N-terminal domain is applied to cancer cells, It is also curious that no changes in the expression levels the targeted claudins should undergo endocytosis and of claudin-15 have been reported in any cancers, disappear from the paracellular TJs (Sonoda et al., considering the relationship between knockout silencing 1999). As a result, the barrier functions of TJs may of the claudin-15 gene and the increased cell prolifera- become defective and the cancer cells could then exhibit tion in murine small intestines. These points remain as greater sensitivity owing to increased accessibility for the future subjects for cancer research. antitumor drugs (Kominsky, 2006; Gonza´ lez-Mariscal To examine the changes in the expression levels of et al., 2007). claudins, the specificities of the probes, RT–PCR, A second possible strategy regarding claudin-based primers and antibodies required to detect each claudin molecular targets for cancer therapy relates to the use of are critically problematic. In fact, some commercially antibodies against claudins. Anti-claudin antibody– available antibodies against particular claudins have antigen reactions may induce TJs into more leaky states been revealed to act as pan-claudin antibodies, thereby or even destroy them, which could sometimes be causing confusion in the field. As multiple combinations accompanied by immune reactions based on inflamma- of claudins may be a critical issue for cancer or other tion. Conversely, the permeability of TJs may be kinds of diseases, the development of distinctive and inhibited by the use of antibodies, which in turn could well-established probes for specific claudins is a subject affect cell proliferation. Taken together, the issue of how of the highest priority in these kinds of studies modulation of the microenvironments around cancer (Moriwaki et al., 2007; Tamura et al., 2008). In any cells regulates cell proliferation is an unexplored field case, changes in the claudin expression levels with the potential to drive the development of novel between cancer and normal cells may be directly or cancer therapies. indirectly suggestive of linkages between the functions of claudins and cell proliferation. The implication that cancer cells might organize the microenvironment around them by modulating the constitution of claudins Conclusions represents a newly arisen question that will need to be addressed. Clarification of the molecular mechanisms for TJ-based regulation of the microenvironment around epithelial cells will require further development of the field of Claudins as molecular targets for cancer ‘barriology’, the science of barriers in multicellular organisms as defined by Shoichiro Tsukita. The issue The biased expression of claudins in cancer cells of barriology is critical not only for the development, compared with normal cells suggests the possibility that differentiation and proliferation of multicellular organ- the transmembrane claudins may represent molecular isms but also for cancer and various other permselective targets for cancer therapy mainly from two points barrier-related diseases, which highlights its importance of view. The second extracellular loops of some in the development of various pathologies.

References

Amieva MR, Vogelmann R, Covacci A, Tompkins LS, Nelson WJ, Behr M, Riedel D, Schuh R. (2003). The Claudin-like megatrachea is Falkow S. (2003). Disruption of the epithelial apicaljunctional essential in septate junctions for the epithelial barrier function in complex by Helicobacter pylori CagA. Science 300: 1430–1434. Drosophila. Dev Cell 5: 611–620. Angelow S, Ahlstrom R, Yu AS. (2008). Biology of claudins. Am J Ben-Yosef T, Belyantseva IA, Saunders TL, Hughes ED, Kawamoto K, Physiol Renal Physiol (in press). Van Itallie CM et al. (2003). Claudin 14 knockout mice, a model for

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6937 autosomal recessive deafness DFNB29, are deaf due to cochlear hair Ikenouchi J, Furuse M, Furuse K, Sasaki H, Tsukita Sa, Tsukita Sh. cell degeneration. Hum Mol Genet 12: 2049–2061. (2005). Tricellulin constitutes a novel barrier at tricellular contacts Chao JC, Liu KY, Chen SH, Fang CL, Tsao CW. (2003). Effect of oral of epithelial cells. J Cell Biol 171: 939–945. Epidermal growth factor on mucosal healing in rats with duodenal Itoh M, Furuse M, Morita K, Kubota K, Saitou M, Tsukita Sh. ulcer. World J Gastroenterol 9: 2261–2265. (1999). Direct binding of three tight junction-associated MAGUKs, Claude P. (1978). Morphological factors influencing transepithelial ZO-1, ZO-2, and ZO-3, with the COOH termini of claudins. J Cell permeability: a model for the resistance of the zonula occludens. Biol 147: 1351–1363. J Membr Biol 39: 219–232. Itoh M, Sasaki H, Furuse M, Ozaki H, Kita T, Tsukita Sh. (2001). Cohen CJ, Shieh JT, Pickles RJ, Okegawa T, Hsieh JT, Bergelson JM. Junctional adhesion molecule (JAM) binds to PAR-3: a possible (2001). The coxsackie virus and adenovirus receptor is a transmem- mechanism for the recruitment of PAR-3 to tight junctions. J Cell brane component of the tight junction. Proc Natl Acad Sci USA 98: Biol 154: 491–497. 15191–15196. Katahira J, Inoue N, Horiguchi Y, Matsuda M, Sugimoto N. Colegio OR, Van Itallie C, Rahner C, Anderson JM. (2003). Claudin (1997). Molecular cloning and functional characterization of the extracellular domains determine paracellular charge selectivity and receptor for Clostridium perfringens enterotoxin. J Cell Biol 136: resistance but not tight junction fibril architecture. Am J Physiol 1239–1247. Cell Physiol 284: 1346–1354. Katsuno T, Umeda K, Matsui T, Hata M, Tamura A, Itoh M et al. Diamond JM. (1977). Twenty-first Bowditch lecture. The epithelial (2008). Deficiency of zonula occludens-1 causes embryonic lethal junction: bridge, gate, and fence. Physiologist 20: 10–18. phenotype associated with defected yolk sac angiogenesis and D’Souza T, Indig FE, Morin PJ. (2007). Phosphorylation of claudin-4 apoptosis of embryonic cells. Mol Biol Cell 19: 2465–2475. by PKCepsilon regulates tight junction barrier function in ovarian Kachar B, Reese TS. (1982). Evidence for the lipidic nature of tight cancer cells. Exp Cell Res 313: 3364–3375. junction strands. Nature 296: 464–466. Farquhar MG, Palade GE. (1963). Junctional complexes in various Kitajiri S, Miyamoto T, Mineharu A, Sonoda N, Furuse K, Hata M epithelia. J Cell Biol 17: 375–412. et al. (2004). Compartmentalization established by claudin-11-based Frizzell RA, Schultz SG. (1972). Ionic conductances of extracellular tight junctions in stria vascularis is required for hearing through shunt pathway in rabbit ileum: influence of shunt on transmural generation of endocochlear potential. J Cell Sci 117: 5087–5096. sodium transport and electrical potential differences. J Gen Physiol 59: Kominsky SL. (2006). Claudins: emerging targets for cancer therapy. 318–346. Expert Rev Mol Med 8: 1–11. Fujita K, Katahira J, Horiguchi Y, Sonoda N, Furuse M, Tsukita Sh. Konrad M, Schaller A, Seelow D, Pandey AV, Waldegger S, Lesslauer A (2000). Clostridium perfringens enterotoxin binds to the second et al. (2006). Mutations in the tight-junction gene claudin 19 extracellular loop of claudin-3, a tight junction integral membrane (CLDN19) are associated with renal magnesium wasting, protein. FEBS Lett 476: 258–261. renal failure, and severe ocular involvement. Am J Hum Genet 79: Furuse M, Furuse K, Sasaki H, Tsukita Sh. (2001). Conversion of 949–957. zonulae occludentes from tight to leaky strand type by introducing Kottra G, Haase W, Fro¨ mter E. (1993a). Tight-junction tightness of Necturus gall bladder epithelium is not regulated by cAMP or claudin-2 into Madin-Darby canine kidney I cells. J Cell Biol 153: intracellular Ca2+. I. Microscopic and general electrophysiological 263–272. observations. Pflugers Arch 425: 528–534. Furuse M, Hata M, Furuse K, Yoshida Y, Haratake A, Sugitani Y Kottra G, Fro¨ mter E. (1993b). Tight-junction tightness of et al. (2002). Claudin-based tight junctions are crucial for the Necturus gall bladder epithelium is not regulated by cAMP or mammalian epidermal barrier: a lesson from claudin-1-deficient intracellular Ca2+. II. Impedance measurements. Pflugers Arch mice. J Cell Biol 156: 1099–1111. 425: 535–545. Furuse M, Hirase T, Itoh M, Nagafuchi A, Yonemura S, Tsukita Sa Krause G, Winker L, Mueller SL, Haseloff RF, Piontek J, Blasig IE. et al. (1993). Occludin: a novel integral membrane protein localizing (2008). Structure and function of claudins. Biochi Biophy Acta 1778: at tight junctions. J Cell Biol 123: 1777–1788. 631–645. Furuse M, Sasaki H, Tsukita Sh. (1999). Manner of interaction of Lioni M, Brafford P, Andl C, Rustgi A, El-Deiry W, Herlyn M et al. Heterogeneous claudin species within and between tight junction (2007). Dysregulation of claudin-7 leads to loss of E-cadherin strands. J Cell Biol 147: 891–903. expression and the increased invasion of esophageal squamous cell Gonza´ lez-Mariscal L, Lechuga S, Garay E. (2007). Role of tight carcinoma cells. Am J Pathol 170: 709–721. junctions in cell proliferation and cancer. Prog Histochem Cytochem Liu Y, Nursat A, Schnell FJ, Reaves TA, Walsh S, Pochet M et al. 42: 1–57. (2000). Human junction adhesion molecule regulates tight junction Gow A, Southwood CM, Li JS, Pariali M, Riordan GP, Brodie SE resealing in epithelia. J Cell Sci 113: 2363–2374. et al. (1999). CNS myelin and Sertoli cell tight junction strands are Loh YH, Christoffels A, Brenner S, Hunziker W, Venkatesh B. (2004). absent in Osp/Claudin-11 null mice. Cell 99: 649–659. Extensive expansion of the claudin gene family in the teleost fish, Gumbiner BM. (1993). Breaking through the tight junction barrier. Fugu rubripes. Genome Res 14: 1248–1257. J Cell Biol 123: 1631–1633. Macarthur M, Hold GL, El-Omar EM. (2004). Inflammation and Hamazaki Y, Itoh M, Sasaki H, Furuse M, Tsukita Sh. (2002). Multi- Cancer II. Role of chronic inflammation and cytokine gene PDZ domain protein 1 (MUPP1) is concentrated at tight junctions polymorphisms in the pathogenesis of gastrointestinal malignancy. through its possible interaction with claudin-1 and junctional Am J Physiol Gastrointest Liver Physiol 286: G515–G520. adhesion molecule. J Biol Chem 277: 455–461. Matsuda M, Kubo A, Furuse M, Tsukita Sh. (2004). A peculiar Hartsock A, Nelson WJ. (2008). Adherens and tight junctions: internalization of claudins, tight junction-specific adhesion mole- structure, function and connections to the actin cytoskeleton. cules, during the intercellular movement of epithelial cells. J Cell Sci Biochim Biophys Acta 1778: 660–669. 117: 1247–1257. Hopfer U. (1977). Isolated membrane vesicles as tools for analysis of Mitic LL, Anderson JM. (1998). Molecular architecture of tight epithelial transport. Am J Physiol 233: E445–E449. junctions. Annu Rev Physiol 60: 121–142. Hou J, Gomes AS, Paul DL, Goodenough DA. (2006). Study Miyamoto T, Morita K, Takemoto D, Takeuchi K, Kitano Y, of claudin function by RNA interference. J Biol Chem 281: Miyakawa T et al. (2005). Tight junctions in Schwann cells of 36117–36123. peripheral myelinated axons: a lesson from claudin-19-deficient Ikenouchi J, Matsuda M, Furuse M, Tsukita Sh. (2003). Regulation of mice. J Cell Biol 169: 527–538. tight junctions during the epithelium-mesenchyme transition: direct Morita K, Sasaki H, Furuse M, Tsukita Sh. (1999a). Endothelial repression of the gene expression of claudins/occludin by Snail. claudin: claudin-5/TMVCF constitutes tight junction strands in J Cell Sci 116: 1959–1967. endothelial cells. J Cell Biol 147: 185–194.

Oncogene TJ-based epithelial microenvironment and cell proliferation S Tsukita et al 6938 Morita K, Sasaki H, Fujimoto K, Furuse M, Tsukita Sh. (1999b). Tamura A, Kitano Y, Hata M, Katsuno T, Moriwaki K, Sasaki H Claudin-11/OSP-based tight junctions of myelin sheaths in brain et al. (2008). Megaintestinein claudin-15-deficient mice. Gastro- and Sertoli cells in testis. J Cell Biol 145: 579–588. enterology 134: 523–534. Morita K, Furuse M, Yoshida Y, Itoh M, Sasaki H, Tsukita Sh et al. Tatum R, Zhang Y, Lu Q, Kim K, Jeansonne BG, Chen YH. (2007). (2002). Molecular architecture of tight junctions of periderm differs WNK4 phosphorylates ser(206) of claudin-7 and promotes para- from that of the maculae occludentes of epidermis. J Invest cellular Cl(-) permeability. FEBS Lett 581: 3887–3891. Dermatol 118: 1073–1079. Toivola DM, Krishnan S, Binder HJ, Singh SK, Omary MB. (2004). Moriwaki K, Tsukita S, Furuse M. (2007). Tight junctions containing Keratins modulate colonocyte electrolyte transport via protein claudin 4 and 6 are essential for blastocyst formation in mistargeting. J Cell Biol 164: 911–921. preimplantation mouse embryos. Dev Biol 312: 509–522. Tsukita Sh, Tsukita Sa. (1989). Isolation of cell-to-cell adherens Nitta T, Hata M, Gotoh S, Seo Y, Sasaki H, Hashimoto N et al. junctions from rat liver. J Cell Biol 108: 31–41. (2003). Size-selective loosening of the blood-brain barrier in claudin- Tsukita Sh, Furuse M. (1999). Occludin and claudins in tight-junction 5-deficient mice. J Cell Biol 161: 653–660. strands: leading or supporting players? Trends Cell Biol 9: 268–273. Okada Y, Irimajiri A, Inouye A. (1977). Electrical properties and Tsukita Sh, Furuse M. (2000). Pores in the wall: claudins constitute tight active Solute transport in rat intestine. II. Conductive properties of junction strands containing aqueous pores. JCellBiol149: 13–16. transepithelial routes. J Membr Biol 31: 221–232. Tsukita Sh, Furuse M, Itoh M. (2001). Multifunctional strands in tight Pan XY, Wang B, Che YC, Weng ZP, Dai HY, Peng W. (2007). junctions. Nat Rev Mol Cell Biol 2: 285–293. Expression of claudin-3 and claudin-4 in normal, hyperplastic, and Turk E, Zabel B, Mundlos S, Dyer J, Wright EM. (1991). Glucose endometrial tissue. Int J Gynecol Cancer 17: 233–241. galactose malabsorption caused by a defect in the Na+/glucose Pinto da Silva P, Kachar B. (1982). On tight junction structure. cotransporter. Nature 350: 354–356. Cell 28: 441–450. Umeda K, Ikenouchi J, Katahira-Tayama S, Furuse K, Sasaki H, Powell DW. (1981). Barrier function of epithelia. Am J Physiol 241: Nakayama M et al. (2006). ZO-1 and ZO-2 independently determine 275–288. where claudins are polymerized in tight-junction strand formation. Reuss L. (1992). Tight junction permeability to ions and water. In: Cell 126: 741–754. Cereijido M (ed). Tight Junctions. CRC Press: London, 49–66. Van Itallie CM, Fanning AS, Anderson JM. (2003). Reversal of Saitou M, Furuse M, Sasaki H, Schulzke JD, Fromm M, charge selectivity in cation or anion-selective epithelial lines by Takano H et al. (2000). Complex phenotype of mice lacking expression of different claudins. Am J Physiol Renal Physiol 285: 1078–1084. occludin, a component of tight junction strands. Mol Biol Cell 11: Van Itallie CM, Anderson JM. (2006a). Claudins and epithelial 4131–4142. paracellular transport. Annu Rev Physiol 68: 403–429. Sa´ nchez-Pulido L, Martı´ n-Belmonte F, Valencia A, Alonso MA. Van Itallie CM, Rogan S, Yu A, Vidal LS, Holmes J, Anderson JM. (2002). MARVEL: a conserved domain involved in membrane (2006b). Two splice variants of claudin-10 in the kidney create apposition events. Trends Biochem Sci 27: 599–601. paracellular pores with different ion selectivities. Am J Physiol Renal Sasaki H, Matsui C, Furuse K, Mimori-Kiyosue Y, Furuse M, Tsukita Physiol 291: 1288–1299. Sh. (2003). Dynamic behavior of paired claudin strands Van Itallie CM, Holmes J, Bridges A, Gookin JL, Coccaro MR, within apposing plasma membranes. Proc Natl Acad Sci USA Proctor W et al. (2008). The density of small tight junction pores 100: 3971–3976. varies among cell types and is increased by expression of claudin-2. Schmidt A, Utepbergenov DI, Mueller SL, Beyermann M, Schneider- J. Cell Sci 121: 298–305. Mergener J, Krause G et al. (2004). Occludin binds to the SH3- Verkleij AJ. (1984). Lipidic intramembranous particles. Biochim hinge-GuK unit of zonula occludens protein 1: potential mechanism Biophys Acta 779: 43–63. of tight junction regulation. Cell Mol Life Sci 61: 1354–1365. Vermeer PD, Einwalter LA, Moninger TO, Rokhlina T, Kern JA, Schultheis PJ, Clarke LL, Meneton P, Miller ML, Soleimani M, Zabner J et al. (2003). Segregation of receptor and ligand Gawenis LR et al. (1998). Renal and intestinal absorptive defects regulates activation of epithelial growth factor receptor. Nature in mice lacking the NHE3 Na+/H+ exchanger. Nat Genet 19: 422: 322–326. 282–285. Xu J, Kausalya PJ, Phua DC, Ali SM, Hossain Z, Hunziker W. (2008). Simon DB, Lu Y, Choate KA, Velazquez H, Al-Sabban E, Praga M Early embryonic lethality of mice lacking ZO-2, but not ZO-3, et al. (1999). Paracellin-1, a renal tight junction protein required for reveals critical and nonredundant roles for individual zonula paracellular Mg2+ resorption. Science 285: 103–106. occludens proteins in mammalian development. Mol Cell Biol 28: Singh AB, Harris RC. (2004). Epidermal growth factor receptor 1669–1678. activation differentially regulates claudin expression and enhances Yap AS, Mullin JM, Stevenson BR. (1998). Molecular analyses of transepithelial resistance in Madin-Darby canine kidney cells. J Biol tight junction physiology: insights and paradoxes. J Membr Biol Chem 279: 3543–3552. 163: 159–167. Song B, Zhao M, Forrester JV, McCaig CD. (2002). Electrical cues Watson CJ, Rowland M, Warhurst G. (2001). Functional modeling of regulate the orientation and frequency of cell division and the rate tight junctions in intestinal cell monolayers using polyethylene of wound healing in vivo. Proc Natl Acad Sci USA 99: 13577–13582. glycol oligomers. Am J Physiol Cell Physiol 281: 388–397. Sonoda N, Furuse M, Sasaki H, Yonemura S, Katahira J, Horiguchi Wattenhofer M, Reymond A, Falciola V, Charollais A, Caille D, Borel Y et al. (1999). Clostridium perfringens enterotoxin fragment C et al. (2005). Different mechanisms preclude mutant CLDN14 removes specific claudins from tight junction strands: evidence for proteins from forming tight junctions in vitro. Human Mutat 25: direct involvement of claudins in tight junction barrier. J Cell Biol 543–544. 147: 195–204. Wu VM, Schulte J, Hirschi A, Tepass U, Beitel GJ. (2004). Sinuous is a Staehelin LA. (1973). Further observation on the fine structure of Drosophila claudin required for septate junction organization and freeze-cleaved tight junctions. J Cell Sci 13: 763–786. epithelial tube size control. J Cell Biol 164: 313–323.

Oncogene