Introduction

Chronic inflammatory diseases of the intestine are increasingly frequent within the last three decades. They can affect the small and the large intestine. The most common diseases are chronic inflammatory bowel diseases (IBD) like Crohn’s disease and ulcerative colitis, as well as microscopic colitis that consists of two subtypes, lymphocytic colitis and collagenous colitis. The most common chronic inflammatory disease of the small intestine is celiac disease. A key feature of all of them is that they impair the epithelial barrier.

The epithelial cell layer as the first “line of defense” defines the epithelial barrier. It is constructed in such a way that uncontrolled passage of antigens, solutes, and water is prevented. Its barrier function consists of two parts, a transcellular and a paracellular one. The morphological correlate of the paracellular part is the tight junction (TJ), a complex meshwork of various strand-like proteins anchored in the apicolateral membrane of epithelial cells that spans across the paracellular space. Its function is to “seal” the paracellular space between adjacent epithelial cells against the outside environment to prevent the uncontrolled paracellular passage of luminal contents and to prevent the loss of solutes and water into the lumen.

The TJ components determining the barrier function are tetraspan transmembrane proteins of two families: the large family of claudins and the smaller family of tight junction-associated Marvel proteins (TAMPs) occludin and marvelD3 as integral components of bicellular TJs [76, 90] and tricellulin, which predominantly localizes at tricellular TJs [42].

In addition, peripheral proteins belong as TJ-associated proteins to the TJ and participate as scaffold proteins in its function like zonula occludens (ZO)-1, ZO-2, and ZO-3 and cingulin. These are connected to the cytoskeleton via F-actin and myosin II [23, 32]. Among these TJ proteins, claudins are the major transmembrane components. They determine the tightness of the epithelial barrier and contribute to the polarity of epithelial cells.

In general, claudins exhibit three different functions: (i) a “gate function” through transepithelial barrier or channel formation; (ii) a “signal function” by which they affect cellular signaling, proliferation, motility, and migration as well as differentiation and receptor function; and (iii) a “fence function” that prevents intermixing of apical and basolateral membrane proteins. All three functions are regulated by stimuli from the cell interior and from the outside.

Chronic inflammatory diseases of the intestine may lead to the dysregulation of TJ proteins that in turn have two distinct effects on barrier function: (i) increased paracellular transport of solutes and water by which ions and water diffuse from the blood to the intestinal lumen and cause leak-flux diarrhea [81] and (ii) increased permeability to large molecules, as e.g., luminal pathogens that induce an immune response and maintain the inflammatory process [2, 16].

This review summarizes our current knowledge about TJ composition and function in chronic immune-mediated intestinal diseases from a clinical perspective, first by looking at the physiological expression and function of intestinal claudins and secondly by discussing their influence and behavior in chronic intestinal diseases, with an active role when the change in a claudin has functional consequences and with a passive role if it is only coincidently affected without functional consequence.

The barrier breakdown occurring in inflammatory diseases is caused for the most part by alterations of claudins and TAMPs. As this gives rise to increased uptake of antigens from the lumen, the tight junction-mediated uptake fires the inflammatory process. In turn, this inflammation releases cytokines which then alters tight junction protein expression and localization. This vicious circle has birthed the idea to focus in our review on the frequently raised question, whether the change in tight junction proteins is actively and passively involved in the chronically inflamed intestine.

Physiology of intestinal claudins—expression, localization, and function

To understand TJ alterations and their implications in chronic intestinal diseases, the physiologic function of claudins have to be elucidated. From their primary function as “gate keeper,” intestinal claudins can be classified as barrier- or channel-forming. In the intestine, the latter is selective paracellular channels, which also determine water permeability. Barrier-forming claudins are subclassified as non-charge- and charge-selective [50]. Charge-selective claudins favor a privileged passage of specifically charged ions by inhibiting passage of the counter-ion. The following enumeration presents the functionally relevant intestinal claudins according to their gate function, and Table 1 summarizes claudin functions and expression patterns in the different intestinal inflammatory diseases.

Table 1 Function and expression changes of claudins in intestinal inflammatory diseases
Full size table

Claudins with barrier function

No charge selectivity

Claudin-1

Claudin-1 and claudin-2 were the first claudins to have been isolated [28]. Claudin-1 is found in many tissues including the intestine [15, 28, 51, 110]. In human sigmoid colon, claudin-1 is detected in the TJ and the subjunctional lateral membrane of surface and crypt enterocytes [110]. The function of claudin-1 is barrier forming, since its overexpression induces an increase of the transepithelial resistance (TER) in cell culture studies [33, 43, 55]. The consequences of a loss of claudin-1 have in particular been shown for barrier function in the epidermis. Homozygous claudin-1 knockout mice die within 1 day after birth and exhibit severe transepidermal water loss and macromolecular diffusion through the epidermal TJ that permits the conclusion that claudin-1 is essential for barrier formation [30]. Other skin diseases that are associated with a downregulation of claudin-1 are atopic dermatitis and psoriasis [19, 34, 103]. Also, claudin-1 is expressed in cholangiocytes and hepatocytes and has been shown to cause neonatal sclerosing cholangitis associated with ichthyosis due to loss-of-function mutations in the claudin-1 gene [33, 35]. The physiologic regulation of claudin-1 within intestinal epithelial cells is largely unknown. Just recently, Saeedi and coworkers demonstrated that hypoxia-inducible factor (HIF) is a mediator that can regulate claudin-1 via prominent hypoxia response elements in the claudin-1 promoter region. Overexpression of claudin-1 in HIF1β-deficient intestinal epithelial cells (IECs) that have lost barrier properties restored exactly these properties [80]. In intestinal inflammation, proinflammatory mediators can cause a dysregulation of claudin-1. Amasheh et al. found upregulation and redistribution of claudin-1 away from the TJ in the presence of TNFα [3], whereas others demonstrated increased paracellular permeability and downregulation of claudin-1 in conjunction with increased levels of TNFα and NFκB [24, 111]. In addition, Bruewer et al. showed that in the presence of IFNγ and/or TNFα, a substantial internalization of claudin-1 occurs after 48 h in T84 cells [14]. Diseases of the intestine that exhibit a dysregulation of claudin-1 are IBD [51, 104, 110], irritable bowel syndrome [12, 72], and HIV enteropathy [24], but not celiac disease or microscopic colitis [11, 15, 86].

Claudin-3

Claudin-3 is expressed in many different epithelia including the gastrointestinal tract [15, 75, 110]. In rat intestine, claudin-3 is expressed ubiquitously throughout the jejunum, ileum, and colon. Its expression is tight junctional, but in the colon, it shows also a lateral membrane localization, both in crypts and at the surface [75]. In human colon, Prasad et al., as well as Schumann et al., localized claudin-3 to the tight junction and lateral cell membrane of crypt and surface epithelial cells [74, 86]. Several studies showed that claudin-3 is barrier-forming. Overexpression of claudin-3 in MDCK II cells led to an increase in TER that was due to a significant increase in paracellular resistance as determined by two-path impedance spectroscopy and reduced permeability for ions of either charge as well as larger molecules like fluorescein (332 Da) and fluorescein isothiocyanate (FITC)-dextran (4 kDa) [60]. When stably expressed in human airway cells (IB3.1), claudin-3 reduced paracellular permeability of dextrans of different sizes (10, 70, 2000 Da) and induced a small increase in resistance across the cell layer that did not reach statistical significance [18]. Further evidence for its barrier-forming properties came from inhibition experiments. siRNA knockdown of claudin-3 in MKN28 gastric epithelial cells resulted in attenuated barrier function similar to ochratoxin A- and TRPV4-induced claudin-3 downregulation in the human colon carcinoma cell line Caco-2 and HC11 mammary cells, respectively [36, 56, 77]. Although these studies clearly showed a barrier-forming effect for claudin-3, conflicting data from alveolar epithelial cells exist in which overexpression of claudin-3 appears to be associated with decreased TER and increased permeability for paracellular markers [62] which could be due to a different TJ protein background within the respective cell model. Information about the physiologic regulation of claudin-3 is limited. Just recently, Dong et al. demonstrated intestinal claudin-3 upregulation in mice to be dependent on activation of intestinal epithelial IGF-1 receptor through glucagon-like peptide-2 (GLP-2). GLP-2 treatment increased epithelial resistance in the jejunum of mice and decreased permeability of 4-kDa FITC-dextran [22]. In addition, zinc depletion reduces TER and claudin-3 expression on a transcriptional level in Caco-2 cells [63].

Some but not all intestinal diseases alter claudin-3 expression and localization. Dysregulation occurs in IBD [31, 74, 95] and celiac disease [86], but not in microscopic colitis [15] or irritable bowel syndrome [49].

Claudin-5

Claudin-5 is expressed in various tissues [65]. It is considered primarily as a barrier-forming claudin that is associated with endothelial TJs and plays a critical role in maintaining the blood-brain barrier [45, 66, 68]. However, claudin-5 localizes to epithelial cells as well, especially to the alveolar epithelium of the lung [65, 100] and to small and large intestine in rat [75] and man [15, 110]. Here, it is strictly located in the TJ of crypts and surface epithelial cells without any gradient along the crypt-to-surface axis [110]. Several studies confirm the barrier properties of claudin-5 in epithelial and endothelial cell models. Stable transfection of a FLAG-tagged claudin-5 into Caco-2 cells increased TER [6]. In cultured rat brain capillary endothelial cells, transfection of claudin-5 reduced the permeability of 14C-inulin [69] and overexpression of claudin-5 in cultured human retinal pigment epithelium cells enhanced barrier function [96]. In addition, the death of homozygous claudin-5-knockout mice within 10 h after birth emphasizes the importance of claudin-5 for barrier integrity [68]. Experiments in endothelial cells delivered most of the information available about claudin-5 regulation. Stamatovic et al. found a lipid-raft-dependent endocytosis mechanism of internalization of claudin-5 into early and recycling endosomes that involves caveolin-1 after addition of the chemokine CCL2 [89]. Other factors that affect claudin-5 (dys)regulation and function are cyclic AMP (cAMP) and protein kinase A [44], the erythroblast transformation-specific (ETS)-related gene [109], as well as sodium caprate, β1-integrin, and bradykinin [20, 53, 70]. However, only little is known about claudin-5 regulation in intestinal epithelial cells. Recently, Watari et al. demonstrated increased TER due to increased claudin-5 expression in Caco-2 cells after checkpoint kinase 1 activation through incubation with daunorubicin and rebeccamycin [102]. Diseases of the gastrointestinal tract in which claudin-5 is dysregulated are Crohn’s disease [110] and celiac disease [86] as well as Campylobacter concisus enteritis [67].

Charge selectivity

Claudin-4

Expression of claudin-4 is not as ubiquitous as the expression of some other members of the claudin family. mRNA and protein expression of claudin-4 has been reported in lung and kidney [65], in rat gastrointestinal tract [75] including enteric neurons [46], and in human colon [15, 110]. In the human colon, claudin-4 is predominantly located in the lateral membrane with additional junctional and cytoplasmic staining in crypt and surface enterocytes [110]. Regarding its function, different effects have been observed depending on the cell model. It can act either as a regular barrier or as a sodium barrier without affecting chloride permeability and in conjunction with claudin-8 as an anion channel [41, 59, 98, 99]. In MDCK II cells, overexpression of claudin-4 increased barrier properties with an increase in TER and a decreased sodium permeability [98]. In mouse collecting duct cells, claudin-4 acts as an anion channel that facilitates NaCl-reabsorption and requires the presence of claudin-8 for correct assembly in the TJ [41]. siRNA knockdown of claudin-8 leads to a delocalization of claudin-4 to the cytoplasm with a consecutive loss of paracellular chloride reabsorption [41]. Hou et al. also demonstrated anion channel properties in mouse collecting duct cells. Knockdown of claudin-4 resulted in decreased TER due to an increase in ion permeability, while a preference change for paracellular reabsorption from anions to cations occurred [40]. In the lung, claudin-4 acts as a sodium barrier to prevent fluid loss into the alveolar space. When claudin-4 is absent, air space fluid clearance was inhibited and ventilator-induced pulmonary edema was exacerbated [107]. In contrast, increased claudin-4 expression levels resulted in better alveolar fluid clearance [78]. In the colon, claudin-4 appears to promote barrier effects based on cell culture studies. Induction of claudin-4 expression in the human colon carcinoma cell line HT-29/B6 through TGF-β-containing whey protein induced a TER increase due to increased claudin-4 promoter activity [38]. In contrast, a combination of TNFα/IFNγ induced a decrease in TER with a concomitant redistribution of claudin-4 from the TJ but no changes in expression in confluent T84 cells [74]. Claudin-4 expression is downregulated in ulcerative colitis [71], collagenous colitis [15], and irritable bowel syndrome [49]. No changes have been reported in Crohn’s disease [31, 110], celiac disease [86], HIV enteropathy [24], and campylobacter enteritis [67].

Claudin-8

Claudin-8 is expressed in lung, liver, skeletal muscle, kidney, testis, and intestine [48, 52, 65, 110]. Expression in the kidney followed a heterogeneous pattern along the aldosterone-responsive parts of the renal tubule, namely, in the entire distal nephron and the late segments of the thin descending limb of Henle [52]. In mouse intestine, claudin-8 is expressed in the more distal parts of the intestine with increasing expression levels from the ileum to the colon [27]. In human intestine, segmental expression differences were not examined so far, but expression is confirmed in sigmoid colon [110]. In HT-29/B6 cells, aldosterone induced claudin-8 upregulation, increased paracellular resistance, and induced electrogenic sodium transport via the epithelial sodium channel (ENaC). The resulting tightened barrier has been proposed to prevent back-leak of sodium into the lumen [5]. This is in accordance with data from Yu et al. who demonstrated in MDCK cells that claudin-8 acts as a sodium barrier [108]. Others showed that claudin-8 recruits claudin-4 to form a TJ-based complex that seals cation passage but mediates anion transport in collecting duct cells [41]. Not much is known about claudin-8 in intestinal diseases. In only two intestinal diseases, claudin-8 was examined. Whereas in Crohn’s disease claudin-8 expression was reduced [110], claudin-8 was unchanged in C. concisus infection [67].

Intestinal claudins forming paracellular cation and water channels

Claudins forming paracellular channels exhibit selectivities for small cations, small anions, or water. However, anion-selective claudins (e.g., claudin-10a and claudin-17) are not abundant in the intestine. The main channel formers of the gut epithelium are claudin-2 and claudin-15. Because other papers in this special issue deal with these two claudins more extensively, we provide only a short roundup here.

Claudin-2

Claudin-2 is constitutively expressed in the most leaky epithelia, e.g., the proximal nephron [48] and the small intestine [75]. In tight epithelia, it is usually not detectable, but in human colon, it is expressed to low amounts in the crypts. Of importance, it can be upregulated in inflammatory conditions like Crohn’s disease, ulcerative colitis, celiac disease, or HIV enteropathy [10]. Claudin-2 forms a paracellular channel containing a common pore for small cations and for water [4, 29, 79, 88].

Only recently, it was discussed that claudin-2 could also have a protective effect on the onset and severity of colitis, e.g., by rinsing off antigens from the apical surface due to a passive leak flux. In line with such a type of explanation, Ahmad et al. demonstrated in villin-claudin-2 transgenic mice that increased expression of claudin-2 augmented colonocyte proliferation and provided protection against colitis-induced colonocyte death in a PI-3Kinase/Bcl-2-dependent manner [1]. However, also a marked suppression of colitis-induced increase in immune activation and associated signaling was observed, suggesting in addition interference with immune tolerance [1].

Claudin-15

Claudin-15 is widely distributed in small and large intestine in both the TJ of villus and crypt cells. Knockout of claudin-15 in mice showed that claudin-15 might act as a paracellular cation channel that facilitates sodium secretion. Loss of claudin-15 results in low intraluminal sodium concentrations that in turn hampers sodium-coupled glucose absorption via SGLT-1 [93].

Interestingly, claudin-15 knockout mice develop congenital enlargement of the small intestine [94]. The exact mechanisms leading to this megaintestine are not known. Data from zebrafish also suggest a role for claudin-15 in congenital development of the small intestine. Here, downregulation of claudin-15 in animals with a loss-of-function mutation in the transcription factor tcf2 resulted in the development of multiple intestinal lumens, a phenotype that was also seen in claudin-15 morphants supporting the central role of claudin-15 in intestine development [7].

Data on claudin-15 expression and localization in human intestinal inflammation are very limited. Only Schumann et al. examined claudin-15 in celiac disease and found an upregulation of protein expression in conjunction with a shift from the membrane into the cytoplasm [86].

Inflammatory diseases of the intestine—expression, localization, and function of intestinal claudins

Ulcerative colitis

Ulcerative colitis belongs, like Crohn’s disease, lymphocytic, and collagenous colitis, to the group of IBDs. Characteristic feature of ulcerative colitis is a chronic inflammation that is, apart from a rare “backwash-ileitis,” restricted to the colon. The inflammation spreads out from the distal to proximal colon segments and always includes the rectum. During disease activity, patients suffer from frequent, often bloody diarrhea and abdominal pain. The disease affects usually adults in their third and fourth decade of life but a second peak occurs during the seventh and eighth life decade. Therapeutic options include various immunosuppressive agents, alone or in combination, and surgery. The therapeutic goal is to reach a long-term remission for the patient.

Many studies exist that examined the pathologic features of ulcerative colitis. Barrier function was very early assumed to be impaired. Sandle and coworkers found a decrease of mucosal resistance in ulcerative colitis to approximately 40% [82]. This finding based upon the total resistance of the tissue. Our group used one-path impedance spectroscopy to discriminate epithelial and subepithelial resistance and found an even more pronounced decrease of the epithelial resistance of approximately 85% and a concomitant increase in subepithelial resistance [83]. This was due to a reduced depth of the TJ meshwork with a reduced number of horizontally oriented TJ strands and increased thickness of the subepithelial layers due to infiltration with immune cells and edema [83]. With the discovery of claudins, the question about their role in TJ alterations in IBD arose.

The first claudin examined in ulcerative colitis was claudin-2 [37]. Heller et al. observed that claudin-2 was not detectable in the sigmoid colon of healthy subjects but was highly upregulated in the inflamed sigmoid colon of ulcerative colitis patients which may be an active event already in early disease stages causing leak flux diarrhea. They demonstrated that upregulation of claudin-2 depends on IL-13. IL-13 is a proinflammatory cytokine that is elevated in active ulcerative colitis as part of the immune response. In addition to claudin-2 upregulation, IL-13 provoked an increase in the epithelial apoptotic rate and an epithelial restitution arrest [37]. This might provide an explanation for the early occurrence of ulcer lesions in ulcerative colitis [39]. Other claudins examined in ulcerative colitis are claudin-1, claudin-3, claudin-4, and claudin-7. Data on claudin-1 are varying. Whereas Oshima et al. did not observe changes in claudin-1 expression in rectal biopsies of patients with active inflammation [71], Poritz and coworkers found increased claudin-1 expression in the inflamed colon in ulcerative colitis [73]. Claudin-1 expression appears to be dependent on the grade of inflammation. Weber et al. found claudin-1 also increased in severely inflamed but not in slightly inflamed colon of ulcerative colitis patients. Interestingly, they observed claudin-1 and claudin-2 expression not to be increased in acute, self-limited colitis but in active ulcerative colitis, in adenomas, and in IBD-associated dysplasia [104]. They speculated that increased claudin-1 and claudin-2 expressions may be involved in neoplastic developments in ulcerative colitis [104], a hypothesis that is supported by others, too [47, 57].

Data on claudin-3 in ulcerative colitis are limited. Claudin-3 expression has been shown to be unaffected [71] or reduced in active ulcerative colitis [74]. Others found increased claudin-3 urinary concentration in active ulcerative colitis and suggested to use urinary claudin-3 concentration as a non-invasive marker of intestinal tight junction deprivation [95]. Data on the expression of claudin-4 were also divergent. In rectal biopsies, Weber et al. demonstrated elevated claudin-4 in severe active ulcerative colitis [104], whereas Oshima et al. found the opposite on protein and mRNA level in active inflammation [71]. Claudin-7, which in human sigmoid colon like claudin-4 is predominantly located in the lateral membrane and in the TJ and the cytoplasm of crypt and surface epithelial cells, was downregulated in active ulcerative colitis on mRNA and protein level [71].

Just recently, claudins were examined as targets for alternative therapeutic approaches. Stio et al. observed that 1,25(OH)2D3, the active form of vitamin D, reversed the increase of claudin-1 and claudin-2 and the decrease of claudin-4 and claudin-7 protein expressions in active ulcerative colitis and proposed vitamin D to be a potential therapeutic agent for the treatment of ulcerative colitis [91]. Devriese et al. reported that activity of Butyricicoccus pullicaecorum, a butyrate-producing bacterium reduced in the stool of patients with active ulcerative colitis, correlates inversely with claudin-1 mRNA-expression and proposed it as a pharmacobiotic in ulcerative colitis [21].

Crohn’s disease

Crohn’s disease differs from ulcerative colitis through dissemination across the entire intestinal tract, although terminal ileum and/or the colon are more abundantly affected. Inflammation is segmentally distributed and can affect more than one area of the gastrointestinal tract at the same time. Often, it does not affect the rectum. Patients suffer from abdominal pain in stenotic disease and mostly from non-bloody diarrhea. The nature of pain and diarrhea depends on the gastrointestinal segments involved. Thus, symptoms in Crohn’s disease are more diverse than in ulcerative colitis. Crohn’s disease tends to start in young people but can occur at any age. The immune-suppressive therapeutic approach is often similar to that for ulcerative colitis.

Zeissig et al. performed the most comprehensive study on claudin expression and localization in Crohn’s disease [110]. The epithelial resistance in mild to moderately inflamed sigmoid colon of Crohn’s disease patients was approximately 60% of that of controls according to one-path impedance spectroscopy analysis suggesting TJ alterations. In fact, tight junction morphology as determined by freeze-fracture electron microscopy showed severe disturbances with reduced and discontinuous TJ strands, suggesting altered claudin expression and/or subcellular localization. They examined claudin-1, claudin-2, claudin-3, claudin-4, claudin-5, claudin-7, claudin-8, claudin-11, claudin-12, claudin-14, claudin-15, and claudin-16 and occludin and found claudin-5 and claudin-8 downregulated and redistributed off the TJ as well as upregulation of claudin-2 in the inflamed colon. Especially, claudin-5 deficiency has to be assumed to play an active role in antigen entry into the mucosa.

In Fig. 1, redistribution of claudins off the TJ is exemplarily shown for claudin-8. The upregulation of claudin-2 is triggered by TNFα but was not as strong as in ulcerative colitis, where it was threefold higher. Claudin-2 upregulation was located in the crypt base and was even higher in patients with lower claudin-8 expression, which is consistent with results from cell culture experiments demonstrating that claudin-2 expression is inversely related to claudin-8 expression and that it can be replaced by claudin-8 [108]. In this manner, claudin-2 and claudin-8 have to be assumed to play a role for the diarrhea (leak flux mechanism). Claudin-1, claudin-4, and claudin-7 were unchanged regarding protein expression and localization within the TJ, whereas protein expression of claudin-11, claudin-12, claudin-14, claudin-15, and claudin-16 was not detectable in sigmoid colon [110]. Others found similar results. Claudin-1 expression for example was unchanged in inflamed duodenum and colon [51, 73] and was proposed as a surrogate marker to differentiate between Crohn’s disease and ulcerative colitis [73].

Fig. 1
figure 1

Distribution and localization of claudin-8 in the sigmoid colon of patients with Crohn’s disease. Claudin-8 shifted from an exclusive TJ localization to the apical cytoplasm in Crohn’s disease, where it stained weakly, primarily not co-localizing with ZO-1. Bar  =  20 μm (reprinted and adapted from [110] with permission from BMJ Publishing Group Ltd. and British Society of Gastroenterology)

Full size image

A recent study by Goswami et al. performed in Crohn’s disease with active duodenal inflammation confirmed that claudin-2 is elevated, whereas claudin-3 and claudin-4 are downregulated and correlated these observations with simultaneous alterations in the TJ ultrastructure [31]. In contrast, treated Crohn patients and patients in remission exhibited no significant changes in epithelial resistance and claudin expression indicating the importance of claudins for the normal function of the epithelial barrier [31, 110]. Proinflammatory cytokines can disturb the integrity of the epithelial barrier. In contrast to ulcerative colitis, a predominant Th17 immune response is observed in Crohn’s disease. High TNFα and IFNγ serum levels are evident and possess the ability to decrease TER and to increase claudin-2 expression as demonstrated in human epithelial HT-29/B6 and Caco-2 cells [26, 110].

Glucocorticoids can improve intestinal barrier function, an effect which can also be studied in cell models. Although the TNFα- and IFNγ-induced TER decline in Caco-2 cells was similar, TER was still higher in glucocorticoid-treated than in non-treated cells because treatment with dexamethasone before TNFα and IFNγ exposure increased TER. This effect was accompanied by decreased claudin-2 and increased claudin-4 protein expression in the glucocorticoid-treated Caco-2 cells and was mediated through activation of MAPK phosphatase-1 [26]. A similar effect is seen with the anti-TNFα agent adalimumab. In the presence of TNFα and IFNγ, adalimumab prevented TER decrease and downregulation/redistribution of claudin-1, claudin-2, and claudin-4 as well as occludin in T84 cells and inhibited the formation of irregular membrane structures [25]. Thus, both glucocorticoids and anti-TNFα agents augment and protect mucosal barrier integrity and function and are involved in the rearrangement of the TJ, inducing “mucosal healing.”

Microscopic colitis

Microscopic colitis consists of two subtypes, lymphocytic and collagenous colitis. The endoscopic appearance is usually normal and histological analysis of endoscopically obtained biopsies is required to establish the diagnosis. Increased numbers of intraepithelial lymphocytes (more than 20 per 100 enterocytes) without mucosal gross lesions are a characteristic feature of both subtypes. However, in contrast to lymphocytic colitis, the histological picture in collagenous colitis offers additional thickening of the apical subepithelial collagenous layer of more than 10-μm thickness due to irregular collagen deposition. Both subtypes usually affect older people. The main symptom is watery, non-bloody diarrhea that also often occurs during the night. Other symptoms include abdominal pain, weight loss, nausea, bloating, and fatigue. Treatment options include antidiarrheal drugs and 5-ASA agents, but most effective are (topic) steroids. Finally, anti-TNFα agents promise relief in a subset of patients with a severe disease course.

The main symptom, watery diarrhea, suggests a mucosal barrier disturbance, but only a few studies exist that examined the epithelial barrier. A study that elucidated mechanisms of diarrhea in collagenous colitis was performed by Bürgel et al. [15]. They found two mechanistic principles that cause diarrhea in collagenous colitis: (i) malabsorption through reduced activity of electroneutral NaCl absorption and (ii) a leak-flux component due to an epithelial barrier dysfunction which could be attributed to TJ alterations. They examined claudin-1, claudin-2, claudin-3, claudin-4, and claudin-5 and occludin and found claudin-4 and occludin protein expression to be downregulated. In parallel, one-path impedance spectroscopy revealed that epithelial resistance is diminished in collagenous colitis by approximately 44%. They suggested that this decrease was attributed to the changes in TJ composition and concluded that a leak-flux mechanism contributes to the watery diarrhea in collagenous colitis.

For lymphocytic colitis, data exist only in abstract form so far. Barmeyer et al. found claudin-4, claudin-5, and claudin-8 downregulation of protein expression and internalization of claudin-5 and claudin-8 from the TJ into the cytoplasm and concluded that in lymphocytic colitis, a leak-flux mechanism contributes to the diarrhea [11]. The importance of this contribution remains unclear. In a small subgroup of patients that were free of symptoms after 6 weeks of oral treatment with budesonide, restauration of electrogenic sodium transport, which is disturbed in lymphocytic and collagenous colitis [8, 9], occurred but epithelial resistance was still reduced indicating an ongoing barrier disturbance [8]. However, this subset of patients was too small for a definite statement. On the other hand, in the absence of apoptotic leaks and gross lesions, the reduction of claudin-5 and claudin-8 in the epithelial tight junction of the colon has to be attributed to play an active role for antigen uptake and perpetuation of inflammation in lymphocytic colitis.

Celiac disease

Celiac disease is an autoimmune disorder that affects the small bowel of genetically susceptible individuals. It appears as enteritis secondary to gluten sensitivity and exhibits one or more of the following pathologic features: villous atrophy, increased intraepithelial lymphocyte count (>30 / 100 enterocytes), chronic inflammatory infiltrate in the lamina propria, and crypt hyperplasia. Duodenal biopsies in combination with serum celiac disease antibodies are relevant for the diagnosis. Affected are all age categories. Symptoms are diverse and include primarily gastrointestinal symptoms, but also atypical symptoms manifesting at the skin, inner organs, joints, or psychiatric and neurological symptoms as well as infertility, recurrent miscarriage, or missed menstrual periods are possible and have to be taken into account. The only therapy is lifelong gluten-free diet, under which the epithelium completely restitutes.

Disease susceptibility genes involved in TJ formation like MYO9B, PARD3, and MAGI2 suggest a primary leaky barrier that promotes disease progression [64, 101]. This hypothesis is supported by the observation of increased intestinal permeability in healthy relatives of celiac disease patients [97] that also exhibit TJ morphology changes despite normal villous architecture and negativity for anti-TTG-IgA [61]. Despite this coherence, the role of disease susceptibility genes in regulating paracellular permeability and gliadin transport is not resolved so far [54, 58].

In active celiac disease, paracellular permeability is increased and TJ morphology is substantially altered [84, 85]. Freeze-fracture electron microscopy demonstrated a thin and tattered TJ meshwork with reduced depth, reduced number of TJ strands, and increased numbers of aberrant strands and strand discontinuities [84]. This is accompanied by molecular changes in the TJ protein composition. Schumann et al. demonstrated in duodenal biopsies of patients with active celiac disease an upregulation of the channel-formers claudin-2 and claudin-15 and downregulation of the barrier-formers claudin-3, claudin-5, and claudin-7 and occludin, whereas claudin-1 and claudin-4 expression was unchanged [86]. Claudin-2 upregulation predominantly localizes to the TJ of crypts and appeared to be posttranscriptional (Fig. 2; [86]). Szakal et al. made similar observations in the duodenum of children with severe celiac disease [92]. But, the role of TJ changes for the diarrhea may be less important than that of the villus atrophy-related malabsorptive mechanism. Claudin-3 did not colocalize with ZO-1 in celiac patients when compared with controls, whereas localization of claudin-1 and claudin-4 and occludin was not altered [86]. In contrast, intense intracellular staining of claudin-5 (Fig. 2) and claudin-15 indicates a relevant protein shift from the TJ into the cytoplasm. While claudin-5 can be linked to barrier function and may have an active role along the leaky gut concept, the redistribution of claudin-15 off the TJ may be more relevant for intestinal glucose malabsorption than for barrier function [93].

Fig. 2
figure 2

Distribution and localization of claudin-2 and claudin-5 in the sigmoid colon of patients with active celiac disease. While not detectable in controls, claudin-2 increased in celiac disease but was restricted to TJs of crypt cells. Claudin-5 was localized to the TJ of both controls and patients with active celiac disease but in addition exhibited an intracellular vesicular pool only in active celiac disease. Bar = 20 μm. (reprinted and adapted from [86] with permission from BMJ Publishing Group Ltd. and British Society of Gastroenterology)

Full size image

The key cytokines in celiac disease are IL-21 and IFNγ [13, 17]. Whereas the role of IL-21 on the TJ in celiac disease has not been evaluated yet, the effect of IFNγ on claudin-2 remains unclear. On the one hand, uptake of the α2-gliadin-33mer was increased in the presence of high IFNγ levels [87] indicating its importance in celiac disease, but on the other hand, IFNγ inhibited the expression of claudin-2 [105, 106]. However, the authors demonstrated this finding only in T84 cells. Therefore, its relevance for celiac disease remains open.

Closing remarks

By presenting the function of intestinal claudins in detail and by describing their changes in four important intestinal diseases with special attention on the functional consequences, we tried to interpret these changes as being causative (active role) or coincident (passive role), in order to intensify the discussion of the functionality of distinct tight junction protein alterations.