Toxicology in Vitro 25 (2011) 45–50

Contents lists available at ScienceDirect

Toxicology in Vitro

journal homepage: www.elsevier .com/locate / toxinvi t

Imaging analysis of the gliadin direct effect on tight junctionsin an in vitro three-dimensional Lovo cell line culture system

Luca Elli a,⇑, Leda Roncoroni a,b, Luisa Doneda c, Michele M. Ciulla d, Roberto Colombo e, Paola Braidotti f,Antonella Bonura a, Maria Teresa Bardella a,b

a Center for Prevention and Diagnosis of Celiac Disease, Fondazione IRCCS ‘‘Cà Granda-Ospedale Maggiore Policlinico”, Via F. Sforza 35, 20100 Milano, Italyb Department of Medical Sciences, Università degli Studi di Milano, Via F. Sforza 35, 20100 Milano, Italyc Department of Biology and Genetic for the Health Sciences, Università degli Studi di Milano, Via Viotti 9, 20100 Milano, Italyd Cardiothoracic Department, Institute of Cardiovascular Medicine, Center of Clinical Physiology and Hypertension, Laboratory of Clinical Informatics and Cardiovascular Imaging,Università degli Studi di Milano, Via F. Sforza 35, 20100 Milano, Italye Department of Biology, Università degli Studi di Milano, Via Viotti 9, 20100 Milano, Italyf Department of Medicine, Surgery and Odontoiatry, Ospedale San Paolo e Fondazione IRCCS ‘‘Cà Granda-Ospedale Maggiore Policlinico”, Università degli Studi di Milano, Via Rudinì5, 20100 Milano, Italy

a r t i c l e i n f o

Article history:Received 11 May 2010Accepted 9 September 2010Available online 17 September 2010

Keywords:DuodenumCeliac diseaseGliadinCell culturesTight junctionIntestinal barrier

0887-2333/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.tiv.2010.09.005

⇑ Corresponding author. Address: Center for PrevenDisease, Fondazione IRCCS ‘‘Cà Granda-Ospedale Magg35, 20122 Milan, Italy. Tel.: +39 02 55033384; fax: +3

E-mail address: lucelli@yahoo.com (L. Elli).

a b s t r a c t

Tight junctions play a pivotal role in maintaining the integrity of the intestinal barrier. Their alteration isinvolved in the pathogenesis of celiac disease. Our aim was to investigate the gliadin effect on the tightjunction proteins in an in vitro three-dimensional cell culture model through imaging analyses.

Lovo multicellular spheroids were treated with enzymatically digested (PT) gliadin 500 lg/mL and itseffect on actin, occludin and zonula occludens-1, was evaluated by means of confocal laser microscopy,transmission electron microscopy and image capture analysis.

Compared to untreated spheroids, PT-gliadin-treated ones showed enlargement of the paracellularspaces (9.0 ± 6.9 vs. 6.2 ± 1.7 nm, p < 0.05) at transmission electron microscopy and tight junction proteinalterations at confocal microscopy and image analyses. In untreated cell cultures thickness of the fluores-cence contour of actin, zonula occludens-1 and occludin appeared significantly larger and more intensethan in the treated ones. In occludin planimetric analysis the lengths of the integral uninterruptedcellular contour appeared longer in untreated than in PT-gliadin treated spheroids (71.8 ± 42.8 vs.23.4 ± 25.9 lm, p < 0.01).

Our data demonstrated that tight junction proteins are directly damaged by gliadin as shown by meansof quantitative imaging analysis.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Celiac disease (CD) is a common chronic enteropathy with aprevalence of about 1/200 in Western countries; it occurs in genet-ically predisposed subjects and is triggered by the ingestion of pro-laminic peptides (gliadin) derived from bread wheat, rye and barley.CD main lesion consists in an architectural rearrangement of theintestinal mucosa characterised by villous atrophy, crypt cell hyper-plasia, and lymphocytic infiltration of the lamina propria andepithelium (Elli and Bardella, 2005; Elli et al., 2009; Freitag et al.,2004; Green and Cellier, 2007; Koning et al., 2005; Schuppan et al.,2005). Although the immunological response appears pivotal in

ll rights reserved.

tion and Diagnosis of Celiaciore Policlinico”, Via F. Sforza9 02 50320403.

the CD pathomechanism, the early steps of these processes are stillunclear, in particular, how gluten peptides cross the intestinal bar-rier (IB) and reach the lamina propria where immunological reactionbegins (Koning et al., 2005). IB is the interface, formed by the entero-cyte layer, between the luminal and submucosal compartments ofthe gastrointestinal tract and its integrity is mainly maintained bythe tight junctions (TJs) (Stevenson et al., 1986; Turner, 2006). Theadjacent cells of the intestinal mucosa are bound together by a com-plex of specialized proteins. The first TJ-associated protein to beidentified was zonula occludens-1 (ZO-1) whose C-terminal halfcontains an actin-binding site and mediates interactions betweentransmembrane proteins and cytoskeleton elements. Occludin, a60 kD integral membrane protein in TJ strands, functions throughits four transmembrane domains and claudin-1 and claudin-2 are23 kD integral membrane proteins that function as major structuralcomponents of TJ strands (Schneeberger and Lynch, 2004; Sears,2000; Stevenson et al., 1986; Tavelin et al., 2003).

46 L. Elli et al. / Toxicology in Vitro 25 (2011) 45–50

In vitro studies showed that integrity of the IB is disrupted in CDpatients, mainly through a damage of the TJ specialized proteins(Clemente et al., 2003, 2000; Montalto et al., 2002; Pizzuti et al.,2004; Schulzke et al., 1998; Sjolander and Magnusson, 1988).

The aim of this study was to investigate and compare by meansof imaging techniques the direct effect of gliadin on TJ proteins andcytoskeleton in an organ-like three-dimensional model of cellculture.

2. Material and methods

2.1. Gliadin digestion

Gliadin was purified from Triticum aestivum flour (HerewardCultivar, UK) according to Capelli (Capelli et al., 1998). Digestionwas performed as previously described by our group (Dolfiniet al., 2002). Briefly, the gliadin was first incubated with pepsinat 37 �C for 24 h, and then with pancreatin at 37 �C for 3 h, adjust-ing to a pH of 8. The digested protein (PT-gliadin) was analyticallycontrolled by means of RP-HPLC, SE-HPLC, and SDS–PAGE, freeze-dried and stored.

2.2. Multicellular three-dimensional cultures (MCSs) and gliadintreatment

Lovo cells from the human colon adenocarcinoma (ATCC,Rockville, USA) were grown in T25 flasks (PBI, Italy) at 37 �C in awater-saturated atmosphere of 5% CO2 and 95% air until conflu-ence and cultured in medium as previously described (Dolfiniet al., 2003; Roncoroni et al., 2008). Cells were weekly removedusing solution containing 0.25% (w/v) trypsin and 0.02% (w/v)EDTA (Sigma–Aldrich, Italy) and the cell suspensions reseeded.Mycoplasma contamination was regularly excluded using the Hoe-chst method (Dolfini et al., 2003).

Three-dimensional cell cultures were initiated by seeding4 � 105 cells/mL in 25 mL of complete medium and incubated in agyratory rotation incubator (60 rev/min) at 37 �C in air (Colaver,Italy) as previously described (Dolfini et al., 2003). Cell clusterswere visible after 4 days of culture, and the MCSs were usually com-plete within the seventh day (average diameter ± SD, 380 ± 49 lm).On the seventh day, the MCSs were treated with PT-digested gliadin(500 lg/mL) in a renewed medium containing PT-gliadin for further4 days and subsequently taken for the evaluation of the endpointparameters. The dose and the timing used were selected from fourdifferent concentrations (125, 500, 750, 1000 lg/mL) and on thebasis of previous time course experiments conducted in our labora-tory (Dolfini et al., 2002). The final dose (500 lg/mL) was chosen onthe basis of its toxic effect without interfering in the formation ofthe MCS three-dimensional structure for a long exposure time.Seven independent series of experiments has been performed.

2.3. Transmission electron microscopy

The Lovo cell line spheroid samples were fixed in 2.5% glutaral-dehyde in phosphate buffer (0.13 mol/L, pH range 7.2–7.4) for 1 h,and then washed in the same buffer. In order to avoid damaging orlosing the MCSs, they were encapsulated in a liquid 2% agar solu-tion solidifying at room temperature. The small MCS-containingcubes (about 1–3 mm) were normally handled and routinely pro-cessed for electron microscopy.

2.4. Intracellular F-actin analysis

Lovo MCSs were washed twice in PBS, fixed in 4% paraformalde-hyde for 1 h, permeabilized with 0.4% Triton X-100 (Sigma–Al-

drich, Italy) for 20 min, washed thrice for 5 min in PBS andstained for immunocytochemistry by means of incubation withTRITC-phalloidin (Sigma–Aldrich, Italy) (1:200 PBS) in a humidchamber at room temperature for 6 h. After washing thrice inPBS each for 5 min, 10 spheroids were transferred onto slides,and each slide was mounted with 90% glycerol in PBS. The excita-tion wavelength was 568 nm and emission 590 nm. The resultswere analyzed using a confocal laser scanning microscope (LeicaTCSNT, Germany) LP590 filter.

2.5. Occludin and zonula occludens-1 analysis

Lovo MCSs were washed twice in PBS and fixed in ethanol for30 min at 4 �C. After the first incubation, the samples were incu-bated with acetone (previously stored at �20 �C) for an additional3 min at room temperature. They were then blocked and incubatedfor immunocytochemistry overnight with anti-occludin-FITC oranti-ZO-1-FITC (Zymed, CA, USA) (excitation wavelength was488 nm, emission 530 nm), before being analyzed by means of con-focal laser scanning microscopy (Leica TCSNT, Germany) BP530/30filter.

2.6. Image capture and analysis

TEM images were randomly and blindly acquired by an expertpathologist (PB). Twenty paracellular spaces were measured eval-uating the orthogonal distances between the external surfaces ofthe lateral plasmatic membranes of two adjacent cells, taken every20 nm.

For confocal microscopy, after localization of fluorescence sig-nals in the cell cultures, multiple adjacent high-power fields ineach section corresponding to the MCSs were selected, acquired,and stored on a personal computer (Power Mac G4, 867 MHz,640 MB RAM, Apple, Cupertino, CA) in JPEG format (5:1) at32 bits/pixel on a 764–560 pixel matrix. Stored images were ana-lyzed to determine the differences between PT-gliadin treated vs.untreated MCSs.

The intensity and fluorescence patterns of the PT-gliadin trea-ted vs. untreated MCSs were evaluated. Before the analysis at leastseven cells were identified in each power field by the experiencedpathologist. Intensities based on a scale of 256 levels (0–255) wereassessed along a linear sample perpendicular to the fluorescentcontour staining of the cell for a total 100 measures.

Since in occludin analysis, differently from actin and ZO-1, sev-eral discontinuities in the pattern of the fluorescent contour stain-ing were present, we measured, in 10 randomly selected images,the lengths in lm of the continuities by planimetry including onceeach fluorescence contour representative of a cellular interplay.

2.7. Statistical analysis

The data were analyzed using the paired Student t-test. The re-sults are expressed as mean values ± SD, and a p-value of <0.05 wasconsidered significant.

3. Results

At phase-contrast microscopy untreated and 4 days PT-gliadin(500 lg/mL) treated MCSs were similar in shape (bright-round)and size (371 ± 46.5 vs. 376 ± 22.2 lm, respectively).

With TEM, untreated MCSs showed a free external cell surfacewith short microvilli and adjacent cells were joined by complexesof TJs. In the cytoplasm there were pseudolumina with microvilli,frequent bundles of cytokeratin tonofilaments, and organelles(mitochondria, rough endoplasmic reticulum, Golgi complex and

Fig. 1. Transmission electron microscopy images of four days PT-gliadin treated(500 lg/mL) (left panel) and untreated spheroids (right panels). In the right panelsbrush border of three adjacent enterocytes (upper panel, 24,750� magnification,scale bar = 250 nm) and a tightly closed paracellular space (lower panel, 265,500�magnification, scale bar = 10 nm) are shown. In the left panel the paracellular spaceof two adjacent enterocytes of a PT-gliadin treated spheroids is represented (scalebar = 100 nm).

L. Elli et al. / Toxicology in Vitro 25 (2011) 45–50 47

lysosomes) were normally represented. In the PT-gliadin treatedMCSs microvilli were only focally maintained on the surface and

Fig. 2. TRITC-phalloidin confocal laser scanning micrographes highlighting the organiz(500 lg/mL) (upper right panel) multicellular tumor spheroids of Lovo cell line. The unwhereas the treated spheroids show a reorganized actin cytoskeleton. In the lower panelsa and b) are reported. The white arrows (a1 and b1), perpendicular to the fluorescence conimage (panels a1 and b1) the corresponding surface plot of the contour is highlighted by mon the y-axis (pseudo-color scale).

TJs were clearly altered with an increase of the paracellular spacefrom 6.2 ± 1.7 to 9.0 ± 6.9 nm (p < 0.05) (Fig. 1).

With confocal laser scanning microscopy the untreated MCSsstained with TRITC-phalloidin showed regular perijunctional actinrings and organized distribution at the cell boundaries whereastreated ones had reorganized and disassembled intracellular actinfilaments (Fig. 2). Moreover, the normal subapical honeycomb pat-tern of TJs, appeared structurally dissociated. The morphologicallycharacteristic ring structure of the immunocolocalised occludinand ZO-1 showed sharp boundaries in the untreated MCSs, butwas partially or completely lost in the PT-gliadin treated MCSs(Figs. 3 and 4).

Analysis of the high density images from confocal laser micros-copy revealed quantitative differences between untreated andPT-gliadin treated MCSs. In fact, the fluorescence intensity of thePT-gliadin treated cell cultures appeared lower than that ofuntreated cultures for all the evaluated proteins (actin, ZO-1 andoccludin) (Table 1). In untreated MCSs the thickness of the fluores-cence contour of actin and ZO-1 appeared larger (Table 2 andFigs. 2–4). In occludin planimetric analysis the lengths of the inte-gral uninterrupted cellular contour appeared longer in untreatedthan in PT-gliadin treated MCSs (71.8 ± 42.8 vs. 23.4 ± 25.9 lm,p < 0.01) (Fig. 4).

4. Discussion

In the present study we demonstrate an alteration of the TJsassociated proteins directly induced by gliadin on an in vitrothree-dimensional cell culture system analyzed by means of imagecapture analysis.

ation of F-actin in untreated (upper left panel) and four days PT-gliadin treatedtreated spheroids have a ‘‘chicken-wire” distribution under the plasma membrane,(a and b), examples of computer analyzed cell obtained from the top panels (inserts

tour staining, indicate where the signal intensity was measured. On the right of eacheans of a three-dimensional surface plot (xyz) showing the color levels of the image

Fig. 3. Confocal microscopy immunolocalisation of zonula occludens-1 (ZO-1) in Lovo multicellular tumor spheroids. In the untreated spheroids ZO-1 localised sharply in theapical region of the lateral membrane visualised en face as a ring pattern (upper left panel), whereas the distribution ZO-1 in the spheroids exposed to a four days PT-gliadintreatment (500 lg/mL) is far from the lateral TJ membrane (upper right panel). In the lower panels (a and b), examples of computer analyzed cell obtained from the top panels(inserts a and b) are reported. The white arrows (a1 and b1), perpendicular to the fluorescence contour staining, indicate where the signal intensity was measured. On the rightof each image (panels a1 and b1) the corresponding surface plot of the contour is highlighted by means of a three-dimensional surface plot (xyz) showing the color levels of theimage on the y-axis (pseudo-color scale).

48 L. Elli et al. / Toxicology in Vitro 25 (2011) 45–50

It is well known that gliadin exerts direct cytotoxic effects onthe enterocytes, represented by apoptotic stimulation, decreaseof vitality and nucleic acid synthesis, and agglutinating activity(Elli et al., 2003, 2004). Using traditional monolayer cell culturesystems, some in vitro studies suggested gliadin as direct responsi-ble of the altered permeability and rearrangement of TJ complexes(Clemente et al., 2003; Sjolander and Magnusson, 1988), and anex vivo study by Pizzuti et al. (2004) demonstrated that glutenexposure alters the intestinal mucosa of celiac patients decreasingZO-1 expression and disrupting F-actin organization. However,these results could represent a secondary effect of the activatedT-cells or inflammatory cytokines as suggested by Ciccocioppoet al. (2006) who demonstrated an IFNc mediated alteration ofTJs in an in vitro model.

The used MCS model simulates biological properties of thein vivo mucosa as cells grow three-dimensionally, oriented andpolarized and there are microvilli plus tightly connected epithelialjunctional complexes similar to those present in human intestinalmucosa (Desoize, 2000). To maintain these characteristics we useda PT-gliadin dose (500 lg/mL) which is able to exert cytotoxic ef-fects without abolish the intercellular connections, (Dolfini et al.,2002, 2003) hampering the formation of MCS three-dimensionalstructure. Moreover, we used a prolonged exposure time (4 days)in order to simulate a long standing exposure, similar to this com-monly observed in humans ingesting gluten in their diet.

Our morphological examinations through TEM and confocalmicroscopy confirm that PT-gliadin directly affects cytoskeletonand TJs. The immunocolocalization of occludin and ZO-1 in the TJring structures is severely disrupted with an almost complete lack

of continuity and breakdown of junctional complexes. Moreover,the present study added a quantitative information of the glia-din-induced damage; in fact, computerised image analysis tech-nique, largely used in other scientific fields and also in clinicalresearch (Caprioli et al., 2006; Ciulla et al., 2007a,b), has been usedfor the first time in this setting. It gives a quantitative evaluation,which correlates with the real protein levels (Maroni et al.,2007), starting from a qualitative method as immunofluorescence.In our study, when applied to confocal and electron microscopy,the image analysis showed that there is a decrease in actin, ZO-1and occludin in the PT-gliadin treated vs. untreated MCSs, an in-creased number of breaks in the occludin fluorescent contourand an increased width of the paracellular spaces.

The disarrangement of the TJ proteins and of the ‘‘belt-like”structure of perijunctional F-actin affiliated with TJs highlightsthe importance of the cytoskeleton network in a three-dimensionalin vitro model of enterocytes. Theoretically, gliadin fragments mayact as modulators on the extracellular loops of TJ transmembraneproteins, mediating TJs opening and the consequent cytoskeletalredistribution. Although the effect demonstrated in our experi-ments is not specific as our model does not involve primary entero-cytes from CD patients, it could mimick one of the early andprimary steps of the celiac cascade which starts the pathomecha-nisms in susceptible subjects carrying particular genetic polymor-phisms (Wapenaar et al., 2008; Wolters et al., 2010). In fact, IBleaks and penetration of gliadin peptides into the lamina propriaare at the basis of the successive immune response and the otherproposed mechanisms of gliadin translocation, the transcellularCD71-mediated retrotranscytosis of gliadin-immunoglobulin A

Fig. 4. Immunolocalisation by confocal laser microscopy of occludin in Lovo multicellular tumor spheroids. In the untreated spheroids occludin is localised in lateralmembrane region as a ring pattern (upper left panel); the distribution of occludin in the four days PT-gliadin treated (500 lg/mL) spheroids is largely interrupted withnumerous discontinuities of the fluorescence contour (upper right panel). The planimetric images of the untreated (a) and PT-gliadin treated (b) spheroids show a series ondiscontinuities in the fluorescence cellular interfaces and an intensity decrease in treated spheroids (b1) in comparison with untreated ones (a1).

Table 1Actin-positive, zonula occludens-1-positive and occludin-positive color levels onrandom locations of confocal laser microscopy sections of untreated (CTR) and PT-gliadin treated multicellular spheroids of Lovo cell line.

Protein Mean color levels (0–255) p

CTR (n = 100) PT-gliadin (n = 100)

Actin 243.01 ± 2.143 229.67 ± 9.12 0.007Zonula occludens-1 254.04 ± 0.83 251.20 ± 0.92 0.002Occludin 253.35 ± 0.36 252.09 ± 0.97 0.03

Table 2Actin, zonula occludens and occluding fluorescence contour thickness on randomlocations of confocal laser microscopy sections of untreated (CTR) and PT-gliadintreated multicellular spheroids of Lovo cell line NS = not statistically significant.

Protein Thickness (lM) p

CTR (n = 100) PT-gliadin (n = 100)

Actin 0.660 ± 0.078 0.466 ± 0.114 0.005Zonula occludens-1 0.579 ± 0.071 0.330 ± 0.049 0.0009Occludin 0.570 ± 0.094 0.651 ± 0.060 NS

L. Elli et al. / Toxicology in Vitro 25 (2011) 45–50 49

complexes (Matysiak-Budnik et al., 2008) and the IFNc effect onthe TJs (Ciccocioppo et al., 2006), are secondary to an already exis-tent inflammatory process involving T and B cells.

In conclusion, as the early steps of gliadin-induced mucosaldamage in patients with CD involve TJ damage, MCSs and imagecapture analysis techniques could become important for testing

the cytotoxic effects of new chemically, enzymatically or geneti-cally modified gliadins studied as alternative therapies to the glu-ten-free diet.

References

Capelli, L., Forlani, F., Perini, F., Guerrieri, N., Cerletti, P., Righetti, P.G., 1998. Wheatcultivar discrimination by capillary electrophoresis of gliadins in isoelectricbuffers. Electrophoresis 19, 311–318.

Caprioli, F., Losco, A., Vigano, C., Conte, D., Biondetti, P., Forzenigo, L.V., Basilisco, G.,2006. Computer-assisted evaluation of perianal fistula activity by means of analultrasound in patients with Crohn’s disease. Am. J. Gastroenterol. 101, 1551–1558.

Ciccocioppo, R., Finamore, A., Ara, C., Di Sabatino, A., Mengheri, E., Corazza, G.R.,2006. Altered expression, localization, and phosphorylation of epithelialjunctional proteins in celiac disease. Am. J. Clin. Pathol. 125, 502–511.

Ciulla, M.M., Cortiana, M., Silvestris, I., Matteucci, E., Ridolfi, E., Giofre, F., Zanardelli,M., Paliotti, R., Cortelezzi, A., Pierini, A., Magrini, F., Desiderio, M.A., 2007a.Effects of simulated altitude (normobaric hypoxia) on cardiorespiratoryparameters and circulating endothelial precursors in healthy subjects. Respir.Res. 8, 58.

Ciulla, M.M., Giorgetti, A., Giordano, R., Silvestris, I., Cortiana, M., Paliotti, R., Lazzari,L., 2007b. Circulating endothelial progenitor cell colony-forming capacity inhealthy subjects: how does an endothelial colony look like? Am. J. Cardiol. 100,559–560.

Clemente, M.G., De Virgiliis, S., Kang, J.S., Macatagney, R., Musu, M.P., Di Pierro, M.R.,Drago, S., Congia, M., Fasano, A., 2003. Early effects of gliadin on enterocyteintracellular signalling involved in intestinal barrier function. Gut 52, 218–223.

Clemente, M.G., Musu, M.P., Frau, F., Brusco, G., Sole, G., Corazza, G.R., De Virgiliis, S.,2000. Immune reaction against the cytoskeleton in coeliac disease. Gut 47, 520–526.

Desoize, B., 2000. Contribution of three-dimensional culture to cancer research. Crit.Rev. Oncol. Hematol. 36, 59–60.

Dolfini, E., Elli, L., Dasdia, T., Bufardeci, B., Colleoni, M.P., Costa, B., Floriani, I., Falini,M.L., Guerrieri, N., Forlani, F., Bardella, M.T., 2002. In vitro cytotoxic effect ofbread wheat gliadin on the Lovo human adenocarcinoma cell line. Toxicol. InVitro 16, 331–337.

50 L. Elli et al. / Toxicology in Vitro 25 (2011) 45–50

Dolfini, E., Elli, L., Ferrero, S., Braidotti, P., Roncoroni, L., Dasdia, T., Falini, M.L.,Forlani, F., Bardella, M.T., 2003. Bread wheat gliadin cytotoxicity: a new three-dimensional cell model. Scand. J. Clin. Lab. Invest. 63, 135–141.

Elli, L., Bardella, M.T., 2005. Motility disorders in patients with celiac disease. Scand.J. Gastroenterol. 40, 743–749.

Elli, L., Bergamini, C.M., Bardella, M.T., Schuppan, D., 2009. Transglutaminases ininflammation and fibrosis of the gastrointestinal tract and the liver. Dig. LiverDis. 41, 541–550.

Elli, L., Dolfini, E., Bardella, M.T., 2003. Gliadin cytotoxicity and in vitro cell cultures.Toxicol. Lett. 146, 1–8.

Elli, L., Dolfini, E., Bardella, M.T., 2004. Direct gliadin cytotoxicity as a cofactor in thepathogenesis of celiac disease. Int. Arch. Allergy Immunol. 134, 88.

Freitag, T., Schulze-Koops, H., Niedobitek, G., Melino, G., Schuppan, D., 2004. Therole of the immune response against tissue transglutaminase in thepathogenesis of coeliac disease. Autoimmun. Rev. 3, 13–20.

Green, P.H., Cellier, C., 2007. Celiac disease. N. Engl. J. Med. 357, 1731–1743.Koning, F., Schuppan, D., Cerf-Bensussan, N., Sollid, L.M., 2005. Pathomechanisms in

celiac disease. Best Pract. Res. Clin. Gastroenterol. 19, 373–387.Maroni, P., Bendinelli, P., Matteucci, E., Desiderio, M.A., 2007. HGF induces CXCR4

and CXCL12-mediated tumor invasion through Ets1 and NF-kappaB.Carcinogenesis 28, 267–279.

Matysiak-Budnik, T., Moura, I.C., Arcos-Fajardo, M., Lebreton, C., Menard, S.,Candalh, C., Ben-Khalifa, K., Dugave, C., Tamouza, H., van Niel, G., Bouhnik, Y.,Lamarque, D., Chaussade, S., Malamut, G., Cellier, C., Cerf-Bensussan, N.,Monteiro, R.C., Heyman, M., 2008. Secretory IgA mediates retrotranscytosis ofintact gliadin peptides via the transferrin receptor in celiac disease. J. Exp. Med.205, 143–154.

Montalto, M., Cuoco, L., Ricci, R., Maggiano, N., Vecchio, F.M., Gasbarrini, G., 2002.Immunohistochemical analysis of ZO-1 in the duodenal mucosa of patients withuntreated and treated celiac disease. Digestion 65, 227–233.

Pizzuti, D., Bortolami, M., Mazzon, E., Buda, A., Guariso, G., D’Odorico, A., Chiarelli, S.,D’Inca, R., De Lazzari, F., Martines, D., 2004. Transcriptional downregulation oftight junction protein ZO-1 in active coeliac disease is reversed after a gluten-free diet. Dig. Liver Dis. 36, 337–341.

Roncoroni, L., Elli, L., Dolfini, E., Erba, E., Dogliotti, E., Terrani, C., Doneda, L.,Grimoldi, M.G., Bardella, M.T., 2008. Resveratrol inhibits cell growth in a humancholangiocarcinoma cell line. Liver Int. 28, 1426–1436.

Schneeberger, E.E., Lynch, R.D., 2004. The tight junction: a multifunctional complex.Am. J. Physiol. Cell Physiol. 286, C1213–1228.

Schulzke, J.D., Bentzel, C.J., Schulzke, I., Riecken, E.O., Fromm, M., 1998. Epithelialtight junction structure in the jejunum of children with acute and treated celiacsprue. Pediatr. Res. 43, 435–441.

Schuppan, D., Dennis, M.D., Kelly, C.P., 2005. Celiac disease: epidemiology,pathogenesis, diagnosis, and nutritional management. Nutr. Clin. Care 8, 54–69.

Sears, C.L., 2000. Molecular physiology and pathophysiology of tight junctions V.Assault of the tight junction by enteric pathogens. Am. J. Physiol. Gastrointest.Liver Physiol. 279, G1129–1134.

Sjolander, A., Magnusson, K.E., 1988. Effects of wheat germ agglutinin on thecellular content of filamentous actin in Intestine 407 cells. Eur. J. Cell Biol. 47,32–35.

Stevenson, B.R., Siliciano, J.D., Mooseker, M.S., Goodenough, D.A., 1986. Identificationof ZO-1: a high molecular weight polypeptide associated with the tight junction(zonula occludens) in a variety of epithelia. J. Cell Biol. 103, 755–766.

Tavelin, S., Hashimoto, K., Malkinson, J., Lazorova, L., Toth, I., Artursson, P., 2003. Anew principle for tight junction modulation based on occludin peptides. Mol.Pharmacol. 64, 1530–1540.

Turner, J.R., 2006. Molecular basis of epithelial barrier regulation: from basicmechanisms to clinical application. Am. J. Pathol. 169, 1901–1909.

Wapenaar, M.C., Monsuur, A.J., van Bodegraven, A.A., Weersma, R.K., Bevova, M.R.,Linskens, R.K., Howdle, P., Holmes, G., Mulder, C.J., Dijkstra, G., van Heel, D.A.,Wijmenga, C., 2008. Associations with tight junction genes PARD3 and MAGI2in Dutch patients point to a common barrier defect for coeliac disease andulcerative colitis. Gut 57, 463–467.

Wolters, V.M., Alizadeh, B.Z., Weijerman, M.E., Zhernakova, A., van Hoogstraten,I.M., Mearin, M.L., Wapenaar, M.C., Wijmenga, C., Schreurs, M.W., 2010.Intestinal barrier gene variants may not explain the increased levels ofantigliadin antibodies, suggesting other mechanisms than alteredpermeability. Hum. Immunol. 71, 392–396.