Proper E-cadherin membrane location in colon requires Dab2 and it modifies by inflammation and cancer

1 Proper E-cadherin membrane location in colon requires Dab2 and it modifies by inflammation and cancer María D. Vazquez-Carretero1*, Pablo García-Miranda1*, María S. Balda2, Karl Matter2, Anunciación A. Ilundáin1, María J. Peral1 1 Departamento de Fisiología, Facultad de Farmacia, Universidad de Sevilla, Spain 2 Department of Cell Biology, Institute of Ophthalmology, University College London, United Kingdom *Corresponding authors: Pablo García Miranda and María D. Vazquez-Carretero Departamento de Fisiología, Facultad de Farmacia, Universidad de Sevilla, Spain C/ Profesor García González, nº 2 41012 Sevilla, Spain. Telephone: 34 954556777 Email address: [email protected]; [email protected] Acknowledgements The work was supported by a Grant from the Junta de Andalucía (CTS 5884) and by a fellowship from The European Molecular Biology Organization (EMBO; ASTF45-2012) to M.D. Vazquez-Carretero. Immunofluorescence images were obtained in the “Centro de Investigación Tecnología e Innovación de la Universidad de Sevilla” (CITIUS), Universidad de Sevilla. We also thank ”Biobanco Hospital Virgen del Rocío-IBIS (Biobanco SSPA y Red de Biobanco-ISCCII-PT13/0010/0056)” for providing us the human samples. 2 ABSTRACT We reported Disabled 2 (Dab2) involvement in milk macromolecules endocytosis in suckling rat intestine. Here, we discovered that Dab2 is mainly at the villi apical cell membrane in the suckling and at both, apical and lateral membrane, along the cryptvillus axis in adult mouse small intestine. In the colon of suckling and adult mouse, and of adult human, Dab2 localized only at the lateral crypt cell membranes, where it colocalized with E-cadherin. In Caco-2 cells, depletion of Dab2, by RNA interference, led to E-cadherin internalization indicating that E-cadherin location at the membrane requires of Dab2. Then, we assessed Dab2/E-cadherin colocalization in mice colon under dextran sulfate sodium (DSS)-induced inflammation. Dab2/E-cadherin colocalization increased by day 3 of DSS-treatment, decreasing thereafter as colitis progressed from 3 to 6 and 9 days. In agreement, in human colon, increased Dab2/Ecadherin colocalization was found in mild and severe ulcerative colitis and in polyps, but reduced and even absent epithelial Dab2 expression was observed in adenocarcinoma. The latter is accompanied by E-cadherin delocalization. The decrease in epithelial Dab2 expression preceded that of E-cadherin. The data suggest that Dab2, by inhibiting Ecadherin internalization, could stabilize AJs and its absence from the epithelial cells may contribute to development of colon inflammation and cancer. Keywords: Dab2, E-cadherin, junctions, colon, inflammation, cancer 3 Introduction The intestinal epithelium acts as selectively permeable barrier allowing the absorption of nutrients and simultaneously preventing the entry of luminal pathogens into the organism. The apical junction complex (AJC), composed of tight (TJ) and adherens (AJ) junctions, connect adjacent cells (see [1] for a review). AJ consist of the transmembrane proteins E-cadherin and nectins and of cytosolic proteins, such as p-120, β-catenin and α-catenin, which mediate E-cadherin anchorage to the actomyosin cytoskeleton. The TJ transmembrane proteins include claudins, occludin, JAM and marvelD3 and among the cytosolic proteins is zonula occludens-1 (ZO-1) that interacts with transmembrane proteins and with the actin cytoskeleton, giving stability to the junction. The formation of TJ is preceded by and depends on AJ formation (see [2] for a review). The AJC are dynamic structures where new proteins are supplied to and recycled from the junctions and its maintenance is crucial for both, keeping the epithelium polarity and the intestinal barrier integrity. The alteration of the intestinal barrier integrity leads to the development of intestinal and extraintestinal diseases (see [3, 4] for review). Disabled 2 (Dab2) is a cytoplasmic adaptor protein [5], that by interacting simultaneously with cytosolic and membrane proteins, plays multiple physiological roles through endocytosis, such as polarized cellular trafficking of membrane receptors, adhesion molecules as E-cadherin, signalling molecules, the cystic fibrosis transmembrane conductance regulator in intestine, among others (see [6] for a review); it participates in Gap junctions endocytosis [7] and in immune responses modulation (see [6] for a review), and it is required for epithelial cells surface positioning in the endoderm [8] and for endosome recycling [9]. Dab2 has also been considered a tumour suppressor because its expression is reduced or lost in several human cancer types (see [6] for a review), including colon cancer [10]. This reduction appears to contribute to epithelialmesenchymal transition (EMT) in breast, cervical and ovarian cancer cells [11–13]. The evidence regarding the role of Dab2 in inflammation is contradictory and focused on immune cells (see [6] for a review). We reported that the Dab2 located at the apical membrane of the suckling rat small intestine epithelium might mediate intestinal endocytosis of milk macromolecules [14]. In the current study, we discovered that Dab2 is involved in E-cadherin localization at cellcell junctions in Caco-2 cells and that Dab2 and E-cadherin colocalization is disrupted in colon inflammation and adenocarcinoma in mouse and human ex vivo. 4 Materials and methods Materials The antibodies and their dilutions used are given in supplementary Materials. Cell culture Stable human epithelial colorectal adenocarcinoma cells (Caco-2) were grown in DMEM (Dulbecco´s Modified Eagle Medium) supplemented with 20% of fetal calf serum and in the presence of 100 U/ml penicillin and 100 µg/ml streptomycin at 37°C and in a 5% CO2 atmosphere. Cells were cultured and plated for experiments as previously described [15]. Depletion of Dab2 in Caco-2 cells RNA-mediated interference (RNAi) experiments were used. Caco-2 cells were cultured to 80% confluence and transfected with either a pool of siRNAs specific for Dab2 (siDab2) or a non-targeting control siRNAs. In both cases, the total final siRNA concentration was 0.2 µM. For the transfection, Interferin transfection reagent (Polyplus Transfection) was used according to the manufacturer’s instructions. The siRNAs were purchased by Sigma-Aldrich. The sequences of the specific siRNAs for Dab2 were (5´…3´): CAAUCGACACCUUCUUCGU and ACGAAGAAGGUGUCGAUUG; CAUGAUGACUUUGAUGCUA and UAGCAUCAAAGUCAUCAUG. 3 days after transfection, samples were collected and processed either for immunofluorescence, western blot or paracellular gate permeability. Analysis of the paracellular gate Ion permeability was determined by measuring transepithelial electrical resistance (TER). A silver/silver-chloride electrode was used to determine the voltage deflection induced by an alternating current square wave current (±20 µA at 12.5 Hz) using an epithelial voltohmmeter (EVOM; World Precision Instruments) as previously described [16]. Briefly, either control or siDab2 Caco-2 cells were plated on transwell filters and allowed to form confluent monolayer in normal medium. TER was measured after 24, 30, 48 and 60 h of culture and results are expressed in ohm x cm2. Paracellular permeability of hydrophilic tracers was monitored by measuring the transmonolayer flux of either 4 kDa FITCor 70 kDa Rhodamine B-conjugated dextran over 2, 4, 6 and 8 h. Fluorescence was then determined with a FLUOstar OPTIMA microplate reader (BMGLabTech, Offenburg, Germany). 5 Animals and experimental colitis C57BL/6 mice of 8 day (suckling) and 3 month-old (adult) were housed in a 12:12 light/dark cycle and fed ad libitum with normal rodent diet (Harlan Ibérica S.L.) and free access to tap water. They were humanely handled and sacrificed by cervical dislocation according with the guidelines of the European Union Council (Directive 2010/63/UE). Colitis was induced in 3 month-old mice by drinking water containing 3% (wt/vol) dextran sulphate sodium (DSS, MW 40 kDa; TdB Consultancy) during 3, 6 or 9 days. Tap water was administered to DSS-untreated animals. Following sacrifice, the colon was removed, washed with ice-cold saline solution and fixed by overnight incubation with phosphate buffer saline (PBS) (in mM, 137 NaCl, 2.7 KCl, 10 Na2HPO4 and 1.8 KH2PO4 pH 7.4) containing 4% para-formaldehyde for histological analysis. The progression of the colon inflammation was assessed by determining disease activity index (DAI), histological score and by the mRNA levels of the pro-inflammatory cytokines IL-1β and TNFα, as described [17]. For DAI evaluation, daily, throughout the DSS-treatment, animals were monitored for weight loss, stool consistency and blood in the faeces (scale of 0-3). Histological score (0–3 scale) was based on destruction of epithelium, dilatation of crypts, loss of goblet cells, inflammatory cell infiltrate and oedema (Supplementary Fig. 2). Human tissue samples Embedded-paraffin sections of: i) colon of patients with mild and severe ulcerative colitis, ii) non-adenomatous colon polyps, iii) colon adenocarcinoma and iv) healthy colon located away from the adenocarcinoma region, were obtained from 25 patients (5 patients per each condition), between 49and 74-year-old, who had undergone colon resection, through the ”Biobanco del Sistema Sanitario Público de Andalucía, Hospital Universitario Virgen del Rocío, Sevilla, Spain”. Tissues were subjected to a pathological examination at the Hospital to confirm the diagnosis of the pathology. The study was approved by the Ethic Committees of Sevilla University and of Virgen del Rocío Hospital. Informed consent was obtained from all patients. Immunofluorescence assays Immunofluorescence was performed on: i) mice ileum and colon, ii) colon from untreated and DSS-treated mice, iii) healthy colon, ulcerative colitis, non-adenomatous polyps and adenocarcinoma samples from patients, and iv) transfected Caco-2 cells. Transfected Caco-2 cells were grown on glass coverslips for 3 days and fixed directly in methanol at −20 °C for 10 min. 5 μm-paraffin embedded intestinal slides were boiled with 10 mM 6 sodium citrate, pH 6, during 10 min. After blocking the intestinal sections for 1h (3% bovine serum albumin, BSA; 3% fetal calf serum, and 0.1% triton x100 in PBS) and the Caco-2 cells for 10 min (0.5% BSA and 20 mM glycine in PBS), they were incubated either without (controls) or with the primary antibody at 4 °C, overnight. In the intestinal samples the primary antibody binding was visualized with the appropriate secondary antibodies: either Alexa Fluor -488 (1:100, Invitrogen) or -546 (1:100, Life technologies), and in Caco-2 cells with either FITC or Cy3 (1:500, Jackson ImmunoResearch Inc.). Nuclei were stained with Hoechst 33258 (Invitrogen). The intestinal slides were mounted (Vectashield, Vector) and photographed with an Olympus BX61 microscope equipped with a DP73 camera. Coverslips containing the Caco-2 cells were mounted (ProLong Gold, Invitrogen) and visualized with a DMIRB Leica microscope with a camera (C474295, Hamamatsu Photonics). Images were analysed using NIH ImageJ program. Quantification of colocalization was done by 3 independent observers by counting the total number of cells in 12 crypts well oriented longitudinally per sample and number of cells with yellow-orange colour signal per crypt. The results are expressed as the number of cells (either total or with yellow-orange colour signal) per crypt. The immunofluorescence intensity signal of Dab2 and E-cadherin was measured in 12 crypts well oriented longitudinally per sample and the results are expressed in arbitrary units. Negative controls without primary antibody were run in parallel (supplementary Fig. 3) Western blot assays Caco-2 cell were collected by washing twice with PBS and after SDS-PAGE sample buffer addition, the cells were homogenized with a 23G needle and heated at 70°C for 10 minutes. Samples were then processed using standard western blotting techniques. Protein-bound antibodies were detected with horseradish peroxidase conjugated secondary antibodies using enhanced chemiluminescence detection system (GE Healthcare Select®). Anti-α-tubulin antibody was used to normalize band density values. Statistical analysis Data are presented as mean ± SEM. The number of animals or patients used for each condition was 5. One way ANOVA followed by Tukey´s test was used for multiple comparisons (GraphPad Prism Program v8.0). For comparisons between two groups, Student´s t test was used. Differences were set to be significant for p<0.05. 7 Results Localization of Dab2 in the small and large intestine of suckling and adult mice In the current work we investigated Dab2 function in adult mouse and human intestine under physiological and physiopathological conditions. We started the work determining by immunofluorescence the location of Dab2 protein in colon of 8 dayand 3 month-old mice and comparing the observations with those obtained in the ileum. As previously, in the suckling ileum the anti-Dab2 antibody specific signal was strong at the apical membrane of the cells lining the villus (Fig.1). In the adult ileum, Dab2 signal was seen at both apical and lateral cell membranes of the villus and crypt cells. The absence of Dab2 signal at the lateral cell membrane of suckling intestine might be due to the strong apical signal that precludes emerging the much less strong signals from the lateral membrane. In the suckling and adult colon, Dab2 signal is mainly at the lateral membranes of the crypt cells, being the signal intensity stronger at the bottom of the crypts. The colonocytes apical membrane was slightly stained mainly in the suckling mice. The results suggest that the main function of Dab2 in adult intestinal epithelial is related to lateral membrane processes. Dab2 depletion disrupts epithelial cell-cell apical junctions As Dab2 was detected at the epithelial lateral membrane we tested whether it regulates the AJC by silencing its gene expression (siDab2) in Caco-2 cells (siRNA technique) and analysing the expression (western blot) and location (immunofluorescence) of TJ (ZO-1 and occludin) and AJ (E-cadherin and β-catenin) proteins. The western blot revealed that the silencing procedure reduced the expression of the two Dab2 isoforms by about 70% (Fig. 2a). Tight junction appearance, as detected by occludin and ZO-1 staining, changed from a linear organization into a wavy structure (Fig. 2b). In Dab2-containing Caco-2 cells, E-cadherin mainly appeared at the cell membrane as a linear structure and as a few E-cadherin spots in the cytoplasm. These spots likely might represent E-cadherin localization in cytosolic vesicles. Dab2 depletion affected E-cadherin membrane localization and increased its staining in the cytosol. This E-cadherin redistribution was accompanied by increases in cytosolic β-catenin (Fig. 2b). Western blots show that Dab2 depletion did not significantly changed the abundance of the AJC proteins assayed (Fig. 2c). These observations suggest that Dab2 is required for proper AJC organization, mainly AJs. 8 Since the observed siDab2-induced alterations in ZO-1 and occludin distribution might modify the barrier properties of the TJ, we next measured the TER and the paracellular dextran permeability in either control or siDab2 Caco-2 cells monolayers. Neither the TER nor the dextran permeability (Supplementary Fig. 1) were affected by siDab2, indicating that the AJC disruption induced by 70% Dab2 depletion is not big enough as to significantly modify the paracellular permeability. Colocalization of Dab2 and E-cadherin in the small and large intestine of suckling and adult mice The most prominent effect of siDab2 on Caco-2 cells AJC was the change in E-cadherin distribution; we next determined Dab2 and E-cadherin colocalization in mouse intestinal epithelium by immunofluorescence. In all the samples examined, the anti-E-cadherin antibody signal was seen at the lateral cell membranes (Fig. 3). Dab2 colocalized with E-cadherin at the cell lateral membrane in both, ileum and colon. In the suckling ileum colocalization was observed at the crypt-villus junction and along the crypt-villus axis in the adult. In both, suckling and adult colon, colocalization was mainly detected at the bottom of the crypts (Fig. 3). These observations suggest that, as in Caco-2 cells, in mouse intestinal epithelium the E-cadherin membrane location is related to Dab2. DSS-treatment effects on the expression and localization of Dab2 and E-cadherin in mice colon Since the available evidence indicates that mucosal inflammation induces E-cadherin downregulation or mislocalization, we evaluated, by immunofluorescence, whether inflammatory conditions (DSS-treatment during 3, 6 and 9 days) affected the colon expression and localization of Dab2 and E-cadherin. Supplementary Fig. 2 summarized the indicators of the colon inflammation progression. Inflammation was relevant by day 3 of DSS-treatment, though some of the inflammatory parameters were mildly increased. Compared with untreated colon, 6and 9days treatment increased the levels of the proinflammatory cytokines by about 100 and 1000 times, respectively. As compared with DSS-untreated mice (see Fig.4), 3 days of DSS-treatment increased: i) the number of cells expressing Dab2 along the entire crypt (Fig. 4a); ii) the intensity of the Dab-2 and E-cadherin signals at the lateral cell membranes, cytosol and apical membrane domain (Fig. 4a and 4b), and iii) the number of crypt cells and colonocytes showing Dab2 and E-cadherin colocalization even at the apical membrane (Fig. 4a and 4c). Colocalization extends to about two-thirds of each crypt, instead of only at the bottom. After 6 days of DSS-treatment the normal structure of the colon epithelium 9 was destroyed (Fig. 4a) and therefore epithelial Dab2 and E-cadherin signal intensity was decreased (Fig. 4b). The remaining epithelium exhibited less Dab2 at the lateral membranes (Fig. 4a) and, consequently, decreased Dab2/E-cadherin colocalization (Fig. 4c). This colocalization was absent after 9 days of DSS-treatment (Fig. 4a). These observations might indicate that the intestine firstly responds to tissue injury by increasing epithelial Dab2 and E-cadherin to strengthen the AJC and hence provide protection against inflammation. After 9 days, the inflammatory inducers lead to depletion of epithelial Dab2 and E-cadherin. Dab2 and E-cadherin localization in human healthy and pathological colon We next determine whether the Dab2 and E-cadherin response to experimental colitis found in mice also occurs in human colon. Immunofluorescence was performed on colon specimens from patients with mild and severe ulcerative colitis, with non-adenomatous colon polyps and colon adenocarcinoma and from healthy colon. As in mouse colon, in human healthy colon Dab2 and E-cadherin are located at cell lateral membrane and both proteins colocalized mainly at the bottom of the crypts (Fig. 5). In mild ulcerative colitis specimens, the E-cadherin signal was found at the lateral cell membrane, at the apical membrane domain in some cells and in the cytosol; Dab2 signal was mainly at lateral membranes (Fig. 5). The intensity of both, Dab2 and E-cadherin signals were greater than in healthy colon (Fig. 7a) and the number of crypt cells showing Dab2 and Ecadherin colocalization was increased (Fig. 7b). In severe ulcerative colitis, expression and colocalization of Dab2 and E-cadherin is still higher than in healthy colon but only seen in those cells that still maintained the epithelial integrity (Fig 5, 7b and 7c). Since ulcerative colitis is associated with increased risk of colon cancer development and Dab2 is considered to act as a tumour suppressor, we examined Dab2 localization in human colon polyps and adenocarcinomas. Polyp specimens (Fig. 6) showed strong Dab2, E-cadherin and Dab2/E-cadherin signals at the crypt cells lateral membranes and cytosol, higher than that in healthy colon and similar to that in colitis (Fig. 7). In the adenocarcinoma regions (Fig. 6) where the epithelium was partially preserved (adenocarcinoma 1), Dab2 and E-cadherin appeared as diffuse membranous and cytosolic signals surrounding the nucleus, being the E-cadherin signal at the membrane more abundant than that of Dab2. Some colocalization of Dab2 and Ecadherin was also observed. In other regions (adenocarcinoma 2), Dab2 was completely absent from the remaining epithelium-like structures whereas E-cadherin was still present. According to these observations, the disappearance of Dab2 from the intestinal epithelium precedes that of E-cadherin and these findings, together with those in a 16 effect of inflammation progression on Dab2 and E-cadherin immunofluorescence intensity and Dab2/E-cadherin colocalization per crypt (p<0.01). Tukey’s test: ap<0.01, DSS-treated vs untreated mice; bp<0.01, 6 days DSS-treated vs 3 days DSS-treated mice. Five mice were used per condition. See Materials and Methods for specific details. Fig. 5 Immunolocalization of Dab2 and E-cadherin in human colon under inflammation conditions. 5-10 µm colon sections from human specimens of: healthy colon, mild ulcerative colitis and severe ulcerative colitis, were incubated with anti-Dab2 (Proteintech, red) and anti-E-cadherin (green) antibodies. Nuclei were visualized with Hoechst (blue). Yellow-orange colour indicates colocalization of the both proteins under study (pointed with arrows). Representative photographs of the immunohistochemistry assay are shown. Scale bars represent 50 µm. Five specimens were used per condition. See Materials and Methods for specific details. Fig. 6 Immunolocalization of Dab2 and E-cadherin in human colon polyps and adenocarcinomas. 5-10 µm colon sections from human specimens of: non-adenomatous polyps and adenocarcinoma were incubated with anti-Dab2 (Proteintech, red) and antiE-cadherin (green) antibodies. Nuclei were visualized with Hoechst (blue). Dab2 localization at the membrane (arrowheads) or at the cytosol (small arrows) are pointed. Yellow-orange colour indicates colocalization of the both proteins under study (pointed with large arrows). Scale bars represent 50 µm. Representative photographs of the immunohistochemistry assay are shown. Healthy colon is shown in Fig. 5. Adenocarcinomas labelled as 1 or 2 indicate two different representative areas within the same adenocarcinoma specimen. Five specimens were used per condition. See Materials and Methods for specific details. Fig. 7 Quantification of immunofluorescence intensity and Dab2/E-cadherin colocalization in human colon specimens. Histograms represent means ± SEM. A, Dab2 or E-cadherin immunofluorescence intensity per crypt and B, Each bar represents the total number of cells per crypt and the number of cells having Dab2/E-cadherin colocalization per crypt. UC: ulcerative colitis. One-way ANOVA showed an effect of inflammation progression on Dab2 and E-cadherin immunofluorescence intensity and Dab2/E-cadherin colocalization per crypt (p<0.01). Tukey’s test: ap<0.01, UC vs healthy colon; bp<0.01, severe UC vs mild UC or polyps vs mild UC. Five specimens were used per condition. See Materials and Methods for specific details. Supplementary Fig. 1. Effects of Dab2 depletion on paracellular and ion permeability. A, Paracellular tracer permeability was measured on Caco-2 cells using fluorescently 17 labelled dextrans of 4 and 70 kDa. B, Transepithelial Electrical Resistance (TER) was measured on Caco-2 cells for a period of 24, 30, 48 and 60 hours culture. Data are means ± SEM, n=3. Supplementary Fig. 2 Evaluation of the DSS-induced colitis progression. DSS (dextran sulphate sodium) was administered in the drinking water during either 3, 6 or 9 days as described in Methods. The following inflammation indicators are displayed: DAI, histological score of colon, representative photographs of H&E stained distal colon sections and mRNA relative abundance of IL-1β and TNF-α. Scale bars represent 100 µm. Data are means ± SEM. The number of animals used in each experimental condition was 5. Student's t-test: *p<0.001, **p<0.05, DSS-treated vs. untreated mice. Supplementary Fig. 3. Negative controls of immunofluorescence assays. 5-10 µm ileum and colon sections from either mouse or patient specimens used in figures 1 and 3-6 were incubated without primary antibody but with the anti-rabbit IgG Alexa fluor-546 secondary antibody (Life technologies). Scale bars represent 50 µm. The photographs are representative of five different assays. Supplementary Fig. 4. Immunolocalization of Dab2 and arginase-1. The colon of 9 days DSS-treated mice and of the human specimens (severe ulcerative colitis or adenocarcinoma) were utilized. 5-10 µm sections of the specimens were incubated with anti-Dab2 (either Santa Cruz for mouse or Proteintech for human samples, red) and antiarginase-1 (green) antibodies. Nuclei were visualized with Hoechst (blue). Yelloworange colour indicates colocalization of the both proteins under study (pointed with arrows). Scale bars represent 50 µm. The photographs are representative of five different. 18 Fig. 1 19 Fig.2 20 Fig. 3 21 22 Fig. 4 23 Fig. 5 24 Fig.6 25 Fig.7