Junctional plakoglobin regulates intestinal epithelial cell proliferation via MAPK/ERK signaling

Abstract

Junctional plakoglobin (JUP) is a desmosomal protein essential for epithelial tissue integrity and epithelial-immune cell crosstalk. Although JUP has been implicated in inflammatory processes, its role in intestinal epithelial repair remains unclear.

In vitro, loss of JUP in Caco-2 monolayers and non-transformed intestinal organoids markedly accelerated wound closure, driven by enhanced epithelial cell migration and proliferation. RNA sequencing-based pathway analysis of organoids revealed activation of MAPK/ERK signaling following JUP deficiency. This effect was corroborated by increased ERK phosphorylation in Western blot analyses. In line with that, pharmacological inhibition of ERK abrogated the accelerated wound closure, confirming that MAPK/ERK signaling is critical for the phenotype. To further examine the relevance in vivo, we used an inducible, intestine-specific JUP knockout mouse model (iVilCreER^T2Jup^fl/fl). Following biopsy-induced intestinal injury, JUP deletion significantly enhanced mucosal repair, accompanied by increased cell proliferation and MAPK signaling, thereby confirming the in vitro results.

Collectively, these findings establish JUP as a previously unrecognized regulator of epithelial cell dynamics and repair through MAPK/ERK signaling.

Graphical Abstract

Loss of epithelial plakoglobin (JUP) activates MAPK/ERK signaling, thus, significantly enhancing epithelial cell motility by increased cell proliferation and migration. The signaling cascade accelerates wound closure in vitro and in vivo, thereby improving intestinal mucosal healing following injury.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12964-026-03256-6.

Introduction

Intestinal epithelial cell dynamics are critical for maintaining local gut homeostasis. Dysregulated epithelial cell motility and proliferation result in bystander tissue injury, a hallmark of intestinal disorders such as inflammatory bowel disease (IBD) [32]. Although adequate mucosal healing has been defined as a critical therapeutic aim for achieving clinical remission in patients with IBD, the mechanisms regulating this process are poorly understood. Consequently, specific therapies to induce mucosal healing are lacking.

Repair of the intestinal epithelium requires a coordinated interplay of epithelial cell proliferation, migration, adhesion, and differentiation to ensure continuous renewal and tissue integrity [15]. Increasing evidence points to a critical role of desmosomes in the regulation of these processes contributing to intestinal homeostasis by largely undefined mechanisms [30, 36]. In the gut, desmosomes consist of the cadherins desmoglein2 (DSG2) and desmocollin2 (DSC2), both of which have been implicated in modulating cell migration and processes associated with epithelial remodeling during repair [7, 35]. Recently, we found that junctional plakoglobin (JUP), an intracellular plaque protein linked to both DSG2 and DSC2, plays a critical role in regulating intestinal homeostasis by activating intestinal inflammation through regulation of the inflammasome [16]. JUP links desmosomal cadherins to the cytoskeleton, thereby contributing to intercellular adhesion and the stability of the junctional complex in intestinal epithelia. Furthermore, JUP has been recognized as a signaling hub mediating key cellular functions in keratinocytes, the heart, the gut and during embryonic development [17, 10, 23]. Given the critical role of JUP as a protein regulating signaling pathways and to balance intercellular adhesion, we aimed to further assess its role in intestinal epithelial integrity and gut homeostasis.

Based on experiments using cell culture models, organoids, and our previously established inducible intestinal epithelial cell-specific Jup knock-out mouse strain, we provide strong evidence that intestinal JUP is a critical regulator of epithelial cell dynamics in the gut. This effect is driven by JUP-associated modulation of mitogen-activated protein kinase/extracellular signal-regulated kinase (MAPK/ERK) pathways, which regulate epithelial proliferation and migration.

Material and methods

Antibodies and test reagents

Antibodies and test reagent used in this study are shown in supplementary Table S1 and Table S2.

iVilCreER^T2 Jup^fl/fl

Conditional knockout mice (iVilCreER^T2 Jup^fl/fl) were generated and characterized as described previously [16]. Induction of Jup deficiency was achieved by intraperitoneal administration of tamoxifen (T5648, MilliporeSigma) for 5 consecutive days (one injection per animal per day). Tamoxifen was dissolved in ethanol (≥ 99,8% purity, MilliporeSigma) and subsequently diluted in corn oil (C8267, MilliporeSigma) to a final concentration of 10 mg/ml (ethanol:corn oil = 1:10), corresponding to a dose of 50 mg/kg administered in a volume of 5 ml/kg. On day 21, the experiment was terminated, and the animals were euthanized under isoflurane anesthesia by cervical dislocation. Throughout the manuscript, mice carrying floxed alleles of Jup (Jup^fl/fl) and expressing tamoxifen-inducible Cre recombinase under the Villin promoter (iVilCreER^T2) are referred to as cKO (conditional knockout), whereas Jup^fl/fl littermates lacking Cre recombinase to as controls (Ctrl) following tamoxifen administration. Experiments were performed on Jup^fl/fl or iVilCreER^T2 Jup^fl/fl mice (n = 26,aged 10–14 weeks, weight: 20–25 g, Janvier, 53,940 Le Genest Saint Isle, France).

Animal experiments

All animal experiments were performed in accordance with the European Directive 2010/63/EU on the protection of animals used for scientific purposes and were approved by the local regulatory authorities, the Government of Lower Franconia, Bavaria, Germany (approval number: RUF-55.2.2–2532-2–1430). Mice (C57BL/6, 8–12 weeks, male and female; ratio 1:1) were housed under specific-pathogen-free (SPF) conditions in standard Makrolon type III cages containing autoclaved wood shavings and autoclaved paper towel as nesting material, enriched with environmental items, under controlled conditions (12-h light/dark cycle, 20–24 °C, and 45–65% humidity). Male mice were housed in pairs or singly in cases of social instability, while female mice were housed in groups of 2–4 animals. Animals were included in the experiments at an age of 10–14 weeks with a weight of 20–25 g.

Murine mucosal wound healing assay

To assess intestinal epithelial wound healing in mechanically generated wounds, the biopsy-based wound healing assay was used in Jup^fl/fl or iVilCreER^T2 Jup^fl/fl mice as already described in detail before [13]. For the in vivo biopsy-based mucosal wound healing model, a high-resolution video endoscope (Coloview Veterinary Endoscope, Karl Storz, Tuttlingen, Germany) equipped with a biopsy forceps was used to induce standardized injuries to the colonic mucosa at seven to eight independent sites along the mesenteric side of the distal colon in anesthetized mice. For anesthesia, isoflurane (cpPharma, Burgdorf, Germany) was used. Since minor variations in wound diameter are unavoidable, the primary outcome parameter is the relative wound closure of each individual wound over time. Wound healing was assessed at 24- and 72-h post-injury by re-endoscopy using high-resolution images (1024 × 768 pixels) on a flat-panel color monitor. Each wound region was photographed, and wound areas were measured using ImageJ (Rasband, W.S., ImageJ, Bethesda, Maryland, USA.) In each experiment, four to five representative wounds per animal were quantified by two independent blinded investigators (MK and MH). After 72 h post injury, experiments were terminated, and colon tissue was harvested by punch biopsies from wounded and non-wounded areas in the same animals.

Caco-2 cell culture

As described previously [24], Caco-2 cells were cultured in DMEM (Dulbecco`s Modified Eagle`s Medium) supplemented with Penicillin–Streptomycin and 10% FCS. In addition, the Caco-2 cell line deficient for Dsg2 was used as described and characterized in detail previously [24]. For passaging, cells at 80% confluency were treated with 1% EDTA in Dulbecco´s Phosphate Buffered Saline (DPBS) and dissolved using Trypsin–EDTA. After centrifugation, the pellet was resuspended in DMEM + + for seeding or experiments.

siRNA-mediated knockdown of JUP in Caco-2 cells

The cells were cultured in 6-well plates for 24 h prior to transfection, as described previously [27]. This was achieved by the introduction of human Jup siRNA (MedChemExpress, USA,Cat. No.: HY-RS07061), or with non-target siRNA serving as the control (Abcam, UK). Transfection was initiated once the cells had reached 80% confluence. The transfection was conducted using Lipofectamine Transfection LTX Reagent (Thermo Fisher Scientific, USA), in accordance with the instructions provided by the manufacturer. Based on preliminary experiments, sufficient knockdown of JUP was observed after 48 h, and all subsequent experiments were conducted at this time point.

Lentivirus production and lentiviral transduction of organoids

Lentivirus production, concentration, and transduction of murine intestinal organoids were performed as previously described [16]. Briefly, lentiviral particles were generated by PEI-mediated transfection of HEK293T cells with shRNA-containing LT3GEPIR or LT3GEBIR together with the packaging plasmids psPAX2 and pMD2.G, concentrated by PEG-8000 precipitation, and used for transduction of dissociated organoid cells. Stable transductants were selected with puromycin, and shRNA expression was induced by doxycycline. For generation of the combined Jup/Erk knockdown, organoids were first transduced with Jup-targeting shRNA and subsequently subjected to a second lentiviral transduction with Erk-targeting shRNA using the same procedure followed by blasticidin selection. The shRNA sequences used in this study were as follows:

• shJup-1, TTCTTGAGTGTGTACTGTCTGC;

• shJup-5, TTGATCGGGATCATACTCTGGG;

• shErk-2; TTCTTTGACAGTAGGTCTGGTG;

• shCTR (Renilla), CAGGAATTATAATGCTTATCT.

Murine organoid generation

Murine organoid generation was carried out as described previously [16, 26]. In brief, the mucosa was mechanically dissected from iVilCreER^T2Jup^fl/fl mice followed by removal of villi using a sterile glass slide. The remaining tissue was transferred into a 50 mL Falcon tube containing 20 mL cold 4 °C Hank's balanced salt solution (HBSS) (Sigma-Aldrich, St. Louis, MO, USA), vortexed for 5 s, and the supernatant was discarded. This washing step was repeated until the supernatant was entirely cleared of cell debris. The tissue was incubated in 4 ^◦C Ethylenediaminetetraacetic acid (EDTA, 2 mM) dissolved in HBSS (Sigma- Aldrich, St. Louis, MO, USA) for 30 min under gentle rotation on a shaker. Subsequently, the tissue was washed manually in 20 mL HBSS by inverting the tube five times. The mucosa was then transferred into a new tube with 10 mL HBSS and shaken five times manually. This shaking procedure was repeated four times. The amount and size of isolated crypts within drops was checked under the microscope. The supernatants containing the most viable-looking crypts were pooled and centrifuged at 350 xg for 3 min at room temperature (RT). Then, the pellet was resuspended in 10 mL basal medium, Dulbecco’s Modified Eagle’s Medium (DMEM)-F12 Advanced (Invitrogen, Carlsbad, CA, USA) supplemented with 1 × N2, 2 × B27, 1 × antibiotic and antimycotic solution, 10 mM 4-(2-hydroxyethyl)−1-piperazineethanesulfonic acid (HEPES), 2 mM GlutaMAX-I (all from Invitrogen, Carlsbad, CA, USA), 1 mM N-acetylcysteine (Sigma-Aldrich, St. Louis, MO, USA), and the crypt number was estimated in a 10 μL drop by microscopic observation. Crypts were transferred to a 1.5 mL tube and centrifuged at 350 g for 3 min at room temperature and the supernatant was removed. The tube with the cell pellet was placed on ice until it was resuspended in cold Cultrex (Corning, Hickory, NC, USA) with 5000 crypts/mL. Drops of 50μL per well were seeded in a 24- well plate and incubated for 10–20 min until Cultrex was solidified. The culture medium contained a mixture of 50% fresh basal medium and 50% Wnt3A-conditioned medium. Furthermore, the following growth factors were added: 500 ng/mL hR-spondin 1, 50 ng/mL human epidermal growth factor (hEGF) (both PeproTech, Rocky Hill, NY, USA), 100 ng/mL rec Noggin (PeproTech, Rocky Hill, NY, USA), 10 mM Nicotinamide, 10 μM SB202190, 10 nM [Leu15]-Gastrin I (all three Sigma-Aldrich, St. Louis, MO, USA), 500 nM A83-01 (Tocris Bioscience, Bristol, UK) and 500 nM LY2157299 (Axon MedChem, Reston, VA, USA). Additionally, 10 μM Rho-kinase inhibitor Y-27632 (Tocris Bioscience, Bristol, UK) was added after seeding and after each splitting. Organoids were maintained in culture for 8–10 weeks.

Beginning at passage 5, organoids were treated with 4-hydroxytamoxifen (10 μM) to induce Jup-deficiency. Previously, we showed that a 10-day incubation with 4-hydroxytamoxifen was sufficient to induce a complete Jup knockout in murine organoids [16]. Tamoxifen was applied at every medium exchange after 48 h. Loss of JUP was confirmed by Western blot analysis.

Plakoglobin overexpression

Wild-type mouse intestinal organoids were dissociated into single cells and seeded in 12-well plates coated with 10% Geltrex. The cells were cultured until they formed confluent organoid-derived monolayers. For transient overexpression of plakoglobin (PG/JUP), wild-type 2D mouse organoids were transfected with a plakoglobin-expressing plasmid (PB-CAG-hJUP-T2A-mNeonGreen) using Lipofectamine LTX with PLUS Reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. Briefly, 1 µg of plasmid DNA was diluted in 50 µL Opti-MEM reduced-serum medium and mixed with 1 µL PLUS Reagent. In parallel, 2 µL Lipofectamine LTX was diluted in 50 µL Opti-MEM. The two solutions were combined, incubated for 5 min at room temperature, and the resulting 100 µL transfection mixture was added dropwise to each well. Control organoids were transfected with the corresponding empty vector (PB-empty).

Once the organoid-derived monolayers had reached confluence, a linear scratch was introduced into the cell layer. Plakoglobin overexpression was confirmed by Western blot analysis. For this purpose, cells were harvested approximately 24 h after transfection and lysed directly in 100 µL SDS sample buffer per well.

Human intestinal organoid cultures

Human intestinal organoids were generated and cultured as previously described [13]. Briefly, full-thickness intestinal tissue specimens were obtained from patients undergoing bowel resection, and healthy tissue from the resection margin was used for organoid generation. Written informed consent was obtained from all patients prior to surgery. The study was approved by the local ethics committee of the University of Würzburg, Germany (approval numbers 46/11, 113/13, 42/16, amendment 2020, and 171/22). Intestinal crypts were isolated by EDTA-based dissociation, embedded in Growth Factor Reduced Basement Membrane Matrix (Corning, Corning, NY, USA), and cultured in IntestiCult™ Organoid Growth Medium Human (STEMCELL Technologies, Vancouver, BC, Canada) supplemented with 10 µM Y-27632 dihydrochloride (Tocris Bioscience, Bio-Techne, Bristol, UK) after seeding. Culture medium was replaced every 2–3 days, and organoids were routinely passaged and maintained as previously described [13].

Scratch-wound assay in organoid-derived and Caco-2 monolayers

2D monolayers were cultured and scratched as described previously [13]. In brief, collected organoids were washed once with BM and resuspended in 1 ml TrypLE Express Enzyme. The resuspended organoids were incubated at 37 °C for 5 min, then mechanically dissociated to obtain a single-cell suspension. The process was halted by adding 1 ml of BM. The cells were washed once more and resuspended in culture medium for murine intestinal organoids supplemented with Y-27632. For seeding, 12-well TPP® tissue culture plates (TPP, Trasadingen, Switzerland) were coated with 10% Geltrex LDEV-Free Reduced Growth Factor Basement Membrane Matrix (Gibco) in BM. The coating solution was incubated in the wells for 30 min, removed and the surface was air-dried at room temperature before adding the single-cell suspension.

Caco-2 cell monolayers or 2D organoid monolayers were cultured in 12 well plates, respectively (Ibidi, Gräfelfing, München). Confluent cell monolayers were scratched using a 10-µm pipette, as described previously [7, 13, 25]. Following scratch wounding, cells were treated with the ERK1/2 inhibitor FR180204 at a final concentration of 5 µM. Treatment was initiated immediately after scratching and maintained for 24 h. All scratches were performed by the same blinded investigator. Images of the wounds were acquired at 0-, 24-, 48-, and 72-h post-scratch, and wound closure was quantified using ImageJ Software (Wayne Rasband) by normalizing the wound area to the initial (t = 0) scratch area.

FITC-dextran permeability assay

As described previously, paracellular permeability was assessed using a FITC-dextran flux assay [3, 25], Caco-2 cells were cultured to confluence on Transwell filter inserts (12-well plates, 0.4 µm pore size). After washing with PBS, monolayers were incubated with phenol red–free DMEM containing FITC-dextran (4 kDa). Samples were collected from the basolateral compartment over 2 h, and fluorescence was measured using a Tecan GENios microplate reader (excitation 485 nm, emission 535 nm). Permeability coefficients were calculated according to established formulas, and experiments were performed in triplicate.

ECIS-based TER measurements in Caco-2 cells

Transepithelial electrical resistance (TER) was measured using the ECIS trans-Filter Adapter connected to the ECIS 1600R system (Applied Biophysics, Ibidi GmbH, Martinsried, Germany). Cells were seeded onto 8-well arrays containing 40 electrodes per well according to established protocols [3]. 24 h after transfection of cells, TER measurements were initiated immediately after reagent application.

Western blot

Western blotting was performed as described previously [18]. Cells were lysed in a buffer, homogenized, and protein quantification was determined using a BCA assay kit (Thermo Scientific, Massachusetts, USA). SDS gel electrophoresis was followed by protein transfer to a nitrocellulose membrane and overnight incubation with primary antibodies (supplementary Table S1). Secondary antibodies were applied for 1 h at room temperature, and signals were detected using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Scientific, Massachusetts, USA). Densitometric analysis and normalization to β-actin were performed using ImageLab (Hercules, California, U.S.).

Immunostaining

Immunostaining was performed as described previously [19]. Standard immunostaining protocols were applied to epithelial monolayers, organoids, and murine specimens. Caco-2 cells were fixed as monolayers on coverslips, whereas murine specimens were paraffin-embedded and sectioned into 1 µm slices. Organoids were released from Cultrex/Matrigel, fixed, and paraffin-embedded. Monolayers, organoid sections, and tissue slides were incubated overnight at 4 °C with the respective primary antibodies specified in supplementary Table S1. Secondary antibodies, including Cy3- or Alexa Fluor 488 labeled goat anti-mouse or goat anti-rabbit, respectively (all diluted 1:600,Dianova, Hamburg, Germany), were utilized. Representative images were documented using a confocal microscope (Leica LSM 780; Zeiss, Oberkochen, Germany).

Ki-67 based assessment of cell proliferation

Cell proliferation was assessed by immunofluorescent Ki-67 staining. To determine the contribution of proliferative activity on epithelial wound closure and under basal conditions (swiss rolls), cells were stained with an anti-Ki-67 antibody and counterstained with DAPI. The fraction of Ki-67-positive cells was quantified relative to the total number of DAPI-positive nuclei in ten representative fields, which were independently analyzed by two different observers. Regarding epithelial wound closure, differentiation between the leading edge and the following cells was performed as demonstrated in Fig. S1A.

Dispase-based dissociation assays

As previously reported [33], images were acquired to assess the number of organoids and Caco-2 cells per well at baseline. Organoid experiments were performed in 2D organoid monolayers. Samples were washed with PBS and incubated with 2.4 U/mL Dispase-II (Roche, Basel, Switzerland) in HBSS for 30 min at 37 °C to release the cells from the well bottom and to dissolve the gel around the organoids. Afterwards, the samples were exposed to shear (mechanical) stress by pipetting up and down 5 times. Images of the entire wells were taken with fluorescence microscope BZ-9000 (Keyence, Osaka, Japan) and numbers of cell fragments before and after the experiments were quantified using Fiji [31, 33].

RNA extraction, bulk sequencing and KEGG pathway analysis

RNA isolation from murine organoids for bulk RNA sequencing was described previously [16]. The original sequencing data are available in GEO (GSE282616) and were re-analysed in the present study for additional pathway analyses in the present study. Differentially expressed genes (DEGs) were identified using DESeq2. Genes with an adjusted p-value < 0.05 and |log₂ fold change|> 1 were considered significant. KEGG pathway enrichment was assessed using two complementary approaches implemented in the clusterProfiler R package. First, over-representation analysis (ORA) was performed on significantly regulated genes, with multiple testing correction using the Benjamini–Hochberg method. Pathways with an adjusted p-value < 0.05 were considered significantly enriched. In parallel, Gene Set Enrichment Analysis (GSEA) was conducted using the complete ranked gene list without prior filtering to identify pathways enriched across the entire transcriptome. Pathways with a false discovery rate (FDR) q-value < 0.05 were considered significant.

In silico analysis of JUP

A search for in silico evidence for physical interactions in volving JUP within the set of sequences from the MAPK, PKC, RAS and ERK families of protein was performed (Figure S4, 5). A STRING score is a composite of experimental, text mining, co-expression and co-evolutionary data for protein–protein interactions, both physical and functional. The IntAct database is a compendium of experimentally-observed PPIs – including co-IPs, Y2Hs, and other methods. While no evidence for interactions with the target set of genes was present in either STRING databases or physical interaction databases, the possibility of transient but meaningful interactions, in a cell-type or protein trafficking-dependant manner untested under experimental conditions to-date, cannot be ruled out. To investigate this effect further, we used Boltz-2. We assumed that interactions would be predominantly single domain-domain interactions, and so we separated the queries into their composite domains and then screened them in the presence of each of a fragmented plakoglobin (N-terminal disordered, central ordered N-term, central ordered C-term, C-terminal disordered). We take the top score of a protein’s domains as the score for that protein. With this screening framework in-place, we now aimed to minimize the risk of false-positives in the screening process. We hypothesised that these would arise primarily from; a variable number of domains and therefore ‘shots-on-target’ between query proteins, the nature of testing a large number of query proteins meaning that the (false discovery rate) FDR requires accounting for, and finally an unknown false-positive rate from the nature of the technique itself. The false-positive rate of the technique almost certainly varies by protein target. To evaluate the uncertainties outlined above, we used the data from the IntAct database search to calibrate our search model. We recognize that the IntAct data itself may carry false-positives and negatives, and so we expected conservative effect sizes, even in a well-calibrated model.

To account for a variable number of domains in the query proteins, which would bias the analysis in favour of proteins with more domains simply by random chance, we adjusted the predicted interaction likelihood by the number of trials per protein. We did this by building a null model from negative control proteins – defined as a random set of (26) human proteins and (176) domain-pairs that are not IntAct hits with plakoglobin. For a protein with m screened domain pairs, we fit a continuous model for expected ipTM values. To account for the non-independence of domains, we adjusted

where m is the number of domain pairs tested per protein, and m_eff is the effective number of independent domains, used for multiple-domain correction. This means that for each number of screened domains, a different null non-binding baseline is used, resulting in a dynamic model that can be applied to proteins containing any number of domains. To create a size-adjusted p-value, we take the probability that a protein fall is a member of the matched null-distribution. While an empirical model was also evaluated to control for a complex true distribution of data, the empirical model was highly sensitive to the exact composition of the negative controls, and was prone to over-estimating the significance of proposed hits. To evaluate the sensitivity and specificity of our approach, we benchmarked the scoring metric against a set of IntAct interacting partners.

Statistics

Statistical analyses were performed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA). Data are presented as mean ± standard error (SE). Statistical significance was assumed for p < 0.05. Paired Student’s t-test or Mann–Whitney-U-test was performed for two-sample group analysis after testing for normality using Shapiro–Wilk-test. Analysis of variance (ANOVA) followed by Tukey’s multiple comparisons or Šίdák’s multiple comparisons test and Bonferroni correction was used for multiple groups. The Kruskal–Wallis test was used for data that did not match the Gaussian distribution.

Results

Loss of plakoglobin promotes epithelial wound healing by increased proliferation and migration in Caco-2 monolayers

To determine the impact of plakoglobin (JUP) on intestinal epithelial cell dynamics, we performed siRNA-mediated Jup knockdown in Caco-2 cells (Fig. 1A, B). Unexpectedly, loss of JUP led to decreased paracellular permeability, as measured by 4 kDa FITC dextran flux across confluent Caco2 monolayers, compared to untransfected controls or non-target (n.t.) siRNA (Fig. 1C). However, JUP deficiency led to loss of intercellular adhesion after mechanical stress, as revealed by dispase-based enterocyte dissociation assays (Fig. 1D, E). Next, we used transepithelial electrical resistance (TER) measurements over time to assess the maturation of epithelial integrity in the presence and absence of JUP. TER measurements were initiated 24 h after seeding and transfection of cells with siRNA targeting JUP or n.t. siRNA, resulting in an accelerated increase in TER in JUP-deficient Caco-2 cells (Fig. 1F). Measurements of TER up to 72 h after initiation did not show convergence of the TER values with Caco2 cells with JUP-knockdown maintained increased barrier function over the time course of measurements. This suggests that JUP affects cell adhesion, migration and proliferation which potentially contributes to enhanced barrier maturation. To further explore the impact of JUP on cell dynamics, scratch wound assays were performed next. After 24 and 48 h, wound closure was significantly accelerated upon JUP knockdown (24 h: 44.60% ± 7.70; 48 h: 70.32% ± 11.54) compared to controls (24 h: 31.02% ± 5.90; 48 h: 58.04% ± 11.32) and n.t.-siRNA (24 h: 34.37% ± 5.74; 48 h: 51.34% ± 7.74; p < 0.001; Fig. 1G, H). Since both cell migration and proliferation contribute to wound healing, we next assessed whether loss of JUP influences these processes. Following mechanical wounding, immunofluorescence staining for Ki-67, a well-established marker of proliferating cells, revealed a significantly higher proportion of Ki-67–positive cells in JUP-deficient monolayers compared to control and n.t.-siRNA monolayers not only at the leading edge but also beyond (control: 13.95 ± 0.85%; n.t.-siRNA: 14.94 ± 0.78%; JUP-siRNA: 21.57 ± 0.82%; p < 0.01; Fig. 1I, J; Figure S1A). In parallel, expression of focal adhesion kinase (FAK), a marker of cell migration, revealed significantly enhanced phosphorylation of FAK following JUP knockdown (p < 0.001), while total FAK levels remained unchanged (Fig. 1K, L). Neither total protein levels nor cellular distribution of beta catenin, a known homologue of JUP, was changed following loss of JUP (Figure S1B-D). Taken together, these findings indicate that JUP deficiency promotes both epithelial proliferation and migration, thus accelerating wound repair in vitro.

Fig. 1.

Loss of plakoglobin (JUP) impairs epithelial barrier integrity and enhances wound healing in vitro. A Representative immunofluorescence images showing junctional localization of plakoglobin (JUP) in control Caco-2 cells and efficient loss of JUP following siRNA-mediated knockdown. Scale bars, 50 μm. B Representative immunoblot confirming siRNA-mediated JUP knockdown in Caco-2 cells. Non-targeting siRNA (n.t.) served as the control, and β-actin was used as a loading control. C Paracellular permeability measured by 4-kDa FITC-dextran flux across Caco-2 monolayers. JUP knockdown significantly increased epithelial permeability. Data are shown as mean ± SEM (n = 8 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. D, E Representative images (D) and quantification (E) of epithelial fragmentation following dispase assay, demonstrating reduced epithelial cohesion after JUP knockdown. Data are shown as mean ± SEM (n = 8 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. F Transepithelial electrical resistance (TER), measured by ECIS, showing impaired barrier formation in JUP-deficient monolayers. Data are shown as mean ± SEM and were analyzed by two-way repeated-measures ANOVA followed by Šidák’s multiple-comparisons test; *p < 0.05. G, H Representative scratch wound images at 0, 24, and 48 h (G) and corresponding quantification of relative wound closure (H), demonstrating accelerated wound healing following JUP knockdown. Scale bars, 50 μm. (n = 5 experiments). Data are shown as mean ± SEM (n = 6 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. I, J Representative immunofluorescence images (I) and quantification (J) of Ki-67-positive cells at the wound edge and the following cells beyond the leading edge 24 h after scratching, indicating enhanced proliferative activity in JUP-deficient cells. Ki-67, red; DAPI, blue. White arrows indicate proliferating cells. Scale bars, 50 μm. (n = 5 experiments). Data are shown as mean ± SEM (n = 12 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. K, L Representative immunoblot (K) and densitometric quantification (L) of phosphorylated FAK (p-FAK) relative to total FAK, demonstrating increased FAK activation following JUP knockdown. β-actin served as loading control. Each dot represents one independent biological experiment. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05

Jup deficiency activates MAPK signaling and accelerates wound closure in intestinal organoids

Intestinal organoids complemented our observations in Caco-2 monolayers, providing a physiologically relevant model without a cancer cell background. We generated intestinal epithelial organoids from VilCreER^T2Jup^fl/fl mice and induced JUP depletion by application of 4-OH-tamoxifen in culture as previously described [16]. Efficient knockout of Jup was confirmed by immunofluorescence and Western blotting (Fig. 2A-C). In addition to the tamoxifen-inducible knockout model, murine intestinal organoids with stable shRNA-mediated knockdown of Jup were generated to independently validate the observed phenotype. Efficient depletion of JUP was confirmed by Western blot analysis (Fig. 2D, E). In both models JUP deficiency led to accelerated wound closure in scratch wound assays of 2D murine intestinal monolayers. To distinguish between the contribution of cell proliferation and migration, proliferation was blocked by mitomycin C. Interestingly, addition of mitomycin removed the effect of Jup deficiency on wound closure in organoids, thus, indicating that JUP mainly regulates cell proliferation rather than migration (Fig. 2F, G).

Fig. 2.

JUP loss activates MAPK signaling in murine intestinal organoids. A Representative immunofluorescence images of murine intestinal organoids showing junctional plakoglobin (JUP) expression under control conditions and efficient loss of JUP following tamoxifen (tam)-induced recombination. Scale bars, 30 μm. B, C Representative immunoblot (B) and densitometric quantification (C) confirming efficient depletion of JUP protein in tamoxifen-treated organoids. β-actin served as loading control. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; **p ≤ 0.01. D, E Representative immunoblot with corresponding densitometric quantification confirms efficient JUP depletion in shRNA-treated organoids. β-actin served as the loading control. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; **p ≤ 0.01. F Representative scratch wound images and quantification of wound closure at 24 and 48 h in control and Jup-shRNA organoids treated with mitomycin C to inhibit cell proliferation. To inhibit cell proliferation, cells were treated with mitomycin C (5 µg/mL) prior to scratch formation. Scale bars, 200 μm. Data represent n = 3 independent experiments. G Quantification of scratch wound closure in control and Jup-shRNA organoids treated with mitomycin C (5 µg/mL) to inhibit cell proliferation. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by two-way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. H Sample distance heatmap of murine organoid profiles illustrating clear clustering and separation between control and tamoxifen-treated organoids. I KEGG pathway analysis (mmu05207) highlighting activation of MAPK-associated signaling pathways in JUP-deficient organoids, including Ras/PKC–Raf–MEK–ERK signaling and downstream regulators of proliferation, apoptosis inhibition, and invasion

RNA sequencing of organoids with and without Jup deficiency revealed clustering between the groups and up- and downregulation of several genes, as reported previously [16]. KEGG pathway analysis identified several pathways significantly enriched by JUP loss. In line with that, the chemical carcinogenesis-receptor activation pathway (mmu05207) demonstrated an enrichment score (ES) of 0.465 and a normalized enrichment score (NES) of 1.565, suggesting a substantial upregulation of genes involved in receptor activation mechanisms related to chemical carcinogenesis (p = 0. 00126, FDR-adjusted q-value of 0.01259). This included the enrichment of Ras and PKC-mediated signaling following JUP-deficiency, which converge into Raf/MEK/ERK1/2 and are collectively referred to as the MAPK/ERK signaling pathway (Fig. 2H, I).

Next, scratch wound assays using organoid-derived 2D monolayers were performed to further assess the relevance of Jup-dependent MAPK activation on a functional level. Consistent with these findings above, Western blot analyses demonstrated markedly increased ERK phosphorylation in Jup-deficient 2D organoids compared to controls following scratching (Fig. 3A, B). In line with this, elevated levels of Ki-67 and c-MYC were detected, both of which are well-established downstream targets of MAPK-driven proliferative activity [4]. As ERK phosphorylation represents a central downstream event of MAPK activation, this indicates robust engagement of MAPK signaling. Interestingly, the effect was completely abolished by ERK1/2 inhibition with FR180204, which also attenuated Ki-67 and c-MYC expression. Furthermore, consistent with our observations in Caco-2 cells, FAK phosphorylation was significantly increased in Jup-deficient organoids, indicating a potential role for JUP in promoting epithelial migration. Again, pharmacological inhibition of ERK1/2 using FR180204 completely abolished the effect of Jup deficiency on FAK phosphorylation, confirming that the observed phenotype is predominantly mediated via JUP-associated MAPK/ERK signaling.

Fig. 3.

JUP deficiency activates MAPK signaling and accelerates wound closure in 2D murine intestinal organoids. A Immunoblot analysis of murine intestinal organoids 24 h after scratching, showing tamoxifen-induced JUP depletion and downstream activation of MAPK/ERK signaling 24 h after scratching. Where indicated, organoids were additionally treated with the ERK inhibitor (FR180204, 5 µM) for a 24 h incubation prior to analysis. β-actin served as a loading control. B Quantification of protein expression in murine organoids 24 h after scratching. Results are shown as Ki-67 and c-MYC protein expression normalized to β-actin, as well as p-ERK/ERK and p-FAK/FAK ratios, demonstrating significant activation of MAPK signaling and proliferative responses upon JUP loss, which are attenuated by ERK inhibition. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01. C, D Representative scratch wound images and quantification of wound closure at 24 and 48 h in control, tamoxifen-treated, and inhibitor-treated organoids (ERK inhibitor, FR180204; Ras inhibitor, Daraxonrasib; Raf inhibitor, Sorafenib; p38-MAPK inhibitor, SB203580). Pharmacological inhibition of the ERK, Ras, and Raf pathways significantly reduced the accelerated wound closure induced by JUP loss, whereas p38-MAPK inhibition showed no significant effect. Scale bars, 200 μm. Data are shown as mean ± SEM (n = 3 experiments) and were analyzed by two- way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01. E, F Representative immunoblot and corresponding densitometric quantification of p-ERK/ERK in control, Jup-shRNA, and Jup-shRNA + Erk-shRNA organoids, with or without Ras or Raf inhibition. Combined knockdown of JUP and ERK abolished ERK activation and did not result in the enhanced wound-healing phenotype induced by JUP deficiency, confirming that JUP-mediated epithelial repair is dependent on ERK signaling. Data are shown as mean ± SEM (n = 3 experiments) and were analyzed by two-way ANOVA followed by Tukey's post hoc testing; **p ≤ 0.01, ***p ≤ 0.001

Quantification of scratch wound assays at 24 and 48 h confirmed accelerated wound healing following Jup knockout in organoids and Caco-2 cells (Fig. 3C, D; Figure S2A/B). Importantly, this effect was significantly blocked following pharmacological inhibition of ERK1/2 using FR180204 as well as of Ras by incubation with Daraxonrasib and Raf inhibitor Sorafenib as known upstream intermediates of the MAPK/ERK signaling pathway. In contrast, specific inhibition of p38-MAPK by SB203580 did not have any effect on cell dynamics following JUP deficiency. Furthermore, an additional knockout of Erk in Jup-deficient organoids removed the effect of Jup-deficiency on wound closure, thus, confirming the specificity of the MAPK/ERK pathway for JUP (Fig. 3E, F).

To investigate the negative regulatory role of JUP and exclude compensatory effects, we next induced overexpression of JUP in organoids (Fig. 4A, B). Overexpression of JUP in wildtype organoids trended to decreased epithelial wound healing compared to wildtype organoids. In comparison to JUP-deficient organoids, overexpression of JUP resulted in significantly reduced wound healing rates. Co-treatment of JUP-overexpressing organoids with ERK1/2 inhibitor or p38MAPK inhibitor had no effect on wound healing compared to wildtype controls (Fig. 4C, D). These findings confirm the specificity of the Jup-deficient phenotype, whereas the inhibitory effect of JUP overexpression appeared not to be modified by ERK or p38MAPK inhibition.

Fig. 4.

JUP regulates epithelial regeneration through MAPK/ERK signaling in murine and human organoids. A, B Representative immunoblot and corresponding densitometric quantification validating efficient JUP overexpression (Jup^OE) and shRNA-mediated Jup depletion (Jup-shRNA) in murine organoids. β-actin served as the loading control. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01. C, D Representative scratch-wound quantification (C) and images (D) of wound healing in murine organoids at 24 h and 48 h. JUP overexpression (Jup^OE) significantly impaired wound closure, whereas JUP depletion (Jup-shRNA) accelerated wound healing compared with control organoids. Pharmacological inhibition of ERK (ERK inh.) or p38-MAPK (SB203580) partially attenuated the enhanced wound-healing phenotype induced by JUP depletion. Data are shown as mean ± SEM (n = 5 experiments) and were analyzed by two-way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01. E, F Quantification (E) and representative bright-field images (F) of organoid fragments per well in a dispase-based enterocyte dissociation assays in 2D murine organoids 24 h after mechanical stress demonstrating increased organoid fragmentation following tamoxifen-induced Jup depletion, which was partially reversed by ERK inhibition. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01. G, H Representative immunoblot and corresponding densitometric quantification confirming efficient shRNA-mediated JUP depletion in human organoids. β-actin served as the loading control. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. I Quantification of scratch-wound healing in human organoids at 24 h and 48 h demonstrating significantly enhanced wound closure following shRNA-mediated JUP depletion. Data are shown as mean ± SEM (n = 4 experiments) and were analyzed by two-way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. J Representative scratch-wound images of human organoids at 24 h and 48 h demonstrating enhanced wound closure following shRNA-mediated JUP depletion

To further test for the role of Jup-deficiency on intercellular adhesion, we performed dispase-based enterocyte dissociation assays in 2D murine organoids 24 h after mechanical stress. Jup-deficient organoids displayed significantly increased fragmentation, consistent with impaired intercellular adhesion (Fig. 4E, F). Inhibition of ERK signaling using FR180204 did not alter the number of fragments in Jup-deficient organoids, thus demonstrating that loss of intercellular adhesion following Jup-deficiency was not modulated by interfering with MAPK/ERK signaling. To further validate these findings in a translational model, we investigated the functional role of JUP in human intestinal organoids. shRNA-mediated JUP depletion was first confirmed by immunoblotting and densitometric quantification (Fig. 4G, H). Consistent with our observations in murine organoids, shRNA-mediated JUP knockdown significantly accelerated wound closure in scratch-wound assays and was associated with enhanced epithelial cell motility (Fig. 4I, J). These findings indicate that the effects of JUP deficiency are conserved across murine and human intestinal epithelial models.

Taken together, these results show that loss of JUP promotes proliferation, migration, and reduces intercellular adhesion in intestinal epithelial cells. While the effects on proliferation/migration are MAPK/ERK-dependent, loss of intercellular adhesion was not restored when interfering with this pathway.

JUP deficiency induces MAPK/ERK-signaling which is associated with increased mucosal healing in vivo

To test the relevance of our findings in vivo, we employed an inducible, intestine-specific JUP knockout mouse model (iVilCreER^T2Jup^fl/fl), in which efficient Jup deletion following tamoxifen administration was confirmed (Fig. 5A, B). First, we assessed whether the proliferation rate of epithelial cells was overall increased as revealed by staining of Ki-67. Interestingly, epithelial cell proliferation was already increased under basal condition for Jup-deficient animals in the colon (Figure S3A, B). The impact of Jup deficiency on mechanically induced colonic wounds was evaluated using a colonoscopy-based mucosal wound-healing assay as previously described [13].The mucosal wound area was monitored up to 72 h after biopsy, and mucosal healing was calculated relative to the baseline measured at 24 h after wounding (Fig. 5C). In mice expressing intestinal epithelial JUP, mucosal restoration reached 39.8 ± 4.7% within 72 h after wounding. In line with the in vitro findings, mice deficient for intestinal epithelial Jup displayed significantly accelerated mucosal healing with a wound closure of 58.4 ± 5.1% of baseline (p < 0.05,Fig. 5D). To test whether MAPK signaling was also involved in wound-healing processes in vivo, we collected tissue samples from outside the wound area (nw) and directly from the wound region (w). Consistent with our findings in organoids, ERK phosphorylation and c-MYC protein levels were markedly increased in the wound area of Jup-deficient animals compared to controls and non-wounded areas (Fig. 5E, F). Furthermore, we observed significantly increased levels of pFAK in JUP-deficient epithelial wounds, while Ki-67 protein expression was increased in wounds with and without JUP, indicating ongoing mucosal healing. In line with our in vitro findings, the effect of JUP loss appeared to be most pronounced during the early phase of repair, particularly within the first 48 h, whereas after 72 h the strong proliferative advantage of JUP-deficient mice diminished. Taken together, these results demonstrate that loss of JUP activates MAPK/ERK signaling in vivo, thus providing the mechanistic background for the enhanced epithelial proliferation and wound healing observed in Jup-deficient mice.

Fig. 5.

Intestinal epithelial JUP deficiency accelerates mucosal wound closure through MAPK/ERK activation in vivo. A Representative immunofluorescence staining of colonic tissue sections from Jup^fl/fl control mice (left) and iVilCreER^T2Jup^fl/fl mice (right) showing epithelial JUP (red) and nuclear DAPI (blue). Efficient loss of epithelial JUP is observed following tamoxifen-induced Cre activation. The lower panels show higher-magnification images of the boxed regions. B Representative immunoblot analysis of JUP expression in heart, small intestine, and colon from Jup^fl/fl and tamoxifen-induced iVilCreER^T2Jup^fl/fl mice with corresponding densitometric quantification. Efficient JUP depletion was restricted to the intestinal epithelium, whereas cardiac JUP expression remained unaffected. β-Actin served as loading control. Data are shown as mean ± SEM (n = 4) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; **p ≤ 0.01. C Representative images of the in vivo biopsy-based mucosal wounding model after 24 h (baseline) and 72 h. Colonic wounds were induced by endoscopic biopsy, followed by re-endoscopy at 24 h (baseline documentation) and 72 h for assessment of wound healing (7–8 wounds/mouse). Scale bars, 1 mm. D Quantification of mucosal wound healing expressed as percentage of baseline wound area, demonstrating significantly enhanced epithelial repair in iVilCreER^T2Jup^fl/fl mice compared with Jup^fl/fl. Data are shown as mean ± SEM (n = 3 experiments, each data point represents one individual mouse) and were analyzed by two‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05. E Immunoblot analysis of colonic tissue lysates from Jup^fl/fl control mice and iVilCreER^T2Jup^fl/fl mice under non-wounded (nw) and wounded (w) conditions. Protein expression of Ki-67, c-MYC, phosphorylated ERK (p-ERK), total ERK, phosphorylated FAK (p-FAK), and total FAK is shown; β-actin served as a loading control. F Quantification of protein expression in vivo in wounded and non-wounded areas. Results are demonstrated as Ki-67 and c-MYC protein expression normalized to β-actin, as well as p-ERK/ERK and p-FAK/FAK ratios. Wounding induces activation of proliferative and MAPK/ERK-associated signaling pathways, which is significantly enhanced in iVilCreER^T2Jup^fl/fl mice compared to Jup^fl/fl controls. Data are shown as mean ± SEM (n = 2 experiments, each data point represents one individual mouse.) and were analyzed by 1‐way ANOVA followed by Tukey's post hoc testing; *p ≤ 0.05, **p ≤ 0.01; n.s., not significant

Discussion

This study highlights a novel role for JUP in regulating epithelial cell dynamics by modulating MAPK/ERK signaling. Using an inducible, intestine-specific JUP knockout mouse model (iVilCreER^T2Jup^fl/fl) together with murine organoids and Caco-2 cells, we demonstrate that loss of JUP induces epithelial wound repair which is predominately driven by increased proliferation through the activation of MAPK/ERK signaling. While these effects were attenuated by a pharmacological inhibitor and by ERK knockdown, impaired cell adhesion caused by JUP deficiency was unaffected by interfering with MAPK/ERK signaling. This highlights the role of JUP as an important stabilizer of intercellular adhesion. In addition, JUP plays a dual role, acting as a crucial molecular scaffold that modulates signaling pathways essential for the regulation of epithelial proliferation and thereby contributing to intestinal mucosal homeostasis.

Loss of JUP in epithelial cells leads to enhanced epithelial cell motility in vitro and in vivo

JUP is a key component of desmosomal junctions and essential for maintaining epithelial tissue integrity [30]. In a previous study, we focused on the role of JUP in the context of inflammation and identified JUP as an important signaling hub for p38MAPK-dependent activation of the inflammasome [16]. Here, we focused on an observation independent from the inflammatory setting, namely the functional consequences of JUP deficiency in intestinal epithelial cells under basal conditions and following mechanical injury. An unexpected result from our in vitro experiments was that barrier function was increased following loss of JUP although we had not detected an effect on intestinal barrier function in the in vivo situation before [16]. This discrepancy may be explained by the high sensitivity of the in vitro assays, which detect cell-intrinsic changes under highly controlled conditions. In vivo, however, epithelial phenotypes are affected by interactions with the microbiota, immune system, extracellular matrix, and surrounding stromal cells which may compensate for or mask subtle epithelial alterations caused by JUP deficiency. The significant observation of this study is that JUP deficiency significantly accelerated wound closure in scratch-wound assays performed in Caco-2 cells as well as in 2D murine and human organoids. Vice versa, overexpression of JUP in wildtype organoids trended to impair wound closure. Those results were confirmed in biopsy-based in vivo wound-healing experiments, in which JUP deficiency improved wound healing. Immunofluorescence images of wounds revealed that the significant finding correlated with the amount of proliferating epithelial cells, similar to that observed in Caco-2 cells and organoids. In line with that, addition of a proliferation inhibitor blunted the observed finding of JUP deficiency on wound closure, thus, confirming the proliferative phenotype after loss of JUP. At first glance, one might assume that these findings are in contrast to previous studies, which demonstrated that loss of desmosomal proteins DSG2 and DSC2 contributes to impaired wound repair and barrier stabilization [7, 9, 11]. However, the observation that reducing JUP elicits the opposite effect highlights a distinct, protein-specific role for individual desmosomal components in regulating epithelial repair and barrier function. In support of our current observation, various studies in non-intestinal tissue such as skin, prostate, and breast revealed that loss of JUP is markedly associated with increased proliferation and migration [1, 12, 14, 22]. Furthermore, a recent review article on the role of JUP on cancer development highlighted its complex and inconsistent role in colorectal cancer. So far, knowledge about the molecular mechanisms of JUP affecting cell proliferation and migration remains limited since only a small number of studies exist with most of the results being based on different cancer cell lines [6]. Thus, previous in vitro investigations reported opposing results which emphasize the relevance of more detailed analyses by in vivo techniques and the investigations of patient specimens. Therefore, as almost all of these studies were conducted in cells with a cancer background, our observations represent the first confirmation of JUP being involved in the regulation of proliferation in intestinal epithelial cells in a non-cancerous context. In conclusion, the presented results demonstrate that JUP is critically involved in epithelial cell dynamics, i.e. cell proliferation and migration in the gut.

JUP regulates epithelial cell motility through MAPK signaling

One of the central signaling modules regulating intestinal epithelial regeneration in response to extracellular cues is the mitogen-activated protein kinase (MAPK)/extracellular signal-regulated kinase (ERK) pathway, which drives wound-induced repair processes [5]. This concept is further supported by recent findings demonstrating that age-dependent activation of interferon-γ (IFN-γ) signaling together with suppression of MAPK/ERK activity establishes the balance between active and quiescent states of Lgr5⁺ intestinal stem cells [28]. A novel finding reported here is that loss of JUP induces strong activation of MAPK/ERK signaling in organoid models in vitro and in vivo, as revealed by biopsy-induced wound-healing assays in animals with inducible, selective deficiency of JUP in the intestinal epithelium. This was evident in unbiased gene-set enrichment analysis of bulk RNA-sequencing data from organoids, as well as in protein levels measured by Western blots, both in vitro and in vivo. Inhibition of the MAPK pathway by an ERK inhibitor blunted enhanced wound closure and activation of Ki-67 as well as phosphorylation of c-MYC in JUP-deficient organoids. Similarly, double knockout of Jup and Erk in organoids removed the effect of JUP deficiency on cell dynamics, thus, confirming the functional relevance of JUP-associated MAPK activation in promoting epithelial repair. These results align with the growing evidence that intestinal desmosomes are not merely structural adhesion complexes but also act as important signaling platforms regulating epithelial homeostasis. This is evident for the epidermis representing a multilayered epithelium and for the heart in which JUP-dependent signaling is crucial for desmosome integrity [2, 37]. Furthermore, it has been shown that DSG2 interacts with EGFR via Src, thus, controlling cell adhesion and proliferation [34]. Furthermore, complex in vitro and in vivo studies revealed the DSC2-associated modulation of integrins and Rap1 as well as the direct effect of DSC2 on Desmoplakin and intermediate filament interaction to regulate cell adhesion and barrier function in the gut [7].

KEGG pathway analysis indicates that this effect is mainly driven by the regulation of Ras activity, as JUP deficiency increased Ras activation and MAPK signaling. Indeed, addition of a Ras inhibitor removed the effect of JUP deficiency on wound closure in organoids while blocking of p38MAPK had no effect. In pancreatic ductal adenocarcinoma, a correlation between changes in plakoglobin expression and modulation of MAPK signaling has been observed, thus supporting our current observation for intestinal epithelial cells [29]. To the best of our knowledge, this is the first description of JUP specifically regulating Ras-ERK-MAPK signaling, a pathway relevant not only to intestinal inflammation but also to colorectal cancer. In previous reports, JUP has been shown to modulate transcriptional complexes in an oncological context [21]. However, this mechanism appears less likely in the context reported here since we observed changes in pathway components predominantly localized in the cytosolic compartment. Instead, it can be speculated that JUP acts as a scaffold protein, similar to the situation we observed with the JUP-p38MAPK interaction in the context of inflammation before [16]. Attempts to identify a comparable direct interaction between MAPK/ERK and JUP were not successful as revealed by an in silico analysis of JUP physical interactions (Figure S4, 5). Based on the observation that erbin is present in the “proximitome” of desmosomes, an indirect regulation of MAPK/ERK signaling by JUP via erbin is conceivable [8]. On the other hand, one could speculate that it might be a receptor-driven effect. Despite the limitation that the precise mechanisms of JUP-dependent MAPK regulation remain to be determined, our current findings support an emerging role for JUP as an important regulatory switch in modulating cell signaling.

MAPK signaling regulates cell proliferation independently of cell adhesion

To further dissect the functional consequences of JUP loss on intercellular adhesion, we employed a dispase-based enterocyte dissociation assay that we have used previously [27]. Consistent with its role as a desmosomal component, JUP-deficient monolayers showed markedly increased fragmentation, reflecting impaired intercellular adhesion following mechanical stress. Notably, inhibition of ERK signaling did not affect fragmentation in 2D murine organoids, indicating that the adhesive function of JUP is independent of MAPK/ERK activity.

These findings highlight a dual role for JUP in the complex epithelial biology of the gut. At the structural level, JUP stabilizes intercellular junctions, thereby maintaining tissue integrity. On the other hand, JUP modulates intracellular signaling via MAPK/ERK, thereby driving epithelial migration and proliferation. The accelerated wound-closure observed in JUP-deficient epithelia is therefore not only the result of weakened adhesion but rather reflects a signaling-driven proliferative phenotype. Importantly, this effect was not only observed in murine but also human organoids, thus, highlighting the potential translational relevance of the detected findings. From a broader perspective, the uncoupling of adhesion and signaling functions is highly relevant, as it positions JUP as a molecular switch that balances epithelial stability with regenerative capacity. In pathological settings such as inflammation or colorectal cancer, loss of JUP may shift this balance towards hyperproliferation, whereas in the context of mucosal injury, JUP deficiency might transiently promote efficient repair. These insights indicate that JUP acts as a critical mediator of epithelial signal transduction, highlighting its dual role as both a structural and regulatory protein. Following that, the observed effects of JUP deficiency on epithelial cell behavior have significant implications for intestinal diseases. Our findings indicate that loss of JUP plays a pivotal role in regulating cellular motility, which not only accelerates wound healing but, when dysregulated, may contribute to tumorigenesis by promoting invasive and metastatic behaviors. Interestingly, a recent study by Kücükköse et al. demonstrated that JUP-dependent cell clusters contribute to lung-specific metastases in colon cancer, further supporting a dual role for JUP in maintaining homeostasis and excessive epithelial cell activity [20]. Therefore, a deeper understanding of the role of JUP could provide valuable insights into therapeutic strategies not only for IBD but also beyond. In this context, identifying MAPK activation as a downstream consequence of JUP deficiency opens new avenues for therapeutic intervention even though we did not find evidence for a direct interaction. The ability of RAS, RAF and ERK1/2 inhibitors to mitigate the accelerated wound-healing phenotype underscores the therapeutic potential of targeting MAPK signaling in conditions where epithelial repair is dysregulated. In line with that, future studies should evaluate and focus on a potential receptor and intermediate molecules as well as therapeutic effects of targeting plakoglobin in human models and intestinal diseases.

Taken together, our study demonstrates that JUP might be a critical regulator for intestinal epithelial cell dynamics and mucosal repair. We identify a novel mechanism by which JUP-deficiency activates the MAPK/ERK signaling pathway to predominantly enhance epithelial cell proliferation. While this finding could have potential therapeutic relevance, future studies will need to explore whether direct targeting of JUP can improve mucosal healing and disease outcomes in conditions of intestinal inflammation, such as IBD and colorectal cancer.

Authors’ contributions

All authors have made substantial contributions to conception or design and approved the submission. Marius Hörner: Writing – original draft, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization. Natalie Burkard: Writing – review & editing, Writing – original draft, Visualization, Validation, Project administration, Methodology, Investigation, Data curation, Conceptualization. Friedrich Forchel: Investigation, Data curation. Andre Schmidt: Investigation, Data curation. Catherine Kollmann: Visualization, Resources, Data curation. Christoph Otto: Validation, Resources, Formal analysis, Data curation. Aarushi Mukherjee: Validation, Formal analysis, Data curation. Stefanie Schmidt: Validation, Formal analysis, Data curation. Christoph-Thomas Germer: Resources, Data curation. Sven Flemming: Resources, Data curation. Timo Kohler: Formal analysis, Investigation, Methodology, Software, Validation, Writing – review & editing. Matthew Penner: Formal analysis, Investigation, Methodology, Software, Validation, Visualization, Writing – review & editing. Franziska Vielmuth: Resources, Investigation, Data curation. Daniela Kugelmann: Formal analysis, Investigation, Methodology. Kai Kretzschmar: Validation, Software, Resources, Conceptualization. Brenda Gerull: Validation, Data curation, Conceptualization. Babak Saravi: Validation, Formal analysis, Data curation. Nicolas Schlegel: Writing – review & editing, Supervision, Resources, Project administration, Investigation, Funding acquisition, Conceptualization. Matthias Kelm: Writing – review & editing, Writing – original draft, Supervision, Resources, Project administration, Investigation, Funding acquisition, Conceptualization.

Funding

This work was supported by the Deutsche Forschungsgemeinschaft (DFG SCHL1962/8–1 to NS, KE 2402/3–1 to MK). This work has been supported by the DFG-funded consortium CRC425 (TPC01, C02, C05 and A05 and CP2).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.