Accepted Article

Original Research Article TNF-related weak inducer of apoptosis (TWEAK) regulates junctional proteins in tubular epithelial cells via canonical NF-B pathway and ERK activation† Sergio Berzala#, Cristian González-Guerreroa#, Sandra Rayego-Mateosb¥, Álvaro Uceroa¥, Carlos Ocañaa, Jesús Egidoa,c, Alberto Ortiza,c, Marta Ruiz-Ortegab, Adrián M. Ramosa* a

Laboratory of Nephrology and Vascular Pathology, Instituto de Investigación Sanitaria-Fundación Jiménez Díaz (IIS-FJD), Av. Reyes Católicos 2, 28040, Madrid, Spain b

Cellular Biology in Renal Diseases Laboratory, Universidad Autónoma de Madrid, Instituto de Investigación Sanitaria-Fundación Jiménez Díaz (IIS-FJD), Av. Reyes Católicos 2, 28040, Madrid, Spain c

Fundación Renal Íñigo Álvarez de Toledo (FRIAT), c/ José Abascal 42, 28003, Madrid, Spain

# ¥ These authors contributed equally to this work

*Address correspondence to: Dr Adrián M. Ramos. Laboratorio de Patología Renal y Vascular (Investigación, 4º planta), Instituto de Investigación Sanitaria-Fundación Jiménez Díaz, Av. Reyes Católicos Nº2, Madrid (CP28040), Spain. Fax: +34-915-442636 E-mail: [email protected]

Running Head: TWEAK modify junctional proteins behavior Keywords: TWEAK, tubular cells, junctional proteins, EMT, NF-B, ERK, VDR

Funding Instituto de Salud Carlos III (Fondos FEDER ISCIII-RETIC REDINREN RD06/0016/0004 and 0003, RD12/0021/0001 and 0002, PI08/1083, PI11/02242, PI11/01854, PI041/00041, PI10/00072 and PS09/00447), Comunidad de Madrid (Fibroteam S2010/BMD-2321, CIFRA S2010/BMD-2378), Instituto Reina Sofia de Investigación Nefrológica (FRIAT).

†

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: [10.1002/jcp.24905] Received 25 July 2014; Revised 10 December 2014; Accepted 18 December 2014 Journal of Cellular Physiology This article is protected by copyright. All rights reserved DOI 10.1002/jcp.24905

This article is protected by copyright. All rights reserved

Accepted Article

ABSTRACT The tubular epithelium may be intrinsically involved in promoting kidney injury by junctional

instability, epithelial-mesenchymal transition (EMT) and extracellular matrix remodelling. In this work, we

investigated whether the pleiotropic and proinflammatory cytokine tumor necrosis factor-like weak inducer of apoptosis (TWEAK), could be able to disturb junctional protein expression and to induce EMT of tubular

cells. In cultured murine proximal tubular cells TWEAK induced phenotypic changes that were accompanied by F-actin redistribution, loss of epithelial adherent (E-cadherin, Cadherin-16, -catenin) and tight junction (ZO-1) proteins, and re-expression of the mesenchymal protein Vimentin. The transcriptional repressors Snail and HNF1 were also modulated by TWEAK. In a murine model of obstructive renal pathology, TWEAK expression correlated with the appearance of the mesenchymal marker SMA in kidney tubular cells. Mechanistically, the epithelial changes induced by TWEAK, including loss of epithelial integrity and EMT, via Fn14 were TGF-1 independent, but mediated by several intracellular signaling systems, including the

canonical NF-B, ERK activation and the vitamin D receptor modulation. These results highlight potential contributions of TWEAK-induced inflammatory mechanisms that

could unveil new pathogenic effects of TWEAK starting tubulointerstitial damage and fibrosis. This article

is protected by copyright. All rights reserved

This article is protected by copyright. All rights reserved

Accepted Article

INTRODUCTION Tumor necrosis factor-related weak inducer of apoptosis (TWEAK) is a pleiotropic cytokine belonging

to the tumor necrosis factor superfamily (TNFSF) (Winkles, 2008; Sanz et al, 2011). The coupling of TWEAK along with its hitherto only known signaling receptor, fibroblast growth factor-inducible 14 (Fn14), has been involved in the pathogenesis of acute and chronic kidney injury (Poveda et al, 2013; Ruiz-Ortega et al, 2014). Thus, it has been shown that TWEAK regulates processes with potential pathophysiological relevance for human kidney damage such as cell death, proliferation and inflammation (Winkles, 2008; Sanz et al, 2011). In reference to this latter process, research from our group established that in tubular cells, TWEAK activates the canonical and non-canonical nuclear factor κB (NF-κB) pathways to promote the synthesis of inflammatory mediators (Sanz et al, 2008, 2010). Moreover, we have recently described that TWEAK knockout mice are protected from renal fibrosis induced by unilateral ureteral obstruction (UUO), mainly by decreasing fibroblast proliferation. By contrast, TWEAK overexpression in mice increased proliferation of renal fibroblasts and fibrosis (Ucero et al, 2013). A sensitive component of tubular epithelium functionality is the maintenance of intercellular

interactions through adherens (AJ) and tight (TJ) junction proteins. Indeed, the loss of these plasmatic membrane structures is thought to be clinically relevant in mediating kidney injury under several pathological conditions. These include nephrotoxicity, ischemia, obstruction and inherited conditions leading to cystic formation (Kwon et al, 1998; Prozialeck et al, 2007; Girshovich et al, 2012; Wilson, 2011). Epithelial

junctional proteins are susceptible targets for endogenous deleterious factors released by the intrinsic renal tissue or by local and infiltrating immune cells during kidney injury. There are many mechanisms by which these factors, namely cytokines, metalloproteinases and caspases, disrupt intercellular junctional bridges. This is because they can either directly interact with particular AJ or TJ proteins or because they trigger signaling pathways resulting in AJ or TJ transcriptional protein regulation. Thus, the breakdown of the kidney epithelial sheets causes loss of functional properties such as permeability and polarity, which in addition may result in decreased viability and apoptosis. The loss of AJ and TJ proteins may also occur when tubular cells react against several harmful or

fibrogenic stimuli undergoing epithelial-mesenchymal transition (EMT), an adaptive process of cell plasticity that forces cells to lose epithelial proteins and re-express mesenchymal proteins (Kalluri and Neilson, 2003).

This article is protected by copyright. All rights reserved

Accepted Article

In kidney, the EMT is induced by many agents and it is mainly regulated at the translational level by a number of transcription factors. Between them, the members of the Snail family of transcriptional repressors may trigger the EMT by integrating signals from numerous intracellular pathways, the most important of which are induced by TGFβ (Boutet el al, 2006; Moustakas and Heldin; 2007). EMT bestows cells with some advantages to favor tissue regeneration, such as an increased motility and the capacity to synthesize extracellular matrix from which new tissue can grow. However, cells in EMT that become matrix-producing myofibroblasts could also contribute to a deregulated tissue repair and fibrosis when the EMT process is chronically fuelled (LeBleu et al, 2013). Furthermore, apart from being potentially involved in kidney renal fibrosis by itself, the EMT has been proposed to be an early biomarker for the occurrence of a pathological fibrotic process (Hazzan et al, 2011). The release of inflammatory cytokines is thought to break down tubular AJ and TJ, thus

compromising tubular cell barrier integrity. TNFα, the conspicuous member of the TNFSF, has been shown to

alter the expression or location of protein complex forming tight junctions in tubular cells (Poritz L et al, 2004). Moreover, TNFα is an important tumorigenic factor and EMT inducer involved in E-cadherin repression, tumor growth and spreading of renal carcinoma cells (Chuang MJ et al, 2008; Ho MY et al, 2012). Aside from TNFα, the role of other TNFSF members in regulating the fate of junctional proteins has not been extensively addressed. Our previous works demonstrated that TWEAK may globally behave as TNF does, although intrinsic differences in molecular mechanisms suggest that both cytokines are not redundant in renal injury. A role of TWEAK disrupting epithelial cell barrier was suggested in inflammatory bowel disease, but it was mostly related to a secondary consequence of cell death rather than a direct effect on cell adhesion complexes (Dohi and Burkly, 2012). Besides, on kidney tubular cells it is also conceivable that as consequence of the inflammatory reaction, TWEAK in combination with other cytokines such TNFα and INFγ, may kill tubular cells with the resulting junctional loss (Justo et al, 2006). However, potential direct actions of TWEAK on protein barrier regulation and function are presently unknown. In the present work, our aim was to evaluate whether TWEAK could also impinge on the behavior of

epithelial proteins involved in cell to cell contacts. We also wanted to know whether TWEAK, in addition to its reported capacity to promote fibrosis through fibroblast proliferation, could act on tubular cells to

This article is protected by copyright. All rights reserved

Accepted Article

transform them into a source of matrix-producing myofibroblasts. Molecular mechanisms focused on inflammatory mediators that could explain these potential new properties were also addressed.

MATERIALS AND METHODS Cells and reagents MCT cells are a cultured line of murine proximal tubular epithelial cells originally obtained from Eric

Neilson (Vanderbildt University, Nashville, TN). Human kidney proximal tubular (HK2) and canine kidney epithelial MDCK-II cell lines were obtained from the American Type Cell Culture (ATCC). MCT and human HK2 cells were cultured in RPMI 1640 (GIBCO, Grand Island, NY) supplemented with 10% decomplemented fetal bovine serum (DFBS), 2mM glutamine, 100 U/ml penicillin, and 10 mg/ml streptomycin, in 5% CO2 at 37ºC. Insulin-Transferrin-Sodium Selenite supplement (100 g/ml) and

hydrocortisone (5 ng/ml) were added to HK2 culture media. Cells were overnight incubated with RPMI without DFBS and stimulated when 70-80% confluent. MDCK-II cells were cultured in DMEM containing 1mM glucose and supplemented with 10% DFBS, 2mM glutamine 100 U/ml penicillin, and 10 mg/ml streptomycin, in 5% CO2 at 37ºC. Recombinant human TWEAK from Chemicon, Millipore (Billerica, Massachusetts) and the following inhibitors (specificity indicated between brackets) were used: neutralizing anti-TGFb1 antibody from R&D Systems (Minneapolis, MN); Bay 11-7082 (IB) and PD98059 (ERK) from Calbiochem (Merck Chemicals, Darmstadt, Germany); blocking ITEM-4 clone antibody (Fn14) from BioLegend and lactacystin (proteasome) from Sigma-Aldrich (Madrid, Spain). Paricalcitol (Zemplar) was obtained from Abbot Laboratories.

Gene expression studies One g RNA isolated with Tripure (Roche, Spain) was reverse transcribed with High

Capacity cDNA Archive Kit and real-time PCR was performed on an ABI Prism 7500 PCR system (Applied Biosystems, Foster City, CA) using the DeltaDelta Ct method. Expression levels are given as ratios to GAPDH. Pre-developed primer and probe assays were all from Applied Biosystems.

This article is protected by copyright. All rights reserved

Accepted Article

Western blot Protein content from cell extracts homogenized in lysis buffer (50 mmol/L Tris, 150 mmol/L NaCl, 2 mmol/L EDTA, 2 mmol/L EGTA, 0.2% Triton X-100, 0.3% NP-40, 0.1 mmol/L PMSF, 25 mmol/L NaF) was determined by the bicinchoninic acid method (Pierce Biotechnology, Rockford, IL). Proteins were separated by 10% SDS-PAGE under reducing conditions and then blotted onto nitrocellulose membranes. Membrane blockade was accomplished with 5% defatted milk in TBS-T (0.05 mol/L Tris, 0.15 mol/L NaCl, 0.05% Tween 20, pH 7.4). Thereafter, membranes were overnight probed at 4ºC with specific primary antibodies made in the same blocking solution or 5% BSA in TBS-T and then incubated with secondary HRP-conjugated antibodies for 1h at room temperature. The following primary antibodies were used for MCT experiments: mouse polyclonal anti E-cadherin and anti Vimentin (BD Pharmingen, San Diego, CA); mouse monoclonal anti Snail (Cell Signaling Technology, Danvers, MA); rabbit polyclonal anti ZO-1 (Zymed-Invitrogen, San Francisco, CA), anti Slug (Abcam, Cambridge, UK), anti -catenin and anti VDR (Santa Cruz Biotechnology, Santa Cruz, CA); goat polyclonal anti N-cadherin (Santa Cruz). In HK2, E-cadherin and Vimentin were detected with rabbit monoclonal antibodies (Cell Signaling Technology). Mouse monoclonal anti -tubulin (Sigma-Aldrich) and mouse polyclonal anti GAPDH (Chemicon International, Temecula, CA) were used to correct minor differences in protein loading. Gel zymography

Supernatants of cultured MCT treated with TWEAK for 48 h were collected and centrifuged at 5000 rpm for 10 minutes at 4°C. MMP-9 activity was analyzed by gelatin zymography using 8% polyacrylamide stacking gel and 8% polyacrylamide resolving gel containing 10 mg/ml gelatin. Samples were mixed with SDS sample buffer under non-reducing conditions. After electrophoresis, gels were incubated in renaturing buffer (2.5% Triton X-100 in distilled water) at RT for 30 min with a gentle shaking. Then, gels were incubated in developing buffer (1.5 M Tris–HCl pH 7.5; 200 mM NaCl; 6.6 mM CaCl2; 0.02 % Igepal) at RT. After 30 min gels were placed at 37ºC overnight in the same buffer. Gels were stained with 0.25% Coomassie Brilliant Blue (R-250) dye in 50% methanol, 10% acetic acid and 40% distilled water and distained in the same solution without the dye. Specificity of metalloprotease activity was corroborated by addition of metalloproteinase inhibitors in the renaturing buffer.

Immunofluorescence

Cells were fixed in 4% paraformaldehyde/PBS, permeabilized in 0.1% Triton X100/PBS, washed in 1% bovine serum albumin (BSA)/PBS, and blocked with 4% BSA/PBS. The following primary antibodies were used: rabbit polyclonal anti-catenin (Santa Cruz Biotechnology), anti-ZO-1 (ZYMED, San Francisco, CA), anti-Snail (Abcam) and mouse polyclonal anti Vimentin (BD Pharmingen) and rat monoclonal

This article is protected by copyright. All rights reserved

Accepted Article

anti E-cadherin (M. Takeichi’s Lab., Kyoto University, Japan). F-actin cytoskeleton was stained with Phalloidin-Alexa568 (Invitrogen, Paisley, U.K.). Cells were incubated with fluorescein isothiocyanate secondary antibodies (Sigma-Aldrich) and nuclei counterstained with propidium iodide. Cells were analyzed using a Confocal System TCS SP5 (Leica, Madrid, Spain).

Transfection assays Cells at 60% of confluence were overnight transfected with 50nmol/L mouse EGFR and VDR small interfering RNA (siRNA) oligonucleotides (Silencer Select® siRNA and Stealth Select RNAiTM, Ambion, respectively) by using lipofectamine RNAi MAX reagent (Invitrogen) in Optimem (Gibco) without DFBS or antibiotics. After transfection, culture medium was replaced by fresh complete RPMI medium and protein expression was evaluated by western blot of total protein extracts after 72 h.

Wound healing assay HK2 cells were grown until confluent and then serum depleted overnight. Plates were scratched with a pipette tip, washed with PBS and placed into fresh medium without DFBS. Cells were treated with the stimulus or left untreated for 72 h. Images were taken with a contrast phase microscope throughout the experiment.

Histological studies Tissue immunofluorescence was carried out in 5 µm thick paraffin-embedded tissue sections. The slides were deparaffinized with xylene and graded concentrations of ethanol and then rehydrated. Buffer citrate was the antigen retrieval method. For immunofluorescence, after antigen retrieval, slides were blocked in PBS with 6% BSA and 10% sheep serum. After the blocking step, tissue sections were incubated with anti- α-smooth muscle actin (αSMA) antibody (1:200, Dako) and further revealed with an anti-αSMA conjugated with Alexa Fluor™-488 (1:500, Invitrogen). Both antibodies were incubated for 1 hour at room temperature. After PBS washes, nuclei were counterstained with DAPI (Sigma) and unspecific staining eliminated with PBS. Slides were mounted in aqueous Mowiol medium (Sigma). Images were obtained with a SP5 confocal microscopy (Leica, Madrid, Spain). Unilateral Ureteral Obstruction (UUO) Studies were conducted according with the NIH Guide for the Care and Use of Laboratory Animals.

All animal procedures were performed with prior approval by the Ethics Committee of Animal Welfare of the IIS-FJD. TWEAK-KO or wild type (WT) C57BL/6 mice provided by Biogen Idec (Campbell et al, 2006; Ucero et al, 2013) to 14-week-old) were anesthetized and the left ureter was ligated (n=5-6/group) or manipulated without ligation (sham, n=4/group). Peri-operative analgesic was administered to the mice.

This article is protected by copyright. All rights reserved

Accepted Article

Animals were sacrificed 48 h or 7 days after surgery. Kidneys were perfused in situ with cold saline before removal. Half kidney was snap-frozen in liquid nitrogen for RNA and protein studies, and the other half was fixed and paraffin-embedded for histological studies.

Statistics Results are expressed as Mean ± SD. Each value corresponds to a minimum of three experiments. Mann-Whitney was applied to detect differences between groups. A p < 0.05 was considered statistically significant. The statistic analysis was performed with Windows SPSS 11.0 package (SPSS Inc, IL). RESULTS TWEAK induces changes in apical junction proteins in renal tubular cells Treatment of murine tubular cells (MCT) with TWEAK for 48 h resulted in the progressive

modification of their normal epithelial architecture, characterized by the shift from a cubical shape into a more elongated one and the loss of intercellular contacts and cluster spatial arrangement. These phenotypic alterations were also accompanied by modifications in the staining pattern of F-actin, which redistributed from the plasma membrane to the cytoplasm to form short and thick stress fibers (Figure 1A). Taken as a whole, these morphological changes induced by TWEAK were suggestive of junctional disassembly. Indeed, TWEAK decreased gene transcription of key epithelial proteins that normally contribute to keep tight intercellular contacts, namely the prototypic adherent junction components E-cadherin and the kidney specific Cadherin-16 (Figure 1B). Moreover, TWEAK modified the mRNA expression levels of representative TJ members, decreasing ZO-1 synthesis and increasing Claudin-1 (Figure 1B). Corresponding changes in protein expression levels of junctional proteins were also observed at 48h. Thus, TWEAK downregulated Ecadherin, β-catenin, N-cadherin (Figure 1C) and ZO-1 (Figures 1C and 1E). E-cadherin changes were further characterized by immunofluorescence in the MDCK cell line, which is prototypically used to study junctional protein behavior. Whereas untreated control cells expressed high levels of E-cadherin along the whole cell membrane and mainly at the intercellular junctions, TWEAK-treated cells showed the protein much less expressed and disposed in a discontinuous pattern (Figure 1F). TWEAK also modulates the expression of members of the Snail family of transcription factors, which are actively involved in repressing

This article is protected by copyright. All rights reserved

Accepted Article

renal E-cadherin and Cadherin 16 (Boutet et al. 2006). Thus, Snail protein expression, but not gene transcription, was increased (Figure 1D and 1G, respectively), whereas Slug protein and mRNA levels were both upregulated (Figures 1D and 1G). TWEAK also downregulated the mRNA trasncription of the Snail

target HNF-1, a differentiation factor that helps to maintain basal Cadherin-16 expression levels (Boutet et al. 2006) (Figure 1G). These data show that TWEAK-induced morphological features in differentiated epithelia paralleled

with the regulation of main molecules involved in building cell to cell contacts.

ERK and NF-kB activation are required for TWEAK-induced junctional proteins

downregulation in tubular cells We have previously demonstrated that ERK engagement and canonical as well as alternative NF-κB

activation contribute to TWEAK inflammatory effects in renal tubular cells and kidney tissue (Sanz et al, 2008, 2010). In this work, blockade of the ERK activation by the pharmacological inhibitor PD98059 reduced TWEAK-induced changes in AJ and TJ proteins, including E-cadherin and ZO-1 downregulation, respectively (Figure 2, A and B). Recently, we have shown that TWEAK-induced ERK activation can be achieved by EGFR transactivation following binding to Fn14 (Rayego-Mateos et al, 2013). To know whether EGFR inhibition could also result in the blockade of TWEAK-induced E-cadherin loss, we transfected HK2

cells with a specific EGFR siRNA. First, we corroborated that in this human cell line, TWEAK downmodulated E-cadherin expression, which correlated with phenotypic changes evocative of the loss of its epithelial nature (Figure 2C). Then, we observed that in cells with silenced EGFR, TWEAK treatment continued inducing the loss of E-cadherin, giving account that EGFR does not participate in the E-cadherin regulation by TWEAK (Figure 2D). TWEAK activates NF-κB to promote inflammation in the kidney, so in the present work we wanted to

know whether NF-B could be also involved in the TWEAK-induced tubular AJ and TJ weakening. Pharmacological inhibition of the canonical NF-B pathway with the IB phosphorylation inhibitor Bay 117082 restrained the TWEAK-induced downregulation of E-cadherin (Figure 3A) and ZO-1(Figure 3B) in tubular cells. However, inhibition of the NF-B2/p100 processing with the proteasome inhibitor lactacystin,

This article is protected by copyright. All rights reserved

Accepted Article

did not prevent TWEAK-induced E-cadherin and ZO-1 loss, pointing out that TWEAK-induced adhesion protein loss is mediated by canonical NF-B activation but not by the alternative pathway (Figure 3C). At last, we assessed the interdependence between both signaling pathways and explored its impact on

the junctional protein expression. When inhibited with PD98058, TWEAK-elicited ERK activity did not affect the simultaneous NF-B nuclear translocation, thus confirming that TWEAK triggered both pathways in an independent manner (Figure 3D). However, concurrent inhibition of ERK and NF-κB activation did not suffice to recover protein expression levels as in control cells, suggesting that additional pathways engaged by TWEAK may contribute to junctional instability (Figure 3E).

TWEAK-induced E-cadherin loss is regulated by the vitamin D receptor (VDR) downstream of NF-B and ERK activation To search for molecular mediators that could account for TWEAK-induced loss of E-cadherin, we

focused on the vitamin D receptor (VDR), since loss of VDR expression was recently involved in mediating tubular EMT induced by members of the TNF superfamily in an UUO murine model (Xiong et al, 2012). In MCT cells, TWEAK reduced VDR gene expression at 3h (Figure 4A) and protein expression at 8 h (Figure 4B). The VDR protein downregulation was detected before the E-cadherin decrease, which as we showed above, occurred 48h after TWEAK treatment. Moreover, both VDR gene and protein downmodulation were reverted by ERK and NF-B activity inhibition (Figure 4, A and B). Treatment with paricalcitol, a known stimulator of VDR function and synthesis, increased VDR protein levels in control tubular cells and significantly recovered E-cadherin levels in the presence of TWEAK (Figure 4C). In addition, TWEAK also inhibited the mRNA expression of the VDR co-factor NCoA-1 and paricalcitol precluded this effect, thus suggesting that NCoA-1 could participate in regulating the E-cadherin fate (Figure 4D). Finally, as proof of the fact that VDR contributes to E-cadherin stabilization, we observed that VDR silencing in human tubular cells, similarly to TWEAK, also resulted in a decreased expression of E-cadherin thus denoting that VDR helps to preserve E-cadherin levels in control untreated cells (Figure 4E). These results show that functional VDR signaling promotes the synthesis of E-cadherin in unstimulated tubular cells, and conversely, suggest that VDR downmodulation by TWEAK could be involved in the loss of E-cadherin.

This article is protected by copyright. All rights reserved

Accepted Article

We took advantage of our UUO model performed in TWEAK-KO mice (Ucero et al, 2013) to study

the in vivo VDR regulation by TWEAK. Remarkably, obstructed wild type animals exhibited decreased kidney VDR gene expression, while in obstructed TWEAK-KO mice, VDR expression remained at normal levels two days after obstruction treatment or even it increased at longer times (Figure 4F).

TWEAK triggers EMT-related changes by common mechanisms leading to the loss of junctional proteins in tubular cells Alpha smooth muscle actin (-SMA) is a prototypical myofibroblast marker also expressed by EMT-

differentiated tubular cells. We evaluated the presence of αSMA in tubular cells using the UUO model developed in TWEAK-KO mice. αSMA positive tubular cells were not found in sham control mice, as expected. After 7 days of UUO, recurring presence of αSMA positive cells was clearly observed in tubules from obstructed WT mice, whereas no positive tubular cells were found in TWEAK-deficient obstructed kidneys (Figure 5A). These results suggested that TWEAK could contribute to the regulation of EMT-related responses in the damaged kidney; therefore we assayed this possibility in cultured tubular cells. Intrinsic downmodulation of epithelial markers occurs during the EMT process. As already shown

above, TWEAK modulates epithelial junctional proteins and transcription factors such as β-catenin and Snail family proteins, which are also involved in EMT development. Murine and human tubular cells treated with TWEAK really exhibited the novo expression of the

mesenchymal marker Vimentin (Figure 5B). In addition, pharmacological inhibition of ERK and NF-B

signaling reduced TWEAK-induced Vimentin upregulation (Figure 5C). Related to this later result, UUO kidneys from WT mice showed tubular αSMA expression in cells with activated NF-B detected through p65 nuclear translocation (Figure 6D). Matrix metalloproteinases (MMPs) contribute to tubular cell EMT, and in

turn, tubular cells in EMT are characterized by their increased MMPs synthesis and secretion and motility (Zhao et al, 2013). TWEAK markedly increased MMP-9 gene expression and enzymatic activity in vitro

(Figure 5E). Further evidence about a role of TWEAK in the MMP-9 regulation was obtained from obstructed kidneys from TWEAK-KO mice, which showed a restrained MMP-9 transcription compared to that of the obstructed kidneys from WT control mice (Figure 5F). Increased cell motility is also associated to

EMT occurrence (Zhao et al, 2013). We implemented a wound healing assay in HK2 cells to assess cell

This article is protected by copyright. All rights reserved

Accepted Article

migration under TWEAK stimulation. Cells treated with TWEAK closed the wound more quickly than control non stimulated cells, suggesting that loss of intercellular adhesion and EMT may help to induce a more motile phenotype resulting in an increased cell migration (Figure 5G).

TWEAK-induced epithelial feature loss and EMT are not related to TGF-β1 and

depend on Fn14 binding TGF-1 is the prototypical EMT inducer in many renal and non-renal epithelia (Iwano, 2010). In the

kidney, TGF-1 is activated in response to a variety of EMT inducers and regulates several transcription factors driving the EMT (Roberts et al, 2006; Cho et al, 2007; Carvajal et al, 2008). We evaluated whether endogenous TGF-1 could be involved in TWEAK-induced EMT changes. Incubation of MCTs cells with TWEAK for 24 hours did not increase TGF-β1 mRNA levels (Figure 6A). Moreover, preincubation of MCTs with a TGF-1 neutralizing antibody, did not prevent TWEAK-mediated E-cadherin and ZO-1 loss and

Vimentin upregulation (Figure 6B). In the UUO model, TGF-β1 mRNA levels were upregulated in obstructed kidneys at 2 and 7 days compared to Sham controls, but there were no significant differences between WT and TWEAK-KO mice at both time points (Figure 6C). These data suggest that TGF-β1 is not involved in the TWEAK-induced overall changes related to cell adhesion and EMT. Involvement of the Fn14 receptor in TWEAK-induced changes in epithelial and mesenchymal markers

was also studied. In this regard, the overall process evaluated through the TWEAK-dependent E-cadherin, ZO-1, Snail and Vimentin modulation was prevented by the Fn14 blocking antibody ITEM-4 (Nakayama et al, 2003), pointing out that TWEAK-induced epithelial feature loss and EMT are mediated by TWEAK binding to Fn14 (Figure 7, A-C).

DISCUSSION Inflammation itself and proinflammatory cytokines have been related to the dismantling of TJ and

hence to the loss of cell polarity and permeability in kidney tubular epithelium, thus highlighting a role of junctional disassembling in kidney damage (Patrick et al, 2006; Eadon el al, 2012; Szaszi and Amoozadeh, 2014). We described here that the inflammatory cytokine TWEAK regulates at the transcriptional and

This article is protected by copyright. All rights reserved

Accepted Article

translational level a set of typical proteins forming the TJ apical complexes of tubular cells. Our in vitro

studies show that TWEAK disrupted the ZO-1 membrane belt, which is one of the most reliable markers

related to TJ disassembly. ZO-1 is also a central scaffolding protein that anchors other TJ proteins such as claudins and occludins to the actin cytoskeleton. In relation, TWEAK treatment also significantly upregulated claudin-1 mRNA expression, previously suggested to be an adaptation to kidney injury (Eadon et al, 2012). In such a way, it could be conceivable that in tubular cells subjected to TWEAK stimulation, the loss of TJ hierarchy and actin cytoskeleton spatial disorganization could be the consequence of ZO-1 fragmentation and reduced localization in the membrane. Moreover, TWEAK also modified the expression levels of important AJ proteins, namely E-cadherin, N-cadherin and Cadherin-16, and also of β-catenin, which links cadherins to the actin cytoskeleton. The loss of cadherins has been implicated in kidney damage by sensitizing tubular cells to acute and chronic kidney injury (Sáenz-Morales et al, 2006; Berzal et al, 2012; Zhou et al, 2012, Hills et al, 2013). Apart from being regulated at the transcriptional level, as it was the case for the TWEAKinduced downregulation, a direct mechanism of E-cadherin shedding in kidney damage has been attributed to metalloproteinase activities (Keller and Nigam, 2003; Zheng et al, 2009). In the present work, we showed that TWEAK upregulated both the MMP-9 expression and activity in cultured cells and that TWEAK deletion restrained MMP-9 synthesis “in vivo”. Thus, the mechanism of MMP-mediated degradation could contribute

to the E-cadherin decreased expression. In tubular cells, the restrained E-cadherin and ZO-1 synthesis along with the upregulation of Snail

family of transcriptional repressors, and the induction of the mesenchymal marker Vimentin, suggest that TWEAK could induced a complete EMT. Remarkably, it has been reported that E-cadherin and ZO-1 loss or Snail overexpression may suffice to trigger the EMT, or in the case of Vimentin, to confer the cells with EMT-related properties such as motility and invasiveness (Reichert et al, 2000; Ryeom et al, 2000; Peinado et al, 2004; Moreno-Bueno et al, 2006; Boutet et al, 2006, Ivaska et al, 2007). We explored different intracellular pathways related to inflammation that may account for the loss of

junctional proteins and EMT in TWEAK-stimulated cells. Of note, we show that pharmacological blockade of ERK1/2 inhibited TWEAK-induced E-cadherin, ZO-1 and Vimentin changes, highlighting that ERK1/2 activation may be a chief mediator of the tubular AJ and TJ disruption and EMT. By contrast, we did not find that EGFR activation had any role in the TWEAK-elicited E-cadherin loss and hence EMT, although this

This article is protected by copyright. All rights reserved

Accepted Article

pathway has been involved in TWEAK-dependent inflammation and ERK1/2 activation (Rayego-Mateos et al, 2013) and also in TJ regulation and EMT (Singh et al, 2004; Dey et al, 2010; Ardura et al, 2010). Results from these studies probably reflect that the engagement of EGFR on different signaling pathways leading to junctional protein loss or EMT is a stimulus specific and cell-dependent event. The observed ERK-mediated TWEAK actions go in the line with previous findings showing MAP kinases involved in regulating epithelial barrier protein expression and function and renal EMT. Thus, pharmacological inhibition of MAPK signaling reversed the deleterious effects of the TNFα/INFγ, hepatocyte growth factor and H 2O2 on TJ proteins (Lipschutz et al, 2005; Patrick et al, 2006; González et al, 2009; Ardura et al, 2010; Dey et al, 2010). The MAPK pathway also mediates the EMT provoked by factors involved in renal damage, such as TGF-β1 and angiotensin II (Ellenrieder et al, 2001; Bakin et al, 2002; Rodríguez-Díez et al, 2008). The transcription factor NF-B is one of the most prominent regulators of kidney inflammation (Sanz

et al, 2010). Studies presented in this paper show that blockade of the canonical NF-B pathway diminished TWEAK-induced junctional protein loss and EMT in cultured tubular cells. Besides, p65 activation in tubular cells from obstructed UUO mice, an episode formerly shown to regulate inflammation in this model (Ucero et al, 2013), along with the heightened expression of αSMA as indicative of EMT induction, strongly suggest

that NF-B could be placed at a crossroad linking both phenomena. Formerly, activation of NF-B by

inflammatory cytokines, mainly TNFα, has been related to altered AJ and TJ disassembly in experimental models of digestive system and cornea damage (He et al, 2012; Fischer et al, 2013; Kimura et al, 2013). In kidney, the question about whether NF-B could influence TJ protein modifications remains elusive.

However, there are many works showing the involvement of NF-κB in regulating TJ proteins during the EMT

process in adult as well as tumoral tubular cells (Barberá et al, 2004; Pantuck el al, 2010; Li et al, 2011). One

of the reported mechanisms by which NF-B promotes the E-cadherin downmodulation is through the Snail

activation and stabilization (Wu and Zhou, 2010). Moreover, it could be possible that NF-κB may directly regulate Vimentin by interacting with previously characterized binding sites (Lilienbaum and Paulin, 1993). Accordingly, our present results show that stimulation with TWEAK both activates Snail proteins and upregulates Vimentin. We confirmed previous data pointing out that NF-B and ERK are independently engaged by TWEAK

(Sanz et al, 2009). Nonetheless, we find that TWEAK-induced VDR loss is a downstream point intersecting

This article is protected by copyright. All rights reserved

Accepted Article

NF-B and ERK signaling, because it was prevented by individual inhibition of NF-κB and ERK1/2. Vitamin D deficiency is associated to undesirable outcomes in CKD patients (Rojas-Rivera, 2010). VDR-KO mice are much more susceptible to kidney obstructive damage and EMT. By contrast, VDR overexpression prevented EMT and VDR knock-down facilitated it in tubular cells (Xiong et al, 2012). VDR deficiency also resulted in E-cadherin content reduction and disruption of intercellular junctions in rat kidney (Migliori et al, 2005). Here, we showed that VDR silencing promotes E-cadherin destabilization. Conversely, the VDR activator

paricalcitol replenished VDR content in cells treated with TWEAK and significantly recovered the E-cadherin expression although without reaching control or paricalcitol stimulated VDR levels. This later result suggests that additional mechanisms of TWEAK-induced E-cadherin loss may be operative further to the VDR decay. VDR interacts with DNA following binding of active vitamin D and subsequent limiting cofactor recruitment. NCoA-1 is a co-factor receptor to which VDR preferentially binds (Carvallo et al, 2007). TWEAK also down-

modulated NCoA-1 expression in our cell culture system and paricalcitol prevented the loss of NCoA-1. The

present results add new insights about the potential contribution of VDR co-activators as mediators of TWEAK-induced tubular damage. TGF-1 is known as the major promoter of EMT during embryogenesis, cancer progression and

fibrosis (Moustakas and Heldin; 2007). However, the present results add TWEAK to the set of endogenous and xenobiotic renal damaging agents that can trigger loss of epithelial integrity and EMT independently of TGF-1, such as Angiotensin II (Carvajal et al, 2008), and cyclosporine (Berzal et al, 2012), among others. In recent years, a pathogenic role of TWEAK in experimental cellular and animal models of renal

damage and disease has been established (Michaelson et al, 2012; Xia et al, 2012; Poveda et al, 2013; RuizOrtega et al, 2014). A relevant fact is that TWEAK promotes inflammation in tubular cultured and kidney cells whereas in vivo TWEAK blockade ameliorates inflammation as well as tubular injury and interstitial fibrosis (Sanz et al, 2010; Ucero et al, 2013). In this work, we have shown that proinflammatory signaling pathways elicited by TWEAK in kidney tubular cells are also involved in the loss of structural junctional proteins and epithelial features and also in the endowment of cells with mesenchymal characteristics (Figure 8). Mechanistically, these findings include the identification of the Fn14/NF-kB and Fn14/ERK axes as well as the decay in the VDR expression downstream both pathways. Additional non defined factors, as indicated by the incomplete recovery of the epithelial character of cells when treated simultaneously with NF-kB and

This article is protected by copyright. All rights reserved

Accepted Article

ERK inhibitors, may account for TWEAK-dependent junctional effects and EMT. These effects of TWEAK may contribute to its observed fibrogenic action in a well characterized model of kidney obstructive disease and open the possibility of disruption of epithelial integrity as a mechanism of TWEAK-mediated renal damage. Thus, targeting TWEAK or TWEAK-induced signaling pathways could be relevant to restrain inflammation and related pathogenic processes in human kidney diseases.

ACKNOWLEDGEMENTS We want to thank to: Dr Carlos Castilla for the veterinary supervision and advice in experimental procedures using mice models; Dr Amparo Cano (Instituto de Investigaciones Alberto Sols, CSIC-UAM, Madrid, Spain)

and Dr Pedro Majano (IIS-Hospital Universitario de la Princesa, Madrid) for kindly providing us with MDCK cell line and E-cadherin antibody for immunofluorescence, and ZO-1 antibody for western blot, respectively. Salary support: AU, SB, CG and CO are fellows of Fundación Renal Conchita Rábago de Jiménez Díaz; Programa Intensificación Actividad Investigadora (ISCIII/Agencia Laín-Entralgo/CM) to AO; Universidad Autónoma de Madrid to JE and MRO; Programa Miguel Servet to AMR. Conflict of Interest Disclosure. The authors declare that no conflicts of interests exist.

REFERENCES Ardura JA, Rayego-Mateos S, Rámila D, Ruiz-Ortega M, Esbrit P. 2010. Parathyroid hormone-related protein promotes epithelial-mesenchymal transition. J Am Soc Nephrol 21(2):237-248.

Bakin AK, Rinehart C, Tomlinson AK, Arteaga CL. 2002. p38 mitogen-activated protein kinase is required for TGFbeta mediated fibroblastic transdifferentiation and cell migration. J. Cell Sci 115:3193–3206.

Barberá MJ, Puig I, Domínguez D, Julien-Grille S, Guaita-Esteruelas S, Peiró S, Baulida J, Francí C, Dedhar S, Larue L, García de Herreros A. 2004. Regulation of Snail transcription during epithelial to mesenchymal

transition of tumor cells. Oncogene 23(44):7345-7354.

This article is protected by copyright. All rights reserved

Accepted Article

Berzal S, Alique M, Ruiz-Ortega M, Egido J, Ortiz A, Ramos AM. 2012. GSK3, snail, and adhesion molecule regulation by cyclosporine A in renal tubular cells. Toxicol Sci 127(2):425-437.

Boutet A, De Frutos CA, Maxwell PH, Mayol MJ, Romero J, Nieto MA.Snail activation disrupts tissue homeostasis and induces fibrosis in the adult kidney. EMBO J. 2006 29;25(23):5603-5613

Carvallo L, Henriquez B, Olate J, van Wijnen AJ, Lian JB, Stein GS, Onate S, Stein JL, Montecino M. 2007. The 1alpha,25-dihydroxy Vitamin D3 receptor preferentially recruits the coactivator SRC-1 during upregulation of the osteocalcin gene. J Steroid Biochem Mol Biol 103(3-5):420-424.

Carvajal G, Rodríguez-Vita J, Rodrigues-Díez R, Sánchez-López E, Rupérez M, Cartier C, Esteban V, Ortiz A, Egido J, Mezzano SA, Ruiz-Ortega M. 2008. Angiotensin II activates the Smad pathway during epithelial mesenchymal transdifferentiation. Kidney Int 74(5):585-595.

Cho HJ, Baek KE, Saika S, Jeong MJ, Yoo J. 2007. Snail is required for transforming growth factor-betainduced epithelial-mesenchymal transition by activating PI3 kinase/Akt signal pathway Biochem Biophys Res Commun 353(2):337-343.

Chuang MJ, Sun KH, Tang SJ, Deng MW, Wu YH, Sung JS, Cha TL, Sun GH. 2008. Tumor-derived tumor necrosis factor-alpha promotes progression and epithelial-mesenchymal transition in renal cell carcinoma cells. Cancer Sci 99(5):905-913.

Dey M, Baldys A, Sumter DB, Göoz P, Luttrell LM, Raymond JR, Göoz M. 2010. Bradykinin decreases

podocyte permeability through ADAM17-dependent epidermal growth factor receptor activation and zonula occludens-1 rearrangement. J Pharmacol Exp Ther 334(3):775-783.

Dohi T, Burkly LC. 2012. The TWEAK/Fn14 pathway as an aggravating and perpetuating factor in inflammatory diseases: focus on inflammatory bowel diseases. J Leukoc Biol 92(2):265-279.

This article is protected by copyright. All rights reserved

Accepted Article

Eadon MT, Haxk BK, Xu C, Ko B, Toback FG, Cunningham PN. 2012. Endotoxemia alters tight junction gene and protein expression in the kidney. Am J Physiol Renal Physiol 303(6):F821-830.

Ellenrieder V, Hendler SF, Boeck W, Seufferlein T, Menke A, Ruhland C, Adler G, Gress TM. 2001. Transforming growth factor beta1 treatment leads to an epithelial–mesenchymal transdifferentiation of pancreatic cancer cells requiring extracellular signalregulated kinase 2 activation. Cancer Res 61:4222–4228.

Fischer A, Gluth M, Pape UF, Wiedenmann B, Theuring F, Baumgart DC. 2013. Adalimumab prevents barrier dysfunction and antagonizes distinct effects of TNF-α on tight junction proteins and signaling

pathways in intestinal epithelial cells. Am J Physiol Gastrointest Liver Physiol 304(11):G970-979.

Girshovich A, Vinsonneau C, Perez J, Vandermeersch S, Verpont MC, Placier S, Jouanneau C, Letavernier E, Baud L, Haymann JP. 2012. Ureteral obstruction promotes proliferation and differentiation of the renal urothelium into a bladder-like phenotype. Kidney Int 82(4):428-435.

Gonzalez JE, DiGeronimo RJ, Arthur DE, King JM. 2009. Remodeling of the tight junction during recovery from exposure to hydrogen peroxide in kidney epithelial cells. Free Radic Biol Med 2009 47(11):1561-1569.

Hazzan M, Hertig A, Buob D, Copin MC, Noël C, Rondeau E, Dubois-Xu YC. 2011. Epithelial-tomesenchymal transition predicts cyclosporine nephrotoxicity in renal transplant recipients. J Am Soc Nephrol 22(7):1375-81.

He F, Peng J, Deng XL, Yang LF, Camara AD, Omran A, Wang GL, Wu LW, Zhang CL, Yin F. 2012. Mechanisms of tumor necrosis factor-alpha-induced leaks in intestine epithelial barrier. Cytokine. 59(2):264272.

This article is protected by copyright. All rights reserved

Accepted Article

Hills CE, Jin T, Siamantouras E, Liu IK, Jefferson KP, Squires PE. 2013. 'Special k' and a loss of cell-to-cell adhesion in proximal tubule-derived epithelial cells: modulation of adherens junction complex by ketamine. Plos One8(8):e71819.

Ho MY, Tang SJ, Chuang MJ, Cha TL, Li JY, Sun GH, Sun KH. 2012. TNF-α induces epithelialmesenchymal transition of renal cell carcinoma cells via a GSK3β-dependent mechanism. Mol Cancer Res 10(8):1109-1119.

Ivaska J, Pallari HM, Nevo J, Eriksson JE. 2007. Novel functions of vimentin in cell adhesion, migration, and signaling. Exp Cell Res 313(10):2050-2062.

Iwano M. 2010. EMT and TGF-beta in renal fibrosis. Front Biosci (Schol Ed) 1;2:229-238.

Justo P, Sanz AB, Sanchez-Niño MD, Winkles JA, Lorz C, Egido J, Ortiz A. 2006. Cytokine cooperation in renal tubular cell injury: the role of TWEAK. Kidney Int 70(10):1750-1758.

Kalluri R, Neilson EG. Epithelial-mesenchymal transition and its implications for fibrosis. 2003. J Clin Invest 112: 1776-1784.

Keller SH, Nigam SK. 2003. Biochemical processing of E-cadherin under cellular stress. Biochem Biophys

Res Commun 307(2):215-223.

Kimura K, Morita Y, Orita T, Haruta J, Takeji Y, Sonoda KH. 2013. Protection of human corneal epithelial cells from TNF-α-induced disruption of barrier function by rebamipide. Invest Ophthalmol Vis Sci. 54(4):2752-2760.

Kwon O, Nelson WJ, Sibley R, Huie P, Scandling JD, Dafoe D, Alfrey E, Myers BD. 1998. Backleak, tight junctions, and cell- cell adhesion in postischemic injury to the renal allograft. J Clin Invest 101(10):20542064.

This article is protected by copyright. All rights reserved

Accepted Article

LeBleu VS, Taduri G, O’Connell J, Teng Y, Cooke VG, Woda C, Sugimoto H, Kalluri R. 2013. Origin and function of myofibroblasts in kidney fibrosis. Nat Med 19(8):1047-1053.

Li Q, Liu BC, Lv LL, Ma KL, Zhang XL, Phillips AO. 2011. Monocytes induce proximal tubular epithelialmesenchymal transition through NF-kappa B dependent upregulation of

ICAM-1. J Cell Biochem

112(6):1585-1592.

Lilienbaum

A, Paulin

D.

1993.

Activation of

the human vimentin gene by

the Tax human T-

cell leukemia virus. I. Mechanisms of regulation by the NF-kappa B transcription factor. J Biol Chem 268(3):2180-2188.

Lipschutz JH, Li S, Arisco A, Balkovetz DF. 2005. Extracellular signal-regulated kinases ½ control claudin-2

expression in Madin-Darby caninekidney strain I and II cells. J Biol Chem 280(5):3780-3788.

Liu Y. 2010. New insights into epithelial-mesenchymal transition in kidney fibrosis. J Am Soc Nephrol 21:212-222.

Michaelson JS, Wisniacki N, Burkly LC, Putterman C. 2012. Role of TWEAK in lupus nephritis: a bench-tobedside review. J Autoimmun 39(3):130-142.

Migliori M, Giovannini L, Panichi V, Filippi C, Taccola D, Origlia N, Mannari C, Camussi G. 2005.

Treatment with 1,25-dihydroxyvitamin D3 preserves glomerular slit diaphragm-associated protein expression in experimental glomerulonephritis. Int J Immunopathol Pharmacol 18:779-790.

Moreno-Bueno G, Cubillo E, Sarrió D, Peinado H, Rodríguez-Pinilla SM, Villa S, Bolós V, Jordá M, Fabra A, Portillo F, Palacios J, Cano A. 2006. Genetic profiling of epithelial cells expressing E-cadherin repressors

This article is protected by copyright. All rights reserved

Accepted Article

reveals a distinct role for Snail, Slug, and E47 factors in epithelial-mesenchymal transition. Cancer Res 66(19):9543-9556.

Moustakas A, Heldin CH. 2007. Signaling networks guiding epithelial-mesenchymal transitions during embryogenesis and cancer progression. Cancer Sci 98(10):1512-1520.

Nakayama M, Ishidoh K, Kojima Y, Harada N, Kominami E, Okumura K, Yagita H. 2003. Fibroblast growth factor-inducible 14 mediates multiple pathways of TWEAK-induced cell death. J Immunol 170(1):341-348.

Pantuck AJ, An J, Liu H, Rettig MB. 2010. NF-kappaB-dependent plasticity of the epithelial to mesenchymal

transition induced by Von Hippel-Lindau inactivation in renal cell carcinomas. Cancer Res 0(2):752-761.

Patrick DM, Leone AK, Shellenberger JJ, Dudowicz KA, King JM. 2006. Proinflammatory cytokines tumor necrosis factor-alpha and interferon-gamma modulate epithelial barrier function in Madin-Darby canine kidney cells through mitogen activated protein kinase signaling. BMC Physiol 21;6:2.

Peinado H, Portillo F, Cano A. 2004. Transcriptional regulation of cadherins during development and

carcinogenesis. Int J Dev Biol 48(5-6):365-375.

Poritz LS, Garver KI, Tilberg AF, Koltun WA. 2004. Tumor necrosis factor alpha disrupts tight junction

assembly. J Surg Res 116(1):14-18.

Poveda J, Tabara LC, Fernandez-Fernandez B, Martin-Cleary C, Sanz AB, Selgas R, Ortiz A, Sanchez-Niño MD. 2013. TWEAK/Fn14 and Non-Canonical NF-kappaB Signaling in Kidney Disease. Front Immunol 4:447.

Prozialeck WC, Edwards JR. 2007. Cell adhesion molecules in chemically-induced renal injury. Pharmacol

Ther 114(1):74-93.

This article is protected by copyright. All rights reserved

Accepted Article

Rayego-Mateos S, Morgado-Pascual JL, Sanz AB, Ramos AM, Eguchi S, Batlle D, Pato J, Keri G, Egido J, Ortiz A, Ruiz-Ortega M. 2013. TWEAK transactivation of the Epidermal growth factor receptor mediates renal inflammation. J Pathol 231(4): 480–494.

Reichert M, Müller T, Hunziker W. 2000. The PDZ domains of zonula occludens-1 an epitelial to mesenchymal transition of Madin-Darby canine kidney I cells. Evidence for a role of beta-catenin/Tcf/Lef signaling. J Biol Chem 275(13):9492-9500.

Roberts AB, Tian F, Byfield SD, Stuelten C, Ooshima A, Saika S, Flanders KC. 2006. Smad3 is key to TGFbeta-mediated epithelial-to-mesenchymal transition, fibrosis, tumor suppression and metastasis. Cytokine Growth Factor Rev 17(1-2):19-27.

Rojas-Rivera J, De La Piedra C, Ramos A, Ortiz A, Egido J. 2010. The expanding spectrum of biological actions of vitamin D. Nephrol Dial Transplant 25(9):2850-2865. Rodrigues-Díez R, Carvajal-González G, Sánchez-López E, Rodríguez-Vita J, Rodrigues Díez R, Selgas R, Ortiz A, Egido J, Mezzano S, Ruiz-Ortega M. 2008. Pharmacological modulation of epithelial mesenchymal transition caused by angiotensin II. Role of ROCK and MAPK pathways. Pharm Res 25(10):2447-2461.

Ruiz-Ortega M, Ortiz A, Ramos AM. 2014. Tumor necrosis factor-like weak inducer of apoptosis (TWEAK) and kidney disease. Curr Opin Nephrol Hypertens 23(1):93-100.

Ryeom SW, Paul D, Goodenough DA. 2000. Truncation mutants of the tight junction protein ZO-1 disrupt corneal epithelial cell morphology. Mol Biol Cell 11(5):1687-1696.

Sáenz- Morales D, Escribese MM, Stamatakis K, García-Martos M, Alegre L, Conde E, Pérez-Sala D, Mampaso F, García-Bermejo ML. 2006. Requirements for proximal tubule epithelial cell detachment in response to ischemia: role of oxidative stress. Exp Cell Res 312(19):3711-3727.

This article is protected by copyright. All rights reserved

Accepted Article

Sanz AB, Justo P, Sanchez-Niño MD, Blanco-Colio LM, Winkles JA, Kreztler M, Jakubowski A, Blanco J, Egido J, Ruiz-Ortega M, Ortiz A. 2008. The cytokine TWEAK modulates renal tubulointerstitial inflammation. J Am Soc Nephrol 19(4):695-703.

Sanz AB, Sanchez-Niño MD, Izquierdo MC, Jakubowski A, Justo P, Blanco-Colio LM, Ruiz-Ortega M, Egido J, Ortiz A. 2009. TWEAK induces proliferation in renal tubular epithelium: a role in uninephrectomy induced renal hyperplasia. J Cell Mol Med 13(9B):3329-3342.

Sanz AB, Sanchez-Niño MD, Izquierdo MC, Jakubowski A, Justo P, Blanco-Colio LM, Ruiz-Ortega M, Selgas R, Egido J, Ortiz A. 2010. TWEAK activates the non-canonical NFkappaB pathway in murine renal tubular cells: modulation of CCL21. PLoS One 5(1):e8955.

Sanz AB, Sanchez-Niño MD, Ortiz A. 2011. TWEAK, a multifunctional cytokine in kidney injury. Kidney Int 80(7):708-718.

Singh AB, Harris RC. 2004. Epidermal growth factor receptor activation differentially regulates claudin

expression and enhances transepithelial resistance in Madin-Darby canine kidney cells. J Biol Chem 279(5):3543-3552.

Szaszi K, Amoozadeh Y. 2014. New insights into functions, regulation, and pathological roles of tight junctions in kidney tubular epithelium. Int Rev Cell Mol Biol 308:205-271.

Ucero AC, Benito-Martin A, Fuentes-Calvo I, Santamaria B, Blanco J, Lopez-Novoa JM, Ruiz-Ortega M, Egido J, Burkly LC, Martinez-Salgado C, Ortiz A. 2013. TNF-related weak inducer of apoptosis (TWEAK) promotes kidney fibrosis and Ras-dependent proliferation of cultured renal fibroblast. Biochim Biophys Acta.

1832(10):1744-1755.

This article is protected by copyright. All rights reserved

Accepted Article

Wilson PD. 2011. Apico-basal polarity in polycystic kidney disease epithelia. Biochim Biophys Acta 1812(10):1239-1248.

Winkles JA. 2008. The TWEAK-Fn14 cytokine-receptor axis: discovery, biology and therapeutic targeting. Nat Rev Drug Discov 7(5):411-425.

Wu Y, Zhou BP. 2010. TNF-alpha/NF-kappaB/Snail pathway in cancer cell migration and invasion. Br J Cancer 102(4):639-644.

Xia Y, Campbell SR, Broder A, Herlitz L, Abadi M, Wu P, Michaelson JS, Burkly LC, Putterman C. 2012. Inhibition of the TWEAK/Fn14 pathway attenuates renal disease in nephrotoxic serum nephritis. Clin Immunol 145(2):108-121.

Xiong M, Gong J, Liu Y, Xiang R, Tan X. 2012. Loss of vitamin D receptor in chronic kidney disease: a

potential mechanism linking inflammation to epithelial-to-mesenchymal transition. Am J Physiol Renal Physiol 303(7):F1107-1115.

Zhao H, Dong Y, Tian X, Tan TK, Liu Z, Zhao Y, Zhang Y, Harris DCh, Zheng G. 2013. Matrix metalloproteinases contribute to kidney fibrosis in chronic kidney diseases. World J Nephrol 2(3):84-89.

Zheng G, Lyons JG, Tan TK, Wang Y, Hsu TT, Min D, Succar L, Rangan GK, Hu M , Henderson BR, Alexander SI, Harris DC. 2009. Disruption of E-cadherin by matrix metalloproteinase directly mediates epithelial-mesenchymal transition downstream of transforming growth factor-beta1 in renal tubular epithelial cells. Am J Pathol 175(2):580-59.

Zhou D, Li Y, Lin L, Zhou L, Igarashi P, Liu Y. 2012. Tubule-specific ablation of endogenous βcatenin aggravates acute kidney injury in mice. Kidney Int 82(5):537-547.

This article is protected by copyright. All rights reserved

Accepted Article

FIGURE LEGENDS Figure 1. TWEAK modificates the actin cytoskeleton structure and induces molecular changes in apical junction proteins and transcriptional regulators in renal tubular cells. A) Murine tubular MCT cells were treated with 100 ng/ml TWEAK for 48 h and then evaluated for the occurrence of morphological changes by phase contrast microscopy (upper panel; original magnification x100) and also stained for F-actin with Phalloidin detected by confocal microscopy (red fluorescence) (lower panel; original magnification x 400). In controls, cells grow clustered in a monolayer which loss its continuity and typical cobblestone epithelial pattern, and display a spindle-shaped, fibroblast like morphology in presence of TWEAK (upper panel). In the lower panel, in cells treated with TWEAK, arrows indicate areas of weak staining indicatives of F-actin loss and arrowheads the cytoplasmic condensation of the fluorescent mark characteristic of stress fiber formation.

Figure shows representative images. B-F) MCT cells were left untreated or treated with 100 ng/ml TWEAK alone for 24 h or 48 h to evaluate gene and protein expression of junctional proteins and transcriptional regulators, respectively. B) TWEAK represses the E-cadherin, Cadherin-16 and ZO-1 mRNA expression and induces Claudin mRNA transcription. The bar graph represents the Mean  SD of three independent experiments measured by qRT-PCR (*p 0.05 vs Control). C) TWEAK diminishes E-cadherin, N-cadherin, -catenin and ZO-1 protein expression. Representative Western blots from a total of three independent experiments are shown. *p 0.05 vs Control. D) TWEAK increases protein levels of Snail and Slug. Representative Western blots from a total of three independent experiments are shown. *p 0.02 vs Control. E) Confocal microscopy showing TWEAK-induced ZO-1 membrane loss, -catenin membrane loss and redistribution, and also the activation of Snail detected by higher cytoplasm and nuclear expression level. Representative experiment is shown. Original magnification x200. F) TWEAK greatly reduced E-cadherin expression levels in MDCK cells treated with 100 ng/ml TWEAK for 48 h. Figure shows representative confocal images taken at an original magnification x100. Arrows point areas that were amplified and showed with more detail in the upper right corner of each picture. In control cells, E-cadherin fluorescence labeling is brighter, especially at the intercellular junctions and placed rounding the entire cell membrane. By contrast, TWEAK-treated cells show an overall pattern exhibiting a weak expression of E-cadherin, with areas showing a complete absence of the protein, and further distributed in a punctate manner. G) Snail, Slug and HNF1 mRNA levels assessed by qRT-PCR in tubular cells treated with TWEAK (n=4, *p 0.01).

This article is protected by copyright. All rights reserved

Accepted Article

Figure 2. MAPK/ERK activation is required for TWEAK induced junctional proteins downregulation

in tubular cells. A-B) Inhibition of the MAP kinase ERK1/2 results in restrained TWEAK-induced E-

cadherin and ZO-1 downmodulation. MCT cells were treated with 100 ng/ml TWEAK alone or pretreated with 20 M the ERK inhibitor PD98059 (PD) for 1 h before TWEAK addition. A) Representative western blot assessing E-cadherin expression. Bar graph depicts the Mean  SD of four independent experiments. (*p  0.02 vs Control; #p  0.05 vs TWEAK). B) Representative confocal images of ZO-1 expression. Original magnification x 200. C) Representative western blot showing E-cadherin expression levels in human tubular HK2 cells treated with TWEAK alone (upper panel). Bar graph depicts western blot quantification (*p  0.02 vs Control;). Phase contrast images were taken after 72 h treatment to detect morphological changes elicited by TWEAK (lower panel). Original magnification x 200. D) EGFR does not participate in the E-cadherin regulation by TWEAK. Representative western blot showing the efficiency of EGFR silencing in HK2 cells transfected with a scrambled siRNA (sc-siRNA) or with a specific EGFR siRNA (siRNA-EGFR) (upper panel). Cells with EGFR silenced or transferred with the scRNA were then treated with TWEAK for 72 h and E-cadherin levels assessed by western blot (lower panel). Internal control of the experiment was carried out in non transfected cells treated with TWEAK in the same conditions. Bar graph depicts the Mean  SD of three independent experiments. (*p  0.05 vs Control; #p  0.05 vs siRNA-EGFR; §p  0.05 vs sc-siRNA).

Figure 3. Canonical NF-kB activation mediates the TWEAK-induced junctional proteins downregulation in tubular cells. Inhibition of the canonical NF-B pathways results in restrained TWEAKmediated E-cadherin and ZO-1 decrease. MCT cells were treated with 100 ng/ml TWEAK alone or pretreated

with 1 M of the NF-B inhibitor Bay 11-708 (Bay) added 1 h before TWEAK addition. A) Representative western blot assessing E-cadherin expression. Bar graph depicts the Mean  SD of four independent experiments (*p  0.02 vs Control; #p  0.05 vs TWEAK). B) Representative confocal images of ZO-1 expression. Original magnification x200. C) TWEAK-induced E-cadherin and ZO-1 loss is not mediated by the NF-B non-canonical pathway. Cells were treated with 100 ng/ml TWEAK for 48 h or pretreated with 10 µM lactacystin (LAC) for 1 h before TWEAK addition. Representative western blot of E-cadherin (upper panel) and confocal microscopy of ZO-1 (lower panel). Bar graph depicts the Mean  SD of four independent

This article is protected by copyright. All rights reserved

Accepted Article

experiments (#p  0.05 vs Control). D) ERK activity resulted dispensable for the TWEAK-induced NF-B nuclear translocation. MCT cells were treated with TWEAK alone for 30 min or pretreated with PD98059 (PD) before TWEAK addition. Representative confocal image of p65 intracellular location. Original magnification x200. E) Effects of simultaneous inhibition of ERK and NF-κB activity on the expression of main AJ and TJ proteins. MCT cells were pretreated with 20 M PD98059 (PD) and 1 M Bay 11-708 (Bay) for 1h before the addition of TWEAK for 48 h. Representative western blots of E-cadherin and ZO-1, from a

total number of three, are shown. Bar chart represents the entire dataset. *p  0.02 vs Control; #p  0.05 vs TWEAK.

Figure 4. VDR loss contributes to the TWEAK-induced E-cadherin loss downstream NF-B and ERK in tubular cells. MCT cells were treated with 100 ng/ml TWEAK for 3 h for the analysis of VDR and NCoA1 transcription, or unless otherwise indicated, for 8 h for the analysis of VDR protein content by western-blot.

Bay 11-708 (Bay) and PD98059 (PD) were used in the same conditions as in figures 3 and 4. A) TWEAK downmodulates VDR transcription and Bay and PD reverse this effect. Figure shows the Mean  SD of four independent experiments (*p  0.02 vs Control; #p  0.05 vs TWEAK). B) TWEAK downmodulates VDR protein expression and Bay and PD inhibit this effect. Figure shows a representative western blot assay of a set of three independent experiments and their corresponding quantification depicted as the Mean  SD (*p  0.05 vs Control; #p  0.05 vs TWEAK). C) TWEAK downmodulates E-cadherin and VDR and paricalcitol opposes to this effect. Cells were treated with 50 g/ml Paricalcitol for 24 h before stimulation with 100 ng/ml TWEAK for 48 h. Figure shows a representative western blot assay of a set of four independent experiments carried out to evaluate VDR and E-cadherin protein levels, and their corresponding quantification depicted as the Mean  SD (*p  0.02 vs Control; #p  0.05 vs TWEAK; §p  0.05 vs PCT ). D) TWEAK downmodulates NCoA-1 mRNA expression and paricalcitol inhibits this effect. Cells were treated with paricalcitol as in C before TWEAK addition for 24 h, then NCoA-1 was assessed by q-RT-PCR. Figure shows the Mean ± SD of four independent experiments (*p  0.02 vs Control; #p  0.05 vs TWEAK. E) VDR contributes to stabilize E-cadherin expression. HK2 cells were transfected with a scrambled siRNA (scsiRNA) or with a specific VDR siRNA (VDR-siRNA) and after 72 h VDR and E-cadherin expression levels were assessed. Representative western blot of a set of three independent experiments is shown (*p  0.02 vs

This article is protected by copyright. All rights reserved

Accepted Article

Control). F) Renal VDR gene expression diminished in UUO WT mice whereas this effect is prevented in TWEAK-KO mice at day 2 postobstruction onward (#p  0.05 vs Sham; *p  0.05 vs TW-KO).

Figure 5. TWEAK promotes EMT-related changes in tubular cells. Cultured MCT and HK2 cells were left untreated or treated with 100 ng/ml TWEAK to evaluate changes in expression of the mesenchymal marker Vimentin or functional changes related to EMT occurrence (B, C, E, G). Expression of EMT markers was also evaluated in kidneys from WT or TWEAK-KO mice subjected to an UUO procedure (A, D, F). A) Expression of αSMA in kidneys from mice 7 days after ureteral obstruction or sham intervention. UUO results in a massive increase of αSMA from interstitial fibroblasts. In addition, αSMA is expressed in some tubular cells from obstructed kidneys in WT mice but not in tubules from sham operated mice. αSMA positive tubules were not observed in obstructed kidneys from TWEAK-KO mice, even in areas specifically selected

for more severe injury, as the one shown in the figure. Tubules showing αSMA expression are contoured with a dashed white line. Arrows point tubular epithelial cells stained for αSMA. Arrowheads indicate αSMA expressed in interstitial fibroblasts. Representative confocal images of paraffin embedded tissue sections. Magnification x 160. B) Representative western blot of Vimentin expression in total protein extracts from MCT (left panel) and HK2 cells (right panel) treated with TWEAK for 72 h (n=3, *p  0.02 vs Control). C) Inhibition of the ERK and NF-B pathways restrains TWEAK-mediated Vimentin upregulation. MCT cells were treated with 100 ng/ml TWEAK alone or pretreated with Bay 11-708 (Bay) and PD98059 (PD) using the same conditions as in figures 3 and 4. Confocal microscopy showing TWEAK-induced increment in Vimentin expression detected by higher cytoplasm expression level. PD and BAY reverse TWEAK-induced Vimentin upregulation. Representative experiment is shown. Original magnification x200. D) In kidneys from obstructed WT mice, αSMA is expressed in tubules showing nuclear p65 staining. Boxed areas in the upper images are displayed at higher magnification in the series of pictures showed below. Dashed white line delimitates areas of individual tubules. Arrows indicate individual tubular cells with high αSMA staining together with p65 nuclear location indicative of NF-B activation. E) TWEAK increases MMP-9 transcription (left panel) and enzymatic activity (right panel) in MCT cells. MMP-9 mRNA synthesis was quantified by qRT-PCR (n=4, *p 0.01). MMP-9 proteolitic activity was assessed by SDS-PAGE gel zymography. Figure shows a representative experiment of a set of three independent experiments. F) Renal

This article is protected by copyright. All rights reserved

Accepted Article

MMP-9 mRNA expression rises in UUO WT mice whereas this effect is prevented in TWEAK-KO mice at 7 days postobstruction (#p  0.01 vs Sham; *p  0.01 vs TW-KO). G) TWEAK increased cell migration and motility. Representative phase contrast images of wound healing assays performed in HK2 cells non-treated or treated with TWEAK for 72 h were obtained at a magnification x100. Yellow lines indicate wound edges which were entirely closed at the end of TWEAK treatment.

Figure 6. Global effects of TWEAK on epithelial feature loss and EMT are independent from TGF-β1

activity. MCT cells were treated with 100 ng/ml TWEAK alone or in the presence of the TGF-1 blocking antibody (10 g/ml) added to cells 1h before the stimulus. A) Transcription of TGF-1 in cells treated with TWEAK for 24 h evaluated by q-RT-PCR. Bar graph represents the Mean  SD of four independent experiments. Results do not show significant differences between cells treated with the stimulus and those left untreated. B) Blockade of the active TGF-1 does not result in prevention of TWEAK-induced E-cadherin (left panel) and ZO-1 (right panel) loss and Vimentin increase (left panel) after 48 h treatment. Figure shows

representative western blots and their corresponding quantification (n=3, *p  0.01 vs Control) and representative confocal microscopy images from a set of three independent experiments. C) Transcription of TGF-1 in kidneys from WT mice and TWEAK-KO mice subjected to UUO assessed by q-RT-PCR at 2 and 7 days after obstruction. Bar graph represents the Mean  SD of TWEAK mRNA expression for each group.

Results do not show significant differences between mice groups.

Figure 7. TWEAK-induced junctional loss and EMT depends on Fn14 binding. MCT cells were treated with 100 ng/ml TWEAK alone or in the presence of the Fn14 blocking antibody ITEM-4 (2.5 g/ml) added to cells 1h before stimulus. Blocking of the Fn14 receptor prevented the E-cadherin (A) and ZO-1 (B) loss and the Snail (B) and Vimentin (C) increase. A-B) Representative western blots and quantifications of three individual experiments for each protein (*p  0.05 vs Control, #p 0.05 vs TWEAK). (C) Vimentin expression was detected by confocal microscopy. Original magnification x 200.

This article is protected by copyright. All rights reserved

Accepted Article

Figure 8. Divergence of TWEAK signaling modulates several pathogenic events in renal tubular cells. TWEAK binding to Fn14 signals to various pathogenic events downstream NF-κB and ERK1/2 activation. The engagement of NF-κB canonical (p65) and non-canonical (p52) pathways promote tubular inflammation (1). Direct activation of ERK1/2 (2) or by the EGFR engagement through Fn14 transactivation (3) may also result in inflammation. Other consequence of TWEAK signaling by the NF-κB clasical via and the EGFR independent ERK1/2 activation is the AJ and TJ regulation and EMT. NF-κB and ERK1/2 are independently activated by Fn14 (4), although both pathways control VDR expression, which is a downstream cross-talk point (5). VDR (6) as well as non identified factors (assigned as “?” in the scheme) commanded by NF-κB and ERK1/2 are involved in AJ and TJ disassembly and EMT. Additional inflammatory factors released by TWEAK action, including cytokines and growth factors, could account for the loss of junctional integrity. Solid lines indicate signaling pathways described in this paper. Dashed lines illustrate additional unknown pathways. Double dotted lines show signaling pathways and pathogenic events previously described.

This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved

Accepted Article This article is protected by copyright. All rights reserved