Myocardin-related Transcription Factor Regulates Nox4 Protein Expression

Background: Generation of myofibroblasts, the culprit of fibrosis, requires cytoskeleton remodeling and Nox4 (NADPH oxidase) expression. The link between these events is unknown.

Results: Down-regulation/inhibition of the cytoskeleton-controlled transcriptional coactivators, myocardin-related transcription factor (MRTF), and TAZ/YAP abrogates Nox4 expression.

Conclusion: MRTF and TAZ/YAP are essential for Nox4 expression.

Significance: We show new mechanisms whereby the cytoskeleton regulates cellular redox state and fibrogenesis.

Abstract

TGFβ-induced expression of the NADPH oxidase Nox4 is essential for fibroblast-myofibroblast transition. Rho has been implicated in Nox4 regulation, but the underlying mechanisms are largely unknown. Myocardin-related transcription factor (MRTF), a Rho/actin polymerization-controlled coactivator of serum response factor, drives myofibroblast transition from various precursors. We have shown that TGFβ is necessary but insufficient for epithelial-myofibroblast transition in intact epithelia; the other prerequisite is the uncoupling of intercellular contacts, which induces Rho-dependent nuclear translocation of MRTF. Because the Nox4 promoter harbors a serum response factor/MRTF cis-element (CC(A/T)₆GG box), we asked if MRTF (and thus cytoskeleton organization) could regulate Nox4 expression. We show that Nox4 protein is robustly induced in kidney tubular cells exclusively by combined application of contact uncoupling and TGFβ. Nox4 knockdown abrogates epithelial-myofibroblast transition-associated reactive oxygen species production. Laser capture microdissection reveals increased Nox4 expression in the tubular epithelium also during obstructive nephropathy. MRTF down-regulation/inhibition suppresses TGFβ/contact disruption-provoked Nox4 protein and mRNA expression, Nox4 promoter activation, and reactive oxygen species production. Mutation of the CC(A/T)₆GG box eliminates the synergistic activation of the Nox4 promoter. Jasplakinolide-induced actin polymerization synergizes with TGFβ to facilitate MRTF-dependent Nox4 mRNA expression/promoter activation. Moreover, MRTF inhibition prevents Nox4 expression during TGFβ-induced fibroblast-myofibroblast transition as well. Although necessary, MRTF is insufficient; Nox4 expression also requires TGFβ-activated Smad3 and TAZ/YAP, two contact- and cytoskeleton-regulated Smad3-interacting coactivators. Down-regulation/inhibition of TAZ/YAP mitigates injury-induced epithelial Nox4 expression in vitro and in vivo. These findings uncover new MRTF- and TAZ/YAP-dependent mechanisms, which link cytoskeleton remodeling and redox state and impact epithelial plasticity and myofibroblast transition.

Introduction

Organ fibrosis is a dysregulated form of tissue repair in response to chronic injury, characterized by excessive extracellular matrix deposition and the consequent loss of tissue architecture, which ultimately leads to organ failure (1, 2). Fibroproliferative diseases, including renal, hepatic, pulmonary, and cardiovascular fibrosis, and systemic sclerosis are thought to contribute to 45% of all deaths in the Western world (2). The major pathogenic culprit of organ fibrosis is the myofibroblast, a contractile, extracellular matrix-producing mesenchymal cell, hallmarked by the expression of α-smooth muscle actin (SMA)² (3). Although myofibroblasts are scarce in normal parenchymal organs, they are abundantly formed during fibrogenesis from various precursors via phenotypic reprogramming. Therefore, both their cellular origins and the molecular mechanisms leading to their formation and activation have become central questions in fibrosis research. Although the identity of myofibroblast progenitors in various fibrotic diseases is a contentious issue, potential sources include resident fibroblasts, the epithelium, the endothelium, pericytes, and bone marrow-derived fibrocytes (4–6). Irrespective of the particular source, the pleiotropic cytokine TGFβ plays a central role in the phenotype shift and the ensuing activation of a myogenic program (SMA expression) in all precursors (6).

Previously, our group has explored the mechanisms through which kidney tubular cells transform to myofibroblasts, a process termed epithelial-myofibroblast transition (EMyT). We have shown that TGFβ, although necessary, is insufficient for EMyT, implying that an intact epithelium is largely resistant to the transforming effect of this cytokine; the other pre-requisite is the disruption or absence of intercellular contacts (7, 8). We denoted this mechanism the two-hit scheme. Contact disruption (e.g. during wounding) facilitates activation of the Rho family small GTPases Rac1 and RhoA, promoting F-actin polymerization, which in turn induces the nuclear translocation of the transcriptional coactivator, myocardin-related transcription factor (MRTF) (9–11). In resting cells MRTF is sequestered in the cytosol through association with G-actin, which masks its nuclear localization sequence. Once in the nucleus, MRTF associates with serum response factor (SRF), and the complex binds to CC(A/T)₆GG elements (CArG box) present in the promoter of a cohort of target genes, including those encoding cytoskeletal proteins (e.g. SMA) (12, 13). Thus, cytoskeleton remodeling is not only an early feature of the phenotypic shift but also a key driver of the ensuing transcriptional reprogramming, thereby linking cell structure to gene expression via the MRTF/SRF pathway (14). Accordingly, we and others have shown that genetic or pharmacological inhibition of MRTF prevents myofibroblast transition (8, 10) and lessens organ fibrosis (15–17).

Recently, redox signaling has emerged as another TGFβ-induced mechanism that plays a central role in myofibroblast transition. Specifically the NADPH oxidase (Nox) isoform Nox4 (18), a downstream target of TGFβ effectors Smad2/3, has been implicated as a chief mediator of fibroblast-myofibroblast transition (19–23). As opposed to other NADPH oxidase variants whose activity is controlled by stimulus-induced assembly with regulatory subunits (p47, p67, and Rac1), Nox4 is thought to be constitutively active and is regulated primarily at the level of its expression (24, 25). Nox4-mediated reactive oxygen species (ROS) production was found to be necessary for the induction of key myofibroblast features, including contractility, matrix production (extra domain A fibronectin, collagen), and SMA expression (19, 22, 26). Accordingly, interventions that suppress ROS signaling or directly target Nox4 have been shown to attenuate fibrosis in mouse models of chronic kidney disease (27–29). Moreover, we found that EMyT is also associated with Smad3-dependent Nox4 expression (30).

Cognizant of the key role of cytoskeleton remodeling and Nox4 expression in myofibroblast transition, we asked whether these processes might be causally linked. Such a hypothesis is supported by the recent finding that RhoA acts upstream of Nox4-mediated ROS generation (31). Moreover, we noted that the Nox4 promoter contains a CArG box. Intriguingly, increased actin polymerization has been proposed to correlate with enhanced ROS production in the context of cellular aging and apoptosis; however, a causal link and the potential mechanisms remained enigmatic (32). In light of this scenario, we sought to determine whether MRTF and the state of the actin skeleton might regulate Nox4 expression and Nox4-dependent ROS production. We investigated this problem in the context of myofibroblast transition, particularly in EMyT, wherein Smad3- and Rho-dependent processes can be effectively dissected. We show that MRTF signaling and thus the state of the actin skeleton are important permissive inputs for Nox4 expression, which in turn is critical for myofibroblast transition. These studies reveal a new mechanistic link between cytoskeleton organization and cellular redox state (ROS generation). Moreover, we also provide evidence that TAZ and YAP, two cell contact-dependent and Smad3-interacting (33) and Hippo pathway-regulated transcriptional co-activators, are also required for Nox4 expression.

Experimental Procedures

Reagents

NADPH oxidase inhibitor VAS2870, dichlorofluorescein diacetate (DCF-DA), jasplakinolide (JK), MRTF/SRF inhibitor CCG-1423, and TAZ/YAP inhibitor verteporfin were purchased from Sigma. TGFβ was purchased from R&D Systems (Minneapolis, MN). Commercially available antibodies were from the following sources: TAZ/YAP and phospho-myosin light chain II (Thr-18/Ser-19), Cell Signaling Technologies (Danvers, MA); tubulin and SMA, Sigma; GAPDH, Santa Cruz Biotechnology (Santa Cruz, CA); and Nox4, Novus Biologicals (Littleton CO). The rabbit polyclonal MRTF antibody (BSAC) has been described previously (34). Given the concerns regarding the specificity of commercially available Nox4 antibodies and the variability of batches (35), experiments were completed only with antibodies that recognized a protein of the proper molecular weight and of which immunoreactivity was lost upon Nox4 siRNA treatment.

Cell Culture

LLC-PK1 (Cl 4) cells, a porcine proximal tubular epithelial cell line (a kind gift from R. C. Harris, Vanderbilt University School of Medicine, Nashville, TN), and C3H-10T1/2 cells, a mouse embryonic mesenchymal cell line (American Type Culture Collection, Manassas, VA), were cultured in low-glucose DMEM (Invitrogen) supplemented with 10% fetal bovine serum and 1% streptomycin/penicillin solution (Invitrogen). NRK-F (rat kidney fibroblasts, American Type Culture Collection) were cultured in high-glucose DMEM (Invitrogen) with similar supplements. To induce contact disruption, cells were washed three times with phosphate-buffered saline (PBS; Invitrogen) and then cultured in nominally calcium-free DMEM (LCM; Invitrogen), as in our earlier studies (8, 30). Where indicated, cells were treated with 5 ng/ml TGFβ. For siRNA transfections, LLC-PK1 cells were cultured in antibiotic-free medium.

ROS Detection

Nox4 activity and ROS production were measured using a DCF-DA assay, as described previously (36–38). Confluent monolayers were washed 1 time with PBS and 1 time with serum-free medium or LCM and then incubated at 37 °C with 50 μm DCF-DA for 30 min. Following the incubation period, cells were washed 2 times with PBS and then one time with serum-free medium or LCM. Plates were immediately read using SpectraMax 5Me fluorescent plate reader (493 nm excitation, 520 nm emission, Molecular Devices, Sunnyvale, CA) or Fluoroskan Ascent FL plate reader (485 nm excitation, 526 nm emission, Lab Systems, Mumbai, India). Two subsequent readings were then made at 30-min intervals followed by cellular lysis. The results were normalized by dividing the obtained fluorescence values by the absorbance from a BCA protein assay (absorbance 562 nm) to generate fluorescence units.

Plasmids and Transfection

The Nox4 luciferase construct was a gift from Dr. Thannickal (University of Alabama, Birmingham) and has been described previously (39). The HA-MRTF B construct has been described previously (40). The SMA luciferase construct was a gift from S. H. Phan (University of Michigan Medical School, Ann Arbor) and contains the proximal 765 bp of the rat SMA promoter cloned into a pGL3-basic vector. The thymidine kinase promoter-driven Renilla was purchased from Promega (Madison, WI). Transfections of the constructs were completed using JetPRIME (Polyplus Transfection SA, New York) in accordance with the manufacturer's guidelines. The inactivation of the CArG box in human Nox4 promoter constructs was achieved by standard PCR-based mutagenesis using a high fidelity proofreading polymerase. The primer pair and subsequent mutations (in parentheses) were as follows: 5′-TATTGGTTTAACAAAAATTTATTAAATATCTATTTTGCAT-3′ and 5′-ATGCAAAATAGATATTTAATAAATTTTTGTTAAACCAATA-3′ (C⁻⁸⁶⁰/A; C⁻⁸⁵⁹/A; G⁻⁸⁵²/A and G⁻⁸⁵¹/A). All new constructs were verified by sequencing.

Luciferase Reporter Assays

Luciferase assays were completed in accordance with our previously published data (8, 30, 40). Briefly, cells were transfected at 60% confluence with luciferase reporter constructs (0.25 μg/ml) and the thymidine kinase promoter-driven Renilla (0.025 μg/ml). Following the indicated treatments, cells were lysed using 1× passive lysis buffer (Promega). Subsequently, luciferase activity was measured using a dual assay luciferase kit (Promega).

RNA Interference

RNA interference was used to target the following porcine-specific sequences: MRTF A, 5′-CCAAGGAGCUGAAGCCAAAUU-3′; MRTF B, 5′-CGACAAACACCGUAGCAAAUU-3′; Yap, 5′-UCAAAGCGCUCCAGUGAAAUU-3′; Taz, 5′-GGAAGAAGAUCCUGCCUGAUU-3′; and Nox4, 5′-CCGCAGACCUGGAUUUUU-3′. siRNA synthesis was completed by Applied Biosystems (Burlington, Ontario, Canada) or by Thermo Scientific (Waltham, MA). Nonrelated (NR) control siRNA was purchased from Applied Biosystems (Silencer Negative Control siRNA number 2). siRNA transfections were completed using either JetPRIME or Lipofectamine RNAiMAX (Invitrogen) in accordance with the manufacturer's guidelines.

Real Time Quantitative PCR (qPCR)

After treatment, mRNA was extracted using the RNeasy mini kit (Qiagen Venlo, Netherlands) according to the manufacturer's instructions as in Ref. 41. mRNA was reverse-transcribed using iScript^TM reverse transcription (Bio-Rad). iQ^TM SYBR® Green qPCR was performed in triplicate on each cDNA sample using 1 μl of cDNA and 10 μl of 2× iQ^TM SYBR® Green Supermix (Bio-Rad), 2 μm forward and reverse primers. Specific primers targeting Nox4 (sense, 5′-GGA ACG CAC TAC CAG GAT GT-3′, and antisense, 5′-CAG GTC TGC GGA AAG TTA GC-3′), MRTF (sense, 5′-CAG ATG GCA CCA CCC ATA TCC TTA-3′, and antisense, 5′-AGC GGA GCG ATT TCA TTA GCC GAA-3′), and GAPDH (sense, 5′-GCA AAG TGG ACA TGG TCG CCA TCA-3′, and antisense, 5′-AGC TTC CCA TTC TCA GCC TTG ACT-3′) were used. qPCR analysis was performed in a Bio-Rad iCycler Thermal Cycler with iQ5 Multicolor real time PCR detection system (95 °C 3 min; 48 × 96 °C 30 s, 55 °C 1 min) with GAPDH as reference gene. Data were analyzed using iQ5 software. Melting curves always confirmed the presence of only one amplicon. Calculation of the relative expression ratio was done using ΔΔCt.

Western Blotting

Following the indicated treatments, cells were first washed two times with PBS and then lysed with Triton lysis buffer (30 mm HEPES, pH 7.4, 100 mm NaCl, 1 mm EGTA, 20 mm sodium fluoride, and 1% Triton X-100) supplemented with 1 mm sodium vanadate, 1 mm phenylmethylsulfonyl fluoride, and Complete Mini Protease Inhibitor (Roche Applied Science, Mississauga, Ontario, Canada). In experiments that measured relative phosphomyosin II levels, a RIPA lysis buffer was used (50 mm Tris, 150 mm NaCl, 0.1% SDS, 0.5% sodium deoxycholate, 1% Triton X-100, pH 7.4) that was supplemented as described above. Protein concentration was determined using a BCA assay (Thermo Scientific). SDS-PAGE and Western blotting was performed as described previously (8). Briefly, blots were blocked in Tris-buffered saline containing 3% BSA and incubated with the primary antibody overnight. Antibody binding was visualized with the corresponding peroxidase-conjugated secondary antibodies and the enhanced chemiluminescence method (GE Healthcare, Mississauga, Ontario, Canada). Where indicated, blots were stripped and reprobed to demonstrate equal loading or to detect levels of down-regulated proteins. Densitometry was performed using a GS800 densitometer and Quantity One software (Bio-Rad).

Immunofluorescence Microscopy

Cells were seeded onto glass coverslips, grown to confluence, and then treated as indicated. The coverslips were fixed using 4% paraformaldehyde (PFA; Canemco & Marivac, Lakefield, Quebec, Canada) for 20 min. Next, PFA was quenched with 100 mm glycine. Cells were then permeabilized with 0.1% Triton X-100 in PBS and then blocked with 3% bovine serum albumin for 1 h. Samples were incubated with the indicated primary antibodies for 1 h, washed with PBS, and then incubated with 4′,6-diamidino-2-phenylindole (DAPI; Lonza, Basel, Switzerland) and fluorescently labeled secondary antibody (Jackson ImmunoResearch, West Grove, PA) for 1 h. Dako (Burlington, Ontario, Canada) mounting medium was used to mount coverslips onto slides. Images were collected using a WaveFX spinning-disk microscopy system (Quorum Technologies, Guelph, Canada) equipped with an ORCA-Flash4.0 digital camera. Z sections were collapsed into a maximum intensity projection using MetaMorph Premiere software (Molecular Devices) to generate the presented images.

Unilateral Ureteric Obstruction

All animal studies were carried out according to a protocol approved by the animal care committee of the St. Michael's Hospital. Male C57BL/6 mice (6–8 weeks old) were purchased from The Jackson Laboratory (Bar Harbor, ME) and underwent either left-sided unilateral ureteral obstruction (UUO) or sham surgery. Briefly, a left-sided flank incision was made in anesthetized mice, and the left kidney and ureter were identified. Subsequently, the left ureter was obstructed via two 4-0 silk suture knots, just distal to the renal pelvis. One or 2 weeks post-surgery, the mice were sacrificed, and the left kidneys were harvested. Half of the kidney was snap-frozen in liquid nitrogen, and the other half was fixed and embedded in paraffin or frozen fresh in Optimum Cutting Temperature (OCT) medium. The samples were stored at −80 °C until processed. For studies using verteporfin, male mice (6–8 weeks old) underwent left-sided UUO or sham surgery. UUO mice were randomized to receive intraperitoneal injection of either verteporfin (100 mg/kg) every other day, dissolved in 10% DMSO/PBS, or 10% DMSO/PBS vehicle control. Treatment began immediately post-surgery and continued until day 7.

Immunofluorescent Staining of Kidney Sections

Snap-frozen kidney sections from sham or UUO animals were embedded on glass coverslips and fixed using 4% PFA for 20 min. PFA was subsequently quenched using 100 mm glycine and then permeabilized with 0.1% Triton X-100 in PBS. Samples were blocked using 5% donkey serum for 1 h and then incubated with primary antibody overnight at 4 °C. The following day, samples were incubated with DAPI and fluorescently labeled secondary antibody for 1 h. Samples were then immersed in Dako mounting medium and mounted on coverslips. Analysis of fixed samples was completed using an Olympus IX81 microscope as above.

Whole Kidney Lysates

Kidney tissues from sham and UUO (7 and 14 days) mice were cut into very small pieces with a razor blade. The samples were homogenized with a rotor-stator homogenizer in ice-cold RIPA buffer containing 50 mm Tris-HCl, pH 7.4, 150 mm NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, and 1 mm EDTA supplemented with fresh 1 mm PMSF, 1 mm sodium vanadate, 1× protease inhibitor mixture. The samples were centrifuged at 12,000 rpm for 20 min at 4 °C. Protein content of the supernatants was determined, and samples were analyzed using SDS-PAGE and Western blotting as above.

Laser Capture Microdissection (LCMi)

Mouse kidneys from sham and unilateral UUO (7 and 14 days)-treated animals were dissected, placed in cryomolds, embedded in Tissue-Tek OCT medium, immediately frozen in liquid nitrogen, and stored at −80 °C. The tissues were sectioned at 2–5 μm in a cryostat and mounted on glass slides. Rapid H&E staining was performed as reported previously (42) with minor modification. Briefly, frozen sections were fixed with 70% ethanol for 1 min and then washed with RNase-free water for 10 s. Sections were stained rapidly with Mayer's hematoxylin solution (Sigma) for 30 s, washed with RNase-free water for 10 s, dehydrated with an ethanol gradient, and counterstained with alcoholic eosin Y solution (Sigma) for 30 s. Sections were washed three times with 100% ethanol (10 s each) and then two times with xylenes (5 min each). After air-drying for 10 min, LCMi was performed using the PixCell II laser capture microdissection system and CapSure LCMi Caps (Arcturus Mountain View, CA). The following parameters were set for dissection: spot size,7.5 μm; power, 5 milliwatts; duration, 650 μs. A total of 100–200 target tubules and interstitial areas were captured for each sample. Total RNA from LCMi-captured samples was isolated using the RNeasy micro RNA isolation kit (Qiagen) and reverse-transcribed using Sens iScript Reverse Transcription kit (Qiagen) using oligo(dT) primer. The PCR was performed using iProof High Fidelity PCR kit (Bio-Rad). The following primers were used to evaluate cell-specific markers: Tamm-Horsfall protein (sense, 5′-TCA GCC TGA AGA CCT CCC TA-3′, and antisense, 5′-TGT GGC ATA GCA GTT GGT CA-3′); basic amino acid transporter (sense, 5′-ACG TCT TCC TCG TGG TTC TG-3′, and antisense, 5′-GCC ATC TCT TAG GGA GCT T-3′); Nephrin (sense, 5′-GTG TTT TTC TTC GGG TGT CA-3′, and antisense, 5′-CCA CTT TCG TCA GGG GAG TA-3′); collagen I (sense, 5′-CCT GG TAA AGA TGG TGC C-3′, and antisense, 5′-CAC CAG GTT CAC CTT TCG CAC C-3′); and GAPDH (sense, 5′-ACC ACA GTC CAT GCC ATC AC-3′, and antisense, 5′-CAC CAC CCT GTT GCT GTA GCC-3′).

Statistical Analysis

Data are presented as representative blots or images from at least three similar experiments or as means ± S.E. of the mean (S.E.) for the number of experiments indicated. Statistical significance was determined by two-tailed Student's t test or one-way analysis of variance (Tukey or Dunn post hoc testing for parametric and nonparametric analysis of variance, as appropriate), using Prism and Instat software. p < 0.05 was accepted as significant. *, p < 0.05; **, p < 0.01.

Results

Fibrogenic Transition of the Epithelium Is Associated with Increased Nox4 Expression, Which Contributes to the Ensuing ROS Production

Although fibroblasts require only TGFβ for myofibroblast transition, the epithelium becomes susceptible to such TGFβ-provoked phenotypic shift only when the integrity of intercellular contacts is compromised (8, 10). To assess whether such synergy between TGFβ and cell contact-dependent signaling is also required for enhanced ROS production and Nox4 expression, we investigated these processes in the context of our previously characterized two-hit scheme (7, 8, 10). To this end, LLC-PK1 tubular cells were exposed to TGFβ or LCM (which uncouples the intercellular contacts) or the combination of these stimuli, which provokes EMyT. To monitor the ensuing intracellular ROS production, cells that had been treated with these stimuli were briefly loaded with DCF-DA, and the extent and rate of change in DCF fluorescence were followed for an hour (Fig. 1, A and B). Untreated cells exhibited slow ROS production, which was not measurably affected by TGFβ. Note that our previous studies have verified that confluent epithelial cells do not show general unresponsiveness to TGFβ, and the TGFβ receptor remains accessible and functional, as evidenced by rapid Smad3 phosphorylation and other downstream events (7). Yet, this response is not sufficient to induce lasting ROS production (above) or SMA expression (8). LCM alone caused a very modest albeit significant increase in the rate of ROS production (Fig. 1B). Importantly, TGFβ + LCM induced a substantial rise in the extent and rate (2.5-fold) of ROS generation (Fig. 1B). These results imply that markedly enhanced ROS generation occurred only under conditions that promote myofibroblast transition. Next, we investigated whether Nox4 expression followed a similar pattern. In untreated epithelial cells basal Nox4 expression was marginal. Again, TGFβ failed to elicit a detectable effect, whereas LCM induced weak Nox4 protein expression. However, simultaneous application of these stimuli led to robust Nox4 expression, indicating strong synergy between these inputs (Fig. 1C). Accordingly, 24 h of combined treatment significantly increased Nox4 mRNA (>10-fold), which continued to grow further by 72 h (Fig. 1D). Thus, in kidney epithelial cells, the two-hit condition induces Nox4 expression at the transcriptional level, the onset of which precedes EMyT.

FIGURE 1.

ROS production, Nox4 protein, and mRNA expression increase upon double hit-induced EMyT. A, LLC-PK1 cells were treated as indicated for 72 h and then loaded with 50 μm DCF-DA as described under “Experimental Procedures.” DCF fluorescence was determined after the indicated times and normalized to the total cellular protein. Lines represent mean fluorescence units (n = 3). B, values from A were used to express the rate of fluorescence increase. Bars represent mean fold relative to untreated cells (n = 3). C, LLC-PK1 cells were treated for 48 h with one or both stimuli of the two-hit scheme, as indicated (control, serum-free; TGFβ, 5 ng/ml; LCM, low calcium medium). Lysates were then subjected to Western blotting for Nox4 and GAPDH. D, LLC-PK1 cells were treated as indicated for either 24 or 72 h. mRNA was extracted, and qPCR was performed to determine relative amounts of Nox4 mRNA. Bars represent mean fold increase in Nox4 mRNA levels (n = 3). Note that the y axis is logarithmic, and data are normalized to 24 h LCM + TGFβ condition to show fold changes, because this was always readily detectable although the basal levels were often near the threshold of detectability.

Next, we wished to examine whether epithelial Nox4 overexpression occurs in an animal model of kidney fibrosis (Fig. 2). UUO appeared particularly relevant, as both TGFβ-induced Smad3 activation (43) and heightened Rho signaling (44) (partly due to ensuing mechanical stretch) have been implicated as key pathogenic mechanisms in this fibrosis model. As expected, UUO induced strong expression of fibrotic markers (e.g. fibroblast specific protein-1 (FSP-1) and SMA), which was accompanied by a remarkable rise in total Nox4 protein levels (Fig. 2A). Importantly, immunofluorescence staining of kidney sections revealed that the most prominent increase in Nox4 expression occurred in the tubular epithelium (Fig. 2B). To substantiate tubular Nox4 changes by an independent method, we performed laser capture microdissection. Using a variety of tubular (Tamm Horsfall protein and basic amino acid transporter), glomerular (nephrin), and interstitial (collagen I) markers, we verified the successful isolation of the tubular compartment (Fig. 2C). Nox4 mRNA was then determined by qPCR in tubular samples obtained from sham animals or mice exposed to UUO for 1, 3, or 7 days (Fig. 2D). A 3-fold significant increase was seen as early as day 3. Thus, although the exact role of Nox4 in the injured tubular epithelium as well as the role of the epithelium in fibrogenesis remain a matter of debate (see “Discussion”), these observations indicate that fibrosis is associated with enhanced tubular Nox4 expression and validate the two-hit model as a relevant in vitro approach to address the underlying mechanisms.

FIGURE 2.

Nox4 protein levels and mRNA are up-regulated in the tubular compartment during renal fibrogenesis. A, lysates were prepared from the left kidney of sham animals or mice challenged with UUO for the indicated times. Western blotting was used to probe for Nox4, SMA, FSP-1, as well as tubulin as a loading control. Sham controls are 14 days post-intervention. B, kidneys were collected from mice subject to the indicated intervention. Sections were mounted and stained for Nox4 and DAPI. T denotes the tubular lumen. Scale bars, 10 μm. C, verification of successful separation of the tubular and interstitial compartments. Approximately 100 laser spots were collected for each sample from Sham (S) or 7-day UUO (U) mice; tub, tubular samples; int, interstitial samples; wk, whole kidney. Samples were analyzed by end-point RT-PCR (35 cycles) for various markers. Tubular markers: Tamm Horsfall protein (THP) and basic amino acid transporter (BAT); glomerular marker, nephrin; interstitial marker, collagen I. Control (Cont), no cDNA was added to the reaction; M, size markers. Note that interstitial areas from UUO mice contain no tubular or glomerular markers. The slight tubular signal in sham animals is due to the fact that there is very little interstitium among the tubules in healthy mice. D, mice were subject to indicated treatments, and subsequently tubular remnants were isolated using LCMi as described under “Experimental Procedures.” Nox4 mRNA levels were then measured using qPCR analysis. Bars represent mean fold changes in expression (n = 3).

Having established robust Nox4 expression under two-hit conditions, we asked whether Nox4 is a significant contributor to ROS production in this setting. To test this, cells were transfected with NR or Nox4-specific siRNA, the latter of which potently (>70%) inhibited Nox4 protein expression (Figs. 3A and also 8C). Subsequently, cells were treated according to the two-hit scheme followed by DCF fluorescence measurements (Fig. 3B). Nox4 siRNA-transfected and LCM + TGFβ-treated cells showed significantly suppressed accumulation of DCF fluorescence compared with their NR siRNA-transfected counterparts (Fig 3B, panels i–iii). Nox4 silencing reduced the rate of ROS generation by ≈50% (Fig. 3C), which nonetheless remained higher than that of the non-stimulated control. Taken together, these results show that ROS production is heightened during EMyT, and a substantial portion of this effect is due to the ensuing Nox4 expression.

FIGURE 3.

Nox4 expression is a major contributor to ROS production during EMyT. A, LLC-PK1 cells were transfected with 50 nm NR or Nox4-specific siRNA at 60% confluence. Forty eight hours post-transfection, lysates were collected and Western blotting was performed for the indicated proteins. Bars represent mean fold Nox4 levels relative to 48 h LCM + TGFβ (n = 4). B, LLC-PK1 cells were transfected with 50 nm NR (panel i) or Nox4 siRNA (panel ii) at 60% confluence followed by the indicated treatment for 24 h. Cells were subsequently incubated with 50 μm DCF-DA, and fluorescence was monitored as in Fig. 1. Data are expressed as mean fluorescence units (n = 3) for each condition. The functions representing the 24-h LCM + TGFβ condition from panels i and ii were plotted on the same graph (panel iii). C, values from (B, panel iii) were used to calculate the rate of fluorescence increase. Bars represent mean fold relative to control (NR siRNA)-transfected cells (n = 3).

FIGURE 8.

EMyT is dependent upon Nox4 expression and ROS production. A, medium of LLC-PK1 cells was supplemented with vehicle (DMSO) or the general NADPH inhibitor VAS2870 (10 μm), and cells were then treated as indicated for 72 h. Lysates were collected, and Western blotting was performed for the indicated proteins. B, LLC-PK1 cells were grown to 60% confluence and transfected with the SMA-Luc promoter construct. Upon reaching confluence, cells were treated as indicated for 24 h, and the medium was supplemented with either DMSO or 10 μm VAS870. Luciferase activity was measured, normalized to Renilla, and expressed as fold change relative to control (n = 3). C, LLC-PK1 cells were transfected with 50 nm NR or Nox4 siRNA at 60% confluence. Upon reaching confluence, cells were treated for 72 h as indicated. Lysates were then collected, and Western blotting was performed on indicated proteins. D, LLC-PK1 cells were co-transfected with either NR or Nox4 siRNA along with the SMA-Luc reporter system. Upon reaching confluence, cells were treated as indicated for 48 h, and luciferase activity was measured as above (n = 3).

Nox4 Expression Is Regulated by Myocardin-related Transcription Factor

TGFβ has been described to stimulate Nox4 expression in mesenchymal cells, presumably through SBE in the Nox4 promoter (39, 45). However, as shown above, TGFβ alone was insufficient to elicit well detected Nox4 protein expression in tubular cells, indicating that Smad3 signaling is qualitatively or quantitatively insufficient for this response in the intact epithelium. Searching for a mediator, MRTF appeared as an attractive candidate because of the following: 1) we have shown that MRTF, downstream to RhoA/Rac1 signaling, is indispensable for EMyT (8, 10, 40) and TGFβ prolongs and augments the contact disruption-induced nuclear accumulation of MRTF (8); 2) Rho has been recently implicated as an important input for fibroblast-myofibroblast transition (46); and 3) we noted a CArG box in the Nox4 promoter (Fig. 4C, red letters). Based on this rationale, we tested whether MRTF might contribute to Nox4 expression in the epithelium. We efficiently down-regulated both isoforms of MRTF (Fig. 4A) and compared stimulus-induced changes in Nox4 expression in NR or MRTF siRNA-transfected cells. MRTF silencing abolished both LCM-induced and LCM + TGFβ-induced Nox4 protein expression (Fig. 4A). To assess whether MRTF down-regulation affected Nox4 mRNA, we performed qPCR measurements. MRTF siRNA caused >60% reduction in the two-hit provoked rise in Nox4 mRNA (Fig. 4B). Next, we sought to assess whether MRTF indeed affects the activity of the Nox4 proximal promoter and whether the putative CArG box is involved in this effect. We compared the activation of the Nox4 promoter in cells transfected with NR or MRTF siRNA (Fig. 4C). Elimination of MRTF significantly mitigated the two-hit-induced activation of the promoter, nearly abolishing the potentiating effect of LCM. Moreover, an inactivating mutation of the CArG box had a similar effect, i.e. it significantly suppressed the synergy between TGFβ and LCM (Fig. 4D). Taken together MRTF, acting through the identified CArG box, is a significant contributor to Nox4 up-regulation.

FIGURE 4.

MRTF is necessary for Nox4 expression and ROS production during EMyT. A, LLC-PK1 cells were transfected with 50 nm NR or MRTF A + B siRNA at 60% confluence. Upon reaching confluence, cells were treated as indicated for 48 h followed by cell lysis and Western blotting for the indicated proteins. Bars represent fold change in Nox4 levels relative to those obtained after 48 h of LCM + TGFβ treatment (n = 3). B, LLC-PK1 cells were transfected with 50 nm NR or MRTF A + B siRNA at 60% confluence and then treated as indicated for 72 h followed by mRNA extraction. qPCR was used to determine Nox4 mRNA levels. Data were normalized to mRNA levels obtained after LCM +TGFβ treatment. Bars represent mean fold (n = 3). C, LLC-PK1 cells were transfected with the Nox4 luciferase reporter (Nox4-Luc) and the indicated siRNA at 60% confluence. Upon reaching confluence, cells were treated according to the two-hit scheme as indicated for 24 h, followed by the determination of luciferase activity. Bars represent mean fold change relative to control (n = 6). The CArG box is shown above the graph. SBE designates the previously described Smad-binding element. D, LLC-PK1 cells were transfected with the wild type (WT) or CArG mutant Nox4 promoter (as shown above the graphs) at 60% confluence and upon reaching confluence were treated as indicated for 24 h followed by measurement of luciferase activity. Bars represent mean fold change in luciferase activity relative to control (n = 9). E, LLC-PK1 cells were transfected with 50 nm NR siRNA (panel i) or 50 nm MRTF A+B siRNA (panel ii) as above. Cells were subsequently incubated with 50 μm DCF-DA, and fluorescence was measured (n = 3) for each condition. For better comparability, values obtained in the LCM + TGFβ-treated NR-transfected (panel i) and MRTF A + B siRNA-transfected (panel ii) samples were plotted on the same graph (panel iii) (n = 3). F, rate of fluorescence increase was determined from E, panel iii (n = 3).

We argued that if MRTF is indeed critical for Nox4 expression, and Nox4 expression is a significant mechanism for the ensuing rise in ROS production, then MRTF down-regulation should reduce ROS generation. To test this assumption, changes in DCF fluorescence were monitored in cells transfected with NR or MRTF-specific siRNA and exposed to the individual and combined stimuli of the two-hit scheme (Fig. 4, E, panels i–iii, and F). In agreement with our hypothesis, MRTF down-regulation significantly inhibited the extent (Fig. 4E, panels i–iii) and rate (Fig. 4F) of ROS production, which was most prominent upon combined stimulation. Remarkably, the extent of the inhibition obtained by MRTF silencing was essentially the same as observed upon Nox4 down-regulation. Together, these results suggest that MRTF is required for enhanced ROS generation in the transitioning epithelium.

Given that MRTF is involved in Nox4 expression and ROS production, we surmised that these processes should be potentiated by enhanced F-actin polymerization, a key regulator of MRTF localization. An intriguing aspect of this possibility is that actin polymerization and augmented oxidative stress have been proposed to be coupled (32); however, the underlying mechanisms remained elusive. To test this assumption, we treated the cells with the actin polymerizing drug, JK, and TGFβ alone or in combination. Although these stimuli had modest effect in themselves, their combination resulted in a significant activation of the Nox4 promoter (Fig. 5A) and a large increase in Nox4 mRNA level (Fig. 5B). Importantly, siRNA-mediated down-regulation of MRTF suppressed the JK + TGFβ-induced rise in Nox4 mRNA by 70% (Fig. 5C). To monitor the functional consequences of JK treatment, DCF fluorescence was followed in the absence and presence of this drug and TGFβ. JK itself promoted ROS production, which was further enhanced by TGFβ (which alone had no effect) (Fig. 5D). Similar results were obtained in cells transfected with NR siRNA (Fig. 5E). In contrast, Nox4-specific siRNA reduced the effect of JK and completely prevented any further stimulation by TGFβ (Fig. 5F). Accordingly, ROS production was significantly higher in NR siRNA-transfected than in Nox4 siRNA-transfected cells (Fig. 5G). Taken together, JK (and thus actin polymerization) and TGFβ synergistically provoke Nox4 expression, and JK augments ROS production partially through Nox4.

FIGURE 5.

Pharmacologically provoked actin polymerization synergizes with TGFβ to induce Nox4 promoter activation, mRNA expression, and ROS production in an MRTF-dependent manner. A, LLC-PK1 cells were transfected with the Nox4-Luc promoter and upon reaching confluence were exposed to either vehicle (DMSO) or the actin polymerization-inducing drug jasplakinolide (1 μm) in the absence or presence of TGFβ for 24 h. Normalized Nox4-luciferase activities were then determined under each condition. Data are expressed as fold change relative to control, mean (n = 3). B, LLC-PK1 cells were treated as in A followed by qPCR analysis of Nox4 mRNA levels. Data are expressed as fold change in mRNA relative to the level obtained upon TGFβ exposure (n = 3). C, LLC-PK1 cells were transfected with 50 nm NR or MRTF A+B siRNA and then treated as indicated for 24 h. Subsequently, Nox4 mRNA was determined by qPCR analysis. Note that MRTF down-regulation effectively suppressed Nox4 mRNA expression induced by JK + TGFβ, mean (n = 3). D–G, Nox4 expression contributes to JK-promoted ROS generation. D, to assess the impact of actin polymerization on ROS production, LLC-PK1 cells were treated for 24 h with DMSO or 1 μm JK in the absence or presence of TGFβ. After DCF-DA loading, fluorescence was followed in the indicated time range. Data are expressed as mean (n = 3). E and F, cells were transfected with either 50 nm NR (E) or Nox4 siRNA (F) and treated and processed as in D. G, LLC-PK1 cells were treated with TGFβ and 1 μm JK for 24 h with the addition of either NR or Nox4 siRNA as in E and F. Following treatment, cells were loaded with DCF-DA as above, and fluorescence was measured for the indicated time range. Date are expressed as mean (n = 6). Note significant inhibition in ROS production in Nox4 down-regulated cells in which the synergy between JK and TGFβ was lost.

Although these results imply that MRTF is necessary for Nox4 expression, it remained an open question whether MRTF activation is sufficient for this process. In fact, previous studies conducted on fibroblasts emphasize the key importance of Smad3-dependent signaling through an SBE in the Nox4 promoter (39), and we found that Smad3 is also required for this process in the transitioning epithelium (30). To address the above question, initially we confirmed the need for the MRTF/SRF axis for the regulation of the Nox4 promoter using yet another (pharmacological) approach. CCG-1423 is a potent inhibitor of MRTF or the MRTF/SRF interaction (47). Pretreatment with this drug strongly reduced the activity of the Nox4 promoter induced by the double hit scheme compared with vehicle control (Fig. 6A). Thus, in agreement with the siRNA data, the full-length promoter is sensitive to the pharmacological inhibition of MRTF. Next, we compared the double hit-induced activation of various truncation constructs of the Nox4 promoter. The shortest construct tested (−741), which lacks both the SBE and the CArG box, showed much less (only marginal) activation (Fig. 6B). Interestingly, the construct that contains the CArG box but not the SBE (−1434) exhibited a similarly weak response (Fig. 6B). Furthermore, overexpression of MRTF (which can strongly induce the SMA promoter) failed to activate the full-length Nox4 promoter (Fig. 6C). Together, these results show that SRF/MRTF, although essential, are not sufficient for Nox4 promoter activation, implying that these factors play an indispensable, permissive role in this process.

FIGURE 6.

MRTF is necessary but not sufficient to drive the Nox4 promoter. A, LLC-PK1 cells were grown to 60% confluence and transfected with the Nox4 Luc construct. Upon reaching confluence, cells were treated as indicated for 24 h in the presence of DMSO or 4 μm CCG-1432. Subsequently, lysates were collected, and luciferase activity was measured. Bars represent mean fold luciferase activity relative to control (DMSO) normalized to Renilla (n = 3). B, LLC-PK1 cells were grown to confluence and transfected with the indicated promoter constructs (WT, −741 and −1434 Nox4-Luc). Upon reaching confluence, cells were treated as indicated for 24 h, and lysates were collected for measurement of luciferase activity. Bars represent mean fold change in luciferase activity relative to control (Cont) normalized to Renilla (n = 3). C, LLC-PK1 cells were grown to 60% confluence and transfected with pCMV5 or HA-MRTF in addition to the Nox4 Luc construct. Cells were subsequently serum-starved for 24 h, and lysates were collected for luciferase activity measurement. Bars represent mean fold luciferase activity relative to control (pCMV5) normalized to Renilla (n = 3).

TAZ and YAP Contribute to the Regulation of Epithelial Nox4 Expression

The fact that TGFβ (and Smad3 signaling) is necessary for Nox4 expression together with our finding that epithelial Nox4 expression also requires contact injury prompted us to ask whether the TGFβ-dependent inputs might also be contact-sensitive. The Hippo pathway effector TAZ and YAP are good candidates to link contact disruption to enhanced Smad3 signaling, because these factors translocate to the nucleus upon the uncoupling of intercellular junctions (41, 48) and can act as nuclear Smad3 retention factors (48). Furthermore, crosstalk between TGFβ and TAZ/YAP signaling has been described, and depending on the target, these pathways may be synergistic or antagonistic (49). To test whether TAZ and YAP might be involved in the up-regulation of Nox4 in the transitioning epithelium, we examined Nox4 expression upon down-regulation of both TAZ and YAP (Fig. 7). In cells transfected with NR siRNA, TGFβ itself caused a substantial increase in TAZ and (to a smaller extent) in YAP protein level, which was further stimulated by contact disruption (Fig. 7A). The mechanism of this effect is subject to ongoing studies and will not be addressed here (see under “Discussion”). Importantly, transfection with TAZ and YAP siRNA efficiently suppressed both the basal level and the stimulus-induced increase in the expression of these proteins and concomitantly prevented the double hit-provoked induction of Nox4 expression (Fig. 7A). To substantiate this finding, we measured the activity of the Nox4 promoter in cells transfected with NR or TAZ and YAP siRNA. Knockdown of TAZ and YAP (TAZ/YAP) expression led to a significant drop in the basal promoter activity, abolished the TGFβ-induced rise, and accordingly strongly suppressed the level achieved by the two-hit stimulation (Fig. 7B). Given that suppression of TAZ/YAP was sufficient for abrogation of Nox4 expression in vitro, we next determined whether this was also the case in a mouse model of obstructive nephropathy using the UUO model. Mice were treated with the TAZ/YAP inhibitor verteporfin (VP) (50). Under control conditions, UUO mice demonstrated up-regulation of tubular Nox4 as assessed by immunofluorescence analysis (Fig. 7C) and also had significantly increased levels of tubular Nox4 mRNA relative to sham mice (Fig. 7D). In comparison, treatment with VP led to substantial reduction in both of these parameters (Fig. 7, C and D). Taken together, these results demonstrate that TAZ and YAP are necessary for efficient Nox4 expression in tubular cells challenged with fibrogenic stimuli both in vitro and in vivo.

FIGURE 7.

TAZ and YAP are necessary for Nox4 expression. A, LLC-PK1 cells were grown to 60% confluence and transfected with either 100 nm NR or TAZ + YAP siRNA. Upon reaching confluence, cells were treated according to the individual or combined stimulation of the two-hit scheme for 48 h. Lysates were collected and analyzed by Western blotting for YAP, TAZ, Nox4, and tubulin. B, LLC-PK1 cells at 60% confluence were transfected with 100 nm NR or TAZ + YAP siRNA along with the Nox4-Luc construct. Upon reaching confluence, cells were treated as indicated for 24 h, and luciferase activity was measured. Data are expressed as fold change in normalized luciferase activity compared with the untreated control (n = 3). C, kidneys were collected from Sham, UUO, and UUO + VP mice 7 days post-surgery. Sections were mounted and stained for Nox4 and DAPI. T denotes the tubular lumen. Scale bars, 10 μm. D, mice were subject to the indicated treatments, and subsequently tubular remnants were isolated using LCMi as described under “Experimental Procedures.” Nox4 mRNA levels were then measured using qPCR analysis. Bars represent mean fold changes in expression (n = 3).

NADPH Oxidase Activity and Nox4 Expression Are Necessary for the Myogenic Shift of the Epithelium

In non-epithelial progenitors, in which TGFβ is sufficient for myofibroblast transition, ROS production was shown to be necessary for the induction of key myofibroblast features, including increased contractility and SMA expression (19, 21, 22). To determine whether ROS are also essential for double hit-provoked epithelial-myofibroblast transition, we initially tested the effect of the general NADPH oxidase inhibitor VAS2870 (Fig. 8, A and B) (51). As expected, LCM + TGFβ triggered SMA expression in vehicle-treated control cells. This effect was entirely blocked by VAS2870 (Fig. 8A). Subsequent experiments using the SMA promoter luciferase reporter revealed that NADPH oxidases are critical for the activation of the SMA promoter, as VAS2870 strongly suppressed both the LCM-induced response and the synergistic action of LCM + TGFβ (Fig. 8B). To assess whether Nox4 is a significant contributor to the transition, similar experiments were performed using NR or Nox4-specific siRNA. We followed two critical features, myosin phosphorylation and SMA protein expression, both of which were strongly induced by LCM + TGFβ and nearly abolished by Nox4 down-regulation (Fig. 8C). In accordance with these findings, silencing of Nox4 significantly suppressed the activation of the SMA promoter in the epithelium (Fig. 8D). Taken together, Nox4 is a chief mediator of EMyT.

MRTF Plays a Critical Role in Nox4 Expression in Mesenchymal Myofibroblast Progenitors

Recent research suggests that the major source of myofibroblasts during experimental fibrogenesis is not the epithelium but various mesenchymal progenitors, such as interstitial fibroblasts and possibly pericytes (52–54). Nox4 is induced in response to TGFβ in these mesenchymal cell types and is necessary for acquisition of key myofibroblast features (23). We therefore sought to assess whether MRTF signaling was also an important regulator of Nox4 in mesenchymal cells. To this end, we first used 10T1/2 fibroblast/pericyte-like cells, in which TGFβ induced substantial up-regulation of the Nox4 protein (Fig. 9A). Importantly, pretreatment of the cells with the MRTF/SRF inhibitor CCG-1423 completely abrogated the TGFβ-provoked increase in Nox4 protein expression (Fig. 9B) and eliminated the activation of the Nox4 promoter (Fig. 9C). To verify that these effects reflect promoter activation in a CArG box-specific manner, we compared the impact of TGFβ on the WT and CArG box-mutated Nox4 promoter (Fig. 9D). The mutation rendered the promoter completely insensitive to TGFβ. Similar results were obtained in normal rat kidney fibroblasts (NRK-F cells) (Fig. 9E). Taken together, these results imply that, similar to the findings in tubular epithelial cells, MRTF and MRTF/SRF-mediated CArG box-dependent Nox4 expression plays an important role also in myofibroblast transition from mesenchymal progenitors.

FIGURE 9.

Nox4 expression also requires MRTF activity in mesenchymal myofibroblast progenitors. A, 10T1/2 cells were treated as indicated for 48 h. Subsequently, Western blotting was performed for Nox4 and tubulin. B, 10T1/2 cells were subjected to indicated treatments for 48 h in the presence of either DMSO or 4 μm CCG-1432. Following treatment, lysates were collected, and Western blotting was performed for Nox4 and β-actin. C, 10T1/2 cells were transfected with the Nox4-luc promoter at 60% confluence. One day later, cells were treated as indicated for 24 h in presence of DMSO or 4 μm CCG-1432. Subsequently luciferase activity was measured (n = 3). D and E, 10T1/2 cells (D) were transfected with the WT or CArG mutant Nox4 promoter at 60% confluence. One day later cells were treated as indicated for 24 h, and luciferase activity was determined (n = 6). E, NRK-F cells were transfected with the WT or CArG mutant Nox4-luc construct at 60% confluence. One day later cells were treated as indicated for 24 h and luciferase activity was measured (n = 3).

Discussion

This study allows new mechanistic insight into the regulation of Nox4, particularly in an epithelial setting. The key new findings include the following: 1) that MRTF is a major permissive regulator of Nox4 expression, a mechanism that unravels a new link between actin polymerization and the regulation of cellular redox state, present both in epithelial cells and fibroblasts; 2) that the Hippo pathway effectors TAZ/YAP are required for epithelial up-regulation of Nox4 both in vitro and in vivo; and 3) that in an epithelial context TGFβ is insufficient to provoke Nox4 expression, and a second input is also required (two hit model) that may emanate from the uncoupling of the intercellular contacts or other mechanical stimuli, which can activate both MRTF and TAZ/YAP signaling.

ROS in general and Nox4 in particular have been shown to play important roles in a variety of physiological and pathological processes (55). Because Nox4 activity is primarily regulated via Nox4 expression (56, 57), it is imperative to understand the inputs and mechanisms that control this process. Several transcription factors and the corresponding cis-elements have been implicated in transcriptional regulation of Nox4, implying the multifactorial nature of this process. These include Smad3 (39, 45), which through AP1/Smad-binding elements (39) couples Nox4 to TGFβ signaling and thereby fibrogenesis; Hif1α (58), Nrf2 (59) and Oct-1 (60), which link Nox4 expression to ROS generation per se under conditions of hypoxia, hyperoxia, and shear stress; NFκB (61) and STAT proteins (62), mediators of inflammatory gene expression; and E2F (63), a regulator of proliferation. This study enriches this picture by two important findings indicating that MRTF and TAZ/YAP, two main cytoskeleton-regulated transcriptional coactivators, are important novel regulators/modulators of Nox4 expression.

The role of the MRTF/SRF pathway is evidenced by the facts that the Nox4 promoter harbors a perfect CArG box and that down-regulation or pharmacological inhibition of MRTF or the MRTF/SRF interaction or mutation of the CArG box abrogates the activation of the Nox4 promoter, suppresses Nox4 mRNA and protein expression, and reduces the ensuing ROS production in response to various fibrogenic stimuli both in epithelial and mesenchymal cells. This recognition is a significant step forward because it directly links the state of the cytoskeleton to the redox state of the cell. Previous studies have revealed that enhanced actin polymerization (reduced actin dynamics) leads to augmented ROS production, which in turn promotes cellular aging and apoptosis (32). Although the underlying mechanisms remained largely undefined, actin polymerization was proposed to alter mitochondrial ROS production in yeast (64). Here, we show that actin polymerization, directly triggered by JK, potentiates TGFβ-induced Nox4 mRNA expression and ROS production in the epithelium in an MRTF-dependent manner. We therefore propose that MRTF, whose nucleocytoplasmic traffic is regulated by actin polymerization (14), serves as a key mediator through which cytoskeletal changes impact the redox state. This represents a novel long term (transcriptional) mechanism of communication between these systems. It is important to note that changes in actin polymerization may have acute and direct effects on ROS production differing from the transcriptional effects described herein. For example, actin depolymerization was shown to promote p47^phox-dependent, Nox2-mediated superoxide production in neutrophils (65) and endothelial cells (66). The underlying mechanism is likely 2-fold as follows: actin depolymerization enhances receptor-induced signaling pathways leading to the assembly of Nox2 from cytosolic and membrane-bound components, and it may also remove physical constraints from the association of these. Indeed, our earlier studies indicated that F-actin depolymerization triggers several tyrosine kinase pathways (67), which may participate in the regulation of NADPH oxidase enzymes. In contrast to Nox2, Nox4 activity does not require the association of the oxidase with cytosolic partners. Thus, cytoskeleton organization may exert differential effects both in terms of timing (acute regulation versus transcription) and in terms of isoform specificity.

Our findings also explain, at least in part, the regulation of Nox4 by Rho GTPase. A recent elegant paper reported that TGFβ-induced Rho/Rho kinase activation is an essential upstream mechanism for Nox4 expression and activity in renal fibroblasts and their transition to myofibroblasts (31). Mechanistically, the authors have implicated polymerase (DNA-directed) δ-interacting protein 2 (Poldip2), a recently described Nox4 enhancer (68), which was also induced by Rho. However, Poldip2 expression does not, a priori, explain the increase in Nox4 protein levels. Because Rho is a major upstream regulator of MRTF and, as our previous studies show, nuclear accumulation of MRTF is abolished by Rho kinase inhibition (10), our data provide a straightforward mechanism linking Rho and Nox4. Moreover, in silico analysis suggests the presence of a putative CArG box in the promoter of Poldip2 as well. Indeed, although the presence of and requirement for an intact CArG box in the Nox4 promoter is highly suggestive of direct action of SRF and presumably MRTF at this locus, MRTF/SRF may also contribute indirectly by controlling other Nox4 regulators. Future studies should validate this intriguing possibility. Our data also imply that MRTF, although necessary, is not sufficient to drive the Nox4 promoter; TGFβ-dependent signaling remains essential. Thus, MRTF is a permissive and potentiating but not directly inductive factor for this promoter. There is precedence for such permissive role of MRTF with regards to VE-cadherin (69) and CCN1 (70) expression. Furthermore, SRF overexpression alone also failed to drive the promoter (data not shown), excluding that the Nox4 promoter is SRF- but not MRTF-sensitive. Future work should clarify how TGFβ-activated transcription factors synergize with SRF/MRTF to induce Nox4 expression.

A characteristic feature of Nox4-related signaling is the presence of positive feedback loops, wherein Nox4 is both a downstream target and an upstream activator of a particular pathway (71). Thus, whereas TGFβ is a prime inducer of Nox4 expression, Nox4-derived ROS facilitated TGFβ signaling both by promoting the liberation of active TGFβ from the latent complex (72) and by enhancing TGFβ-induced Smad2/3 phosphorylation (19). Similarly, Nox4 can be driven by and activate NFκB (73) as well as Hif1α and -2α (74, 75). Our findings together with earlier reports extend this notion for the Rho/Nox4 relationship as well. Nox4-derived ROS has been suggested to promote Rho activation (68), whereas active Rho can induce MRTF translocation, which in turn stimulates Nox4 expression.

Interestingly, the vast majority of the literature has examined the regulation of Nox4 expression in vascular smooth muscle, endothelium, and fibroblasts, although information in the epithelium is scarce, despite the fact that the tubules were reported to possess 80% of all Nox4 activity in the kidney (35). Two additional reasons lend importance to the epithelial context, namely, the epithelium differs from the other tissues both in terms of the stimuli required for Nox4 induction and in terms of the ensuing (patho)physiological consequences (see below). Considering the first point, our data show that, similar to the induction of EMyT, epithelial Nox4 expression requires two-hit conditions; although TGFβ is sufficient in other tissues, it does not provoke a major increase in Nox4 in the epithelium. The other prerequisite is the injury of the intercellular contacts, which translates into enhanced contractility (30). This finding is concordant with the fact that TGFβ is a weak and transient Rho activator in the intact epithelium and that prolonged nuclear MRTF accumulation requires both stimuli (8, 10). At the same time, MRTF overexpression (or Rho activation, data not shown) is also insufficient, and this implies that TGFβ is also indispensable. This is in agreement with our earlier data showing that Smad3 down-regulation abrogates Nox4 expression under double hit conditions (30). The dependence of Nox4 expression both on TGFβ and the loss of cell contacts prompted us to investigate the potential involvement of the Hippo pathway effectors TAZ and YAP, which, similar to MRTF, are also regulated by these inputs (41, 48). Indeed, knockdown of TAZ/YAP suppressed Nox4 induction. Similar results were obtained in vivo, using the TAZ/YAP inhibitor verteporfin. This is likely due to the fact that TAZ/YAP are Smad3 nuclear retention factors, which can facilitate TGFβ-induced gene expression (49, 76). Thus, the mechanism underlying the synergy between contact disruption and TGFβ is 2-fold: TGFβ prolongs MRTF nuclear accumulation induced by contact disassembly-provoked Rho activation (8), and contact injury translocates TAZ/YAP to the nucleus, which in turn prolongs TGFβ-induced Smad3 accumulation and gene expression. In addition, we noted that TGFβ (in itself) strongly elevates TAZ/YAP levels. Our ongoing studies address the mechanism of this intriguing phenomenon. In any case, this finding further confirms that the intact epithelium remains responsive to TGFβ (although EMyT does not develop) and that TGFβ may prime the cells for injury-induced Nox4 expression through this indirect pathway, too. Importantly, TAZ/YAP are regulated by many fibrogenic conditions, including matrix stiffness (77, 78) and hyperglycemia (79). Clearly, the involvement of TAZ and YAP in the regulation of the redox state of the cell warrants further mechanistic studies.

Finally, the pathophysiological role of Nox4 should be considered. There is strong evidence that Nox4 is up-regulated during and is required for fibroblast-myofibroblast transition in vitro (19, 21–23). Moreover, Nox4 down-regulation by siRNA or antisense oligonucleotides proved to be protective in a lung fibrosis model (22) and reduced matrix deposition in diabetic kidney disease (DKD) (27), whereas new Nox1/4-selective pharmacological inhibitors exerted potent kidney-protective and antifibrotic effects in DKD models (28, 29). All these findings suggest that Nox4 protein levels rise under fibrogenic (predominantly diabetic) conditions, and this contributes to the ensuing nephropathy and fibrosis. Consistent with this scenario and our new findings, a recent paper reports that MRTF-deficient mice are largely protected against DKD (16). However, results obtained with whole body Nox4 knock-out mice are controversial or more complex. One study found that Nox4 deletion indeed mitigated the progression of DKD on the apoE^−/− background (29). However, Babelova et al. (35) using four chronic kidney injury models detected no protective effect of Nox4 deletion; in fact, they reported slight deterioration of obstructive nephropathy. Similar results were obtained in another study, claiming that Nox4 protects against kidney fibrosis (80). How can these contradictory findings be reconciled? To solve this riddle, a biphasic model has been proposed (81). According to this, limited early Nox4-mediated ROS production may be protective because, presumably through Nrf2 and Hif1α, it may induce antioxidant genes (80). In addition, Nox4 is a mediator of cellular senescence (82) and autophagy (83), both of which were proposed to lessen fibrosis (84, 85). However, sustained activation of Nox4 likely contributes to tissue damage and fibrosis. In addition, whether epithelial Nox4 (at least initially) serves more protective roles, while Nox4 expression in fibroblasts is pathogenic, remains an open question, which can be addressed by tissue-specific knockouts.

Although most studies found increased renal Nox4 expression during chronic injury (27, 86, 87), progressive loss was also reported (35). We found that epithelial Nox4 increases during UUO as detected by immunofluorescence as well as by qPCR performed on tubular samples obtained by laser capture microdissection. In the UUO model, the primary insult is a pressure-induced stretch of the tubular compartment, which then can lead to local cytoskeleton remodeling and thus MRTF activation. This in turn can potentiate tubular Nox4 expression. It is conceivable that initially this mechanism is adaptive and renoprotective, although upon sustained overwhelming injury it becomes fibrogenic. It is noteworthy that the epithelium is only a minor source of myofibroblasts via EMyT in vivo (52). This view assigns a key but indirect role to the epithelium in fibrogenesis; upon injury it assumes a secretory phenotype and primes the underlying mesenchyme in a paracrine fashion (53, 88–90). We propose that MRTF is a key mediator of both the partial (secretory) and the full-blown (EMyT) phenotype shift. In fact, our ongoing studies suggest that a subtle interplay between the MRTF and TAZ/YAP pathways may be responsible for preventing the shift to go “all the way” to a myogenic (SMA-expressing) stage. In any case, the recognition of the role of the cytoskeleton-MRTF/SRF axis in the regulation of Nox4 expression both in epithelial and mesenchymal cells should facilitate our understanding of the biology of this intriguing enzyme, which seems to be involved both in preserving tissue integrity and promoting fibrosis.

Author Contributions

M. R. designed, performed, and analyzed experiments shown in Figs. 1 and 3–9, wrote part of the manuscript, and prepared the figures; J. F. B. designed, conducted, and analyzed qPCR experiments shown in Figs. 1, 4, and 5 with the participation of T. E. T. K. P. S. generated construct by mutagenesis an provided experimental advice; Q. D. performed the laser capture microdissection studies; S. G. S. and D. A. Y. were involved in the in vivo experiments and data analysis; K. S. and S. F. P. contributed to the design of studies, critical interpretation of the data, and the final assembly of the manuscript; A. K. conceived and conceptualized the work and wrote the majority of the manuscript. All authors reviewed the results and approved the final version of the manuscript.