Cross talk between lysine methyltransferase Smyd2 and TGF-β-Smad3 signaling promotes renal fibrosis in autosomal dominant polycystic kidney disease

Abstract

Autosomal dominant polycystic kidney disease (ADPKD) is an inherited genetic disorder that is caused by mutations in PKD1 or PKD2 genes and is characterized by renal fluid-filled cyst formation and interstitial fibrosis. PKD1 gene mutation results in the upregulation of SET (suppressor of variegation, enhancer of zeste, trithorax) and MYND (myeloid-nervy-DEAF1) domain-containing lysine methyltransferase 2 (SMYD2) in kidneys from Pkd1 mutant mice and patients with ADPKD. However, the role and mechanism of Smyd2 in the regulation of renal fibrosis in ADPKD remains elusive. In the present study, we showed that 1) expression of Smyd2 can be regulated by transforming growth factor (TGF)-β-Smad3 in normal rat kidney 49F (NRK-49F) cells and mouse fibroblast NIH3T3 cells; 2) knockdown of Smyd2 and inhibition of Smyd2 with its specific inhibitor, AZ505, decreases TGF-β-induced expression of α-smooth muscle actin, fibronectin, collagen type 1 and 3, and plasminogen activator inhibitor-1 in NRK-49F cells; 3) Smyd2 regulates the transcription of fibrotic marker genes through binding on the promoters of those genes or through methylating histone H3 to indirectly regulate the expression of those genes; and 4) knockout and inhibition of Smyd2 significantly decreases renal fibrosis in Pkd1 knockout mice, supporting that targeting Smyd2 can not only delay cyst growth but also attenuate renal fibrosis in ADPKD. This study identified a cross talk between TGF-β signaling and Smyd2 in the regulation of fibrotic gene transcription and activation of fibroblasts in cystic kidneys, suggesting that targeting Smyd2 with AZ505 is a potential therapeutic strategy for ADPKD treatment.

NEW & NOTEWORTHY Here, we identified a cross talk between SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) and transforming growth factor (TGF)-β-Smad3 signaling and a synergistic feedback loop between them, in which TGF-β stimulates expression of Smyd2 in a Smad3-dependent manner, and upregulation of Smyd2 regulates the transcription of TGF-β and other fibrotic marker genes through direct binding on their promoters or methylating histone H3 indirectly to regulate the transcription of those genes in fibroblasts. Thus, the Smyd2-TGF-β-Smad3-Smyd2 signaling axis plays an important role in promoting renal fibrosis, and targeting Smyd2 with its specific inhibitor should not only delay cyst growth but also ameliorate renal fibrosis in ADPKD.

INTRODUCTION

Autosomal dominant polycystic kidney disease (ADPKD) is the most common monogenic inherited kidney disease and is caused by the mutations in multiple genes, among which mutations of PKD1 account for ∼85% and mutations of PKD2 account for 10–15% of clinical cases. ADPKD is typically diagnosed in adults, has an incidence of 1:400–1,000 (1), and is characterized by multiple fluid-filled and progressive enlarged tubular cysts accompanied with interstitial inflammation and fibrotic changes (2, 3). The progression of renal cysts in ADPKD eventually results in a medical condition in which a person’s kidneys cease functioning, named as end-stage renal disease, leading to the need for a regular course of long-term dialysis or a kidney transplant to maintain life.

Renal fibrosis is commonly observed in ADPKD kidneys, which contributes to PKD mutation-mediated end-stage renal disease. Various inflammatory factors, including TNF-α, macrophage migration inhibitory factor, and transforming growth factor (TGF)-β, can be secreted and accumulated in cyst fluid to promote cyst progression in ADPKD (4–7), in which TGF-β has been identified as the most important regulator of renal fibrosis (8–10). The canonical pathway that transduces the TGF-β signal through the membrane receptor to its target genes has been well studied. Activated extracellular TGF-β brings together its serine/threonine kinase receptors, including TGF-β type I receptor (TβRI) and type II receptor (TβRII), to form a complex (11, 12) in which TβRII autophosphorylates and activates TβRI to facilitate the interaction and phosphorylation of Smad2/3 by TβRI. The phosphorylated Smad2/3 complex translocates to the nucleus and forms a complex with Smad4 to initiate transcription of fibrotic genes with other coactivators or corepressors (11). The progression of renal fibrosis relies on both matrix synthesis and matrix degradation, and TGF-β is the most critical regulator of the synthesis and deposition of the extracellular matrix, which is required for tissue repair (13, 14). In addition, TGF-β is able to inhibit epidermal growth factor-dependent proliferation in normal human renal tubule epithelia, which is lost in ADPKD (15). The dual roles of TGF-β in tissue repair and cell proliferation suggest that TGF-β may play a protective role at the early stage of ADPKD kidneys with abnormal proliferation and tissue injury. However, it has been reported that expression of TGF-β and activation of Smad signaling are increased in cyst-lining epithelial cells in polycystic kidney disease (PKD), leading to tissue scar formation and resulting in renal fibrosis and atrophy (6). The mechanism for the elevation of TGF-β signaling in ADPKD remains unknown.

The involvement of an epigenetic mechanism in the regulation of renal fibrosis has been investigated in the past decade (16–18). A previous study has shown that TGF-β can induce expression of the lysine methyltransferase SET domain-containing lysine methyltransferase 7/9 (SET7/9) and that use of siRNA to target SET7/9 or TGF-β-Smad2/3 components suppresses injury-induced renal fibrosis in the unilateral ureter obstruction (UUO) mouse model and in other chronic kidney diseases (16). Other studies have found that histone demethylase LSD1 is increased in mouse kidneys with UUO and in cultured NRK-52E cells undergoing TGF-β1-induced epithelial-mesenchymal transition. Inhibition of LSD1 with its specific inhibitor ORY1001 attenuated renal epithelial-mesenchymal transition and fibrosis, which was associated with decreased deposition of extracellular matrix proteins and expression of fibrotic genes, including α-smooth muscle actin (α-SMA) and fibronectin (18). SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) regulates global gene transcription through chromatin modifications on the methylation of histone H3 at lysine 4 (H3K4) and lysine 36 (H3K36; 19, 20). Smyd2 can also regulate the cytokine network through two independent and synergistical feedback loops (the Smyd2-STAT3-IL-6-Smyd2 and Smyd2-NFκB-p65-TNFα-Smyd2 axes) to promote cyst growth in ADPKD (20, 21). However, whether and how Smyd2 regulates renal fibrosis in ADPKD remains elusive.

In the present study, we found a dysregulation of fibrotic genes in fibroblasts even before cyst initiation in collecting duct cells in Pkd1 mutant mouse kidneys with single-cell RNA-sequencing analysis, suggesting that fibroblasts can be activated through unknown mechanisms at the early stage and contribute to the progression of ADPKD. We also found that both Smyd2 and TGF-β were upregulated in Pkd1 mutant kidneys. Knockout and inhibition of Smyd2 with its specific inhibitor AZ505 ameliorated renal fibrosis and delayed cyst growth. We further found that TGF-β induction induced expression of Smyd2 and fibrotic genes in a Smad3-dependent manner and that inhibition of Smyd2 blocked expression of those genes induced by TGF-β. In addition, we found that Smyd2 transcriptionally regulates the expression of TGF-β and fibrotic genes through either binding to the promoters of those genes or methylation of histone H3K4 in response to TGF-β stimulation. This study suggests that the Smyd2-TGF-β-Smad3-Smyd2 feedback loop promotes renal fibrosis in ADPKD.

METHODS

Cell Culture and Reagents

Normal rat kidney 49F (NRK-49F) cells were purchased from the American Type Culture Collection (Manassas, VA), cultured in DMEM/high-glucose medium, and supplemented with 1% penicillin, streptomycin, and 10% FBS in a humidified incubator containing 95% air and 5% CO₂ at 37°C. NIH/3T3 cells from the American Type Culture Collection were cultured in normal 10% FBS and DMEM at 5% CO₂ at 37°C. Recombinant TGF-β were obtained from R&D Systems (Minneapolis, MN). AZ505 was purchased from MedChem Express and dissolved in DMSO (Sigma) at a stock solution of 20 mM. The Smad3 inhibitor SIS3 was purchased from Cayman. All stock solutions were stored at −20°C.

Western Blot Analysis and Immunoprecipitation

Cell pellets were collected and resuspended in lysis buffer [20 mM Tris·HCl (pH 7.4), 150 mM NaCl, 10% glycerol, 1% Triton X-100, 1 mM Na₃VO₄, 25 mM β-glycerol-phosphate, 0.1 mM PMSF, Roche complete protease inhibitor set, and Sigma phosphatase inhibitor set]. The resuspended cell pellet was vortexed for 20 s, then incubated on ice for 30 min, and centrifuged at 20,000 g for 30 min. Supernatants were collected for Western blot analysis or immunoprecipitation. Renal homogenates (50 μg) or cell lysates (20 or 40 μg) were solubilized in 3× Laemmli sample buffer supplemented with 400 mM dithiothreitol (Thermo Fisher Scientific). Each sample was denatured by boiling at 100°C for 5 min, resolved with corresponding polyacrylamide SDS-PAGE according to molecular weights, and thereafter transferred to a PVDF membrane (Immobilon-P). Nonspecific protein binding was blocked by incubation in 5% fat-free milk powder in Tris-buffered saline for 60 min at room temperature. PVDF membranes were next incubated overnight at 4°C with primary antibodies against α-SMA (1:2,000, No. A2547, Sigma-Aldrich; 22), Smyd2 (1:500, sc-79084, Santa Cruz Biotechnology; 20), fibronectin (1:1,000, No. 63779, Cell Signaling Technology; 18), β-actin (1:30,000, A5316, Sigma-Aldrich; 20), Smad2 (D43B4, 1:1,000, No. 5339, Cell Signaling Technology; 23), Smad3 (1:1,000, No. 9523, Cell Signaling Technology; 23), phospho-Smad2-S465/467 (1:1,000, No. 3108, Cell Signaling Technology; 23), phospho-Smad3-S423/425 (1:1,000, No. 9520, Cell Signaling Technology; 23), collage type I-α₁ (Col1a1; 1:500, No. 91144, Cell Signaling Technology; 23), H3k4me1 (1:500, No. ab8895, Abcam; 20), H3k4me2 (1:500, No. ab7766, Abcam; 20), H3k36me1 (1:500, No. ab9048, Abcam; 20), and H3k36me2 (1:500, No. ab9049, Abcam; 20). After the primary antibody was washed away and a further blocking step was performed, membranes were incubated for 1 h at room temperature with the appropriate secondary antibody including donkey anti-rabbit IgG-horseradish peroxidase (sc-2313), and goat anti-mouse IgG-horseradish peroxidase (sc-2005) purchased from Santa Cruz Biotechnology. Donkey anti-goat IgG (H + L) horseradish peroxidase (No. A15999) was purchased from Thermo Fisher Scientific. After being washed, membranes were incubated with chemiluminescent horseradish peroxidase substrate (Millipore) for 2 min, and the signal was developed with a Kodak X-OMAT processor.

Immunohistochemistry and Immunofluorescence Staining

Paraffin-embedded sections (5 μm) were subjected to the staining. For Smyd2 staining, polyclonal goat anti-Smyd2 antibody (1:100 dilution, Santa Cruz Biotechnology), biotinylated secondary antibody (1:100 dilution, Santa Cruz Biotechnology), and the diaminobenzidine (DAB) substrate system were used. Kidney sections were counterstained by hematoxylin. Images were analyzed with a Nikon TI2E microscope.

Quantitative Analysis of Protein Expression in Immunohistochemistry, Immunofluorescence Staining, and Western Blot Analysis

ImageJ software was used to calculate the expression levels of protein in the Western blot analysis and immunohistochemistry staining, and immunofluorescence staining was calculated according to previously published protocols (24, 25).

Quantitative RT-PCR

Total RNA was extracted using the RNeasy plus mini kit (Qiagen). Total RNA (1 μg) was used for reverse transcription reactions in a 20 μL reaction to synthesize cDNA using Iscript cDNA Synthesis Kit (Bio-Rad). RNA expression profiles were analyzed by real-time PCR using iTaq SYBER Green Supermix with ROX (Bio-Rad) in a CFX Connect Real-time PCR detection system. The complete reactions were subjected to the following program of thermal cycling: 40 cycles of 10 s at 95°C and 20 s at 60°C. A melting curve was run after the PCR cycles, followed by a cooling step. Each sample was run in triplicate in each experiment, and each experiment was repeated three times. Expression levels of target genes were normalized to the expression level of actin. All primers used are provided in Supplemental Table S1.

RNA Interference

The RNA oligonucleotides that specifically target rat Smyd2 (Dharmacon) and Smad3 (Dharmacon) as well as mouse Smad3 (Santa Cruz Biotechnology) were purchased. RNA oligonucleotides were transfected with DharmaFECT siRNA transfection reagent (Dharmacon). Twenty-four and forty-eight hours after transfection, cells were harvested and analyzed by Western blot analysis.

Mouse Strains

All animal protocols were approved and conducted in accordance with Laboratory Animal Resources of the Mayo Clinic and Institutional Animal Care and Use Committee regulations. Smyd2^fl/fl mice were kindly provided by Dr. Julien Sage (Department of Pediatrics and Genetics, Stanford University Medical Center, Stanford, CT). Smyd2^fl/fl mice developed normally in size and behavior and were fertile. Hypomorphic Pkd1^nl/nl mice were generated by crossbreeding Pkd1^nl/+ female mice with Pkd1^nl/+ male mice. Pkd1^fl/fl:Pkhd1-Cre mice were generated by crossbreeding Pkd1^fl/+:Pkhd1-Cre female mice with Pkd1^fl/+:Pkhd1-Cre male mice. Pkd1^fl/fl:Pkhd1-Cre mice are referred to as homozygous (HOMO) animals, and Pkd1^fl/+:Pkhd1-Cre mice are referred to as heterozygous (HET) animals in this study. Pkd1^fl/fl:tamoxifen-Cre mice were generated by crossbreeding Pkd1^fl/fl female mice with Esr1⁺-Cre transgenic male mice [B6N.129S6(Cg)-Esr1^{tm1.1(cre)And}/J, Stock 017911, Jackson Laboratory], in which Cre is driven by the Esr1⁺ promoter. These mice were injected intraperitoneally with tamoxifen (125 mg/kg body wt, formulated in corn oil) on two sequential postnatal days (postnatal days 31 and 32) to induce Pkd1 deletion in the test group. Genotyping was confirmed by tail PCR using previously published primers. The treatment of AZ505 in these mouse strains was as previously described. Pkd1^nl/nl pups were injected intraperitoneally with AZ505 [5 mg/kg, dissolved in DMSO, with a final DMSO concentration of 10% (vol/vol) in PBS] or DMSO (control) daily from postnatal day 7 to postnatal day 27, and kidneys from six female mice and six male mice per group were harvested and analyzed at postnatal day 28. Pkd1^fl/fl: tamoxifen-Cre mice were treated with AZ505 for ∼5 mo, starting 10 days (postnatal day 42) after Pkd1 deletion with a dosage of 10 mg/kg body wt. Treatment began with a 3 times/wk schedule up 3.5 mo of age, followed by a once per week schedule from 3.5 to 6 mo of age. We collected blood samples 24 h after the last dose and harvested kidneys from 6-mo-old AZ505- and vehicle (DMSO)-treated mice for histopathological and biochemical analysis (20). Both male and female animals were used in all experiments. Statistical analysis was performed in age-matched and sex-matched manners, respectively, and at least three animals of each sex were included.

Chromatin Immunoprecipitation Assay

A chromatin immunoprecipitation (ChIP) assay was performed according to the protocol previously described in Ref. 26. Chromatin DNA was subjected to immunoprecipitation with anti-Smyd2 and anti-H3K4me2 antibody or normal rabbit IgG and then washed, after which the DNA-protein cross links were reversed. The recovered DNA was analyzed by PCR for the binding of fibronectin, α-SMA, Col1α1, and plasminogen activator inhibitor-1 (PAI-1) promoters by Smyd2 or H3K4me2 antibodies. All primers used are provided in Supplemental Table S2.

In Vitro Isolation of Stress Fibers

Stress fibers were isolated following methods previously described in the literature (27, 28). NRK-49F cells were stimulated with TGF-β in the presence or absence of the Smyd2 inhibitor AZ505 for 24 h, and cells were washed with PBS and then extracted with buffer containing 2.5 mM triethanolamine (pH 8.2) for 30 min. The buffer was changed every 5 min, followed by extraction with 0.05% Nonidet P-40 (pH 7.2) for 5 min and subsequent extraction with 0.5% Triton X-100 (pH 7.2) for an additional 5 min. Cells were then immediately washed with cold PBS, scraped off from the dish, and suspended in PBS, followed by centrifugation at 100,000 g for 1 h. The supernatant was removed, and the pellet was sonicated in 0.5% Triton X-100, 50 mM NaCl, 20 mM HEPES (pH 7.0), and 1 mM EDTA; 3× SDS sample buffer was added, and samples were boiled for 5 min before further Western blot analysis.

Statistics

All data presented were repeated at least three times independently with similar results. All statistical data are presented as means ± SE. P values were calculated by a two-tailed unpaired Student’s t test (two independent groups), one-way ANOVA (one independent variable, multiple groups), and two-way ANOVA (two independent variables, multiple groups), and a P value of <0.05 was considered significant. All graphs were generated with GraphPad Prism.

Study Approval

All animal protocols were approved and conducted in accordance with Laboratory Animal Resources of Mayo Clinic and Institutional Animal Care and Use Committee regulations (Protocol No. A00003756-18-R21).

RESULTS

Upregulation of Smyd2 Is Positively Correlated With Disease Progression in a Pkd1 Conditional Knockout Mouse Model of Renal Fibrosis

Renal fibrosis is a key characteristic and common complication of late-stage ADPKD. To understand how renal fibrosis is regulated during cyst progression, we compared the expression of fibrotic genes in different cell types in Pkd1^fl/fl:Pkhd1-Cre (Pkd1-HOMO) mouse kidneys versus Pkd1^fl/+:Pkhd1-Cre (Pkd1-HET) mouse kidneys at three different stages, including postnatal days 7, 14, and 21, with single-cell RNA-sequencing analysis. In Pkd1 conditional knockout Pkd1-HOMO kidneys, cysts were initiated at postnatal days 8 or 9, rapidly developed around postnatal day 14, and then reached late stage around postnatal day 21 (7). To support that activation of fibroblasts is the main mechanism for renal fibrosis, we found that expression of most of the fibrotic genes was increased in fibroblasts in postnatal day 21 Pkd1-HOMO kidneys compared with age-matched Pkd1-HET kidneys (Fig. 1A). We also found that expression of some of those fibrotic genes was upregulated in collecting duct cells, in which Pkd1 was specifically deleted (Fig. 1A). We further found that expression of some fibrotic genes, such as Col1α1, collagen type I-α₂ (Col1α2), collagen type III-α₁ (Col3α1), collagen type V-α₃ (Col5α3), and collagen type VI-α₃ (Col6α3), was increased in fibroblasts in postnatal day 7 Pkd1-HOMO kidneys, even before cyst initiation in collecting duct cells, suggesting that Pkd1 deletion in collecting duct cells may result in the activation of fibroblasts at the early stage with an uncertain mechanism. In addition to fibrotic genes, we found that expression of the histone methyltransferase Smyd2 was increased in Pkd1-HOMO kidneys compared with Pkd1-HET kidneys at all the stages, especially in postnatal day 21 collecting duct cells and fibroblasts (Fig. 1A). Upregulation of Smyd2 was confirmed with immunofluorescence staining (Fig. 1B). We found that the cellular localization of Smyd2 gradually changed during cyst development. Before cyst initiation (postnatal day 7), expression of Smyd2 was at a similar level as that in Pkd1-HET kidneys and mostly localized in the cytosol and partially in the nucleus. During the stage of cyst progression, upregulated Smyd2 was observed in the interstitial area in postnatal day 14 Pkd1-HOMO kidneys and was largely accumulated at cyst-lining cells marked with Dolichos biflorus agglutinin to indicate collecting ducts (29) in postnatal day 21 Pkd1-HOMO kidneys (Fig. 1B). We further found that cyst progression in Pkd1-HOMO kidneys was accompanied by increased renal fibrosis in both male (Fig. 1, C and D) and female (Supplemental Fig. S1, A and B) Pkd1-HOMO kidneys at postnatal day 7, 14, and 21 compared with age-matched Pkd1-HET kidneys as examined by picrosirius red staining and Masson’s trichrome staining. We did not observe a significant difference in renal fibrosis between male and female mice. We further found that the accumulation of collagen was started in the interstitial area in Pkd1-HOMO kidneys at postnatal day 14 and constantly increased afterward, resulting in deposition of collagen in both the interstitial area and cyst-lining area at the later stage. Upregulation of Smyd2 was further confirmed in Pkd1 mutant kidneys by immunohistochemical staining (Fig. 2, A and B). Expression of fibrotic genes, including α-SMA and fibronectin, was also increased in Pkd1-HOMO kidneys, which was correlated with the upregulation of Smyd2 protein as examined by Western blot (Fig. 2C). Upregulation of SMYD2 and the increase of renal fibrosis were also found in human ADPKD kidneys compared with normal human kidneys (Supplemental Fig. S2, A and B).

Figure 1.

Upregulation of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) is positively correlated with renal fibrosis in an autosomal dominant polycystic kidney disease mouse model. A: heatmap showing the alterations of fibrotic genes and Smyd2 expression during the progression of polycystic kidney disease pathology in collecting duct cells and fibroblasts. B: representative images of immunofluorescence staining of Smyd2 in day 7, day 14, and day 21 Pkd1 homozygous (Pkd1-HOMO) kidneys costained with Dolichos biflorus agglutinin. Scale bars = 20 μm. One of at least three independent experiments is shown. C and D, top: picrosirius red and Masson’s trichrome blue staining revealed that renal fibrosis was gradually increased in kidneys of Pkd1^fl/fl:Pkhd1-Cre mice. Scale bars = 50 μm. Only data from male animals are shown in B–D. C and D, bottom: statistical analysis of the area (%) of picrosirius red staining (C) and Masson’s trichrome staining (D) in Pkd1 heterozygous (Pkd1-HET) mice at days 7, 14, and 21. n = 3. *P < 0.01 compared with age-matched Pkd1-HET mice; #P < 0.05 compared with day 7 Pkd1-HOMO mice. P values by two-way ANOVA followed by Tukey’s post hoc test are indicated.

Figure 2.

Expression of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) is increased and accompanied by the progression of renal fibrosis in kidneys from Pkd1 mutant mice. Immunohistochemistry analysis indicating that Smyd2 expression was increased in cyst-lining epithelia in kidneys from male (A) and female (B) Pkd1 mutant mice (right) compared with normal mouse kidneys (left). The higher-magnification image insets show that Smyd2 was mainly localized to the cytoplasm and partially in the nuclear of cyst-lining epithelia cells. Scale bars = 20 μm. C, top: Western blot analysis of Smyd2, fibronectin, and α-smooth muscle actin (α-SMA) levels in kidneys of male Pkd1^fl/fl:Pkhd1-Cre mice. Representative data from one of two independent experiments are shown. C, bottom: quantitative and statistical analysis (n = 3). *P < 0.01 compared with day 7 Pkd1 heterozygous (Pkd1-HET) mice; #P < 0.05 compared with day 21 Pkd1 homozygous (Pkd1-HOMO21) mice as calculated by a one-way ANOVA test.

Expression of Smyd2 Is Positively Regulated by TGF-β in Renal Cells

TGF-β plays an important role in the development of renal fibrosis in ADPKD (15, 30). To determine the role of TGF-β in the regulation of Smyd2 expression, we treated a rat kidney interstitial fibroblast cell line (NRK-49F) and a mouse fibroblast cell line (NIH3T3) with TGF-β. We found that treatment with TGF-β increased expression of Smyd2 in a dose- and time-dependent manner (Fig. 3, A–D). To further determine whether TGF-β regulates expression of Smyd2 through canonical TGF-β-Smad3 signaling (11, 31), we knocked down Smad3 with siRNA in NRK-49F cells and NIH3T3 cells treated with or without TGF-β. Expression of Smad3 was reduced by 80–90% in both cells (Fig. 4, A–C). We found that knockdown of Smad3 significantly decreased expression of Smyd2 in NRK-49F cells treated with TGF-β (Fig. 4, A and B). In addition, knockdown of Smad3 decreased expression of Smyd2 in NRK-49F cells even without induction of TGF-β (Fig. 4C), suggesting that Smad3 might contribute to maintain the level of endogenous Smyd2. We further found that treatment with the Smad3 inhibitor SIS3 also decreased expression of Smyd2 in NRK-49F and NIH3T3 cells, which were induced with or without TGF-β (Fig. 4, D and E). These results suggested that the upregulation of Smyd2 is dependent on TGF-β-Smad3 signaling.

Figure 3.

Transforming growth factor (TGF)-β stimulated the expression of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) in NRK-49F and NIH3T3 cells. A and B, top: expression of Smyd2 was regulated by TGF-β in a time-dependent (A) and dose-dependent manner (B) in NRK-49F cells. C and D, top: expression of Smyd2 was regulated by TGF-β in a time-dependent (C) and dose-dependent manner (D) in NIH3T3 cells. Representative data from three independent experiments are shown. A–D, bottom: quantitative and statistical analysis (n = 3). *P < 0.05 compared with no treatment as calculated by a one-way ANOVA test. NRK-49F, normal rat kidney 49F.

Figure 4.

Transforming growth factor (TGF)-β induced the expression of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) in a Smad3-dependent manner. A and B, left: Western blot analysis of Smyd2 and Smad3 expression in NRK-49F (A) and NIH3T3 cells (B) treated with or without siRNA of Smad3 (siSmad3) in the presence or absence of TGF-β. Representative data from three independent experiments are shown. A and B, right: statistical analysis of the expression (n = 3) of Smad3 and Smyd2 by normalization with the loading control (actin). *P < 0.05 compared with no treatment; #P < 0.05 compared with the TGF-β-treated group as indicated by a one-way ANOVA test. C, left: Western blot analysis of Smyd2 and Smad3 expression in NRK-49F cells with or without siRNA of Smad3. C, right: quantitative and statistical analysis (n = 3). *P < 0.01 compared with each control as calculated by a one-way ANOVA test. D and E, left: Western blot analysis of Smad3, phosphorylation of Smad3, and Smyd2 expression in NRK-49F (D) and NIH3T3 cells (E) with or without the Smad3 inhibitor SIS3 in the presence or absence of TGF-β. Representative data from three independent experiments are shown. D and E, right: quantitative and statistical analysis (n = 3). *P < 0.01 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells as calculated by a one-way ANOVA test. NRK-49F, normal rat kidney 49F, p-Smad3, phosphorylated Smad3.

Deletion of Smyd2 Attenuates Renal Fibrosis in Early-Onset Pkd1 Conditional Knockout Mice

Given the correlation of upregulation of Smyd2 with renal fibrosis and that expression of Smyd2 was regulated by TGF-β-Smad3 signaling in fibroblasts, we investigated whether knockout and inhibition of Smyd2 attenuates renal fibrosis in Pkd1 mutant mouse models. We first found that knockout of Smyd2 decreased mRNA and protein expression of fibrotic genes, including α-SMA, Col1α1, and fibronectin, as well as phosphorylation of Smad2 and Smad3 in Pkd1 conditional knockout kidneys compared with age- and sex-matched Pkd1 single knockout kidneys (Fig. 5, A and B). There was no significant difference in the expression of those genes between male and female mice. Expression of α-SMA was strikingly decreased in Pkd1 and Smyd2 double conditional knockout kidneys compared with Pkd1 single knockout kidneys as examined with immunohistochemistry staining (Fig. 5C and Supplemental Fig. S3A). We also observed a reduction of collagen deposition in Pkd1 and Smyd2 double knockout kidneys compared with Pkd1 single knockout kidneys with no sex difference as examined with picrosirius red staining (Fig. 5D and Supplemental Fig. S3B). These results suggested that deletion of Smyd2 should attenuate renal fibrosis in Pkd1 mutant mouse kidneys.

Figure 5.

Knockout of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) suppressed the progression of renal fibrosis in Pkd1 knockout mice. A: quantitative RT-PCR analysis of the expression of fibronectin, collagen type I (Col1), α-smooth muscle actin (α-SMA), transforming growth factor (TGF)-β, plasminogen activator inhibitor-1 (PAI-1), and Smyd2 mRNAs in kidneys from Pkd1 single knockout mice compared with Pkd1 and Smyd2 double knockout mice. All data were analyzed from at least three experiments. All statistical data are presented as means ± SE. *P < 0.01 compared with sex-matched Pkd1 single knockout mice as calculated by two-way ANOVA followed by a Tukey’s post hoc test. No significant difference was found between male or female animals within the same genotype. B: Western blot analysis of the expression of Smyd2 and phosphorylated (p-)Smad2, p-Smad3, Smad2, Smad3, collagen type I-α₁ (Col1α1), fibronectin, and α-SMA in day 7 Pkd1 single knockout, Pkd1 and Smyd2 double knockout, and control mouse kidneys. Littermates with the same sex were used, and one of three independent experiments from male pups is shown. C, top: representative images of immunohistochemistry staining of α-SMA in male Pkd1 single knockout, Pkd1 and Smyd2 double knockout, and wild-type control mouse kidneys at day 7. Scale bars = 100 μm. C, bottom: quantitative and statistical analysis (n = 3) of the area (%) of α-SMA staining in male (left) and female (right) Pkd1 heterozygous, Pkd1 single knockout, and Pkd1/Smyd2 double knockout kidneys at day 7. *P < 0.05 compared with sex-matched Pkd1 heterozygous mice; #P < 0.05 compared with sex-matched Pkd1 single knockout mice as calculated by a one-way ANOVA test. D, top: picrosirius red staining revealed that renal fibrosis was decreased in kidneys from Pkd1 and Smyd2 double knockout mice compared with Pkd1 single knockout male mice. Scale bars = 50 μm. D, bottom: quantitative and statistical analysis (n = 3) of the area (%) of picrosirius red staining in male (left) and female (right) Pkd1 heterozygous, Pkd1 single knockout, and Pkd1-Smyd2 double knockout kidneys at day 14. *P < 0.05 compared with sex-matched Pkd1 heterozygous mice; #P < 0.05 compared with sex-matched Pkd1 single knockout mice as calculated by a one-way ANOVA test.

Inhibition of Smyd2 Delays Renal Fibrosis in Pkd1 Hypomorphic and Milder Pkd1 Conditional Knockout Mouse Models

To determine whether targeting Smyd2 is a potential treatment for renal fibrosis in ADPKD kidneys, we treated two Pkd1 mouse models, Pkd1^nl/nl mice and Pkd1^fl/fl:Tam-Cre mice, with a specific SMYD2 inhibitor, AZ505 (32). Pkd1^nl/nl mice harbor a hypomorphic Pkd1 allele, yielding only 13–20% normally spliced Pkd1 transcripts and resulting in bilaterally enlarged polycystic kidneys (33). The Pkd1^nl/nl mouse is a moderately progressive model compared with highly aggressive Pkd1^fl/fl:Ksp-Cre mice and long-lasting milder Pkd1^fl/fl:Tam-Cre mice in which Pkd1 deletion is induced by an injection of tamoxifen at postnatal day 31 and postnatal day 32 (20). We found that treatment with AZ505 decreased the expression of Col1α1, fibronectin, α-SMA, and Smyd2 as well as phosphorylation of Smad2 and Smad3 in kidneys from Pkd1^nl/nl and Pkd1^fl/fl:Tam-Cre mice compared with kidneys from age-matched control mice treated with DMSO (Fig. 6, A–C). Importantly, treatment with AZ505 significantly decreased renal fibrosis characterized by a reduction of the areas of α-SMA and collagen in Pkd1^nl/nl male (Fig. 6D) and female (Supplemental Fig. S4A) kidneys as well as in long-lasting milder Pkd1^fl/fl:Tam-Cre kidneys (Fig. 6E and Supplemental Fig. S4B). These results suggest that inhibition of Smyd2 attenuates the progression of renal fibrosis in PKD mouse models.

Figure 6.

Treatment with the SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) inhibitor AZ505 delayed renal fibrosis in Pkd1 knockout mice. Western blot analysis of the phosphorylation of Smad2 and Smad3, collagen type I-α₁ (Col1α1), fibronectin, and α-smooth muscle actin (α-SMA) in male Pkd1^nl/nl mice (A) and male Pkd1^fl/fl:Tam-Cre mice (B). Inhibition of Smyd2 with AZ505 decreased the phosphorylation of Smad2 and Smad3 but did not affect their expression, inhibition of Smyd2 with AZ505 decreased the expression of Col1α1, fibronectin, and α-SMA. C: quantitative and statistical analysis (n = 3) of band intensities as shown in the graphs corresponding to A (top) and B (bottom). The density of each protein band was normalized to actin, and the value was then divided by the value of the corresponding vehicle band density to actin. The value of the vehicle band to actin was set as 1. *P < 0.01 compared with vehicle (DMSO)-treated mice as indicated by a one-way ANOVA test. D and E, top: AZ505 treatment reduced the areas of fibrosis in kidneys from male Pkd1^nl/nl (D) and male Pkd1^fl/fl:Tam-Cre mice (E), as detected with α-SMA staining (left) and picrosirius red staining (right) compared with those treated with vehicle (DMSO). *P< 0.05. Scale bars = 100 μm. n = 3 for each group. D and E, bottom: quantitative and statistical analysis (n = 3) of the area (%) of α-SMA staining and picrosirius red staining. *P < 0.01 compared with vehicle (DMSO)-treated mice as calculated by a one-way ANOVA test. p-Smad3, phosphorylated Smad3; p-Smad2, phosphorylated Smad2.

Knockdown or Inhibition of Smyd2 Decreases TGF-β-Induced Activation of Renal Fibroblasts

To investigate the mechanisms mediated by Smyd2 in the regulation of renal fibrosis, we targeted Smyd2 with siRNA or its inhibitor AZ505 in fibroblasts in vitro. First, we found that knockdown of Smyd2 abrogated the TGF-β-induced expression of fibrotic genes including α-SMA, fibronectin, Col1α1, TGF-β, and PAI-1 as well as expression of Smyd2 in NRK-49F cells (Fig. 7, A and B). Next, we found that treatment with AZ505 also abrogated the TGF-β-induced expression of those fibrotic genes in NRK-49F cells (Fig. 7, C and D). These results suggest a role for Smyd2 in the regulation of fibrogenesis through targeting the expression of fibrotic genes. As a histone methyltransferase, Smyd2 may regulate gene transcription through binding to gene promoters and/or through methylating histone H3 at lysine 4 (K4) and lysine 36 (K36; 20, 21). We found that treatment with TGF-β increased the methylation of histone H3 at K4 and K36 (Fig. 8A), suggesting that Smyd2 may regulate the expression of fibrotic genes through chromatin modification. We further found that Smyd2 bound with the promoters of Col1α1, fibronectin, and PAI-1 but not with the promoter of α-SMA as examined with a ChIP assay (Fig. 8B). In addition, we found that treatment with TGF-β enhanced the binding of Smyd2 with the fibronectin promoter, but not the binding of Smyd2 with the promoters of Col1α1 and PAI-1 (Fig. 8B). These results suggest that Smyd2 may regulate the transcription of fibronectin in a TGF-β-dependent manner and the transcription of Col1α1 and PAI-1 in a TGF-β-independent manner. Furthermore, our ChIP assay with H3K4me2 antibody indicated that H3K4me2 bound with the promoters of α-SMA, Col1α1, and PAI-1 and that inhibition of Smyd2 decreased those bindings in NRK-49F cells (Fig. 8C). Taken together, these results suggested that Smyd2 could regulate the transcription of fibrotic genes through direct binding with the promoter of some fibrotic genes or indirectly through Smyd2-mediated histone methylation, resulting in the activation of fibroblasts and then renal fibrosis.

Figure 7.

Inhibition or knockdown of SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) suppressed the expression of fibrotic genes. A, left: Western blot analysis of the expression of Smyd2, fibronectin, and α-smooth muscle actin (α-SMA) in NRK-49F cells. Knockdown of Smyd2 with siRNA (siSmyd2) blocked the transforming growth factor (TGF)-β-induced upregulation of Smyd2, fibronectin, and α-SMA. One of three independent experiments is shown. A, right: statistical analysis of the expression (n = 3) of Smyd2, fibronectin, and α-SMA proteins. *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells. B: quantitative RT-PCR analysis of fibrotic gene expression in the presence or absence of TGF-β and Smyd2 siRNA treatment in NRK-49F cells. Statistical analysis of the mRNA levels of collagen type I-α₁ (Col1α1), collagen type III-α₁ (Col3α1), fibronectin, and plasminogen activator inhibitor-1 (PAI-1) is shown in the graphs. *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells. n = 3 for each group. C, left: Western blot analysis of the expression of Smyd2, fibronectin, and α-SMA in NRK-49F cells. Treatment with the Smyd2 inhibitor AZ505 blocked the TGF-β-induced upregulation of Smyd2, fibronectin, and α-SMA. C, right: statistical analysis of the expression (n = 3) of Smyd2, fibronectin, and α-SMA proteins. *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells. D: quantitative RT-PCR analysis of fibrotic genes and Smyd2 expression in the presence or absence of TGF-β and AZ505 treatment in NRK-49F cells. Statistical analysis of the mRNA levels of Col1α1, Col3α1, fibronectin, and PAI-1 is shown in the graphs. *P < 0.05 compared with no treatment; #P < 0.05 compared with TGF-β-treated cells. n = 3 for each group. P values were calculated by a one-way ANOVA test. NRK-49F, normal rat kidney 49F.

Figure 8.

SET and MYND domain-containing lysine methyltransferase 2 (Smyd2) transcriptionally regulates the expression of fibrotic genes. A, left: treatment with transforming growth factor (TGF)-β induced the expression of Smyd2 and the mono- and dimethylation of histone H3 at lysine K4 and K36 in NRK-49F cells. A, right: quantitative and statistical analysis (n = 3) of the band intensities. The density of each protein band was normalized to actin, and the value was then divided by the value of the corresponding vehicle band density to actin. The value of the vehicle band to actin was set as 1. *P < 0.05 compared with cells with no treatment. B and C, left: Smyd2 and H3K4me2 bound to the promoters of fibrotic genes. Chromatin immunoprecipitation assays were performed with anti-Smyd2, anti-H3K4me2 antibody, or normal rabbit IgG in NRK-49F cells treated with or without TGF-β and AZ505. B and C, left: statistical analysis of the enrichment of Smyd2 (B) and H3K4me2 (C) in treated compared with cells with no treatment or TGF-β-treated NRK-49F cells. *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β treated cells. P values were calculated by one-way ANOVA test. α-SMA, α-smooth muscle actin; H3K4, histone H3 at lysine 4; H3K36, histone H3 at lysine 36; NRK-49F, normal rat kidney 49F; PAI-1, plasminogen activator inhibitor-1.

Inhibition of Smyd2 Suppresses TGF-β-Induced Stress Fiber Formation

TGF-β has been reported to be able to stimulate myofibroblast differentiation, in that when fibroblasts were treated with TGF-β, their ultrastructures would be changed with an increase of cytoskeletal stress fiber formation, mainly referred to the formation of actin stress fibers with a rapid incorporation of α-SMA (34). Meanwhile, this process is accompanied by an increase of the expression of cytoskeletal proteins and components of the extracellular matrix, including α-SMA, Col1α1, and fibronectin, as termed as TGF-β-meditated pathogenesis of tissue fibrosis (32, 35). To determine the involvement of Smyd2 in this process, we investigated the effects of Smyd2 on TGF-β-induced stress fiber formation in NRK-49F cells by immunofluorescence staining with α-SMA and phalloidin, a highly selective bicyclic peptide that is used for staining actin filaments (F-actin). We found that treatment with TGF-β promoted the formation of stress fibers as seen by an incorporation of α-SMA that was upregulated by TGF-β treatment, whereas cotreatment with TGF-β and AZ505 decreased the formation of stress fibers through reducing the expression and assembly of α-SMA in NRK-49F cells (Fig. 9A), supporting the involvement of Smyd2 in the regulation of TGF-β-induced formation of stress fibers. In addition, we found that treatment with TGF-β increased the expression and formation of fibronectin-containing stress fibers in NRK-49F cells, which was abrogated in those cells treated with TGF-β and AZ505 (Fig. 9B). By isolation of stress fibers from NRK-49F cells treated with TGF-β alone or cotreatment with TGF-β and AZ505, we further confirmed that treatment with TGF-β increased stress fiber actin and that treatment with AZ505 abrogated TGF-β-induced stress fiber actin in NRK-49F cells, in which stress fiber actin was normalized to the actin in total cell lysates (Fig. 9C).

Figure 9.

AZ505 treatment blocked transforming growth factor (TGF)-β induced stress fiber formation in NRK-49F cells. A, left: representative images of immunofluorescence staining of stress fibers in NRK-49F cells treated with TGF-β alone or TGF-β and AZ505, as detected by α-smooth muscle actin (α-SMA; red) costained with F-actin (green) antibodies and DAPI (blue). Scale bar = 10 μm. B, left: representative images of immunofluorescence staining of stress fibers in NRK-49F cells treated with TGF-β alone or TGF-β and AZ505, as detected by fibronectin (red) costained with DAPI (blue). Scale bar = 10 μm. A and B, right: statistical analysis of fluorescent density of α-SMA and F-actin (A) and fibronectin (B) in NRK-49F cells treated with TGF-β alone or cotreatment with AZ505. *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells calculated by a one-way ANOVA test. C, left: Western blot analysis of actin stress fibers in NRK-49F cells. C, right: statistical analysis of actin stress fibers normalized to total actin (n = 3). *P < 0.05 compared with cells with no treatment; #P < 0.05 compared with TGF-β-treated cells as calculated by a one-way ANOVA test. NRK-49F, normal rat kidney 49F.

DISCUSSION

In ADPKD, cyst enlargement to a certain extent may be restricted by fibrosis around the cysts, and fibrosis becomes the dominant pathogenic factor at this stage. However, the role of the epigenetic mechanism and its cross talk with TGF-β signaling in renal fibrosis in ADPKD remains unclear. In this study, we determined the role and mechanisms of Smyd2 in the regulation of renal fibrosis in ADPKD. We showed that expression of Smyd2 is highly associated with renal fibrosis during the disease progression in Pkd1 mutant mouse models and patients with ADPKD. Expression of Smyd2 is regulated through the canonical TGF-β-Smad3 signaling pathway in fibroblast NRK-49F cells. Smyd2 and its histone substrates bind on the promoters of fibrotic genes to regulate their transcription and then the activation of fibroblasts in response to TGF-β. Inhibition of Smyd2 with its inhibitor AZ505 and knockdown of Smyd2 suppress the TGF-β-induced expression of fibrotic genes and activation of renal fibroblasts as well as the formation of stress fibers. We further showed that treatment with Smyd2 inhibitor ameliorates renal fibrosis in Pkd1 mutant mice, suggesting that targeting Smyd2 may be a promising therapeutic strategy for renal fibrosis in patients with ADPKD and chronic kidney diseases.

The progression of renal fibrosis relies on the balance of matrix synthesis and matrix degradation, in that increased matrix synthesis and/or inadequate matrix degradation will lead to the fibrogenesis (36). Matrix synthesis includes the fibrotic deposition of α-SMA, collagens, and fibronectin (37), whereas matrix degradation is mainly regulated by matrix metalloproteinases (MMPs), tissue inhibitors of MMPs, and PAI-1, a negative regulator of MMPs (38). We showed that TGF-β-induced phosphorylation of Smad3 is responsible for the upregulation of Smyd2. TGF-β-Smad3 signaling is the canonical pathway, in which TGF-β-induced phosphorylation of Smad3 serves as a component of transcriptional factors to regulate the expression of fibrotic genes and promote renal fibrosis (39). We show, for the first time, that TGF-β-Smad3 signaling is responsible for the upregulation of Smyd2 in Pkd1 mutant renal cells and tissues, and TGF-β also induces the methylation of H3K4me2 at the promoters of fibrotic genes, including α-SMA, Col1α1, fibronectin, and PAI-1. These results support the existence of a synergistic feedback loop between TGF-β and Smyd2 and that Smyd2-TGF-β-Smad3-Smyd2 signaling plays an important role in promoting renal fibrosis in ADPKD.

The effects of epigenetic modifications on renal fibrosis have been reported in past years, such as histone demethylase LSD1 and histone methyltransferase Set7/9 (16, 18). Our previous study reported that deletion of Smyd2 delays cyst growth and extends the lifespan of Pkd1 knockout mice (20). In this study, we show that Smyd2 regulates renal fibrosis in Pkd1 mutant mouse kidneys. It has been reported that inhibition of Smyd2 with AZ505 ameliorated renal fibrosis in a UUO model where the renal fibrosis was induced by kidney injury (17). In UUO mouse kidneys, the role of Smyd2 in renal fibrosis is associated with regulation of the transition of renal epithelial cells through Snail and Twist and activation of intracellular signaling pathways, including STAT3, ERK, NF-κB, and AKT (17), in which those pathways have been associated with Smyd2-mediated cyst renal epithelial cell proliferation in our previous publication (20). In this study, we identified novel Smyd2-TGF-β-Smad3-Smyd2 signaling in the regulation of renal fibrosis in Pkd1 mutant mouse kidneys. Furthermore, we show that Smyd2 is involved in the formation of stress fibers through regulating the expression and assembly of α-SMA and fibronectin. Stress fibers are higher-order cytoskeletal structures composed of cross-linked actin filament bundles and are physiologically important in processes that require cellular contraction, such as tissue repair and gland secretion, etc. (40). Pathologically excessive stress fiber formation caused by elevated TGF-β might promote scar formation and renal fibrosis (32). We show an unknown role of Smyd2 in the regulation of the formation of stress fibers in ADPKD. Thus, we and others have provided a full spectrum of the functional roles and mechanisms of Smyd2 in kidney injury and Pkd1 mutation-induced renal fibrosis as well as cyst formation in ADPKD.

We should point out that the multiple Smyd2-mediated signaling pathways may not work independently to promote the progression of ADPKD, which makes it impossible to clearly attribute the effects of Smyd2 inhibition on a specific signaling pathway on delaying cyst growth and fibrosis. It is likely that increased cystic cell proliferation promotes cyst enlargement and progression, resulting in secondary events, such as the secretion of factors, including TGF-β, into the microenvironment adjacent to the cysts to promote the progression of renal fibrosis. In advanced stages of PKD, fibrosis becomes the dominant pathogenic factor, and fibrosis around the cysts may constrain cyst enlargement despite the proliferation of cyst-lining epithelial cells. Therefore, targeting a factor that can regulate both cell proliferation and fibrosis may be a better therapeutic strategy for ADPKD treatment.

Perspectives and Significance

In this study, we identified a novel mechanism and cross talk between Smyd2 and TGF-β signaling. We show that the expression of Smyd2 can be upregulated by TGF-β-Smad3 in NRK-49F cells, in which TGF-β can be secreted in cyst fluid in ADPKD kidneys. The elevated Smyd2 is able to positively regulate the transcription of fibrotic genes through direct binding to the promoters of those genes or through methylation of H3K4me1/2 and H3K36 me1/2. Importantly, targeting Smyd2 suppresses TGF-β-induced expression of fibrotic genes and ameliorates renal fibrosis in Pkd1 mutant mice. Taken together, this and our previous study have demonstrated a critical role of Smyd2 in the regulation of cyst and renal fibrosis progression in ADPKD kidneys, and targeting Smyd2 with its specific inhibitor, AZ505, should be a potential therapeutic strategy for ADPKD treatment.

SUPPLEMENTAL DATA

Supplemental Tables S1 and S2 and Supplemental Figs. S1–S4: https://doi.org/10.6084/m9.figshare.19915015.

GRANTS

X.L. was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants R01DK129241 and R01DK126662 and by a PKD Foundation research grant.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.X.L. conceived and designed research; L.X.L., L.Z., E.A., and X.Z. performed experiments; L.X.L., J.X.Z., and X.L. analyzed data; L.X.L. interpreted results of experiments; L.X.L. prepared figures; L.X.L. drafted manuscript; L.X.L. and X.L. edited and revised manuscript; X.L. approved final version of manuscript.