Atrial Natriuretic Peptide Inhibits Transforming GrowthFactor �–Induced Smad Signaling and Myofibroblast

Transformation in Mouse Cardiac FibroblastsPeng Li,* Dajun Wang,* Jason Lucas, Suzanne Oparil, Dongqi Xing, Xu Cao, Lea Novak,

Matthew B. Renfrow, Yiu-Fai Chen

Abstract—This study tested the hypothesis that activation of atrial natriuretic peptide (ANP)/cGMP/protein kinase Gsignaling inhibits transforming growth factor (TGF)-�1–induced extracellular matrix expression in cardiac fibroblastsand defined the specific site(s) at which this molecular merging of signaling pathways occurs. Left ventricularhypertrophy and fibrosis, collagen deposition, and myofibroblast transformation of cardiac fibroblasts in response topressure overload by transverse aortic constriction were exaggerated in ANP-null mice compared with wild-typecontrols. ANP and cGMP inhibited TGF-�1–induced myofibroblast transformation, proliferation, collagen synthesis,and plasminogen activator inhibitor-1 expression in cardiac fibroblasts isolated from wild-type mice. Followingpretreatment with cGMP, TGF-�1 induced phosphorylation of Smad3, but the resultant pSmad3 could not betranslocated to the nucleus. pSmad3 that had been phosphorylated with recombinant protein kinase G-1� was analyzedby use of Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) and ion trap tandem massspectrometry. The analysis revealed phosphorylation of Ser309 and Thr388 residues, sites distinct from the C-terminalSer423/425 residues that are phosphorylated by TGF-� receptor kinase and are critical for the nuclear translocation anddown-stream signaling of pSmad3. These results suggest that phosphorylation of Smad3 by protein kinase G is apotential molecular mechanism by which activation of ANP/cGMP/protein kinase G signaling disrupts TGF-�1–induced nuclear translocation of pSmad3 and downstream events, including myofibroblast transformation, proliferation,and expression of extracellular matrix molecules in cardiac fibroblasts. We postulate that this process contributes to theantifibrogenic effects of the natriuretic peptide in heart. (Circ Res. 2008;102:185-192.)

Key Words: atrial natriuretic factor � transforming growth factor � cardiac fibroblast � cardiac fibrosis andremodeling � signal transduction

Our previous studies have shown that endogenous atrialnatriuretic peptide (ANP) and transforming growth fac-

tor (TGF)-� play important counterregulatory roles in pres-sure overload–induced cardiac hypertrophy, remodeling, andfibrosis.1–4 ANP and TGF-� expression are upregulated inheart with pressure-overload stress.1 The functional signifi-cance of this stress-induced increase in ANP is supported byobservations of exaggerated left ventricular hypertrophy,fibrosis, and remodeling in ANP-deficient (Nppa�/�) micecompared with wild-type controls in response to pressureoverload.1,2,4 Our recent studies in a novel mouse model thatexpresses an inducible dominant negative mutation of theTGF-� receptor type II gene (DnTGF�RII), and thus cannotactivate the TGF-�/Smad signaling cascade, demonstratedthat disruption of TGF-� signaling greatly attenuated thepressure overload–induced MF transformation and interstitial

fibrosis in heart, supporting a critical role for TGF-� signal-ing in the pathogenesis of pressure overload–induced cardiachypertrophy and remodeling.3

ANP increases intracellular cGMP levels and activatescGMP-dependent protein kinase (PK)G, with resultantgrowth inhibiting and antigrowth/-proliferative effects in avariety of cell types, including cardiac fibroblasts (CFs),5

whereas activated TGF-� stimulates cellular differentiation,transformation, proliferation, migration, and extracellular ma-trix (ECM) expression.6–8 TGF-� signals through membrane-bound heteromeric type I (TGF�RI) and type II (TGF�RII)receptor kinases that transduce intracellular signals via phos-phorylation and nuclear translocation of receptor-activatedSmad2 and Smad3 proteins, which modulate the transcriptionof many genes.9 Phosphorylation of the C-terminal Ser423/425 of Smad3 by the TGF-� receptor kinase is critical for its

Original received June 13, 2007; revision received October 22, 2007; accepted October 31, 2007.From the Vascular Biology and Hypertension Program (P.L., D.W., J.L., S.O., D.X., Y.-F.C.), Department of Medicine; Department of Pathology

(X.C., L.N.); and University of Alabama Biomedical FT-ICR MS Laboratory (M.B.R.), Department of Biochemistry and Molecular Genetics, Universityof Alabama at Birmingham.

*The first two authors contributed equally to this work.Correspondence to Yiu-Fai Chen, PhD, Department of Medicine, University of Alabama at Birmingham, Birmingham, AL 35294. E-mail

yfchen@uab.edu© 2008 American Heart Association, Inc.

Circulation Research is available at http://circres.ahajournals.org DOI: 10.1161/CIRCRESAHA.107.157677

185 by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

nuclear translocation and downstream signaling. The molec-ular mechanisms of the counterregulatory effects of ANP/cGMP/PKG signaling on activated TGF-�–induced Smadsignaling in cardiac cells have not been studied.

The present study tested the hypotheses that TGF-�1accelerates myofibroblast (MF) transformation and ECMproduction in CFs and that activation of the ANP signalingcascade inhibits these processes by interrupting TGF-�1signaling. Specifically, we tested whether ANP/cGMP/PKGsignaling interrupts specific downstream events in theTGF-�1 signaling pathway, including phosphorylation andnuclear translocation of Smad2 and Smad3. We demonstratedthat, following pretreatment with cGMP, TGF-�1 inducedphosphorylation of Smad2 and Smad3, but the resultantphosphorylated (p)Smads could not be translocated to thenucleus. We then tested the hypothesis that PKG has thepotential to phosphorylate Smad3 at sites different from thoserequired for its TGF-�1–induced nuclear translocation, thusdisrupting its entry into the nucleus and downstream signal-ing. Two novel sites of Smad3 phosphorylation were identi-fied (Ser309 and Thr388) by use of high-resolution massspectrometry (MS). These differ from the C-terminal Ser423/425 residues that are substrates for phosphorylation by theTGF-� receptor kinase, suggesting a novel mechanism bywhich the ANP/cGMP/PKG signaling pathway can inhibitthe profibrotic effects of TGF-�.

Materials and Methods

Animal PreparationMale ANP null (Nppa�/�) mice1 and wild-type controls (Nppa�/�) ofthe C57BL/6 strain were studied. Mice were fed a standard diet(Harlan-Teklad) and were housed in rooms maintained at constanthumidity (60�5%), temperature (24�1°C), and light cycle (6:00 AM

to 6:00 PM). All protocols were approved by the Institutional AnimalCare and Use Committee at the University of Alabama at Birming-ham and were consistent with the Guide for the Care and Use ofLaboratory Animals published by the NIH (Department of Health,Education, and Welfare Publication No. 96-01, revised in 2002).

Surgical ProceduresNppa�/�, 9 to 10 weeks of age, and age-matched wild-type Nppa�/�

mice underwent transverse aortic constriction (TAC) or sham sur-gery under ketamine/xylazine (8 mg/1.2 mg per 100 g IP) anesthe-sia.1,2,4 Using this methodology, we have previously demonstratedreproducible pressure gradients across the TAC of 50 to 65 mm Hg.1

Effects of TAC on Collagen Deposition and MFTransformation in Hearts of Nppa�/� andNppa�/� MiceOne week after TAC, mice were killed with an overdose ofpentobarbital and by cervical dislocation. Heart were dissected andweighed. The left ventricle (LV) was fixed with 4% paraformalde-hyde, paraffin embedded, and sectioned for morphological andimmunohistochemical examination of collagen deposition and MFtransformation. Collagen was assessed using Picrosirius red staining,and MFs were identified by �-smooth muscle actin (�-SMA) (usingclone1A4 anti–�-SMA antibody, Dako) immunochemical stain-ing.2,4,10 Adjacent cross-sections (5 �m) from LVs were examined toassess colocalization of collagen and �-SMA in or near the same celltypes. In a subgroup of wild-type mice, TGF-�1 (total and active)protein levels in LVs were measured using an ELISA kit (R&DSystems).

CF PreparationCFs were isolated from hearts of adult male C57BL/6 mice weighing18 to 25 g. Hearts were excised, rinsed in cold Hank’s balanced saltsolution, minced, and digested with collagenase type 4 (100 U/mL)and trypsin (0.6 mg/mL) at 37°C for 30 minutes. The first digestionwas discarded. The collagenase medium from the second digestioncontaining the CFs was centrifuged for 10 minutes at 180g andresuspended in DMEM with 15% FBS. The digestion was repeateduntil the digestion fluid became clear (5 to 6 times). Cells wereplated in laminin-coated 60-mm dishes (Becton Dickinson) andallowed to attach for 45 minutes before the first media change, whichremoved weakly adherent cells, including myocytes and endothelialcells. Passage 1 CFs were used for the experiments.

Effects of cGMP on TGF-�1–Induced MFTransformation in Cultured Mouse CFsCFs isolated from wild-type mice were grown in 15% FBS-DMEMto 85% confluence on 18-mm2 glass cover slides and then madequiescent by culturing them in 0.1% FBS medium for 48 hours.Quiescent CFs were pretreated with cGMP (8-bromo-cGMP) (acGMP analog, 10�3 mol/L) (Sigma) or vehicle for 30 minutes andthen exposed to TGF-�1 (1 ng/mL) (Sigma) or vehicle for anadditional 24 hours. Cells were then fixed with 4% paraformalde-hyde and immunostained for �-SMA as a marker of MF transfor-mation and counterstained with hematoxylin for visualization ofnuclei. Quantitative analysis for �-SMA–positive cell/total cell ratioswas performed by light microscopy with a Qimaging QiCam colorcamera (Qimaging) interfaced with a computer system runningMetaMorph software (Universal Imaging Corp).

Effects of ANP on cGMP Levels and Effectsof ANP/cGMP on Collagen Synthesis, CellProliferation, and Plasminogen ActivatorInhibitor-1 Expression in Mouse CFsQuiescent mouse CFs were treated with ANP (1 �mol/L) (Sigma) for30 minutes and then harvested for cGMP measurement using astandard radioimmunoassay kit (Amersham). CFs isolated from adultmale Sprague–Dawley rats were isolated and treated similarly andused as controls.

Separate groups of quiescent CFs were pretreated with ANP(1 �mol/L) or cGMP (1 mmol/L) for 30 minutes before exposure toTGF-�1 (5 ng/mL) for an additional 24 hours, and de novo collagensynthesis in CFs was evaluated by measuring [3H]-proline incorpo-ration into cells using the method of Maki et al.11

To test the hypothesis that cGMP inhibits TGF-�1–stimulated cellproliferation, quiescent CFs were pretreated with cGMP (1 mmol/L)for 30 minutes before exposure to TGF-�1 (5 ng/mL). Cell prolif-eration was measured using a CellTiter 96 Aqueous Non-RadioactiveCell Proliferation Assay (Promega) at 30, 90, and 240 minutes afterTGF-�1 treatment.

The effects of cGMP on TGF-�1–stimulated plasminogen activa-tor inhibitor (PAI)-1 mRNA expression (a biomarker of TGF-�action in cells) was tested in quiescent CFs treated with TGF-�1 (5ng/mL) for 24 hours with or without pretreatment with cGMP(1 mmol/L for 30 minutes) and/or the PKG inhibitor KT5823(1 �mol/L, 15 minutes before cGMP) (Calbiochem). Northern blotanalysis for PAI-1 and GAPDH (internal control) mRNA levels wasperformed.1,12

Effects of ANP/cGMP on TGF-�1–InducedNuclear Translocation of pSmad3 in Mouse CFsTo test the hypothesis that inhibition of TGF-�1–stimulated MFtransformation and ECM expression by ANP/cGMP/PKG signalingis dependent on events downstream from phosphorylation of Smad2and Smad3, we examined the effects of ANP and cGMP onTGF-�1–stimulated nuclear translocation of pSmad2 and pSmad3 inCFs. Quiescent mouse CFs were pretreated with ANP (1 �mol/L),cGMP (1 mmol/L), or vehicle for 30 minutes and then exposed toTGF-�1 (1 ng/mL) for an additional 30 minutes. Subgroups of CFs

186 Circulation Research February 1, 2008

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

were pretreated with the PKG inhibitor KT5823 (1 �mol/L) for 15minutes before ANP or cGMP. CFs were fixed in 4% paraformal-dehyde and permeabilized in 0.1% Triton X-100 in PBS. The fixedCFs were stained with selective anti–pSmad2, anti-pSmad3 oranti-Smad2/3 primary antibodies (1:400�, Cell Signaling Technol-ogy) overnight at 4°C and then with a Texas-conjugated donkeyanti-rabbit IgG secondary antibody (1:500�, Jackson ImmunoRe-search Laboratory) for 1 hour at room temperature to assess nucleartranslocation of pSmad2 and pSmad3 using confocal fluorescencemicroscopy with a computerized Zeiss Axioskop system.

Effects of TGF-�1 and cGMP on Phosphorylationof Smad ProteinsQuiescent CFs were pretreated with cGMP (1 mmol/L) or vehicle for30 minutes before addition of TGF-�1 (1 ng/mL) to the medium andincubated for an additional 30 minutes and then harvested forassessment of pSmad2 and pSmad3 using Western blot analysis.12

We then tested the hypothesis that cGMP/PKG activation canphosphorylate Smad3 protein at sites other than the C-terminal SSXSmotif, in which the 2 end serines (Ser423/425) are phosphorylated bythe TGF-� type I receptor kinase.13 A plasmid with a glutathioneS-transferase (GST)-Smad3 fusion gene was transfected into Esch-erichia coli and expression of fusion proteins was induced byisopropyl-B-D-thiogalactoside. Proteins were harvested with gluta-thione beads and cleaved with thrombin to yield purified Smad3 asdescribed previously.14 Purified Smad3 (10 �g) was first phosphor-ylated with a recombinant PKG-1� (1000 U for 20 minutes,Calbiochem) using �-32P-ATP or nonradioactive ATP as substrateand then digested with carboxypeptidase Y (a C-terminal peptidase,1 ng/�L, Sigma) for 1, 30, and 60 minutes. The reaction mixture wassubjected to 10% SDS-PAGE. The gel with 32P-Smad3 was used forradioautography to identify the phosphorylated peptides, and thegel with nonradioactive pSmad3 was used for Western blotanalysis to identify the Smad3 peptides. The Western blot wasprobed with either a selective anti–N-terminal Smad3 antibody or ananti–C-terminal Smad3-pSer423/S425 antibody.

Identification of the PKG Phosphorylated Sites onSmad3 Protein by Fourier Transform IonCyclotron Resonance Mass SpectrometryTwo micrograms of PKG-1�-phosphorylated Smad3 (n�6) andunphosphorylated (reaction without PKG in the buffer, as a negativecontrol, n�2) were dissolved in 100 �L of 50 mmol/L ammoniumbicarbonate solution. pSmad3 was first reduced in 10 mmol/Ldithiothreitol (by adding 5 �L of 200 mmol/L dithiothreitol to thereaction mixture) for 30 minutes, alkylated in 50 mmol/L iodoacet-amide (by adding 5 �L of 1 mol/L iodoacetamide to the reactionmixture) for 30 minutes, and then digested with 0.2 �g (by adding 5�L of 40 �g/mL enzyme to the reaction mixture) of trypsin(Promega) or GluC (Roche) for 24 hours at room temperature. Thetrypsin- or GluC-digested pSmad3 peptides were then analyzed byuse of reversed-phase C18 liquid chromatography (RP-C18 LC)-Fourier transform ion cyclotron resonance (FT-ICR) mass spectrom-etry (MS) analysis and ion trap tandem MS (MS/MS), as describedpreviously.15,16

Statistical AnalysisResults are expressed as means�SE. Statistical analyses wereperformed using the SigmaStat package (Jandel Scientific Software,San Rafael, Calif) on a personal computer. The primary statisticaltest was ANOVA. If ANOVA results were significant, a post hoccomparison among groups was performed with the Newman–Keulstest. Differences in mean values were reported as significant if theprobability value was �0.05.

ResultsMF Transformation and Collagen Deposition AreIncreased in Hearts of Nppa�/� Mice ComparedWith Nppa�/� Mice Subjected to TACIn response to TAC, Nppa�/� mice developed significantcardiac hypertrophy (left ventricular weight [adjusted by

body weight]: 209�6 mg in TAC-Nppa�/� mice [n�10]versus 124�3 mg in Sham-Nppa�/� mice [n�8]), withgreatly increased numbers of MFs (�-SMA antibody–stainedcells that are not vascular smooth muscle cells or cardiomyo-cytes [CMs]) and more robust collagen protein expression(Picrosirius stain) compared with wild-type (Nppa�/�) mice(left ventricular weight [adjusted by body weight]: 113�3 mgin TAC-Nppa�/� mice [n�9] versus 99�2 mg in Sham-Nppa�/� mice [n�9]) (Figure 1). MFs were colocalized withcollagen in the hypertrophic LVs in both genotypes (Figure1A versus 1B and 1C versus 1D). No MFs (data not shown)or significant collagen deposition was observed in sham-operated Nppa�/� or Nppa�/� mice (Figure 1E). Higher activeTGF-�1 protein levels were observed in LVs of wild-typeTAC mice than in sham-operated controls (Figure 1F).

Expression of �-SMA was also induced in some CMs inLVs of TAC mice (data not shown). The �-SMA in CMs wasnot colocalized with collagen, suggesting that these cells arelikely not synthesizing ECM.

These results suggest that ANP is a negative modulator ofMF transformation in response to pressure-overload stressand that the exaggerated cardiac hypertrophy/remodelingobserved in Nppa�/� mice in response to TAC is related toincreased MF transformation and ECM deposition.

Figure 1. A through D, Representative micrographs (�100) ofcollagen (Picrosirius red staining) (A and C) and �-SMA–positiveMFs (B and D) in LVs of male Nppa�/� and Nppa�/� mice 1week after TAC. Arrows indicate MFs. E, Interstitial collagen vol-ume calculated in Picrosirius red–stained cross-sections of theLVs below the mitral valve in control or TAC (1 week) Nppa�/�

and Nppa�/� mice. F, Effects of 1 week of TAC on total andactive TGF-�1 protein levels in LVs of Nppa�/� mice. Results aremeans�SE; numbers in parentheses indicate numbers of mice.*P�0.05 compared with respective Nppa�/� groups; #P�0.05compared with respective sham-operated controls by 2-wayANOVA.

Li et al PKG Blocks TGF-� Action in Cardiac Fibroblasts 187

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

cGMP Inhibits TGF-�1–Induced MFTransformation of Mouse CFs In VitroMFs in vehicle-treated cultures (0.1% FBS medium) weresparse and characterized by disorganized intracellular�-SMA (percentage of �-SMA–positive cells, 27�5%; n�4plates/group) (Figure 2A and 2E), whereas most (95�3%;n�4) cells in TGF-�1 treated cultures were �-SMA–positiveand contained well-organized �-SMA filaments, indicatingnearly complete transformation of CFs to MFs (Figure 2B).cGMP completely inhibited TGF-�1–induced MF transfor-mation (percentage of �-SMA–positive cells, 21�5%; n�4)but did not change basal (% �-SMA positive cells�23�5%,n�4) MF numbers (Figure 2C through 2E). Together withour in vivo data, these results support the hypothesis thatANP (through cGMP) inhibits TGF-�–induced MF transfor-mation and ECM deposition.

ANP Increases cGMP Expression and ANP and/orcGMP Inhibits TGF-�–Stimulated CollagenSynthesis, Cell Proliferation, and PAI-1 mRNAExpression in CFs In VitroExposure to ANP (10�7 mol/L) for 30 minutes increasedcGMP levels in both mouse and rat CFs (Figure 3A). TGF-�increased collagen synthesis (3[H]-proline incorporation intoCFs) (Figure 3B) and cell proliferation (total dehydrogenaseactivity) (Figure 3C), and pretreatment with ANP or cGMP

decreased basal collagen synthesis and inhibited TGF-�–induced collagen synthesis and proliferation of mouse CFs.

Pretreatment with cGMP decreased baseline levels ofPAI-1 and significantly attenuated TGF-�1–stimulated PAI-1mRNA expression (Figure 3D). Pretreatment with the PKGinhibitor KT5823 blocked the inhibitory effects of cGMP onTGF-�1–stimulated PAI-1 mRNA expression in CFs, sug-gesting that cGMP was acting through activation of PKG.These data support the hypothesis that ANP/cGMP/PKGsignaling has antifibrogenic effects that antagonize TGF-�–induced stimulation of ECM expression in CFs.

ANP and cGMP Inhibit TGF-�–Induced NuclearTranslocation of pSmad3 in Mouse CFsIn vehicle-treated cells, immunostaining of pSmad3 (Figure4A and 4E) and pSmad2 (data not shown) was weak anddistributed evenly in cytoplasm and nucleus, suggestingthat pSmad3 and pSmad2 levels were low and withoutsignificant nuclear translocation. TGF-�1 (1 ng/mL for 30minutes) treatment significantly stimulated nuclear trans-location of pSmad3, indicated by strong pSmad3 stainingin the nucleus (Figure 4B and 4E). We did not observesignificant pSmad2 nuclear translocation after 30 minutesof TGF-�1 treatment in mouse CFs (data not shown).Nearly 100% of TGF-�1–treated CFs had accumulatedpSmad3 in their nuclei, and �30% of CFs had doublenuclei, indicating cell proliferation. Pretreatment withANP (1 �mol/L for 30 minutes) or cGMP (1 mmol/L for30 minutes) inhibited TGF-�1–induced pSmad3 transloca-tion, with substantial levels of Smad3 staining remainingin the cytoplasm in most cells (Figure 4C through 4E).

To test whether activation of PKG mediates the inhibitoryeffects of ANP and cGMP on TGF-�1–induced nucleartranslocation of pSmad3, pSmad3 nuclear translocation wasmeasured in the presence of the PKG inhibitor KT5823.Pretreatment with KT5823 blocked the inhibitory effects ofANP and cGMP on TGF-�1–stimulated pSmad3 nucleartranslocation (Figure 4F).

cGMP Does Not Inhibit TGF-�1–Induced pSmad3and pSmad2 PhosphorylationWestern blot analysis demonstrated that TGF-�1 treatmentsignificantly increased pSmad3 and pSmad2 levels in CFsand that pretreatment with cGMP did not inhibit thisprocess (Figure 5A). Neither cGMP nor TGF-� alteredtotal Smad2/3 levels in these cells. Thus, disruption ofSmad3 and Smad2 phosphorylation does not account forthe inhibitory effects of ANP and cGMP on TGF-�1–induced ECM expression in CFs.

PKG Phosphorylates Smad3 Protein at Sites OtherThan the C-Terminal ResiduesAs an initial test of the hypothesis that PKG can phosphor-ylate Smad3 protein at novel sites other than the C-terminalSer423/425 residues that are phosphorylated by the TGF-�type I receptor kinase, PKG-phosphorylated Smad3 wasdigested with carboxypeptidase Y and size fractionated bySDS-PAGE. Autoradiographic analysis and Western blots ofpSmad3 fragments probed with either a selective anti–N-

Figure 2. A through E, Representative micrographs (�400) (Athrough D) and bar graphs of means�SE (E) of �-SMA–stainedcultured mouse CFs treated with TGF-�1 (1 ng/mL for 24 hours)and/or cGMP (8-bromo-cGMP, a cGMP analog, 1 mmol/L for 25hours or 1 hour before TGF-�1). Slides were counterstainedwith hematoxylin to show nuclei. A total of �500 cells werecounted in 4 slides per group in 2 experiments. *P�0.05 com-pared with the vehicle control group; #P�0.05 compared withthe TGF-�1 alone group.

188 Circulation Research February 1, 2008

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

terminal Smad3 antibody or an anti–C-terminal Smad3-pSer423/S425 antibody indicated that PKG did phosphorylateSmad3 when the C-terminal residues were removed by aC-terminal peptidase (Figure 5B), supporting the hypothesisthat PKG can phosphorylate Smad3 at novel sites other thanits C-terminal residues.

RP-C18-LC FT-ICR-MS/MS analysis of trypsin- or Gluc-digested pSmad3 that was phosphorylated by PKG confirmedthat Ser309 and Thr388 on pSmad3 was phosphorylated byPKG (Figure 6). Figure 6B shows the FT-ICR MS spectrumof a doubly charged ion species at m/z 822.3542, correspond-ing to the mass of the pSmad3 tryptic peptide L296-S309 plusthe addition of a phosphate group (theoretical, 1643.7063 m/z;3.18 ppm mass error). Figure 6C shows the linear quadrupoleion trap (LTQ) MS/MS product ion spectrum of the same ionafter fragmentation with the dominant fragment correspond-ing to the neutral loss of 80 Da, confirming the phosphory-lated pSmad3 peptide. Phosphorylation of Thr388 was con-firmed in similar fashion (data not shown).

DiscussionThis study yields novel insights into precise sites of“molecular merging” of pro- and antifibrogenic pathwaysin heart and characterizes CFs as the target cell in whichpro- and antifibrogenic signaling cascades converge andregulate responses of the heart to hemodynamic stress. Wedemonstrate crosstalk between ANP and TGF-� signalingpathways in CFs in vitro such that ANP and cGMP,through a PKG-dependent mechanism, block induction ofECM and PAI-1 expression by TGF-�1. Our observationthat cGMP inhibits TGF-�1–induced nuclear translocationof pSmad2 and pSmad3 in CFs defines for the first time aprecise molecular mechanism by which ANP/cGMP/PKG

signaling interferes with downstream signaling from TGF-�and thus protects against cardiac remodeling/fibrosis andfailure in response to hemodynamic stress. These findingssuggest the intriguing hypothesis that ANP signaling resultsin phosphorylation of Smad proteins on sites other than theirC-terminal residues, thus blocking their nuclear translocationand binding to TGF-�–Smad responsive elements in thepromoter regions of ECM genes.

TGF-�1, the major isoform in heart,17 is produced in CFsand CMs under stressful conditions and stimulates CF trans-formation (to MFs) and proliferation, as well as ECMproduction in response to hypertrophic stimuli.17,18 Thus, itplays a major role in cardiac remodeling under stressconditions.

Activated TGF-� ligands bind to a heteromeric complex oftype II (TGF�RII) and type I (TGF�RI) receptors thattransduce intracellular signals via phosphorylation ofTGF�RI-associated Smad2 and Smad3.14,18 Phosphorylationof TGF�RI-associated Smad2 and Smad3 and nuclear trans-location of pSmad2 and pSmad3 are critical steps in TGF-�signaling.14,18 Smad2 (467 aa) and Smad3 (425 aa) containpredominantly serine, with some threonine residues, in theC-terminal, linker, and MH1 regions that are accessible forphosphorylation. After ligand binding, phosphorylation byTGF�R1 kinase of the 2 most C-terminal serine residuesdrives the activation of Smad2 and Smad3 is required fornuclear translocation and subsequent binding of pSmad2 andpSmad3 to nuclear transcriptional factors and DNA thatregulate the transcriptional expression of downstreamgenes.14,18

The interplay between Smads and other signaling pathwaysin the cytoplasm and nucleus is a critical mechanism bywhich the activities and expression of Smad proteins are

Figure 3. A, Effects of ANP (0.1 �mol/Lfor 30 minutes) on intracellular cGMPlevels in mouse and rat CFs. B throughD, Effects of ANP (1 �mol/L) or cGMP(1 mmol/L) on TGF-�1–stimulated (5ng/mL) collagen synthesis, cell prolifera-tion, and PAI-1 mRNA expression inmouse CFs. In B and D, quiescent CFswere pretreated with ANP or cGMP for30 minutes before TGF-�1 for an addi-tional 24 hours before being harvested.In C, proliferation of CFs (n�24 wells/group) was measured using the Non-Radioactive Cell Proliferation Assay(Promega) at 30, 90, and 240 minutesafter TGF-�1 treatment. MTS (3-[4,5-dimethylthiazol-2-yl]-5-[3-carboxy-methoxyphenyl]-2-[4-sulfophenyl]-2H-tetrazolium), a dehydrogenase substrate,and phenazine-methosulfate (PMS), anelectron-coupling reagent, were addedto medium at 30 minutes after TGF-�1treatment for measurement of formazanformation as an index of cell proliferation.In D, subgroups of CFs were pretreatedwith KT5823 (1 �mol/L) for 15 minutesbefore cGMP or TGF-�1; Northern blotanalysis was performed with 15 �g oftotal RNA, and PAI-1 mRNA data were

normalized to GAPDH mRNA. Results are means�SE (n�dishes/group). *P�0.05 compared with the vehicle control group; #P�0.05compared with the TGF-�1 alone group.

Li et al PKG Blocks TGF-� Action in Cardiac Fibroblasts 189

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

modulated and is responsible for the diverse effects of theTGF-� family of proteins.14 Mice overexpressing TGF-�have cardiac hypertrophy and interstitial fibrosis,8 and studiesin rat and human CFs have shown that B-type natriureticpeptide (BNP) and NO attenuate expression of TGF-�through activation of cGMP. However, the signaling cascadesinvolved remain to be elucidated.8,19

Both inhibition and overphosphorylation of Smad2 orSmad3 have been reported to disrupt their heterodimerizationwith Smad4 and nuclear translocation, resulting in repressionof transcriptional activation of TGF-� responsive promot-ers.20,21 Thus, phosphorylation not only activates Smad pro-teins but also modulates their activity. This provides apotential mechanism for integration of the Smad pathwaywith ANP/cGMP/PKG signaling pathway that modulatesTGF-� signal transduction. The current study demonstratesthat the PKG acts as a stronger protein kinase than TGF�RIand phosphorylates additional serine or threonine residues onSmad3 (and/or Smad2), thus disrupting their nuclear translo-cation, resulting in repression of transcriptional activation ofTGF-� response promoters, eg, on collagen and PAI-1 genes.This represents a novel molecular merging mechanism bywhich ANP/cGMP/PKG signaling intercepts the TGF-� sig-

naling cascade and may contribute to the antifibrogeniceffects of ANP in the stressed heart.

We recognize that overphosphorylation of Smads is not theonly cellular function of PKG activation. PKG has diverseintracellular actions, including integrin signal transduction,modulation of Ca2� release and uptake into SR, alteration ofmembrane K� fluxes, and nuclear protein phosphorylationand translocation.22 An alternative (to overphosphorylation)explanation for the observation that ANP signaling preventsnuclear translocation of Smads is that cGMP and PKG mayalter the affinity of Smads for cytoplasmic anchoring mole-cules or nuclear export proteins. Subcellular localization ofSmads has been shown to be controlled by interaction withthese cytoplasmic and nuclear retention factors.14 The precisemolecular basis for retention of pSmads in the cytosolfollowing ANP or cGMP treatment remains to be identifiedand is a topic for current investigation in our laboratory.

Several recent studies have demonstrated that variouscytoplasmic protein kinases and cyclic nucleotides participatein regulating responses to TGF-�.23 Extracellular signal-regulated mitogen-activated protein kinase inhibits TGF-�signaling via inhibition of nuclear accumulation of Smad2 innormal epithelial cells.24 PKC directly phosphorylates Smad3

Figure 4. A through D, Representativemicrographs show that that cGMP inhib-its TGF-�1–stimulated nuclear transloca-tion of pSmad3 in mouse CFs. CFs werecultured on cover slides in 0.1% FBS-DMEM. CFs were treated with TGF-�1(1 ng/mL) for 30 minutes (B1 through B3)and cGMP (1 mmol/L) for 60 minutes (C1through C3), respectively. D1 through D3,CFs were treated with cGMP (1 mmol/L)for 30 minutes before TGF-�1 (1 ng/mL)for an additional 30 minutes. Fixed CFswere stained with a anti-pSmad3 anti-body and a Texas red–labeled secondaryantibody. Arrows indicate representativeCFs with accumulation of pSmad3 incells with double nuclei. A1 through D1,Images taken with a fluorescent micro-scope (�100). A2 through D2, Imagestaken with a confocal fluorescencemicroscope (�400). A3 through D3,Nuclei of A2 through D2 were stainedwith DAPI. E, Relative intensity ofpSmad3 staining measured in nucleusand cytoplasm of CFs treated with vehi-cle, cGMP (1 mmol/L), and/or TGF-�1 (1ng/mL) as in A2 through D2. Results aremeans�SE (numbers in parentheses indi-cate number of cells measured; a total of�20 fields [10 in nucleus and 10 in cyto-plasm] in each cell were scanned andaveraged). F, Quiescent CFs were pre-treated with ANP (1 �mol/L) or cGMP(1 mmol/L) before TGF-�1 (1 ng/mL) foran additional 30 minutes before stainedfor pSmad3. Subgroups of CFs were pre-treated with KT5823 (1 �mol/L) for 15minutes before ANP or cGMP. Y-axisscale represents percentage of total cellswith pSmad3 nuclear translocation.Results are means�SE (numbers inparentheses indicate number of slides/group; a total of �200 cells were

counted per group). *P�0.05 compared with the vehicle control group; #P�0.05 compared with the TGF-�1 alone group.

190 Circulation Research February 1, 2008

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

and abrogates the ability of Smad3 to bind directly to DNA,leading to impairment of transcriptional responses dependenton the direct binding of Smad3 to DNA.21 Activation ofCa2�/calmodulin-dependent protein kinase II preventsSmad2/4 heterodimerization and nuclear translocation andconcomitant transcriptional responses in HEK-293 humankidney fibroblasts.25 Further, intracellular cAMP-elevatingagents, such as prostaglandin E2 and the adenylate cyclaseactivator forskolin, inhibit TGF-�–induced Smad3/4-dependent gene expression via a cAMP-dependent, PKA-dependent mechanism in human keratinocytes.26 The currentstudy shows that ANP and cGMP inhibit TGF-�–inducednuclear translocation of pSmad2 and pSmad3 in CFs, defin-ing a new role for cyclic mononucleotide phosphate secondmessenger in regulating profibrogenic responses to TGF-�.Pretreatment with the PKG inhibitors KT5823 prevented theinhibitory effects of ANP and cGMP on TGF-�–stimulatednuclear translocation of pSmad3, supporting the hypothesisthat these effects are mediated through activation of PKG.

Our previous in vivo studies have validated the importanceof ANP and TGF-� and ANP as opposing influences in the

pathogenesis of pressure overload–induced cardiac hypertro-phy, fibrosis, and remodeling.1 We have shown that Nppa�/�

mice develop cardiac enlargement and remodeling in re-sponse to pressure-overload stress, compared with Nppa�/�

mice.1,2 Similarly, other investigators have shown thatNpr1�/� mice carrying gene-targeted disruption of Npr1(encoding for the NPRA receptor) exhibit the same pheno-type of cardiac hypertrophy and fibrosis as Nppa�/� mice andare associated with reduced guanylyl cyclase activity andcGMP levels and increased expression of angiotensin-converting enzyme and angiotensin II, as well as proinflam-matory cytokines such as tumor necrosis factor-�,interleukin-6, and TGF-�1 in the heart, suggesting thatdisruption of ANP/NPRA/cGMP signaling leads to increasesin cardiac hypertrophic stimuli.27,28 We have shown that theexcess cardiac enlargement and remodeling in Nppa�/� miceare a consequence of increased MF transformation andproliferation and deposition of ECM components, rather thanexcess CM hypertrophy.1 In contrast, using a novelDnTGF�RII mouse model, these processes are markedlyattenuated by disruption of TGF-� signaling in heart.3–4

Taken together, these data support the hypothesis that endog-enous ANP and TGF-� play important counterregulatoryroles in regulating MF transformation and proliferation, ECMproduction, and cardiac remodeling in response to pressure-overload stress. An imbalance in the normal relationshipsbetween the antigrowth/antifibrogenic effects of ANP and

Figure 5. A, Representative Western blots and averaged bargraphs of pSmad3 demonstrating that cGMP does not blockTGF-�1–induced phosphorylation of Smad3 and Smad2 inmouse CFs. Quiescent CFs were stimulated with TGF-�1 (1ng/mL) for 30 minutes with or without pretreatment with cGMP(1 mmol/L for 30 minutes). Cell lysates (25 �g) were size frac-tionated by SDS-PAGE, and Western analysis was performedwith selective anti-pSmad3, anti-Smad2, and �-actin antibodies.�-Actin protein levels were measured to show protein loading.Veh indicates vehicle control: CFs cultured in 0.1% FBSmedium. Results are means�SE (numbers in parentheses indi-cate samples per group). *P�0.05 compared with the vehiclecontrol group; #P�0.05 compared with the TGF-�1 alonegroup. B, Radioautographs and Western blots of purified Smad3protein that was phosphorylated with PKG (20 minutes) andthen digested with carboxypeptidase Y for 1, 30, and 60 min-utes. Western blot analysis of carboxypeptidase Y–digestedpSmad3 was performed with selective anti-Smad3 N-terminal oranti-Smad3 C-terminal antibodies.

Figure 6. MS/MS analysis of Smad3 that has been phosphory-lated with PKG (n�6 for either trypsin or Gluc digestion duringthe analysis). The solid underlined portion indicates the MH1domain; dashed underlined portion, the MH2 domain. More than99% of Smad3 peptide sequence (except V424 and S425) wasdetected by MS/MS analysis. The asterisks mark the phosphor-ylation sites (Ser309 and Thr388) identified by MS/MS in bothtrypsin- and Gluc-digested pSmad3 fragments. These 2 siteswere not phosphorylated in negative controls (unphosphorylatedSmad3; n�2). B, LTQ-FT-ICR spectrum of single ion isolatedfrom a LC-MS of trypsin-digested Smad3 that had been incu-bated with PKG. The ion species matches the theoretical massof the Smad3 tryptic peptide�HPO3. C, LTQ MS/MS spectrumof the same ion shows a neutral loss of 80 Da, corresponding tothe loss of HPO3 that is characteristic of phosphopeptides, con-firming phosphorylation of Smad3.

Li et al PKG Blocks TGF-� Action in Cardiac Fibroblasts 191

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

mitogenic/profibrogenic effects of TGF-� results in pressureoverload–induced cardiac fibrosis and remodeling.

Sources of FundingThis work was supported in part by NIH grants HL-080017,HL-044195, HL-075614, HL-07457, HL-64614, HL-075211, CA-101955, and DK60913 and American Heart Association Grant0455197B GIA.

DisclosuresNone.

References1. Wang D, Oparil S, Feng JA, Li P, Perry GJ, Chen LB, Dai M, John SWM,

Chen YF. Effects of pressure overload on extracellular matrix expressionin heart of the atrial natriuretic peptide null mice. Hypertension. 2003;42:88–95.

2. Franco V, Chen YF, Oparil S, Feng JA, Wang D, Hage F, Perry G. Atrialnatriuretic peptide dose dependently inhibits pressure overload-inducedcardiac remodeling. Hypertension. 2004;44:746–750.

3. Mehra A, Wrana JL. TGF-beta and the smad signal transduction pathway.Biochem Cell Biol. 2002;80:605–622.

4. Franco V, Chen YF, Feng JA, Wang D, Oparil S, Perry G. Eplerenoneprevents adverse cardiac remodelling induced by pressure overload inatrial natriuretic peptide-null mice. Clin Exp Pharmacol Physiol. 2006;33:773–779.

5. Redondo J, Bishop JE, Wilkins MR. Effect of atrial natriuretic peptideand cyclic GMP phosphodiesterase inhibition on collagen synthesis byadult cardiac fibroblasts. Br J Pharmacol. 1998;124:1455–1462.

6. Calderone A, Takahashi N, Izzo NJ Jr, Thaik CM, Colucci WS. Pressure-and volume-induced left ventricular hypertrophies are associated withdistinct myocyte phenotypes and differential induction of peptide growthfactor mRNAs. Circulation. 1995;92:2385–2390.

7. Lijnen PJ, Petrov VV, Fagard RH. Induction of cardiac fibrosis bytransforming growth factor-�1. Mol Genet Metab. 2000;71:418–435.

8. Leask A, Abraham DJ. TGF-beta signaling and the fibrotic response.FASEB J. 2004;18:816–827.

9. Shi X, Yang X, Chen D, Chang Z, Cao X. Smad1 interacts withhomeobox DNA-binding proteins in bone morphogenetic protein sig-naling. J Biol Chem. 1999;274:13711–13717.

10. Chen YF, Feng JA, Li P, Xing D, Zhang Y, Serra R, Ambalavanan N,Majid-Hassan E, Oparil S. Dominant negative mutation of the TGF-betareceptor blocks hypoxia-induced pulmonary vascular remodeling. J ApplPhysiol. 2006;100:564–571.

11. Maki T, Horio T, Yoshihara F, Suga SI, Takeo S, Matsuo H, Kangawa K.Effect of neutral endopeptidase inhibitor on endogenous atrial natriureticpeptide as a paracrine factor in cultured cardiac fibroblasts. Br JPharmacol. 2000;131:1204–1210.

12. Li P, Oparil S, Novak L, Cao X, Shi W, Lucas J, Chen YF. ANP signalinginhibits TGF-beta-induced Smad2 and Smad3 nuclear translocation and

extracellular matrix expression in rat pulmonary arterial smooth musclecells. J Appl Physiol. 2007;102:390–398.

13. Massague J. TGF� signal transduction. Annu Rev Biochem. 1998;67:753–791.

14. Shi Y, Massague J. Mechanisms of TGF-beta signaling from cellmembrane to the nucleus. Cell. 2003;113:685–700.

15. Chalmers MJ, Kolch W, Emmett MR, Marshall AG, Mischak H. Identi-fication and analysis of phosphopeptides. J Chromatogr B Analyt TechnolBiomed Life Sci. 2004;803:111–120.

16. Eliuk S, Renfrow MB, Shonsey EM, Barnes S, Kim H. Active sitemodifications of creatine kinase brain isoform by 4-hydroxy-2-nonenalcorrelate with reduced enzyme activity; mapping of modified sites byFourier transform-ion cyclotron resonance mass spectrometry. Chem ResToxicol. 2007;20:1260–1268.

17. Calderone A, Murphy RJ, Lavoie J, Colombo F, Beliveau L. TGF-�1 andprepro-ANP mRNAs are differentially regulated in exercise-inducedcardiac hypertrophy. J Appl Physiol. 2001;91:771–776.

18. Petrov VV, Fagard RH, Lijnen PJ. Stimulation of collagen production bytransforming growth factor-beta1 during differentiation of cardiac fibro-blasts to myofibroblasts. Hypertension. 2002;39:258–263.

19. Rosenkranz AC, Woods RL, Dusting GJ, Ritchie RH. Antihypertrophicactions of the natriuretic peptides in adult rat cardiomyocytes: importanceof cyclic GMP. Cardiovasc Res. 2003;57:515–522.

20. Abdelaziz N, Colombo F, Mercier I, Calderone A. Nitric oxide attenuatesthe expression of transforming growth factor-�3 mRNA in rat cardiacfibroblasts via destabilization. Hypertension. 2001;38:261–266.

21. Yakymovych I, Ten Dijke P, Heldin CH, Souchelnytskyi S. Regulation ofSmad signaling by protein kinase C. FASEB J. 2001;15:553–555.

22. Pan D, Estevez-Salmeron LD, Stroschein SL, Zhu X, He J, Zhou S, LuoK. The integral inner nuclear membrane protein MAN1 physicallyinteracts with the R-Smad proteins to repress signaling by the TGFbetasuperfamily of cytokines. J Biol Chem. 2005;280:15992–16001.

23. Lincoln TM, Dey N, Sellak H. cGMP-dependent protein kinase signalingmechanisms in smooth muscle: from the regulation of tone to geneexpression. J Appl Physiol. 2001;91:1421–1430.

24. Liu F. Smad3 phosphorylation by cyclin-dependent kinases. CytokineGrowth Factor Rev. 2006;17:9–17.

25. Wicks SJ, Lui S, Abdel-Wahab N, Mason RM, Chantry A. Inactivation ofsmad-transforming growth factor beta signaling by Ca(2�)-calmodulin-dependent protein kinase II. Mol Cell Biol. 2000;20:8103–8111.

26. Schiller M, Verrecchia F, Mauviel A. Cyclic adenosine 3�,5�-monophosphate-elevating agents inhibit transforming growth factor-beta-induced SMAD3/4-dependent transcription via a protein kinaseA-dependent mechanism. Oncogene. 2003;22:8881–8890.

27. Vellaichamy E, Kaur K, Pandey KN. Enhanced activation of pro-inflammatory cytokines in mice lacking natriuretic peptide receptor-A.Peptide. 2007;28:893–899.

28. Vellaichamy E, Zhao D, Somanna N, Pandey KN. Disruption of guanylylcyclase/natriuretic peptide receptor-A gene upregulates AT1 receptorsignaling in null mutant mice: role in cardiac hypertrophy. PhysiolGenomics. 2007;31:193–202.

192 Circulation Research February 1, 2008

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from

Matthew B. Renfrow and Yiu-Fai ChenPeng Li, Dajun Wang, Jason Lucas, Suzanne Oparil, Dongqi Xing, Xu Cao, Lea Novak,

Signaling and Myofibroblast Transformation in Mouse Cardiac FibroblastsInduced Smad−βAtrial Natriuretic Peptide Inhibits Transforming Growth Factor

Print ISSN: 0009-7330. Online ISSN: 1524-4571 Copyright © 2007 American Heart Association, Inc. All rights reserved.is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Circulation Research

doi: 10.1161/CIRCRESAHA.107.1576772008;102:185-192; originally published online November 8, 2007;Circ Res. 

http://circres.ahajournals.org/content/102/2/185World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://circres.ahajournals.org//subscriptions/

is online at: Circulation Research Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer about this process is available in the

located, click Request Permissions in the middle column of the Web page under Services. Further informationEditorial Office. Once the online version of the published article for which permission is being requested is

can be obtained via RightsLink, a service of the Copyright Clearance Center, not theCirculation Researchin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on June 9, 2015http://circres.ahajournals.org/Downloaded from