NAD(P)H Oxidase 4 Mediates Transforming Growth Factor-1eCInduced Differentiation of Cardiac Fibroblasts Into Myofibroblasts
http://www.100md.com
循环研究杂志 2005年第10期
The Department of Medicine, Division of Cardiology, Emory University, Atlanta, Ga.
Abstract
Human cardiac fibroblasts are the main source of cardiac fibrosis associated with cardiac hypertrophy and heart failure. Transforming growth factor-1 (TGF-1) irreversibly converts fibroblasts into pathological myofibroblasts, which express smooth muscle -actin (SM -actin) de novo and produce extracellular matrix. We hypothesized that TGF-1eCstimulated conversion of fibroblasts to myofibroblasts requires reactive oxygen species derived from NAD(P)H oxidases (Nox). We found that TGF-1 potently upregulates the contractile marker SM -actin mRNA (7.5±0.8-fold versus control). To determine whether Nox enzymes are involved, we first performed quantitative real time polymerase chain reaction and found that Nox5 and Nox4 are abundantly expressed in cardiac fibroblasts, whereas Nox1 and Nox2 are barely detectable. On stimulation with TGF-1, Nox4 mRNA is dramatically upregulated by 16.2±0.8-fold (n=3, P<0.005), whereas Nox5 is downregulated. Small interference RNA against Nox4 downregulates Nox4 mRNA by 80±5%, inhibits NADPH-driven superoxide production in response to TGF-1 by 65±7%, and reduces TGF-1eCinduced expression of SM -actin by 95±2% (n=6, P<0.05). Because activation of small mothers against decapentaplegic (Smads) 2/3 is critical for myofibroblast conversion in response to TGF-1, we also determined whether Nox4 affects Smad 2/3 phosphorylation. Depletion of Nox4 but not Nox5 inhibits baseline and TGF-1 stimulation of Smad 2/3 phosphorylation by 75±5% and 68±3%, respectively (n=7, P<0.0001). We conclude that Nox 4 mediates TGF-1eCinduced conversion of fibroblasts to myofibroblasts by regulating Smad 2/3 activation. Thus, Nox4 may play a critical role in the pathological activation of cardiac fibroblasts in cardiac fibrosis associated with human heart failure.
Key Words: Nox4 human cardiac fibroblasts transforming growth factor reactive oxygen species Smad 2/3
Introduction
Heart failure remains the leading cause of hospital admissions in the United States, with more than 550 000 new patients diagnosed each year.1 Regardless of etiology, cardiac fibrosis is a major contributor to cardiac remodeling associated with cardiomyopathies. It is characterized by expansion of the interstitial compartment due to increased deposition of extracellular matrix by activated myofibroblasts.2 Cardiac myofibroblasts are specialized contractile fibroblasts formed by irreversible acquisition of contractile proteins such as smooth muscle -actin (SM -actin) in response to potent fibrogenic cytokines.3 The expression of SM -actin is regulated by transforming growth factor-1 (TGF-1), a primary fibrogenic growth factor in heart failure that is downstream of many of the pro-fibrotic actions of other fibroblast growth factors, such as angiotensin II, aldosterone, and norepinephrine.4 TGF-1 is upregulated in failing human hearts and various experimental models of cardiac hypertrophy,4 and functional blockade of TGF-1 prevents cardiac interstitial fibrosis induced by pressure overload in rats.5
There is increasing evidence that oxidative stress plays a critical role in the development and progression of cardiac remodeling associated with heart failure.6 Oxidative stress is increased in human heart failure7,8 and animal models of cardiac hypertrophy and fibrosis (induced by angiotensin II,9,10 aldosterone,11 myocardial infarction,12 tachycardia-induced cardiomyopathy,13,14 and aortic coarctation10,15). The first evidence that elevated oxidative stress can cause cardiomyopathy was provided by Li et al,16 who showed that deletion of superoxide dismutase (a main scavenger of intracellular superoxide) caused dilated cardiomyopathy and cardiac fibrosis in mice.16 Reactive oxygen species (ROS) modulate extracellular matrix remodeling by mediating cardiac fibroblast function and also by stimulating collagen turnover via activation of matrix metalloproteinases, enzymes critical for extracellular matrix remodeling.17 ROS also stimulate release and activation of cytokines. For example, in cell culture, exposure of cardiac fibroblasts to superoxide stimulates their proliferation by increasing the production of TGF-1.18
NADPH oxidases (Nox) are multisubunit enzymes that generate superoxide by transferring electrons from NADPH to molecular oxygen.19 Although the prototype catalytic subunit gp91phox (Nox2) was originally discovered in neutrophils, there are also multiple homologues in nonphagocytic cells with wide range tissue distribution: Nox1, Nox3, Nox4, Nox5, and the dual oxidases Duox1 and Duox2.20 In phagocytes, ROS originating from Nox participate in bacterial killing, whereas in nonphagocytic cells they are required for cellular responses to cytokines and growth factors.21 Fibroblasts also produce ROS via Nox in response to growth factors. For example, lung fibroblasts endogenously release ROS from Nox in response to TGF-1.22 An important consequence of fibroblast stimulation with TGF-1 is a phenotypical conversion into myofibroblasts. Although NADPH oxidases are activated by TGF-1, it is unknown whether their activation is required for upregulation of SM -actin, the key element that defines the acquisition of contractile phenotype by myofibroblasts.
In the current study, we sought to (1) determine whether ROS derived from NADPH oxidases are important for TGF-1eCinduced myofibroblast differentiation, (2) identify the catalytic Nox subunit that is required for SM -actin expression in response to TGF-1, and (3) discover the mechanism by which Nox-derived ROS mediate the conversion to the myofibroblast phenotype. These data provide new insights into the mechanisms underlying the pathological activation of fibroblasts in cardiac fibrosis induced by TGF-1.
Materials and Methods
Cell Culture
Primary human cardiac fibroblast cells (catalog number ACBRI 5118) were purchased at passage 2 from Cell Systems (Kirkland, WA) and grown in fibroblast growth medium-2 (Cambrex, Baltimore, Md). All cells were harvested at passage 3 to 6 and used at 50% to 70% confluence for transfection experiments and 80% to 90% confluence for time course experiments (see online data supplement at http://circres.ahajournals.org for characterization of human cardiac fibroblasts).
Western Blotting
Human cardiac fibroblasts were sonicated in lysis buffer. After separation by SDS-PAGE and transfer to nitrocellulose membranes, signals were detected with specific corresponding antibodies, visualized with enhanced chemiluminescence, and quantified by densitometry (see detailed Methods in online data supplement).
Immunofluorescent Cytochemistry
Single and double-label fluorescent immunocytochemistry was performed on cells plated onto glass coverslips as described previously.23 The antibodies used were rabbit polyclonal anti-Nox4 (1:50 dilution) and mouse monoclonal anti-smooth muscle actin (clone 1A4, 1: 100 dilution, Sigma). Similar coverslips treated with secondary antibodies alone did not show specific staining.
Quantitative Real-Time Polymerase Chain Reaction
Quantification of human Nox 1, Nox 2, Nox 4, Nox 5, human smooth muscle specific- actin and 18S rRNA was performed by amplification of cDNA using the LightCycler real-time thermocycler as described previously23 (see detailed Methods in online data supplement).
Detection of Intracellular Superoxide Using High Performance Liquid Chromatography
To measure intracellular release superoxide, we uses the superoxide-specific fluorescent dihydrodethidium and quantified the superoxide specific signal by using high performance liquid chromatography as previously described24,25 (see detailed Methods in online data supplement).
Measurement of NADPH-Dependent Superoxide Production
Membrane samples from fibroblasts were prepared as described previously26 (see detailed Methods in online data supplement).
Transfection With Small Interference RNA
Human cardiac fibroblasts were trypsinized and plated on 100-mm dishes at 30% to 50% confluence 24 hours before transfection. To downregulate Nox4, Nox5, and small mothers against decapentaplegic (Smad) 2, scrambled small interference RNA (siScr) and small interference RNA (siRNA) against human Nox4 (siNox4, sense sequence: 5'-ACUGAGGUACAGCUGGAUGUU-3', anti-sense sequence:3' CAUCCAGCUGUACCUCAGUUU-3'), human Nox5 (siNox5: sense sequence 5'-GGUGGACUUUAUCUGGAUCTT-3', anti-sense sequence 5'-GAUCCAGAUAAAGUCCACCTT-3'), and Smad 2 (sense sequence: 5'-GUCCCAUGAAAAGACUUAATT-3', anti-sense sequence 5'-UUAAGUCUUUUCAUGGGACTT-3'), were purchased from Ambion. Individual siRNAs (at 25 to 50 nmol/L), oligofectamine and Opti-MEM were mixed and incubated at room temperature for 20 minutes. siRNA-oligofectamine complexes were added to cells for 24 hours, after which siRNA-oligofectamine complexes were removed and cells were washed and placed in serum free fibroblast growth medium (Cambrex) for 72 hours. Subsequently, cells were treated with or without TGF-1 (10 ng/mL) for the indicated time and harvested for RNA or protein extraction, immunocytochemistry, or superoxide assay. To control for possible nonspecific effects of siRNAs, multiple scrambled siRNAs were used, including sequence specific Nox4 scrambled (which has the same nucleotide content but in random order), siLuc (siRNA against luciferase), and the universal negative control siRNA commercially available from Ambion.
Statistical Analysis
Results are expressed as mean±standard error of mean (SEM). Statistical significance was assessed by using paired Student’s t test or 1-way or 2-way ANOVA with appropriate post-hoc analysis (Scheffe) to exclude possible interactions between various variables within subgroups using Origin software version 7.5. A value of P<0.05 was considered to be statistically significant.
Results
TGF-1 Upregulation of SM -Actin in Human Cardiac Fibroblasts Is ROS Sensitive
TGF-1 is the most potent stimulus for differentiation of fibroblasts into myofibroblasts. This phenotype is characterized by upregulation of SM -actin, which confers contractile behavior to myofibroblasts. To determine whether ROS are necessary for expression of SM -actin in response to TGF-1, we exposed cardiac fibroblasts to the ROS inhibitors diphenylene iodonium (DPI; an inhibitor of flavin-containing oxidoreductases such as NADPH oxidases), ebselen (a nonspecific H2O2 scavenger), or N-acetyl cysteine (NAC; a glutathione precursor and scavenger of H2O2) and then examined the expression of SM -actin at the RNA and protein levels after stimulation with TGF-1 (10 ng/mL) for 24 hours. As shown in Figure 1A, TGF-1 upregulated SM -actin 7.5±0.8 fold (n=4, P=0.001), and all ROS inhibitors potently abrogated the expression of SM -actin mRNA as quantified by real-time polymerase chain reaction (RT-PCR). This effect was also observed at the protein level after exposure to TGF-1 (Figure 1B). It has been previously shown that when fibroblasts differentiate into myofibroblasts, SM -actin is assembled into thicker, stronger stress fibers.3 The antioxidant DPI not only inhibited the expression of SM -actin, but also completely prevented the appearance of the myofibroblast phenotype at 72 hours in primary human cardiac fibroblasts isolated from normal hearts (Figure 1B, lower panel) and from failing human hearts (data not shown). The treatment with antioxidants inhibited SM -actin expression and organization in cells already converted to myofibroblasts by prolonged TGF-1 treatment (data not shown), suggesting that continuous ROS production is necessary for maintenance of the typical myofibroblast phenotype. These experimental data suggest that ROS released in response to TGF-1 are required for SM -actin expression and phenotypic conversion to myofibroblasts.
TGF-1 Stimulates NADPH Oxidase Activity in Human Cardiac Fibroblasts
NADPH oxidases are important sources of ROS in human fibroblasts.22 It is well established that TGF-1 activates NADPH oxidases and thereby stimulates ROS release from fetal lung fibroblasts.22 Therefore, we examined whether TGF-1 stimulates NADPH-driven superoxide production in human cardiac fibroblasts by performing a 24-hour time course. Indeed, TGF-1 induces NADPH-driven superoxide production from cardiac fibroblast membranes (Figure 2). After an initial stimulation at 1 hour (64±13%), the membrane-derived NADPH-driven superoxide production returns to baseline and increases again at 18 and 24 hours (2.4±0.4-fold, n=3, P=0.002). These results confirm that TGF-1 chronically increases NADPH-driven superoxide production from membranes of human cardiac fibroblasts.
TGF-1 Upregulates Nox4, but not Nox1, Nox2, or Nox5, in Human Cardiac Fibroblasts
Because TGF-1 induced an increase in NADPH-driven ROS production in cardiac fibroblasts that was sustained over 24 hours, we hypothesized that it is mediated by increased expression of a NADPH oxidase. To identify which Nox is upregulated, we stimulated cardiac fibroblasts with TGF-1 and quantified the expression of various Nox enzymes at the level of RNA or protein. At baseline, Nox4 and Nox 5 mRNA were abundantly expressed, whereas Nox 1 or Nox 2 (gp91phox) were expressed at very low levels (limit of detection of the assay, data not shown). After stimulation with TGF-1, Nox 4 mRNA was potently upregulated by 16.2±0.8-fold at 24 hours (n=3, P<0.005) starting as early as 2 hours, with a peak at 24 hours. In contrast, Nox 5 mRNA was progressively reduced by 75±13% at 24 hours (n=4, P<0.005; Figure 3A) whereas Nox 1 and Nox 2 were unaffected (data not shown). These data suggest that upregulation of Nox4 may be important for TGF-1 effects on cardiac fibroblasts.
If Nox4-mediated release of ROS is required for SM -actin expression, induction of Nox4 protein should precede the expression of SM -actin on stimulation with TGF-1. Therefore, we stimulated cells with TGF-1 and harvested them at various time points for Western blotting of Nox4 or SM -actin proteins. To identify Nox4 protein, we performed Western blotting using a rabbit polyclonal antibody. This antibody identifies an &112 to 120 KDa band that is specifically blocked by preincubation with the peptide against which this antibody was raised (supplemental Figure I). Furthermore, this band was specifically upregulated by TGF-1 and was inhibited by depleting fibroblasts of Nox4 using siRNA (see below). These experiments showed that after stimulation with TGF-1, Nox4 expression increases as early as 4 hours (2.1±0.1-fold), whereas SM -actin increase starts at 16 hours (1.88±0.3-fold). Maximum upregulation of both Nox4 and SM -actin occurred at 24 hours (3.9±0.2-fold for Nox4 and 3.4±0.8-fold, n=4, for SM -actin; Figure 3B). The fact that expression of Nox4 preceded the increase in SM -actin by 12 hours suggests that Nox4-derived ROS are important for upregulation of SM -actin after treatment with TGF-1.
Nox4 Mediates TGF-1 NADPH-Driven Superoxide Production and Is Required for the Expression of SM -Actin, Fibronectin, Collagen A1, and Connective Tissue Growth Factor
On the basis of these observations, we performed additional studies to determine whether Nox4 is necessary for TGF-1 expression of SM- actin. We downregulated the expression of Nox4 protein using transfection with small interference oligonucleotide RNA directed against Nox4 (siNox4). First, we confirmed that siNox4 is able to decrease Nox4 mRNA levels. On transfection with siNox4, both basal and TGF-1eCinduced expression of Nox4 mRNA were dramatically reduced (74±2% and 78±3% respectively; n=4; P<0.05; Figure 4B). Similarly, Nox4 protein was potently inhibited by siNox4 (Figure 4A). Transfection with fluorescein-labeled siScr showed that siRNA entered more than 95% of cells (supplemental Figure II). We also found that Nox5 mRNA was not inhibited by transfection with siNox4, demonstrating the specificity of siNox4 sequence (data not shown). To confirm the functional effect of siNox4, we measured NADPH-driven superoxide production from membranes of cells in which Nox4 was depleted. Transfection with siNox4 reduced basal and TGF-1eCstimulated superoxide production at 24 hours by 32±4% and 44±5%, respectively (n=3, P<0.01 versus siScr; Figure 4C), demonstrating that Nox4 participates in NADPH-driven superoxide production from fibroblast membranes both basally and on chronic stimulation with TGF-1. Furthermore, deletion of Nox4 and not Nox5 also inhibited intracellular superoxide in response to acute TGF-1 stimulation (30 minutes; supplemental Figure III). Because at 30 minutes the absolute TGF-1eCinduced superoxide production was blocked acutely by &50%, it is possible that other sources of superoxide may account for residual superoxide production.
To assess whether Nox4 is required for SM -actin expression, we quantified SM -actin at the RNA and protein levels after transfection of cells with siNox4. SM -actin mRNA was reduced by more than 90% by the siRNA against Nox4, whereas SM -actin protein was inhibited by 66±8% under basal conditions and by 75±7% after stimulation with TGF-1 (n=3, P<0.001 versus siScr; Figure 5A). Next, to determine whether Nox4 mediates the conversion to the myofibroblast phenotype, we performed immunofluorescent cytochemistry for Nox4 and SM -actin after transfection with siNox4. Transfection with siNox4 abolished the myofibroblast phenotype (data not shown). In contrast to Nox4, downregulation of Nox5 did not prevent TGF-1 stimulation of SM -actin, demonstrating that Nox4 but not Nox5 is required for this response (supplemental Figure IV).
Another important effect of TGF-1 is stimulation of production of extracellular matrix proteins such as fibronectin and collagen 1. Because it has been demonstrated both in vitro and in vivo that cytokine connective tissue growth factor (CTGF) production is potently increased in response to TGF-1 and in turn stimulates collagen secretion by the fibroblast,27 we tested the effect of Nox4 on these responses. We found that depletion of Nox4 inhibited basal and TGF-1eCinduced CTGF protein by 51±6% (n=3, P<0.01; Figure 5B), fibronectin production by 61±5% (n=3, P<0.003; Figure 5B), and collagen 1A1 mRNA upregulation by 47±7% (n=3, P<0.01; data not shown). Taken together, these results demonstrate that Nox4-derived ROS are critical for differentiation of cardiac fibroblasts in myofibroblasts and extracellular matrix protein production in response to TGF-1.
Nox4-Derived ROS Are Required for Chronic Smad 2/3 Activation in Response to TGF-1
Recent studies have demonstrated that Smads mediate TGF-1 induction of SM -actin in human lung fibroblasts.28 On phosphorylation by TGF-1 type I receptor, Smad2 and Smad3 become active, form a heterotrimeric complex with Smad4, and translocate to the nucleus, where the complex activates target gene transcription.29 To verify that Smad 2/3 phosphorylation mediates TGF-1-induction of SM -actin in our cells, we reduced Smad 2 levels by transfecting fibroblasts with siSmad2 and performed Western blotting for phospho-Smad 2/3, total Smad 2/3, and SM -actin. The siSmad2 effectively blocked not only phosphorylation of Smad 2/3 and total Smad 2 by >90%, but also SM -actin induction in response to TGF-1 at 24 hours (n=3; Figure 6A) and Nox4 mRNA at 4 hours (data not shown).
To identify a mechanism by which Nox4-derived ROS are important for Smad 2/3 activation in response to TGF-1, we first tested the ROS sensitivity of Smad activation. As previously reported, TGF-1 potently stimulated Smad 2/3 activation with a time course of activation that started as early as 5 minutes and was sustained up to 72 hours (data not shown). Pretreatment with DPI, ebselen, or NAC partially inhibited Smad 2/3 phosphorylation by 51±8%, 52±6%, and 71±5% (n=3, P=0.001 versus control; Figure 6B), respectively. To specifically test whether Nox4 is the source of ROS responsible for TGF-1eCinduced Smad 2/3 activation, we depleted cells of Nox4 using transfection with siNox4 and measured Smad 2/3 phosphorylation after 1 hour and 24 hours of stimulation with TGF-1. At 24 hours, Smad 2/3 phosphorylation was potently inhibited by 75±5% and 68±3% in basal and TGF-1 stimulated cardiac fibroblasts, respectively (n=7, P<0.0001). At 1 hour, siNox4 inhibited TGF-1eCinduced phosphorylation of Smad 2/3 by 52±2% (n=3, P<0.01), whereas depletion of Nox5 did not have any effect (supplemental Figure V). Together with the requirement of Smads for SM -actin expression, these data show that TGF-1 induces ROS production from Nox4 as early as 30 minutes that subsequently participate in the phosphorylation and activation of Smad 2/3, leading to the long-term induction of Nox4, SM -actin expression, and conversion to myofibroblasts.
Discussion
In the present study, we provide evidence that Nox4 is essential for differentiation of human cardiac fibroblasts into myofibroblasts in response to TGF-1. We demonstrate that (1) TGF-1 upregulates Nox4, ROS production, and SM -actin in human cardiac fibroblasts; (2) the change in SM -actin expression and the development of the myofibroblast phenotype in response to TGF-1 require Nox4-derived ROS; and (3) Nox4 modulates SM -actin expression by controlling long-term activation of Smad 2/3. These new findings support the notion that Nox4 is critical for modulation of contractile phenotype in response to TGF-1.
The excessive interstitial fibrosis from failing hearts is produced by the activated myofibroblast in response to regulatory signals such as angiotensin II, aldosterone, or stretch via paracrine release of TGF-1.3,30 Therefore, it has been hypothesized that the phenotypic conversion of fibroblasts into specialized myofibroblasts is a key process mediating cardiac fibrosis. In the present study, we used TGF-1 as a stimulus for myofibroblast differentiation because this cytokine is a key element that mediates the excessive fibrogenic reaction both in cardiac fibroblast culture and in animals and humans with heart failure. It is well known that TGF-1 stimulates NADPH oxidase activity and ROS release in human lung fibroblasts. The link between ROS and the acquisition of the myofibroblast phenotype was first suggested by Vozenin-Brotons et al,31 who showed that superoxide mediates the conversion of skin fibroblasts into myofibroblasts via paracrine release of TGF-1 in a skin model of wound healing. The source and mechanisms by which ROS mediate myofibroblast differentiation, however, had not been determined before our current study.
The NADPH oxidases are multisubunit enzymes originally discovered in neutrophils. The neutrophil oxidase consists of 5 subunits: 2 membrane-spanning components (the small subunit p22phox and the large catalytic subunit Nox2) and 3 cytosolic components (rac1, p67phox and p47phox).32 Similar oxidase systems have now been identified in nonphagocytic cells and have been shown to be the primary source of ROS that mediates angiotensin II-induced vascular myocyte hypertrophy (reviewed in Griendling et al19). The activation and structure of each cardiovascular NADPH oxidase is determined by the type of catalytic subunit (Nox homologue). For example, Nox1 and Nox2 require recruitment of the cytosolic factors p47phox and p67phox (or their homologues NoxO1 and NoxA1, respectively); Nox5 requires intracellular calcium release, and Nox4 appears to be intrinsically active and requires only p22phox and possibly Rac for activity.21 Using RT-PCR and Western blot analysis, we demonstrated that 2 catalytic subunits expressed in cardiac fibroblasts are regulated by TGF-1 (Nox4 is upregulated whereas Nox5 is downregulated). The fact that each cell contains multiple catalytic subunits suggests that each Nox is coupled with different cellular functions. For example, Nox1 is involved in mitogenic stimulation of vascular myocytes33 and Nox5 has been implicated in growth and apoptosis,34 whereas Nox4 has been associated with growth inhibition, because overexpression of Nox4 in NIH 3T3 cells causes senescence and growth arrest.35 Importantly, TGF-1 is also known to inhibit mitogenic growth by upregulating cell cycle inhibitors,36 which would suggest that TGF-1 and Nox4 participate in similar cellular processes.
NADPH oxidase subunit expression and activity are increased in various models of cardiac hypertrophy and heart failure.37,38 Recently, several studies have established a role for NADPH oxidases in angiotensin II-induced cardiac hypertrophy in vitro and in vivo.9,10,15,39 Very little is known, however, about the role of NADPH oxidases in fibroblast function. In the current study, we found that downregulation of Nox4 prevented myofibroblast formation, CTGF expression, and production of extracellular matrix proteins such as fibronectin and collagen 1A1. Because fibronectin, collagen, and CTGF are essential for extracellular matrix remodeling in response to TGF-1,27 the current data support the notion that Nox4 activation in fibroblast in response to TGF-1 is vital to development of cardiac fibrosis.
Other Nox proteins have also been implicated in cardiac fibrosis. For example, one study demonstrated that in vivo deletion of Nox2 reduced angiotensin II-induced cardiac fibrosis (subpressor concentration).9 Nox2 does not appear to mediate the cardiac fibroblast response, however, because transfection with siNox2 had no effect on SM -actin protein induction in response to TGF-1 (D.S., I.C., unpublished observations, 2004). The hypothesis that different Nox enzymes couple with different agonists is also supported by 2 other in vivo studies. Maytin et al15 and Byrne et al10 showed that Nox2 knockout did not prevent cardiac fibrosis or hypertrophy in a pressure-overload model of cardiac hypertrophy, although this model was shown to require ROS and resulted in increased cardiac levels of Nox4. Our observations, taken together with these studies, suggest that Nox4 may mediate TGF-1eCinduced fibrosis and pressure-overload (ie, stretch), whereas Nox2 is required for angiotensin II-induced cardiac fibrosis. Because we found that Nox2 was not involved in TGF-1 differentiation of human cardiac myofibroblasts, this may suggest that Nox4 substitutes for Nox2 in TGF-1eCinduced fibrosis in humans in contrast to murine models of cardiac fibrosis. The distribution of Nox subunits in mouse cardiac fibroblasts is currently unknown. Further studies are required to clarify whether Nox4 or Nox2 involvement in cardiac fibrosis is a function of agonist (angiotensin II versus TGF-1), species (human versus rodents), or model (in vitro versus in vivo).
Although it is well known that TGF-1 stimulates differentiation of fibroblasts into myofibroblasts, little is known about the signaling mechanisms that mediate this response. Classically, TGF-1 uses the Smad proteins to activate cellular functions. After ligand activation, the type II receptor associates with the type I receptor (TGF-1RI). Next, the TGF-1RI serine-threonine kinase phosphorylates receptor-Smads (R-Smads) Smad2 to form a complex with Smad3, which in turn associates with a Co-Smad (Smad4) and is translocated into the nucleus, where it acts as a transcription factor.29 After initial phosphorylation, Smad 2/3 activation in response to TGF-1 is prolonged (hours to days depending on the cell type). In our model, activation of Smad 2/3 occurs as early at 5 minutes and is maintained up to 72 hours after stimulation with TGF-1 (data not shown). In our cells, this long-term activation of Smad 2/3 mediates the chronic TGF-1 upregulation of SM -actin. Interestingly, downregulation of Nox4 leads to a marked inhibition of Smad2/3 phosphorylation at 24 hours. These data imply that Nox4 mediates SM -actin upregulation in response to TGF-1 by stimulating prolonged phosphorylation and activation of Smads 2/3.
In summary, we have identified a novel role for Nox4 as essential mediator of Smad2/3 transcription factor activation in response to TGF-1 and demonstrated its significance in differentiation of fibroblasts into myofibroblasts. This study identifies for the first time a role for Nox4 in acquisition and maintenance of contractile phenotype via modulation of transcription factors Smad2/3 activation. Because of the critical role of TGF-1eCinduced myofibroblast activation in cardiac fibrosis, these data provide insight into novel mechanisms with potential therapeutic implications for heart failure. Further studies are necessary to determine whether Nox4-derived ROS are also involved in vivo in models of cardiac fibrosis.
Acknowledgments
This work was supported by an American Heart Association Scientist Development Grant to D.S. and by a National Institutes of Health cardiovascular training grant to I.C. We are indebted to Drs Bernard Lasseegue, Kathy Griendling, David G. Harrison, and A. Maziar Zafari for their critical reading of the manuscript, and Marcie Burnham and Tranise Coryell for excellent secretarial assistance.
References
American Heart Association. Heart Disease and Stroke Statistics-2005 Update. Dallas, Tex: American Heart Association; 2005.
Sun Y, Kiani MF, Postlethwaite AE, Weber KT. Infarct scar as living tissue. Basic Res Cardiol. 2002; 97: 343eC347.
Gabbiani G. Evolution and clinical implications of the myofibroblast concept. Cardiovasc Res. 1998; 38: 545eC548.
Rosenkranz S. TGF-beta1 and angiotensin networking in cardiac remodeling. Cardiovasc Res. 2004; 63: 423eC432.
Kuwahara F, Kai H, Tokuda K, Kai M, Takeshita A, Egashira K, Imaizumi T. Transforming growth factor-beta function blocking prevents myocardial fibrosis and diastolic dysfunction in pressure-overloaded rats. Circulation. 2002; 106: 130eC135.
Sawyer DB, Colucci WS. Mitochondrial oxidative stress in heart failure: "oxygen wastage" revisited. Circ Res. 2000; 86: 119eC120.
Keith M, Geranmayegan A, Sole MJ, Kurian R, Robinson A, Omran AS, Jeejeebhoy KN. Increased oxidative stress in patients with congestive heart failure. J Am Coll Cardiol. 1998; 31: 1352eC1356.
Mallat Z, Philip I, Lebret M, Chatel D, Maclouf J, Tedgui A. Elevated levels of 8-iso-prostaglandin F2alpha in pericardial fluid of patients with heart failure: a potential role for in vivo oxidant stress in ventricular dilatation and progression to heart failure. Circulation. 1998; 97: 1536eC1539.
Bendall JK, Cave AC, Heymes C, Gall N, Shah AM. Pivotal role of a gp91(phox)-containing NADPH oxidase in angiotensin II-induced cardiac hypertrophy in mice. Circulation. 2002; 105: 293eC296.
Byrne JA, Grieve DJ, Bendall JK, Li JM, Gove C, Lambeth JD, Cave AC, Shah AM. Contrasting roles of NADPH oxidase isoforms in pressure-overload versus angiotensin II-induced cardiac hypertrophy. Circ Res. 2003; 93: 802eC805.
Sun Y, Zhang J, Lu L, Chen SS, Quinn MT, Weber KT. Aldosterone-induced inflammation in the rat heart: role of oxidative stress. Am J Pathol. 2002; 161: 1773eC1781.
Kinugawa S, Tsutsui H, Hayashidani S, Ide T, Suematsu N, Satoh S, Utsumi H, Takeshita A. Treatment with dimethylthiourea prevents left ventricular remodeling and failure after experimental myocardial infarction in mice: role of oxidative stress. Circ Res. 2000; 87: 392eC398.
Ide T, Tsutsui H, Kinugawa S, Suematsu N, Hayashidani S, Ichikawa K, Utsumi H, Machida Y, Egashira K, Takeshita A. Direct evidence for increased hydroxyl radicals originating from superoxide in the failing myocardium. Circ Res. 2000; 86: 152eC157.
Ide T, Tsutsui H, Kinugawa S, Utsumi H, Kang D, Hattori N, Uchida K, Arimura K, Egashira K, Takeshita A. Mitochondrial electron transport complex I is a potential source of oxygen free radicals in the failing myocardium. Circ Res. 1999; 85: 357eC363.
Maytin M, Siwik DA, Ito M, Xiao L, Sawyer DB, Liao R, Colucci WS. Pressure-overload-induced myocardial hypertrophy in mice does not require gp91phox. Circulation. 2004; 109: 1168eC1171.
Li Y, Huang TT, Carlson EJ, Melov S, Ursell PC, Olson JL, Noble LJ, Yoshimura MP, Berger C, Chan PH, Wallace DC, Epstein CJ. Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat Genet. 1995; 11: 376eC381.
Siwik DA, Pagano PJ, Colucci WS. Oxidative stress regulates collagen synthesis and matrix metalloproteinase activity in cardiac fibroblasts. Am J Physiol Cell Physiol. 2001; 280: C53eCC60.
Li PF, Dietz R, von Harsdorf R. Superoxide induces apoptosis in cardiomyocytes, but proliferation and expression of transforming growth factor-beta1 in cardiac fibroblasts. FEBS Lett. 1999; 448: 206eC210.
Griendling KK, Sorescu D, Ushio-Fukai M. NAD(P)H oxidase: role in cardiovascular biology and disease. Circ Res. 2000; 86: 494eC501.
Lambeth JD. Nox/Duox family of nicotinamide adenine dinucleotide (phosphate) oxidases. Curr Opin Hematol. 2002; 9: 11eC17.
Krause KH. Tissue distribution and putative physiological function of NOX family NADPH oxidases. Jpn J Infect Dis. Oct 2004; 57: S28eCS29.
Thannickal VJ, Fanburg BL. Activation of an H2O2-generating NADH oxidase in human lung fibroblasts by transforming growth factor b1. J Biol Chem. 1995; 270: 30334eC30338.
Sorescu D, Weiss D, Lassegue B, Clempus RE, Szocs K, Sorescu GP, Valppu L, Quinn MT, Lambeth JD, Vega JD, Taylor WR, Griendling KK. Superoxide production and expression of nox family proteins in human atherosclerosis. Circulation. 2002; 105: 1429eC1435.
Zhao H, Joseph J, Fales HM, Sokoloski EA, Levine RL, Vasquez-Vivar J, Kalyanaraman B. Detection and characterization of the product of hydroethidine and intracellular superoxide by HPLC and limitations of fluorescence. Proc Natl Acad Sci U S A. 2005; 102: 5727eC5732.
Fink B, Laude K, McCann L, Doughan A, Harrison DG, Dikalov S. Detection of intracellular superoxide formation in endothelial cells and intact tissues using dihydroethidium and an HPLC-based assay. Am J Physiol Cell Physiol. 2004; 287: C895eCC902.
Sorescu D, Somers MJ, Lasseegue B, Grant S, Harrison DG, Griendling KK. Electron spin resonance characterization of the NAD(P)H oxidase in vascular smooth muscle cells. Free Radic Biol Med. 2001; 30: 603eC612.
Matsui Y, Sadoshima J. Rapid upregulation of CTGF in cardiac myocytes by hypertrophic stimuli: implication for cardiac fibrosis and hypertrophy. J Mol Cell Cardiol. 2004; 37: 477eC481.
Evans RA, Tian YC, Steadman R, Phillips AO. TGF-beta1-mediated fibroblast-myofibroblast terminal differentiation-the role of Smad proteins. Exp Cell Res. 2003; 282: 90eC100.
Massague J. How cells read TGF-beta signals. Nat Rev Mol Cell Biol. 2000; 1: 169eC178.
Sun Y, Weber KT. Angiotensin converting enzyme and myofibroblasts during tissue repair in the rat heart. J Mol Cell Cardiol. 1996; 28: 851eC858.
Vozenin-Brotons MC, Sivan V, Gault N, Renard C, Geffrotin C, Delanian S, Lefaix JL, Martin M. Antifibrotic action of Cu/Zn SOD is mediated by TGF-beta1 repression and phenotypic reversion of myofibroblasts. Free Radic Biol Med. 2001; 30: 30eC42.
Cross AR, Parkinson JF, Jones OTG. The superoxide-generating oxidase of leucocytes. Biochem J. 1984; 223: 337eC344.
Suh Y, Arnold RS, Lasseegue B, Shi J, Xu X, Sorescu D, Chung AB, Griendling KK, Lambeth JD. Cell transformation by the superoxide-generating oxidase mox1. Nature. 1999; 401: 79eC82.
Brar SS, Corbin Z, Kennedy TP, Hemendinger R, Thornton L, Bommarius B, Arnold RS, Whorton AR, Sturrock AB, Huecksteadt TP, Quinn MT, Krenitsky K, Ardie KG, Lambeth JD, Hoidal JR. NOX5 NAD(P)H oxidase regulates growth and apoptosis in DU 145 prostate cancer cells. Am J Physiol Cell Physiol. 2003; 285: C353eCC369.
Geiszt M, Kopp JB, Varnai P, Leto TL. Identification of renox, an NAD(P)H oxidase in kidney. Proc Natl Acad Sci U S A. 2000; 97: 8010eC8014.
Reynisdottir I, Polyak K, Iavarone A, Massague J. Kip/Cip and Ink4 Cdk inhibitors cooperate to induce cell cycle arrest in response to TGF-beta. Genes Dev. 1995; 9: 1831eC1845.
Li JM, Gall NP, Grieve DJ, Chen M, Shah AM. Activation of NADPH oxidase during progression of cardiac hypertrophy to failure. Hypertension. 2002; 40: 477eC484.
Heymes C, Bendall JK, Ratajczak P, Cave AC, Samuel JL, Hasenfuss G, Shah AM. Increased myocardial NADPH oxidase activity in human heart failure. J Am Coll Cardiol. 2003; 41: 2164eC2171.
Nakagami H, Takemoto M, Liao JK. NADPH oxidase-derived superoxide anion mediates angiotensin II-induced cardiac hypertrophy. J Mol Cell Cardiol. 2003; 35: 851eC859.(Ioan Cucoranu, Roza Clemp)
Abstract
Human cardiac fibroblasts are the main source of cardiac fibrosis associated with cardiac hypertrophy and heart failure. Transforming growth factor-1 (TGF-1) irreversibly converts fibroblasts into pathological myofibroblasts, which express smooth muscle -actin (SM -actin) de novo and produce extracellular matrix. We hypothesized that TGF-1eCstimulated conversion of fibroblasts to myofibroblasts requires reactive oxygen species derived from NAD(P)H oxidases (Nox). We found that TGF-1 potently upregulates the contractile marker SM -actin mRNA (7.5±0.8-fold versus control). To determine whether Nox enzymes are involved, we first performed quantitative real time polymerase chain reaction and found that Nox5 and Nox4 are abundantly expressed in cardiac fibroblasts, whereas Nox1 and Nox2 are barely detectable. On stimulation with TGF-1, Nox4 mRNA is dramatically upregulated by 16.2±0.8-fold (n=3, P<0.005), whereas Nox5 is downregulated. Small interference RNA against Nox4 downregulates Nox4 mRNA by 80±5%, inhibits NADPH-driven superoxide production in response to TGF-1 by 65±7%, and reduces TGF-1eCinduced expression of SM -actin by 95±2% (n=6, P<0.05). Because activation of small mothers against decapentaplegic (Smads) 2/3 is critical for myofibroblast conversion in response to TGF-1, we also determined whether Nox4 affects Smad 2/3 phosphorylation. Depletion of Nox4 but not Nox5 inhibits baseline and TGF-1 stimulation of Smad 2/3 phosphorylation by 75±5% and 68±3%, respectively (n=7, P<0.0001). We conclude that Nox 4 mediates TGF-1eCinduced conversion of fibroblasts to myofibroblasts by regulating Smad 2/3 activation. Thus, Nox4 may play a critical role in the pathological activation of cardiac fibroblasts in cardiac fibrosis associated with human heart failure.
Key Words: Nox4 human cardiac fibroblasts transforming growth factor reactive oxygen species Smad 2/3
Introduction
Heart failure remains the leading cause of hospital admissions in the United States, with more than 550 000 new patients diagnosed each year.1 Regardless of etiology, cardiac fibrosis is a major contributor to cardiac remodeling associated with cardiomyopathies. It is characterized by expansion of the interstitial compartment due to increased deposition of extracellular matrix by activated myofibroblasts.2 Cardiac myofibroblasts are specialized contractile fibroblasts formed by irreversible acquisition of contractile proteins such as smooth muscle -actin (SM -actin) in response to potent fibrogenic cytokines.3 The expression of SM -actin is regulated by transforming growth factor-1 (TGF-1), a primary fibrogenic growth factor in heart failure that is downstream of many of the pro-fibrotic actions of other fibroblast growth factors, such as angiotensin II, aldosterone, and norepinephrine.4 TGF-1 is upregulated in failing human hearts and various experimental models of cardiac hypertrophy,4 and functional blockade of TGF-1 prevents cardiac interstitial fibrosis induced by pressure overload in rats.5
There is increasing evidence that oxidative stress plays a critical role in the development and progression of cardiac remodeling associated with heart failure.6 Oxidative stress is increased in human heart failure7,8 and animal models of cardiac hypertrophy and fibrosis (induced by angiotensin II,9,10 aldosterone,11 myocardial infarction,12 tachycardia-induced cardiomyopathy,13,14 and aortic coarctation10,15). The first evidence that elevated oxidative stress can cause cardiomyopathy was provided by Li et al,16 who showed that deletion of superoxide dismutase (a main scavenger of intracellular superoxide) caused dilated cardiomyopathy and cardiac fibrosis in mice.16 Reactive oxygen species (ROS) modulate extracellular matrix remodeling by mediating cardiac fibroblast function and also by stimulating collagen turnover via activation of matrix metalloproteinases, enzymes critical for extracellular matrix remodeling.17 ROS also stimulate release and activation of cytokines. For example, in cell culture, exposure of cardiac fibroblasts to superoxide stimulates their proliferation by increasing the production of TGF-1.18
NADPH oxidases (Nox) are multisubunit enzymes that generate superoxide by transferring electrons from NADPH to molecular oxygen.19 Although the prototype catalytic subunit gp91phox (Nox2) was originally discovered in neutrophils, there are also multiple homologues in nonphagocytic cells with wide range tissue distribution: Nox1, Nox3, Nox4, Nox5, and the dual oxidases Duox1 and Duox2.20 In phagocytes, ROS originating from Nox participate in bacterial killing, whereas in nonphagocytic cells they are required for cellular responses to cytokines and growth factors.21 Fibroblasts also produce ROS via Nox in response to growth factors. For example, lung fibroblasts endogenously release ROS from Nox in response to TGF-1.22 An important consequence of fibroblast stimulation with TGF-1 is a phenotypical conversion into myofibroblasts. Although NADPH oxidases are activated by TGF-1, it is unknown whether their activation is required for upregulation of SM -actin, the key element that defines the acquisition of contractile phenotype by myofibroblasts.
In the current study, we sought to (1) determine whether ROS derived from NADPH oxidases are important for TGF-1eCinduced myofibroblast differentiation, (2) identify the catalytic Nox subunit that is required for SM -actin expression in response to TGF-1, and (3) discover the mechanism by which Nox-derived ROS mediate the conversion to the myofibroblast phenotype. These data provide new insights into the mechanisms underlying the pathological activation of fibroblasts in cardiac fibrosis induced by TGF-1.
Materials and Methods
Cell Culture
Primary human cardiac fibroblast cells (catalog number ACBRI 5118) were purchased at passage 2 from Cell Systems (Kirkland, WA) and grown in fibroblast growth medium-2 (Cambrex, Baltimore, Md). All cells were harvested at passage 3 to 6 and used at 50% to 70% confluence for transfection experiments and 80% to 90% confluence for time course experiments (see online data supplement at http://circres.ahajournals.org for characterization of human cardiac fibroblasts).
Western Blotting
Human cardiac fibroblasts were sonicated in lysis buffer. After separation by SDS-PAGE and transfer to nitrocellulose membranes, signals were detected with specific corresponding antibodies, visualized with enhanced chemiluminescence, and quantified by densitometry (see detailed Methods in online data supplement).
Immunofluorescent Cytochemistry
Single and double-label fluorescent immunocytochemistry was performed on cells plated onto glass coverslips as described previously.23 The antibodies used were rabbit polyclonal anti-Nox4 (1:50 dilution) and mouse monoclonal anti-smooth muscle actin (clone 1A4, 1: 100 dilution, Sigma). Similar coverslips treated with secondary antibodies alone did not show specific staining.
Quantitative Real-Time Polymerase Chain Reaction
Quantification of human Nox 1, Nox 2, Nox 4, Nox 5, human smooth muscle specific- actin and 18S rRNA was performed by amplification of cDNA using the LightCycler real-time thermocycler as described previously23 (see detailed Methods in online data supplement).
Detection of Intracellular Superoxide Using High Performance Liquid Chromatography
To measure intracellular release superoxide, we uses the superoxide-specific fluorescent dihydrodethidium and quantified the superoxide specific signal by using high performance liquid chromatography as previously described24,25 (see detailed Methods in online data supplement).
Measurement of NADPH-Dependent Superoxide Production
Membrane samples from fibroblasts were prepared as described previously26 (see detailed Methods in online data supplement).
Transfection With Small Interference RNA
Human cardiac fibroblasts were trypsinized and plated on 100-mm dishes at 30% to 50% confluence 24 hours before transfection. To downregulate Nox4, Nox5, and small mothers against decapentaplegic (Smad) 2, scrambled small interference RNA (siScr) and small interference RNA (siRNA) against human Nox4 (siNox4, sense sequence: 5'-ACUGAGGUACAGCUGGAUGUU-3', anti-sense sequence:3' CAUCCAGCUGUACCUCAGUUU-3'), human Nox5 (siNox5: sense sequence 5'-GGUGGACUUUAUCUGGAUCTT-3', anti-sense sequence 5'-GAUCCAGAUAAAGUCCACCTT-3'), and Smad 2 (sense sequence: 5'-GUCCCAUGAAAAGACUUAATT-3', anti-sense sequence 5'-UUAAGUCUUUUCAUGGGACTT-3'), were purchased from Ambion. Individual siRNAs (at 25 to 50 nmol/L), oligofectamine and Opti-MEM were mixed and incubated at room temperature for 20 minutes. siRNA-oligofectamine complexes were added to cells for 24 hours, after which siRNA-oligofectamine complexes were removed and cells were washed and placed in serum free fibroblast growth medium (Cambrex) for 72 hours. Subsequently, cells were treated with or without TGF-1 (10 ng/mL) for the indicated time and harvested for RNA or protein extraction, immunocytochemistry, or superoxide assay. To control for possible nonspecific effects of siRNAs, multiple scrambled siRNAs were used, including sequence specific Nox4 scrambled (which has the same nucleotide content but in random order), siLuc (siRNA against luciferase), and the universal negative control siRNA commercially available from Ambion.
Statistical Analysis
Results are expressed as mean±standard error of mean (SEM). Statistical significance was assessed by using paired Student’s t test or 1-way or 2-way ANOVA with appropriate post-hoc analysis (Scheffe) to exclude possible interactions between various variables within subgroups using Origin software version 7.5. A value of P<0.05 was considered to be statistically significant.
Results
TGF-1 Upregulation of SM -Actin in Human Cardiac Fibroblasts Is ROS Sensitive
TGF-1 is the most potent stimulus for differentiation of fibroblasts into myofibroblasts. This phenotype is characterized by upregulation of SM -actin, which confers contractile behavior to myofibroblasts. To determine whether ROS are necessary for expression of SM -actin in response to TGF-1, we exposed cardiac fibroblasts to the ROS inhibitors diphenylene iodonium (DPI; an inhibitor of flavin-containing oxidoreductases such as NADPH oxidases), ebselen (a nonspecific H2O2 scavenger), or N-acetyl cysteine (NAC; a glutathione precursor and scavenger of H2O2) and then examined the expression of SM -actin at the RNA and protein levels after stimulation with TGF-1 (10 ng/mL) for 24 hours. As shown in Figure 1A, TGF-1 upregulated SM -actin 7.5±0.8 fold (n=4, P=0.001), and all ROS inhibitors potently abrogated the expression of SM -actin mRNA as quantified by real-time polymerase chain reaction (RT-PCR). This effect was also observed at the protein level after exposure to TGF-1 (Figure 1B). It has been previously shown that when fibroblasts differentiate into myofibroblasts, SM -actin is assembled into thicker, stronger stress fibers.3 The antioxidant DPI not only inhibited the expression of SM -actin, but also completely prevented the appearance of the myofibroblast phenotype at 72 hours in primary human cardiac fibroblasts isolated from normal hearts (Figure 1B, lower panel) and from failing human hearts (data not shown). The treatment with antioxidants inhibited SM -actin expression and organization in cells already converted to myofibroblasts by prolonged TGF-1 treatment (data not shown), suggesting that continuous ROS production is necessary for maintenance of the typical myofibroblast phenotype. These experimental data suggest that ROS released in response to TGF-1 are required for SM -actin expression and phenotypic conversion to myofibroblasts.
TGF-1 Stimulates NADPH Oxidase Activity in Human Cardiac Fibroblasts
NADPH oxidases are important sources of ROS in human fibroblasts.22 It is well established that TGF-1 activates NADPH oxidases and thereby stimulates ROS release from fetal lung fibroblasts.22 Therefore, we examined whether TGF-1 stimulates NADPH-driven superoxide production in human cardiac fibroblasts by performing a 24-hour time course. Indeed, TGF-1 induces NADPH-driven superoxide production from cardiac fibroblast membranes (Figure 2). After an initial stimulation at 1 hour (64±13%), the membrane-derived NADPH-driven superoxide production returns to baseline and increases again at 18 and 24 hours (2.4±0.4-fold, n=3, P=0.002). These results confirm that TGF-1 chronically increases NADPH-driven superoxide production from membranes of human cardiac fibroblasts.
TGF-1 Upregulates Nox4, but not Nox1, Nox2, or Nox5, in Human Cardiac Fibroblasts
Because TGF-1 induced an increase in NADPH-driven ROS production in cardiac fibroblasts that was sustained over 24 hours, we hypothesized that it is mediated by increased expression of a NADPH oxidase. To identify which Nox is upregulated, we stimulated cardiac fibroblasts with TGF-1 and quantified the expression of various Nox enzymes at the level of RNA or protein. At baseline, Nox4 and Nox 5 mRNA were abundantly expressed, whereas Nox 1 or Nox 2 (gp91phox) were expressed at very low levels (limit of detection of the assay, data not shown). After stimulation with TGF-1, Nox 4 mRNA was potently upregulated by 16.2±0.8-fold at 24 hours (n=3, P<0.005) starting as early as 2 hours, with a peak at 24 hours. In contrast, Nox 5 mRNA was progressively reduced by 75±13% at 24 hours (n=4, P<0.005; Figure 3A) whereas Nox 1 and Nox 2 were unaffected (data not shown). These data suggest that upregulation of Nox4 may be important for TGF-1 effects on cardiac fibroblasts.
If Nox4-mediated release of ROS is required for SM -actin expression, induction of Nox4 protein should precede the expression of SM -actin on stimulation with TGF-1. Therefore, we stimulated cells with TGF-1 and harvested them at various time points for Western blotting of Nox4 or SM -actin proteins. To identify Nox4 protein, we performed Western blotting using a rabbit polyclonal antibody. This antibody identifies an &112 to 120 KDa band that is specifically blocked by preincubation with the peptide against which this antibody was raised (supplemental Figure I). Furthermore, this band was specifically upregulated by TGF-1 and was inhibited by depleting fibroblasts of Nox4 using siRNA (see below). These experiments showed that after stimulation with TGF-1, Nox4 expression increases as early as 4 hours (2.1±0.1-fold), whereas SM -actin increase starts at 16 hours (1.88±0.3-fold). Maximum upregulation of both Nox4 and SM -actin occurred at 24 hours (3.9±0.2-fold for Nox4 and 3.4±0.8-fold, n=4, for SM -actin; Figure 3B). The fact that expression of Nox4 preceded the increase in SM -actin by 12 hours suggests that Nox4-derived ROS are important for upregulation of SM -actin after treatment with TGF-1.
Nox4 Mediates TGF-1 NADPH-Driven Superoxide Production and Is Required for the Expression of SM -Actin, Fibronectin, Collagen A1, and Connective Tissue Growth Factor
On the basis of these observations, we performed additional studies to determine whether Nox4 is necessary for TGF-1 expression of SM- actin. We downregulated the expression of Nox4 protein using transfection with small interference oligonucleotide RNA directed against Nox4 (siNox4). First, we confirmed that siNox4 is able to decrease Nox4 mRNA levels. On transfection with siNox4, both basal and TGF-1eCinduced expression of Nox4 mRNA were dramatically reduced (74±2% and 78±3% respectively; n=4; P<0.05; Figure 4B). Similarly, Nox4 protein was potently inhibited by siNox4 (Figure 4A). Transfection with fluorescein-labeled siScr showed that siRNA entered more than 95% of cells (supplemental Figure II). We also found that Nox5 mRNA was not inhibited by transfection with siNox4, demonstrating the specificity of siNox4 sequence (data not shown). To confirm the functional effect of siNox4, we measured NADPH-driven superoxide production from membranes of cells in which Nox4 was depleted. Transfection with siNox4 reduced basal and TGF-1eCstimulated superoxide production at 24 hours by 32±4% and 44±5%, respectively (n=3, P<0.01 versus siScr; Figure 4C), demonstrating that Nox4 participates in NADPH-driven superoxide production from fibroblast membranes both basally and on chronic stimulation with TGF-1. Furthermore, deletion of Nox4 and not Nox5 also inhibited intracellular superoxide in response to acute TGF-1 stimulation (30 minutes; supplemental Figure III). Because at 30 minutes the absolute TGF-1eCinduced superoxide production was blocked acutely by &50%, it is possible that other sources of superoxide may account for residual superoxide production.
To assess whether Nox4 is required for SM -actin expression, we quantified SM -actin at the RNA and protein levels after transfection of cells with siNox4. SM -actin mRNA was reduced by more than 90% by the siRNA against Nox4, whereas SM -actin protein was inhibited by 66±8% under basal conditions and by 75±7% after stimulation with TGF-1 (n=3, P<0.001 versus siScr; Figure 5A). Next, to determine whether Nox4 mediates the conversion to the myofibroblast phenotype, we performed immunofluorescent cytochemistry for Nox4 and SM -actin after transfection with siNox4. Transfection with siNox4 abolished the myofibroblast phenotype (data not shown). In contrast to Nox4, downregulation of Nox5 did not prevent TGF-1 stimulation of SM -actin, demonstrating that Nox4 but not Nox5 is required for this response (supplemental Figure IV).
Another important effect of TGF-1 is stimulation of production of extracellular matrix proteins such as fibronectin and collagen 1. Because it has been demonstrated both in vitro and in vivo that cytokine connective tissue growth factor (CTGF) production is potently increased in response to TGF-1 and in turn stimulates collagen secretion by the fibroblast,27 we tested the effect of Nox4 on these responses. We found that depletion of Nox4 inhibited basal and TGF-1eCinduced CTGF protein by 51±6% (n=3, P<0.01; Figure 5B), fibronectin production by 61±5% (n=3, P<0.003; Figure 5B), and collagen 1A1 mRNA upregulation by 47±7% (n=3, P<0.01; data not shown). Taken together, these results demonstrate that Nox4-derived ROS are critical for differentiation of cardiac fibroblasts in myofibroblasts and extracellular matrix protein production in response to TGF-1.
Nox4-Derived ROS Are Required for Chronic Smad 2/3 Activation in Response to TGF-1
Recent studies have demonstrated that Smads mediate TGF-1 induction of SM -actin in human lung fibroblasts.28 On phosphorylation by TGF-1 type I receptor, Smad2 and Smad3 become active, form a heterotrimeric complex with Smad4, and translocate to the nucleus, where the complex activates target gene transcription.29 To verify that Smad 2/3 phosphorylation mediates TGF-1-induction of SM -actin in our cells, we reduced Smad 2 levels by transfecting fibroblasts with siSmad2 and performed Western blotting for phospho-Smad 2/3, total Smad 2/3, and SM -actin. The siSmad2 effectively blocked not only phosphorylation of Smad 2/3 and total Smad 2 by >90%, but also SM -actin induction in response to TGF-1 at 24 hours (n=3; Figure 6A) and Nox4 mRNA at 4 hours (data not shown).
To identify a mechanism by which Nox4-derived ROS are important for Smad 2/3 activation in response to TGF-1, we first tested the ROS sensitivity of Smad activation. As previously reported, TGF-1 potently stimulated Smad 2/3 activation with a time course of activation that started as early as 5 minutes and was sustained up to 72 hours (data not shown). Pretreatment with DPI, ebselen, or NAC partially inhibited Smad 2/3 phosphorylation by 51±8%, 52±6%, and 71±5% (n=3, P=0.001 versus control; Figure 6B), respectively. To specifically test whether Nox4 is the source of ROS responsible for TGF-1eCinduced Smad 2/3 activation, we depleted cells of Nox4 using transfection with siNox4 and measured Smad 2/3 phosphorylation after 1 hour and 24 hours of stimulation with TGF-1. At 24 hours, Smad 2/3 phosphorylation was potently inhibited by 75±5% and 68±3% in basal and TGF-1 stimulated cardiac fibroblasts, respectively (n=7, P<0.0001). At 1 hour, siNox4 inhibited TGF-1eCinduced phosphorylation of Smad 2/3 by 52±2% (n=3, P<0.01), whereas depletion of Nox5 did not have any effect (supplemental Figure V). Together with the requirement of Smads for SM -actin expression, these data show that TGF-1 induces ROS production from Nox4 as early as 30 minutes that subsequently participate in the phosphorylation and activation of Smad 2/3, leading to the long-term induction of Nox4, SM -actin expression, and conversion to myofibroblasts.
Discussion
In the present study, we provide evidence that Nox4 is essential for differentiation of human cardiac fibroblasts into myofibroblasts in response to TGF-1. We demonstrate that (1) TGF-1 upregulates Nox4, ROS production, and SM -actin in human cardiac fibroblasts; (2) the change in SM -actin expression and the development of the myofibroblast phenotype in response to TGF-1 require Nox4-derived ROS; and (3) Nox4 modulates SM -actin expression by controlling long-term activation of Smad 2/3. These new findings support the notion that Nox4 is critical for modulation of contractile phenotype in response to TGF-1.
The excessive interstitial fibrosis from failing hearts is produced by the activated myofibroblast in response to regulatory signals such as angiotensin II, aldosterone, or stretch via paracrine release of TGF-1.3,30 Therefore, it has been hypothesized that the phenotypic conversion of fibroblasts into specialized myofibroblasts is a key process mediating cardiac fibrosis. In the present study, we used TGF-1 as a stimulus for myofibroblast differentiation because this cytokine is a key element that mediates the excessive fibrogenic reaction both in cardiac fibroblast culture and in animals and humans with heart failure. It is well known that TGF-1 stimulates NADPH oxidase activity and ROS release in human lung fibroblasts. The link between ROS and the acquisition of the myofibroblast phenotype was first suggested by Vozenin-Brotons et al,31 who showed that superoxide mediates the conversion of skin fibroblasts into myofibroblasts via paracrine release of TGF-1 in a skin model of wound healing. The source and mechanisms by which ROS mediate myofibroblast differentiation, however, had not been determined before our current study.
The NADPH oxidases are multisubunit enzymes originally discovered in neutrophils. The neutrophil oxidase consists of 5 subunits: 2 membrane-spanning components (the small subunit p22phox and the large catalytic subunit Nox2) and 3 cytosolic components (rac1, p67phox and p47phox).32 Similar oxidase systems have now been identified in nonphagocytic cells and have been shown to be the primary source of ROS that mediates angiotensin II-induced vascular myocyte hypertrophy (reviewed in Griendling et al19). The activation and structure of each cardiovascular NADPH oxidase is determined by the type of catalytic subunit (Nox homologue). For example, Nox1 and Nox2 require recruitment of the cytosolic factors p47phox and p67phox (or their homologues NoxO1 and NoxA1, respectively); Nox5 requires intracellular calcium release, and Nox4 appears to be intrinsically active and requires only p22phox and possibly Rac for activity.21 Using RT-PCR and Western blot analysis, we demonstrated that 2 catalytic subunits expressed in cardiac fibroblasts are regulated by TGF-1 (Nox4 is upregulated whereas Nox5 is downregulated). The fact that each cell contains multiple catalytic subunits suggests that each Nox is coupled with different cellular functions. For example, Nox1 is involved in mitogenic stimulation of vascular myocytes33 and Nox5 has been implicated in growth and apoptosis,34 whereas Nox4 has been associated with growth inhibition, because overexpression of Nox4 in NIH 3T3 cells causes senescence and growth arrest.35 Importantly, TGF-1 is also known to inhibit mitogenic growth by upregulating cell cycle inhibitors,36 which would suggest that TGF-1 and Nox4 participate in similar cellular processes.
NADPH oxidase subunit expression and activity are increased in various models of cardiac hypertrophy and heart failure.37,38 Recently, several studies have established a role for NADPH oxidases in angiotensin II-induced cardiac hypertrophy in vitro and in vivo.9,10,15,39 Very little is known, however, about the role of NADPH oxidases in fibroblast function. In the current study, we found that downregulation of Nox4 prevented myofibroblast formation, CTGF expression, and production of extracellular matrix proteins such as fibronectin and collagen 1A1. Because fibronectin, collagen, and CTGF are essential for extracellular matrix remodeling in response to TGF-1,27 the current data support the notion that Nox4 activation in fibroblast in response to TGF-1 is vital to development of cardiac fibrosis.
Other Nox proteins have also been implicated in cardiac fibrosis. For example, one study demonstrated that in vivo deletion of Nox2 reduced angiotensin II-induced cardiac fibrosis (subpressor concentration).9 Nox2 does not appear to mediate the cardiac fibroblast response, however, because transfection with siNox2 had no effect on SM -actin protein induction in response to TGF-1 (D.S., I.C., unpublished observations, 2004). The hypothesis that different Nox enzymes couple with different agonists is also supported by 2 other in vivo studies. Maytin et al15 and Byrne et al10 showed that Nox2 knockout did not prevent cardiac fibrosis or hypertrophy in a pressure-overload model of cardiac hypertrophy, although this model was shown to require ROS and resulted in increased cardiac levels of Nox4. Our observations, taken together with these studies, suggest that Nox4 may mediate TGF-1eCinduced fibrosis and pressure-overload (ie, stretch), whereas Nox2 is required for angiotensin II-induced cardiac fibrosis. Because we found that Nox2 was not involved in TGF-1 differentiation of human cardiac myofibroblasts, this may suggest that Nox4 substitutes for Nox2 in TGF-1eCinduced fibrosis in humans in contrast to murine models of cardiac fibrosis. The distribution of Nox subunits in mouse cardiac fibroblasts is currently unknown. Further studies are required to clarify whether Nox4 or Nox2 involvement in cardiac fibrosis is a function of agonist (angiotensin II versus TGF-1), species (human versus rodents), or model (in vitro versus in vivo).
Although it is well known that TGF-1 stimulates differentiation of fibroblasts into myofibroblasts, little is known about the signaling mechanisms that mediate this response. Classically, TGF-1 uses the Smad proteins to activate cellular functions. After ligand activation, the type II receptor associates with the type I receptor (TGF-1RI). Next, the TGF-1RI serine-threonine kinase phosphorylates receptor-Smads (R-Smads) Smad2 to form a complex with Smad3, which in turn associates with a Co-Smad (Smad4) and is translocated into the nucleus, where it acts as a transcription factor.29 After initial phosphorylation, Smad 2/3 activation in response to TGF-1 is prolonged (hours to days depending on the cell type). In our model, activation of Smad 2/3 occurs as early at 5 minutes and is maintained up to 72 hours after stimulation with TGF-1 (data not shown). In our cells, this long-term activation of Smad 2/3 mediates the chronic TGF-1 upregulation of SM -actin. Interestingly, downregulation of Nox4 leads to a marked inhibition of Smad2/3 phosphorylation at 24 hours. These data imply that Nox4 mediates SM -actin upregulation in response to TGF-1 by stimulating prolonged phosphorylation and activation of Smads 2/3.
In summary, we have identified a novel role for Nox4 as essential mediator of Smad2/3 transcription factor activation in response to TGF-1 and demonstrated its significance in differentiation of fibroblasts into myofibroblasts. This study identifies for the first time a role for Nox4 in acquisition and maintenance of contractile phenotype via modulation of transcription factors Smad2/3 activation. Because of the critical role of TGF-1eCinduced myofibroblast activation in cardiac fibrosis, these data provide insight into novel mechanisms with potential therapeutic implications for heart failure. Further studies are necessary to determine whether Nox4-derived ROS are also involved in vivo in models of cardiac fibrosis.
Acknowledgments
This work was supported by an American Heart Association Scientist Development Grant to D.S. and by a National Institutes of Health cardiovascular training grant to I.C. We are indebted to Drs Bernard Lasseegue, Kathy Griendling, David G. Harrison, and A. Maziar Zafari for their critical reading of the manuscript, and Marcie Burnham and Tranise Coryell for excellent secretarial assistance.
References
American Heart Association. Heart Disease and Stroke Statistics-2005 Update. Dallas, Tex: American Heart Association; 2005.
Sun Y, Kiani MF, Postlethwaite AE, Weber KT. Infarct scar as living tissue. Basic Res Cardiol. 2002; 97: 343eC347.
Gabbiani G. Evolution and clinical implications of the myofibroblast concept. Cardiovasc Res. 1998; 38: 545eC548.
Rosenkranz S. TGF-beta1 and angiotensin networking in cardiac remodeling. Cardiovasc Res. 2004; 63: 423eC432.
Kuwahara F, Kai H, Tokuda K, Kai M, Takeshita A, Egashira K, Imaizumi T. Transforming growth factor-beta function blocking prevents myocardial fibrosis and diastolic dysfunction in pressure-overloaded rats. Circulation. 2002; 106: 130eC135.
Sawyer DB, Colucci WS. Mitochondrial oxidative stress in heart failure: "oxygen wastage" revisited. Circ Res. 2000; 86: 119eC120.
Keith M, Geranmayegan A, Sole MJ, Kurian R, Robinson A, Omran AS, Jeejeebhoy KN. Increased oxidative stress in patients with congestive heart failure. J Am Coll Cardiol. 1998; 31: 1352eC1356.
Mallat Z, Philip I, Lebret M, Chatel D, Maclouf J, Tedgui A. Elevated levels of 8-iso-prostaglandin F2alpha in pericardial fluid of patients with heart failure: a potential role for in vivo oxidant stress in ventricular dilatation and progression to heart failure. Circulation. 1998; 97: 1536eC1539.
Bendall JK, Cave AC, Heymes C, Gall N, Shah AM. Pivotal role of a gp91(phox)-containing NADPH oxidase in angiotensin II-induced cardiac hypertrophy in mice. Circulation. 2002; 105: 293eC296.
Byrne JA, Grieve DJ, Bendall JK, Li JM, Gove C, Lambeth JD, Cave AC, Shah AM. Contrasting roles of NADPH oxidase isoforms in pressure-overload versus angiotensin II-induced cardiac hypertrophy. Circ Res. 2003; 93: 802eC805.
Sun Y, Zhang J, Lu L, Chen SS, Quinn MT, Weber KT. Aldosterone-induced inflammation in the rat heart: role of oxidative stress. Am J Pathol. 2002; 161: 1773eC1781.
Kinugawa S, Tsutsui H, Hayashidani S, Ide T, Suematsu N, Satoh S, Utsumi H, Takeshita A. Treatment with dimethylthiourea prevents left ventricular remodeling and failure after experimental myocardial infarction in mice: role of oxidative stress. Circ Res. 2000; 87: 392eC398.
Ide T, Tsutsui H, Kinugawa S, Suematsu N, Hayashidani S, Ichikawa K, Utsumi H, Machida Y, Egashira K, Takeshita A. Direct evidence for increased hydroxyl radicals originating from superoxide in the failing myocardium. Circ Res. 2000; 86: 152eC157.
Ide T, Tsutsui H, Kinugawa S, Utsumi H, Kang D, Hattori N, Uchida K, Arimura K, Egashira K, Takeshita A. Mitochondrial electron transport complex I is a potential source of oxygen free radicals in the failing myocardium. Circ Res. 1999; 85: 357eC363.
Maytin M, Siwik DA, Ito M, Xiao L, Sawyer DB, Liao R, Colucci WS. Pressure-overload-induced myocardial hypertrophy in mice does not require gp91phox. Circulation. 2004; 109: 1168eC1171.
Li Y, Huang TT, Carlson EJ, Melov S, Ursell PC, Olson JL, Noble LJ, Yoshimura MP, Berger C, Chan PH, Wallace DC, Epstein CJ. Dilated cardiomyopathy and neonatal lethality in mutant mice lacking manganese superoxide dismutase. Nat Genet. 1995; 11: 376eC381.
Siwik DA, Pagano PJ, Colucci WS. Oxidative stress regulates collagen synthesis and matrix metalloproteinase activity in cardiac fibroblasts. Am J Physiol Cell Physiol. 2001; 280: C53eCC60.
Li PF, Dietz R, von Harsdorf R. Superoxide induces apoptosis in cardiomyocytes, but proliferation and expression of transforming growth factor-beta1 in cardiac fibroblasts. FEBS Lett. 1999; 448: 206eC210.
Griendling KK, Sorescu D, Ushio-Fukai M. NAD(P)H oxidase: role in cardiovascular biology and disease. Circ Res. 2000; 86: 494eC501.
Lambeth JD. Nox/Duox family of nicotinamide adenine dinucleotide (phosphate) oxidases. Curr Opin Hematol. 2002; 9: 11eC17.
Krause KH. Tissue distribution and putative physiological function of NOX family NADPH oxidases. Jpn J Infect Dis. Oct 2004; 57: S28eCS29.
Thannickal VJ, Fanburg BL. Activation of an H2O2-generating NADH oxidase in human lung fibroblasts by transforming growth factor b1. J Biol Chem. 1995; 270: 30334eC30338.
Sorescu D, Weiss D, Lassegue B, Clempus RE, Szocs K, Sorescu GP, Valppu L, Quinn MT, Lambeth JD, Vega JD, Taylor WR, Griendling KK. Superoxide production and expression of nox family proteins in human atherosclerosis. Circulation. 2002; 105: 1429eC1435.
Zhao H, Joseph J, Fales HM, Sokoloski EA, Levine RL, Vasquez-Vivar J, Kalyanaraman B. Detection and characterization of the product of hydroethidine and intracellular superoxide by HPLC and limitations of fluorescence. Proc Natl Acad Sci U S A. 2005; 102: 5727eC5732.
Fink B, Laude K, McCann L, Doughan A, Harrison DG, Dikalov S. Detection of intracellular superoxide formation in endothelial cells and intact tissues using dihydroethidium and an HPLC-based assay. Am J Physiol Cell Physiol. 2004; 287: C895eCC902.
Sorescu D, Somers MJ, Lasseegue B, Grant S, Harrison DG, Griendling KK. Electron spin resonance characterization of the NAD(P)H oxidase in vascular smooth muscle cells. Free Radic Biol Med. 2001; 30: 603eC612.
Matsui Y, Sadoshima J. Rapid upregulation of CTGF in cardiac myocytes by hypertrophic stimuli: implication for cardiac fibrosis and hypertrophy. J Mol Cell Cardiol. 2004; 37: 477eC481.
Evans RA, Tian YC, Steadman R, Phillips AO. TGF-beta1-mediated fibroblast-myofibroblast terminal differentiation-the role of Smad proteins. Exp Cell Res. 2003; 282: 90eC100.
Massague J. How cells read TGF-beta signals. Nat Rev Mol Cell Biol. 2000; 1: 169eC178.
Sun Y, Weber KT. Angiotensin converting enzyme and myofibroblasts during tissue repair in the rat heart. J Mol Cell Cardiol. 1996; 28: 851eC858.
Vozenin-Brotons MC, Sivan V, Gault N, Renard C, Geffrotin C, Delanian S, Lefaix JL, Martin M. Antifibrotic action of Cu/Zn SOD is mediated by TGF-beta1 repression and phenotypic reversion of myofibroblasts. Free Radic Biol Med. 2001; 30: 30eC42.
Cross AR, Parkinson JF, Jones OTG. The superoxide-generating oxidase of leucocytes. Biochem J. 1984; 223: 337eC344.
Suh Y, Arnold RS, Lasseegue B, Shi J, Xu X, Sorescu D, Chung AB, Griendling KK, Lambeth JD. Cell transformation by the superoxide-generating oxidase mox1. Nature. 1999; 401: 79eC82.
Brar SS, Corbin Z, Kennedy TP, Hemendinger R, Thornton L, Bommarius B, Arnold RS, Whorton AR, Sturrock AB, Huecksteadt TP, Quinn MT, Krenitsky K, Ardie KG, Lambeth JD, Hoidal JR. NOX5 NAD(P)H oxidase regulates growth and apoptosis in DU 145 prostate cancer cells. Am J Physiol Cell Physiol. 2003; 285: C353eCC369.
Geiszt M, Kopp JB, Varnai P, Leto TL. Identification of renox, an NAD(P)H oxidase in kidney. Proc Natl Acad Sci U S A. 2000; 97: 8010eC8014.
Reynisdottir I, Polyak K, Iavarone A, Massague J. Kip/Cip and Ink4 Cdk inhibitors cooperate to induce cell cycle arrest in response to TGF-beta. Genes Dev. 1995; 9: 1831eC1845.
Li JM, Gall NP, Grieve DJ, Chen M, Shah AM. Activation of NADPH oxidase during progression of cardiac hypertrophy to failure. Hypertension. 2002; 40: 477eC484.
Heymes C, Bendall JK, Ratajczak P, Cave AC, Samuel JL, Hasenfuss G, Shah AM. Increased myocardial NADPH oxidase activity in human heart failure. J Am Coll Cardiol. 2003; 41: 2164eC2171.
Nakagami H, Takemoto M, Liao JK. NADPH oxidase-derived superoxide anion mediates angiotensin II-induced cardiac hypertrophy. J Mol Cell Cardiol. 2003; 35: 851eC859.(Ioan Cucoranu, Roza Clemp)