Genetic Deficiency of Plasminogen Activator Inhibitor-1 Promotes Cardiac Fibrosis in Aged MiceClinical Perspective
Involvement of Constitutive Transforming Growth Factor-β Signaling and Endothelial-to-Mesenchymal Transition
Background—Elevated levels of plasminogen activator inhibitor-1 (PAI-1), a potent inhibitor of urokinase plasminogen activator and tissue plasminogen activator, are implicated in the pathogenesis of tissue fibrosis. Paradoxically, lack of PAI-1 in the heart is associated with the development of cardiac fibrosis in aged mice. However, the molecular basis of cardiac fibrosis in aged PAI-1–deficient mice is unknown. Here, we investigated the molecular and cellular bases of myocardial fibrosis.
Methods and Results—Histological evaluation of myocardial tissues derived from aged PAI-1–deficient mice revealed myocardial fibrosis resulting from excessive accumulation of collagen. Immunohistochemical characterization revealed that the levels of matrix metalloproteinase-2, matrix metalloproteinase-9, and transforming growth factor-β1/2 and the number of Mac3-positive and fibroblast specific protein-1–positive cells were significantly elevated in aged PAI-1–deficient myocardial tissues compared with controls. Zymographic analysis revealed that matrix metalloproteinase-2 enzymatic activity was elevated in PAI-1–deficient mouse cardiac endothelial cells. Real-time quantitative polymerase chain reaction analyses of RNA from myocardial tissues revealed the upregulation of profibrotic markers in aged PAI-1–deficient mice. The numbers of phosphorylated Smad2–, phosphorylated Smad3–, and phosphorylated ERK1/2 MAPK–, but not pAkt/PKB–, positive cells were significantly increased in PAI-1–deficient myocardial tissues. Western blot and immunocytochemical analysis revealed that PAI-1–deficient mouse cardiac endothelial cells were more susceptible to endothelial-to-mesenchymal transition in response to transforming growth factor-β2.
Conclusions—These results indicate that spontaneous activation of both Smad and non-Smad transforming growth factor-β signaling may contribute to profibrotic responses in aged PAI-1–deficient mice hearts and establish a possible link between endothelial-to-mesenchymal transition and cardiac fibrosis in PAI-1–deficient mice.
Cardiac fibrosis, a common consequence of cardiovascular diseases, is characterized by an excessive synthesis and pathological accumulation of extracellular matrix (ECM) proteins such as collagen in myocardial tissue. Cardiac fibrosis contributes to the development of ventricular dysfunction (systolic and diastolic), heart failure, and arrhythmias. Cardiac fibrosis is generally preceded by inflammation and is associated with elevated levels of profibrotic cytokines.1 Rate of synthesis and deposition of collagen and its degradation by proteolytic activities contribute to tissue homeostasis. Along with matrix metalloproteinases (MMPs), urokinase-type plasminogen activator (uPA), and tissue-type plasminogen activator (tPA), serine proteases play a major role in controlling the levels of ECM proteins. Plasminogen activator inhibitor-1 (PAI-1), a member of the serine protease inhibitor gene family, inhibits both uPA and tPA, thus regulating plasmin-mediated activation of MMPs and protecting ECM proteins from proteolytic degradation. Additionally, uPA/tPA and plasmin are known to facilitate macrophage infiltration and to activate transforming growth factor-β (TGF-β).2,–,4 Nonphysiological levels of PAI-1 are implicated in several diseases, possibly because of the altered regulation of uPA/tPA and fibrinolysis or altered cellular migration, adhesion, and proliferation.5 While elevated levels of PAI-1 are associated with atherosclerosis and thrombosis resulting from impaired fibrinolysis, the absence of PAI-1 is associated with uncontrolled bleeding and impaired wound healing.5
Clinical Perspective on p 1209
While the TGF-β signaling is involved in multiple cellular processes, and its deregulation is associated with different diseases, including tissue fibrosis.6 The implication of TGF-β in cardiac fibrosis has been further evidenced by the presence of interstitial fibrosis in transgenic mice overexpressing TGF-β1, the decreased levels of age-related cardiac fibrosis in heterozygous TGF-β1 mice, and the lack of cardiac fibrosis in mice lacking functional TGF-β1.7 Treatment of normal fibroblasts with TGF-β causes increased collagen synthesis, further implicating TGF-β signaling in tissue fibrosis.1,8 TGF-β is known to transduce its signal from receptors to the nucleus via activation of Smad molecules. On ligand binding, activated TGF-β receptors I and II transphosphorylate Smad2 and Smad3, which heterodimerize with Co-Smad4 and translocate to the nucleus. Here, they interact directly with DNA and transcriptional coactivator p300 to activate the expression of target genes like type I collagen. In contrast, the inhibitory Smad 7 interacts with the receptors to block further phosphorylation of R-Smads, thus controlling the magnitude of stimulation by TGF-β.8,9 Furthermore, TGF-β transduces its signal via activation of non-Smad pathways such as PI3-kinase, protein kinase B, and Extracellular Signal-Regulated Kinases 1 and 2 Mitogen-Activated Protein Kinase (ERK1/2 MAPK) These non-Smad signaling pathways are known to play a significant role in the TGF-β regulation of type I collagen synthesis directly or via cross-talk with the Smad pathway.1
Although fibroblasts are the major contributor to fibrosis development, the origin of participating fibroblasts in fibrosis is still controversial. Originally, it was thought that adult fibroblasts are derived only from embryonic mesenchymal cells. However, recent studies suggest that adult fibroblasts also originate from endothelial cells by endothelial-to-mesenchymal transition (EndMT). EndMT-derived fibroblasts contribute to ECM protein synthesis and thus play a significant role in the development of fibrosis in organs like lung and heart. Interestingly, TGF-β is also implicated in the induction of EndMT.10,11
TGF-β stimulates the expression of PAI-1 in a variety of cells, and elevated levels of PAI-1 play a significant role in the development of fibrosis by inhibiting the tissue collagenolytic activities.1,4,5 Paradoxically, present and other findings demonstrate that mice lacking PAI-1 develop age-dependent cardiac fibrosis.12,13 However, the molecular basis of fibrosis development in aged PAI-1–deficient mice has not been investigated. Here, we investigate the histological changes in fibrotic hearts and the molecular mechanism governing the spontaneous development of cardiac fibrosis in aged PAI-1–deficient mice. We also studied the role of PAI-1 in EndMT in vitro. The results demonstrate the presence of dramatic myocardial fibrosis in aged PAI-1–deficient mice, which is associated with inflammation, elevated levels of Smad- and non–Smad-driven TGF-β signaling, and increased numbers of fibroblasts. Furthermore, the present study establishes that PAI-1–deficient cardiac endothelial cells are more susceptible to EndMT in response to TGF-β compared with wild-type controls, indicating that PAI-1 deficiency–associated EndMT may play a pivotal role in cardiac fibrosis. The implication of these findings is discussed.
PAI-1–deficient (PAI-1−/−) and wild-type (PAI-1+/+) mice were on a C57BL/6 background (Jackson laboratory, Bar Harbor, Me). All mouse protocols were approved by the Animal Care and Use Committee of Vanderbilt University (Nashville, Tenn) and Northwestern University (Chicago, Ill).
Histology and Collagen Staining
Hearts were sectioned in the short axis along the mid ventricle. The levels of collagen deposition were determined by Masson trichrome staining. Photographs were taken with an Olympus DP71 camera. To quantify the extent of fibrosis, stained sections (n=3 to 5) were analyzed with Image-Pro Plus 6.3 software (Media Cybernetics, Bethesda, Md).
Immunohistochemistry and Immunofluorescence
Tissue sections were incubated with primary antibody overnight at 4°C. The antibodies used included MMP2, MMP9 (Abcam Inc, Cambridge, Mass), fibroblast specific protein-1 (FSP1; gift from Dr Eric Neilson, Vanderbilt University), phosphorylated (p) Smad2, pSmad3, pAkt, and pERK1/2 (all from Cell Signaling, Beverly, Mass). A 3-layer peroxidase antiperoxidase method was used for antigen detection. For immunofluorescence, antibodies against TGF-β1, TGF-β2, and Mac3 (Pharmingen, San Diego, Calif) were used. Photographs were taken with an Olympus DP71 camera.
Quantification of Immunohistochemistry With Image-Pro Plus 6.3
Masson trichrome–stained and immunostained areas were determined with Image-Pro Plus 6.3 software (Media Cybernetics). The sum of 5 fields per mouse (n=3 to 5 mice) was used for statistical analysis to determine significant difference.
RNA Extraction From Myocardial Tissues and Real-Time Reverse-Transcription Polymerase Chain Reaction Analysis
Total RNA was extracted from myocardium derived from PAI-1–deficient and wild-type mice with TRIZOL (Invitrogen, Carlsbad, Calif). The messenger RNA (mRNA) levels were quantified by complementary DNA synthesis (iScript cDNA Synthesis Kit, Bio-Rad, Hercules, Calif) and quantitative polymerase chain reaction with SYBR Green SuperMix for IQ (Quanta Bioscience, Gaithersburg, Md). We measured 18S ribosomal mRNA and used it as an internal standard. Q-gene software was used for quantification.14
Isolation of Mouse Cardiac Endothelial Cells
Mouse cardiac endothelial cells were isolated following the protocols as described with modifications.15 The endothelial cells were isolated with 2 endothelial cell-specific markers: anti-mouse CD31 Dynabeads and anti-mouse CD102 Dynabeads (Invitrogen). Confluent cultures (passages 2 to 6) were used for subsequent EndMT studies.
Cell Proliferation Assay
Mouse cardiac endothelial cells derived from wild-type and PAI-1–null mice were cultured in 96-well plates (3×103 cells per well). At different time points, cell proliferation was determined following company protocol (BioVision, Mountain View, Calif).
Primary PAI-1–deficient and wild-type mouse cardiac endothelial cells were cultured in Dulbecco modified Eagle medium with 2% FBS and treated with TGF-β2 (10 ng/mL) for 2 to 7 days.
The levels of CD31 and α-smooth muscle actin (α-SMA) were measured by immunostaining with FITC-conjugated CD31 (Chemicon, Temecula, Calif) or FITC-conjugated α-SMA (Sigma, St Louis, Mo).
Western Blot Analysis
Cultures of endothelial cells in the presence and absence of TGF-β2 were harvested, and equal amounts of protein were used for Western blot using specific antibodies against CD31 (GenScript, Piscataway, NJ), type I collagen (Southern Biotech, Birmingham, Ala), DDR2, PAI-1 (Santa Cruz Biotechnology, Inc, Santa Cruz, Calif), FSP1 (gift from Dr Eric Neilson, Vanderbilt University), pSmad2, pSmad3, pERK1/2 (Cell Signaling), and actin (Abcam).
Zymography of MMP2 and MMP9
Conditioned media were subjected to electrophoresis and zymography following company protocol (Invitrogen) and photographed in a Molecular Imager (Bio-Rad).
Data are presented as mean±SEM. The significance of differences between controls and experimental groups was estimated by ANOVA, and a value of P<0.05 by Student t test was considered statistically significant. Statistical analyses were performed with GraphPad Prism 3.0 (GraphPad Software Inc, San Diego, Calif). For histological analysis, a power calculation was performed showing that a sample size of 3 would give 90% power to identify a statistically significant mean difference (0.12−0.02=0.10) of experimental and control groups and SD (0.02).
Elevated Levels of Collagen in PAI-1–Deficient Cardiac Tissue: Development of Spontaneous Cardiac Fibrosis
Myocardial tissues stained with Masson's trichrome revealed excessive collagen deposition in PAI-1–deficient heart with pericardial, perivascular, and interstitial distribution. Wild-type myocardium showed no evidence of such random accumulation of collagen. Quantification of deposited collagen revealed significantly more collagen deposition in PAI-1–deficient hearts (≈15% of total area; P<0.001) compared with age-matched controls (<2% of total area) (Figure 1A). The levels of collagen deposition in myocardial tissues derived from 24-month-old wild-type and PAI-1–deficient mice were not significantly different. The levels of collagen-1α (Ι) mRNA expression were significantly elevated in PAI-1–deficient mice compared with wild-type mice, suggesting that elevation of type I collagen is at the levels of transcripts (Figure 1B).
Levels of MMP2 and MMP9 Are Elevated in PAI-1–Deficient Myocardial Tissue
Immunohistochemistry results revealed that the levels of MMP2 and MMP9 protein were elevated in PAI-1–deficient myocardial tissue compared with age-matched wild-type controls (Figure 2 and Figure I in the online-only Data Supplement). Zymography revealed that TGF-β–induced MMP2 activity and both basal and TGF-β-induced MMP2 proteolytic activities were higher in PAI-1–deficient cardiac endothelial cells compared with wild-type cells (Figure II in the online-only Data Supplement). These results are consistent with the previous studies showing elevated levels of MMP2 and MMP9 in myocardial tissues derived from other animal models of cardiac fibrosis. Although the levels of basal MMP9 activity were modestly elevated in PAI-1–deficient cells, TGF-β failed to stimulate MMP9 (Figure II in the online-only Data Supplement).
Increased Inflammation in PAI-1–Deficient Myocardial Tissues
Generally, tissue fibrosis is preceded by inflammation, and inflammatory cells contribute to the secretion of profibrotic cytokines that activate fibroblasts.6,16 Results revealed the presence of a significantly increased number of Mac3-positive cells in the myocardial tissues derived from aged PAI-1–deficient mice compared with wild-type controls. This increase suggests the existence of inflammation and indicates that infiltrating macrophages may contribute to the progression and pathogenesis of cardiac fibrosis in PAI-1–deficient mice (Figure 3).
Elevated Levels of FSP1-Positive Cells in PAI-1–Deficient Myocardial Tissues
Because fibroblasts are the major source of collagen and other ECM proteins and are implicated in fibrosis, the levels of FSP1 were measured. The results revealed that the number of FSP1-positive cells was significantly increased in cardiac tissues derived from PAI-1–deficient mice compared with wild-type controls, which suggests the existence of more collagen-producing fibroblasts in PAI-1–deficient myocardial tissues (Figure 4).
PAI-1 Deficiency Is Associated With Elevated TGF-β, Fibroblast Growth Factor 2, and Fibroblast Growth Factor Receptor 2 Levels in Myocardial Tissues
PAI-1 is known to control the activity of pleiotrophic cytokine TGF-β.17 The present study tested the hypothesis that increased accumulation of collagen in a PAI-1–deficient heart is due to elevated levels of TGF-β signaling. The levels of TGF-β1 and TGF-β2 in myocardial sections were measured by immunohistochemical analysis. PAI-1–deficient myocardial tissues derived from 24-month-old mice showed increased levels of TGF-β1 and TGF-β2 immunostaining in the myocardial and perivascular tissues (Figure 5). The levels of TGF-β1 and TGF-β2 were modestly elevated only in the vascular walls of 12-month-old PAI-1–deficient mouse hearts (data not shown). In addition, we measured the TGF-β2 mRNA levels by quantitative polymerase chain reaction analysis, revealing upregulation of TGF-β2 expression in PAI-1–deficient mice compared with wild-type controls (Figure 1B). Because fibroblast growth factor (FGF) and its receptor (FGFR) are also known to play an important role in fibrogenic processes,18 the levels of these factors were measured. The results of the quantitative polymerase chain reaction analyses revealed that the levels of FGF2 and FGFR2 mRNA were significantly elevated in PAI-1–deficient myocardial tissues (Figure 1B), indicating that FGF2 signaling may also contribute to the progression of cardiac fibrosis in the absence of PAI-1.
Elevated Levels of Nuclear pSmad2/pSmad3 in PAI-1–Deficient Myocardial Tissue
Because PAI-1 deficiency in the heart is associated with elevated levels of TGF-β, which transduces its signal via activation of the Smad pathway, the activation of Smad signaling pathway was examined. The results revealed that although PAI-1–deficient myocardial tissues from 12-month-old mice displayed modest positive nuclear staining for pSmad2, myocardial tissue from 24-month-old mice displayed significantly increased levels and intensity of pSmad2 molecules, located predominantly in the nucleus. In contrast, myocardial tissues from age-matched wild-type controls showed almost undetectable levels of nuclear pSmad2 (Figure 6A). The levels of pSmad3-positive cells were also elevated in PAI-1–deficient myocardial tissues (Figure III in the online-only Data Supplement). Western blot analysis of cell extracts revealed elevated levels of TGF-β–induced pSmad2 and pSmad3 in PAI-1–deficient mouse cardiac endothelial cells and mouse cardiac fibroblasts compared with wild-type cells (Figures IV and V in the online-only Data Supplement). Results also revealed that there was no difference in immunostaining intensity of pAkt/PKB in myocardial tissue derived from PAI-1–deficient mice and wild-type controls (data not shown).
Levels of Active pERK1/2 Are Elevated in PAI-1–Deficient Myocardial Tissues
The influence of PAI-1 deficiency on activation of ERK1/2 MAP kinase pathway was examined. Results showed that a lack of PAI-1 was associated with significantly elevated levels of pERK1/2 MAPK in myocardial tissues derived from 24-month-old mice compared with age-matched wild-type controls (Figure 6B). pERK1/2-positive cells were detected in the cardiac vascular walls in both 12-month-old wild-type mice and 24-month-old wild-type mice; however, the intensity of pERK1/2 specific immunostaining in the vascular wall of 12-month-old mice was stronger than that of 24-month-old mice. Western blot analysis of cell extracts revealed that the levels of pERK1/2 MAPK were elevated in PAI-1–deficient mouse cardiac endothelial cells and mouse cardiac fibroblasts compared with wild-type cells (Figures IV and V in the online-only Data Supplement).
PAI-1–Deficient Cardiac Endothelial Cells Are More Susceptible to EndMT
The effect of PAI-1 deficiency on the TGF-β–induced EndMT was examined. Endothelial cells were isolated from PAI-1–knockout and wild-type mouse hearts. The purity of the isolated cardiac endothelial cells was characterized by labeling with DiL-tagged acetylated low-density lipoprotein. Results revealed that both wild-type and PAI-1–deficient endothelial cells (>90%) were labeled with DiL-tagged acetylated low-density lipoprotein. Immunofluorescence studies with the endothelial marker CD31 and the fibroblast marker α-SMA revealed that >90% of the isolated cultured cells were endothelial cells (Figure 7A). Proliferation assay revealed that the growth rate of PAI-1–deficient mouse cardiac endothelial cells was modestly faster than wild-type mouse cardiac endothelial cells (Figure VI in the online-only Data Supplement), consistent with a previous report.19
To study the progress of EndMT, PAI-1–deficient and wild-type cardiac endothelial cells were treated with TGF-β2 for 2 to 7 days and then harvested. Equal amounts of whole-cell proteins were subjected to Western blot. Results showed that both PAI-1–deficient and wild-type endothelial cells express CD31 (endothelial marker). On day 7, the expression of CD31 was significantly reduced (P<0.05) in the presence of TGF-β. TGF-β–untreated endothelial cell populations expressed very low levels of type I collagen. In sharp contrast, TGF-β–treated cells expressed elevated levels of this fibroblastic marker. Similar results were obtained with other fibroblast/myofibroblast markers, DDR2 and FSP1 (Figure IV in the online-only Data Supplement), indicating the transition of cardiac endothelial cells to fibroblast-like cells. The levels of type I collagen in PAI-1–deficient EndMT-derived cells (TGF-β treated) were significantly higher compared with wild-type controls (Figure 7B; compare lanes 2 and 4). Next, EndMT was studied by immunofluorescence. Cells were fixed and the levels of endothelial cell marker (CD31) and fibroblastic marker (α-SMA) were measured. Results revealed that TGF-β2 induced EndMT in both PAI-1 wild-type and PAI-1–deficient cells as evidenced by the reduced levels of CD31 expression. PAI-1–deficient mouse endothelial cells were more susceptible to EndMT in response to TGF-β2 at 48 hours as evidenced by the presence of a significantly higher number of myofibroblasts in the absence of PAI-1 compared with wild-type controls (Figure 7C). However, the number of myofibroblasts in the presence and absence of PAI-1 was not significantly different on day 7 (Figure VII in the online-only Data Supplement).
Fibrosis is characterized by excessive synthesis and deposition of ECM proteins such as type I collagen and altered levels of PAI-1, a key regulator of tissue remodeling.1,20,–,25 Excessive PAI-1 is implicated in tissue fibrosis such as kidney, skin, lung, heart, and liver.21,26,–,28 Paradoxically, the present study and other previous reports demonstrated that deficiency of PAI-1 or excess uPA caused spontaneous cardiac fibrosis in aged mice.12,13,29,30 The present study was undertaken to understand the cellular and molecular events that may cause excessive accumulation of collagen in PAI-1–deficient heart tissue. Here, we demonstrated that 12-month-old PAI-1–deficient mice developed modest spontaneous cardiac fibrosis as reported previously.12,30 However, 24-month-old mice developed very prominent cardiac fibrosis as evidenced by significantly elevated levels of perivascular and myocardial collagen accumulation. Elevated levels of collagen α1(I) mRNA in myocardial tissue derived from PAI-1–deficient hearts indicated that elevated levels of collagen accumulation were due to increased collagen gene expression.
Like the PAI-1/tPA/uPA system, the collagenase/gelatinase/MMP system is also known to play a significant role in myocardial tissue remodeling, and numerous reports indicate that certain MMPs play a pivotal role in cardiac fibrosis. The present study showed that a lack of PAI-1 in myocardial tissue was associated with elevated levels of MMP2 and MMP9, indicating the possible role of MMPs in myocardial tissue remodeling and progression of fibrosis in this murine model of cardiac fibrosis. Moreover, PAI-1 deficiency in mouse cardiac endothelial cells was also associated with increased MMP2 and MMP9 enzymatic activity, consistent with previous observations. For example, elevated levels of MMP expression were noticed in ventricular tissues derived from patients with dilated cardiomyopathy.31 Additionally, targeted deletion of MMP9 blunted left ventricular dilation and myocardial fibrosis in an experimental model of myocardial infarction, suggesting a pivotal role of MMP9 in myocardial remodeling.32 A recent study demonstrated that the levels of MMP2 and MMP9 are elevated in PAI-1–deficient myocardial tissues compared with wild-type controls.30 Generally, accumulation of ECM proteins like collagen in a particular tissue depends on the levels of synthesis of collagen and the rate of degradation of collagen by MMP and other collagenolytic activities. Therefore, the logical question is, How are elevated levels of collagens accumulated while the levels of MMPs are also increased under pathological condition of myocardial fibrosis? This can be explained in light of several studies suggesting that under pathological conditions, physiological cross-linked collagens are digested by increased proteolytic activities of MMP2/MMP9, and poorly cross-linked collagens are synthesized and accumulated in the myocardial tissues, leading to the development of cardiac fibrosis.33,34
The present study also demonstrated the presence of inflammation in aged PAI-1–deficient myocardial tissues and established the cellular signals governing the development of cardiac fibrosis in PAI-1–deficient mice. A recent study also demonstrated the presence of increased vascular permeability and local inflammation in PAI-1–deficient myocardial tissues.30 The increased inflammation may be due to increased uPA activity in the absence of PAI-1, as has been suggested by Moriwaki et al,12 who demonstrated that overexpressed uPA induced macrophage infiltration.12 That study also demonstrated that the levels of uPA activity were significantly higher in PAI-1–deficient mice, which might augment macrophage infiltration and accumulation in the myocardial tissues.12 Furthermore, studies demonstrated that macrophage ablation blocks myofibroblast differentiation and interstitial fibrosis in a murine model of glomerulonephritis.35 Because TGF-β–stimulated type I collagen synthesis is implicated in tissue fibrosis,1,8 we asked whether cardiac fibrosis in aged PAI-1–deficient mice is linked to elevated TGF-β signaling. Indeed, the levels of TGF-β1 and TGF-β2 are significantly elevated in aged PAI-1–deficient myocardial tissues compared with wild-type controls, suggesting that increased TGF-β signaling may drive the profibrotic responses in aged PAI-1–deficient mice. This result is consistent with recent findings that PAI-1 controls the activity of TGF-β in PAI-1–deficient mouse embryonic fibroblasts17 and that the levels of TGF-β are elevated in PAI-1–deficient heart30 and PAI-1–deficient kidney with glomerulonephritis.4 It is reasonable to speculate that infiltrating macrophages secrete profibrotic cytokines like TGF-β, which in turn activate fibroblasts and other cells for paracrine synthesis of TGF-β. The elevated TGF-β activity may be due to elevated uPA activity in PAI-1–null mice, which is known to play a role in the activation of TGF-β.2,–,4,36
Because TGF-β is known to activate the profibrotic marker type I collagen via activation of Smad signaling molecules, Smad activation was examined by measuring the levels of Smad2/Smad3 phosphorylation and their subcellular localization in PAI-1–deficient and wild-type myocardial tissues. PAI-1–deficient cardiac tissues showed intensely phosphorylated Smad2/Smad3-positive cells, and most important, pSmad2 and pSmad3 were localized predominantly in the nuclei, suggesting the presence of a strong spontaneously activated Smad pathway. This result is consistent with the recent observation of constitutive activation of Smad2/3 in PAI-1–null mouse embryonic fibroblasts in vitro.17 The present study demonstrated that in response to TGF-β2, the levels of phosphorylated Smad2 and pSmad3 are elevated in PAI-1–null mouse cardiac endothelial cells compared with wild-type mouse cardiac endothelial cells. Most important, Smad-dependent TGF-β signaling is implicated in the elevated synthesis of collagen,1,8 indicating that activated Smads may be contributors in the excessive synthesis of collagen and progression of cardiac fibrosis in aged PAI-1–deficient mice.
Besides Smad-dependent signaling, TGF-β transduces its profibrotic signal through non-Smad pathways or through cross-talk of Smad and non-Smad pathways, including ERK1/2 MAPK and PI3K/PKB pathways.1,8 To delineate the role of a non-Smad pathway in cardiac fibrosis, we asked whether non-Smad signaling like the Akt/PKB or ERK1/2 MAPK pathway was also activated in PAI-1–deficient cardiac tissues that might contribute to the development of cardiac fibrosis. Although activation of the Akt/PKB pathway was implicated in tissue fibrosis, there was no alteration of the Akt/PKB axis in the myocardial tissue derived from PAI-1–deficient mice compared with controls. However, the results of the present study clearly indicate that cardiac fibrosis in PAI-1–deficient mice not only was associated with constitutive Smad activation but also might be due to strong activation of the ERK1/2 MAPK pathway. Several reports indicated that TGF-β–activated Smad cross-talks with ERK1/2 and that activated ERK1/2 MAPK is known to phosphorylate Smad, leading to elevated collagen synthesis. Furthermore, pharmacological inhibition of ERK1/2 leads to suppression of Smad-dependent TGF-β signaling and downstream target gene expression.1,8 Therefore, it is reasonable to conclude that physical and functional cooperation of activated Smads and ERK1/2 MAPK may effectively contribute to excessive collagen synthesis and fibrosis in PAI-1–deficient hearts.
The present study shows increased numbers of fibroblasts and elevated levels of collagen in the myocardial tissues in aged PAI-1–deficient mice, suggesting that an increased number of fibroblasts may contribute to excessive collagen synthesis and fibrosis in aged PAI-1–deficient mice. The increased number of fibroblasts may be due to increased proliferation of resident fibroblasts or may be originating from endothelial cells by EndMT as reported by Zeisberg et al10 in a murine model of cardiac fibrosis. Our in vitro data indicated that PAI-1–deficient cardiac endothelial cells were more susceptible to EndMT in response to exogenous TGF-β. This may be due to increased cellular uPA/tPA activities in the PAI-1–deficient cells, which activate endogenous latent TGF-β (induced by exogenous TGF-β or paracrine signaling) to active TGF-β (autocrine signaling). Increased autocrine TGF-β signaling may contribute to increased EndMT in PAI-1–deficient mouse cardiac endothelial cells.37 The present study also demonstrates that during EndMT the levels of Smad-dependent and ERK1/2 MAPK-dependent TGF-β signaling are increased in PAI-1–deficient murine cardiac endothelial cells. Therefore, the increased number of fibroblasts in fibrotic myocardial tissues in PAI-1–deficient hearts may be EndMT derived. The origin of cardiac fibroblasts contributing to spontaneously developing cardiac fibrosis in aged PAI-1–deficient mice requires further in vivo investigation using relatively large sample sizes.
This novel investigation provides substantial evidence for a model demonstrating the molecular basis of cardiac fibrosis development in aged PAI-1–deficient mice. According to the proposed model, cardiac fibrosis development in PAI-1–deficient mice is associated with inflammation, increased number of fibroblasts, elevated TGF-β signaling, and activation of Smad2/3 and ERK1/2 MAPK pathways (Figure 8). Because PAI-1–deficient cardiac endothelial cells are more susceptible to TGF-β–induced EndMT compared with wild-type controls, it is reasonable to conclude that EndMT-derived myofibroblasts synthesize excessive collagen in the myocardial tissues and contribute to cardiac fibrosis in aged PAI-1–deficient mice.
Sources of Funding
This work is supported by grants from the National Institutes of Health, National Heart, Lung, and Blood Institute (HL051387) and Specialized Centers of Clinically Oriented Research (SCCOR) (5P50HL081009–04-0001).
Dr Vaughan has received support from the National Institutes of Health, National Heart, Lung, and Blood Institute. The other authors report no conflicts.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.110.955245/DC1.
- Received November 30, 2009.
- Accepted July 1, 2010.
- © 2010 American Heart Association, Inc.
- Ruiz-Ortega M,
- Rodríguez-Vita J,
- Sanchez-Lopez E,
- Carvajal G,
- Egido J
- Lyons RM,
- Keski-Oja J,
- Moses HL
- Lyons RM,
- Gentry LE,
- Purchio AF,
- Moses HL
- Hertig A,
- Berrou J,
- Allory Y,
- Breton L,
- Commo F,
- Costa De Beauregard MA,
- Carmeliet P,
- Rondeau E
- Bujak M,
- Frangogiannis NG
- Ghosh AK
- Arciniegas E,
- Frid MG,
- Douglas IS,
- Stenmark KR
- Moriwaki H,
- Stempien-Otero A,
- Kremen M,
- Cozen AE,
- Dichek DA
- Weisberg AD,
- Albornoz F,
- Griffin JP,
- Crandall DL,
- Elokdah H,
- Fogo AB,
- Vaughan DE,
- Brown NJ
- Colgan SP
- Lim YC,
- Luscinskas FW
- Petrov VV,
- Fagard RH,
- Lijnen PJ
- Pedroja BS,
- Kang LE,
- Imas AO,
- Carmeliet P,
- Bernstein AM
- Chaudhary NI,
- Roth GJ,
- Hilberg F,
- Müller-Quernheim J,
- Prasse A,
- Zissel G,
- Schnapp A,
- Park JE
- Takeshita K,
- Hayashi M,
- Iino S,
- Kondo T,
- Inden Y,
- Iwase M,
- Kojima T,
- Hirai M,
- Ito M,
- Loskutoff DJ,
- Saito H,
- Murohara T,
- Yamamoto K
- Xu Z,
- Castellino FJ,
- Ploplis VA
- Ducharme A,
- Frantz S,
- Aikawa M,
- Rabkin E,
- Lindsey M,
- Rohde LE,
- Schoen FJ,
- Kelly RA,
- Werb Z,
- Libby P,
- Lee RT
- Vanhoutte D,
- Schellings M,
- Pinto Y,
- Heymans S
- Li YY,
- McTiernan CF,
- Feldman AM
- Vaughan DE
Cardiac fibrosis, defined as the proliferation of interstitial fibroblasts and accumulation of extracellular matrix components in the heart, is a common consequence of cardiovascular disease, including acute myocardial infarction and hypertension. Cardiac fibrosis contributes to the development of ventricular dysfunction (systolic and diastolic), heart failure, and arrhythmias. Although fibrosis is predictably identified in end-stage heart disease, the origin of fibroblasts contributing to the excessive synthesis of collagen in the fibrotic heart is controversial. At present, there is no effective treatment to prevent or to reverse cardiac fibrosis. The present study elucidates the molecular basis of cardiac fibrosis using a murine model of age-dependent spontaneous cardiac fibrosis that develops in the absence of the plasminogen activator inhibitor-1 gene. The present study suggests that the cardiac fibrosis in plasminogen activator inhibitor-1–deficient mice is due to increased inflammation, elevated levels of transforming growth factor-β, and induction of transforming growth factor-β–induced profibrotic responses. Plasminogen activator inhibitor-1–deficient endothelial cells appear to be more susceptible to the phenomenon called endothelial-mesenchymal transition in response to transforming growth factor-β via induction of both Smad and ERK1/2 MAPK pathways. These findings provide new insights into molecular mechanisms of cardiac fibrosis and suggest that physiological plasminogen activator inhibitor-1 levels help to protect the heart from age-dependent fibrogenesis. Thus, specific disruption of activated Smad and ERK1/2 MAPK signaling pathways with small-molecule inhibitor(s) may be useful in limiting endothelial-mesenchymal transition and may inform a novel therapeutic approach to prevent and treat cardiac fibrosis in humans.