- Email: [email protected]
Fibroblast activation protein alpha expression identifies activated fibroblasts after myocardial infarction
Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Contents lists available at ScienceDirect
Journal of Molecular and Cellular Cardiology journal homepage: www.elsevier.com/locate/yjmcc
Fibroblast activation protein alpha expression identifies activated fibroblasts after myocardial infarction☆ Jochen Tillmanns a,⁎, Daniel Hoffmann b,1, Yasmin Habbaba a,2, Jan D. Schmitto c, Daniel Sedding a, Daniela Fraccarollo a, Paolo Galuppo a, Johann Bauersachs a a b c
Department of Cardiology and Angiology, Hannover Medical School, 30625 Hannover, Germany Department of Medicine, University Hospital Wurzburg, 97080 Wurzburg, Germany Department of Cardiothoracic Surgery, Hannover Medical School, 30625 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 16 April 2015 Received in revised form 3 August 2015 Accepted 20 August 2015 Available online 28 August 2015 Keywords: Fibroblast activation protein α Inflammation Myofibroblast Wound healing Heart Myocardial infarction
a b s t r a c t Introduction: Fibroblast activation protein α (FAP) is a membrane-bound serine protease expressed by activated fibroblasts during wound healing in the skin. Expression of FAP after myocardial infarction (MI) and potential effects on cardiac wound healing are largely unknown. Methods: MI was induced in rats and FAP expression was analyzed at 3, 7 and 28 days post-MI by microarray, Western blot and immunohistochemistry. In human hearts after MI, a FAP+ fibroblast population was identified, and characterized by immunohistochemistry for prolyl-4-hydroxylase β, α-smooth muscle actin, Thy-1 and vimentin. Signaling pathways leading to FAP expression were studied in human cardiac fibroblasts by Western blot and ELISA using TGFβ1, TGF-beta type I-receptor (TGFbR1)-inhibitor SB431542 or the MAPK-inhibitor U0126 as well as siRNA targeting SMAD2 and SMAD3. Finally, fibroblasts were assayed for FAP-dependent migration (modified Boyden-chamber), proliferation (BrdU-assay) and gelatinolytic activity by gelatin zymography. Results: In rats, FAP expression was increased after MI especially in the peri-infarct area peaking at 7 days post-MI. Co-localization analysis identified the majority of FAP+ cells as activated proto-myofibroblasts and myofibroblasts. Concordantly, FAP+ fibroblasts were abundant in ischemic tissue of human hearts after MI, but not in healthy control hearts. In vitro, FAP was induced by TGFβ1 via the canonical SMAD2/SMAD3 pathway. Depletion of FAP in fibroblasts reduced migratory capacity, while proliferation was not affected. Gelatin zymography revealed gelatinase activity by fibroblast-derived FAP. Conclusion: In this study, we show for the first time the expression of FAP in activated fibroblasts after MI and its activation by TGFβ1. Effects of FAP on fibroblast migration and gelatinolytic activity indicate a potential role in cardiac wound healing and remodeling. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Cardiac fibroblasts are central mediators of the reparative response after myocardial infarction (MI) [1,2]. Following MI, fibroblasts are activated, resulting in dynamic phenotypic changes and differentiation into collagen-secreting proto-myofibroblasts, mediated by various Abbreviations: FAP, fibroblast activation protein α; APCE, antiplasmin-cleaving enzyme; rh, recombinant human; DPPIV, dipeptidyl peptidase IV; MI, myocardial infarction; LV, left ventricular; CF, cardiac fibroblast. ☆ The authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors declare no conflict of interest. ⁎ Corresponding author at: Department of Cardiology and Angiology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany. E-mail address: [email protected] (J. Tillmanns). 1 Present address: Department of Trauma and Reconstructive Surgery, University Hospital, 37099 Goettingen, Germany. 2 Present address: Department of Psychiatry and Psychotherapy, University Hospital Wurzburg, 97080 Wurzburg, Germany.
http://dx.doi.org/10.1016/j.yjmcc.2015.08.016 0022-2828/© 2015 Elsevier Ltd. All rights reserved.
factors including mechanical stress and TGFβ-signaling [3,4]. Protomyofibroblasts express β- and γ-cytoplasmic actins, and can further differentiate into mature myofibroblasts, characterized by the expression of α-smooth muscle actin (SMA) [3,5]. Myofibroblasts migrate into the injured tissue and secrete increased amounts of growth factors, cytokines, as well as peri-cellular proteases and their inhibitors, thereby regulating the composition of the extracellular matrix and promoting healing of the infarct scar [6]. However, persistence of myofibroblasts within the surviving myocardium contributes to detrimental cardiac fibrosis, leading to systolic and diastolic heart failure by decreased contraction and increased cardiac muscle stiffness [3]. Because of their central role in cardiac remodeling, fibroblasts are under investigation as therapeutic targets in ischemic cardiomyopathy [7]. Beneficial effects of fibroblasts in early infarct healing and detrimental effects of excessive amounts of fibroblasts in uninjured tissue require careful timing for safe and effective anti-fibrotic treatments. Therefore, a better understanding of the tempro-spatial activation and function of fibroblasts within the heart after MI is needed.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fibroblast activation protein α (FAP, also known as Seprase) is a type II integral membrane glycoprotein with dipeptidyl-peptidase and type I collagenase activity [8]. FAP is transiently expressed in some fetal mesenchymal tissues, but is absent or expressed at low levels in most adult tissues under baseline conditions [9,10]. Elevated production of FAP is associated with active tissue remodeling in mesenchymal tissues during embryogenesis [11], cancer [9,12], idiopathic pulmonary fibrosis [13], liver cirrhosis [14] and rheumatoid arthritis [15]. Of note, FAPexpressing fibroblasts have been detected in healing wounds [9], and its expression level correlates with fibrosis severity in liver cirrhosis [14]. Since FAP is expressed at sites of active tissue remodeling, we hypothesized that it is expressed in activated cardiac fibroblasts, and contributes to healing and matrix turnover of the infarcted tissue after MI. 2. Materials and methods Additional materials and methods are presented as Supplementary online material. 2.1. Experimental myocardial infarction in rats Myocardial infarction (MI) was induced in male Wistar rats weighing 200–250 g, as described previously [16,17]. Briefly, under 1.5–2% isoflurane anesthesia (induction with 5% isoflurane), the thorax was opened and the proximal left anterior descending coronary artery was occluded using a 5–0 suture. Sham-operated control rats underwent the same surgical procedure except that the suture around the coronary artery was not tied. Animals were kept warm with a heating pad. Depth of anesthesia was tested using the pedal withdrawal reflex. Analgesia was maintained using buprenorphine (0.05 mg/kg BW i.p.). Before and after surgery, animals were housed in the animal facility and monitored daily for activity and signs of pain. Animals were euthanized and hearts removed at time points 3 days (n = 7), 7 days (n = 7) and 28 days after MI (n = 8), and from sham-operated animals (n = 6). Animal studies were conducted in accordance with the principles and procedures outlined in the Guide for the Care and Use of Laboratory Animals and were approved by the local government (Regierung Unterfranken AZ: K 55.2-2531.01-6409). 2.2. Immunohistochemistry of rat and human myocardial tissue Frozen sections from rat hearts were stained with antibodies against FAP, α-smooth muscle actin (SMA), Ki67, prolyl-4-hydroxylase β (P4H), CD31, CD68, CD11b/c, and vimentin, and visualized by fluorescent labeling as described previously [18] and in Appendix B. Quantitative analysis of staining for FAP, SMA and P4H was analyzed in rat hearts at 3 days (n = 3), 7 days (n = 3) and 28 days (n = 4) after MI, and in sham-operated animals (n = 4). Pictures of the infarct borderzone area were taken at 63× magnification using a Zeiss Axiovert microscope with Axiovison 4.0 software (Carl Zeiss MicroImaging). Counting was performed by two independent individuals using Image Pro Plus software (Media Cybernetics, Bethesda, USA), and results were averaged. FAP expression was evaluated in paraffin-embedded formalin-fixed tissue from the left ventricular apex obtained from discarded tissue of patients receiving a left ventricular assist device (LVAD) due to acute MI (n = 5) as described previously [18] and in Appendix B. Biopsies obtained from non-failing donor hearts served as controls (n = 4). The number of FAP+ cells was quantified in each sample using large images taken at 10 × magnification. To minimize sampling error, quantified areas ranged from 2.8 to 9.8 mm2, depending on the number of positive cells. The study was approved by the ethics committee of Hannover Medical School and conforms to the ethical guidelines of the 1975 Declaration of Helsinki.
195
2.3. In vitro studies with cardiac fibroblasts Adult human cardiac fibroblasts (CFs) were obtained commercially (Cell Applications, no. 306-05a) and cultured according to the manufacturer instructions. FAP concentrations in cell lysates were quantified using a commercial ELISA (R&D Systems) according to manufacturer instructions and as described previously [19] and in Appendix B. For FAP knockdown experiments, CF were transfected with siRNA against FAP, SMAD2 or SMAD3 (50 nM, Ambion) using oligofectamine or lipofectamine RNAiMAX reagent (Invitrogen). The efficiency of FAP silencing was documented 4 days later. Fibroblasts transfected with scrambled siRNA (Ambion) were used as control in all studies. Before the experiments, cells were starved in serum-free medium consisting of DMEM high glucose (Lonza) supplemented with 10 mM HEPES and 0.2% bovine serum albumin (Sigma-Aldrich) for 24 h. Cells were studied in passages 3 to 6. For stimulation experiments, fibroblasts at 70% confluence were supplemented with fresh serum-free medium containing various concentrations of recombinant human TGFβ1 (R&D Systems) and cultured for 48 h to 72 h. In selected experiments, the TGFβ-receptor-inhibitor SB431542 (10 μM, Sigma-Aldrich) or the MAP kinase inhibitor U0126 (10 μM, Sigma-Aldrich) was added to the culture prior to the addition of TGFβ1. Cultured fibroblasts were analyzed for proliferation and migration as described in Appendix B. 2.4. Gelatin zymography Fibroblasts were homogenized in ice-cold cacodylic acid buffer, mixed with loading buffer and electrophoresis was performed under nonreducing conditions on a 10% SDS polyacrylamide gel containing 1 mg/ml of gelatin as described previously [20]. Zymography was also performed with EDTA to inhibit Ca++-dependent matrix-metalloproteinase activity. FAP gelatinolytic activity resulted in a band at 170 kDa, corresponding to the enzymatically active FAP dimer [21]. Gelatinolytic bands were quantified by densitometry using NIH ImageJ software. 2.5. Statistics Data are expressed as mean ± SE. We used the unpaired Student's t test for analysis of differences between two groups, and one-way ANOVA with multiple comparisons for analysis of differences between more than two groups. A value of P less than 0.05 was considered statistically significant. Statistical analysis was performed with Prism 5 (GraphPad). 3. Results 3.1. Expression levels of FAP are increased after MI Analysis of gene expression by microarray was performed in the ischemic tissue of rat hearts explanted 3 and 7 days after MI. From a total of 31,099 genes, 1287 genes (4%) were induced N2 fold at 3 days after MI, and 878 genes (3%) were induced N 2 fold at 7 days after MI,
Fig. 1. A marked increase of FAP protein expression was present in the left ventricular ischemic tissue and scar until day 28 after MI. FAP expression after MI was maximal in scar and border zone at 7 days after MI, and persisted within the infarct scar until at least 28 days after MI. In contrast, FAP expression was low within non-infarcted interventricular septum of the same animals. MI 3 days, n = 2; MI 7 days, n = 2; MI 28 days, n = 2.
196
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fig. 2. (A) Immunohistological analysis revealed abundant FAP+ fibroblasts infiltrating the infarcted myocardium especially in the border zone. In hearts of sham-operated animals, FAP+ cells were scarce. Abundant co-localization of FAP and P4H was present, indicating increased collagen synthesis of FAP+ cells. (B) FAP+ fibroblasts have an activated fibroblast phenotype, as indicated by co-localization with P4H and vimentin. A part of FAP+ cells also co-localized with SMA, identifying them as differentiated myofibroblasts. FAP: red, P4H, Vimentin, SMA: green, DAPI: blue. Scale bars: A: 10 μm, B: 100 μm. Image processing included changes in brightness, contrast and tonal range, and was applied equally across the entire image. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
compared to expression levels in hearts of sham-operated animals. Thirty six of the induced genes were classified as proteolytic, including FAP (Supplementary data Table A.1). FAP gene expression was increased 2.0 ± 0.1 fold at 3 days and 2.8 ± 0.0 fold at 7 days after MI, respectively. These results were confirmed on the protein level using Western blot analysis of rat hearts at 3, 7 or 28 days after MI (Fig. 1).
197
myofibroblast phenotypes: We analyzed the frequency of SMA+ expression as a marker indicating differentiation of fibroblasts into mature myofibroblasts. The majority of FAP+ cells showed strong co-expression of P4H at days 7 and 28 after MI, while a minor part of FAP+ fibroblasts co-expressed SMA, indicating that the FAP+ fibroblast subpopulation consisted predominantly of proto-myofibroblasts and to a minor part of myofibroblasts (Fig. 3B).
3.2. FAP is expressed in activated fibroblasts after MI Morphometric analysis demonstrated induction of large myocardial infarctions in rat hearts. Infarct sizes were 44 ± 3%, 52 ± 3% and 44 ± 5% on days 3, 7 and 28 after MI, respectively (Supplementary data Fig. C.1). To characterize the expression of FAP in cardiac cells, we performed multi-color fluorescence immunohistochemistry analyses in infarcted rat hearts. On days 3 and 7 after MI, FAP was increasingly expressed in cells located within the border zone of the infarcted area. While there were few FAP+ cells in the necrotic area on day 3 and day 7 after MI, FAP+ cells were frequently detected clustered within the scar tissue 28 days after MI (Fig. 2A). Only few FAP+ cells were found in the remote area and interventricular septum. In myocardium of sham-operated animals, FAP+ cells were only rarely found within connective tissue. We then aimed to further characterize the phenotype of FAP+ cells. We used prolyl-4-hydroxylase β (P4H) as marker for activated, collagen-synthesizing fibroblasts, and α-smooth muscle actin (SMA) as marker to identify differentiated, mature myofibroblasts. The fibroblast-phenotype was further supported by expression of vimentin. After MI, FAP+ cells were indeed identified as activated fibroblasts, as evidenced by co-expression of FAP with P4H, SMA and vimentin (Fig. 2B). FAP+ activated fibroblasts were actively replicating, indicated by expression of Ki67 (Supplementary data Fig. C.2). Next, we used P4H and FAP co-expression analysis to quantify the fraction of activated, collagen-synthesizing FAP+ cells after MI. The density of activated P4H+ fibroblasts peaked on day 7 and declined thereafter, and the density of FAP+ cells showed a similar time course (Fig. 3A). On days 7 and 28 after MI, the majority of FAP+ cells co-expressed P4H while there were few FAP+/P4H− cells, indicating an activated fibroblastic phenotype of FAP+ cells at these time points. At day 3 after MI, a significant number of FAP+ cells did not express P4H. However, using co-expression analysis we found that FAP was only rarely expressed in non-fibroblast cells such as inflammatory cells, endothelial cells or vascular smooth muscle cells (Supplementary data Fig. C.3A), indicating that FAP+/P4H− cells are fibroblasts in an early activation state with low collagen production. This was supported by further analyses to establish whether FAP+ fibroblasts exhibited proto-myofibroblast or
3.3. FAP is expressed in human hearts after acute MI To investigate a potential role for FAP in human hearts after MI, we examined heart tissue sections from individuals with severe heart failure due to acute MI. As reference, tissue from donor organs for heart transplantation was used. Patient characteristics are detailed in Supplementary data Table A.4. Indeed, immunohistochemical staining revealed accumulation of spindle-shaped, FAP expressing cells in areas of acute injury after MI, whereas FAP staining was rarely detected within normal myocardium (Fig. 4A). To establish the phenotype of FAP+ cells after MI in human hearts, we characterized FAP+ cells in human tissues by multi-color fluorescence immunohistochemistry. We identified fibroblasts by spindle-shaped morphology, expression of Thy-1 (CD90) [22] and α-smooth muscle actin, inflammatory cells by expression of CD68, and endothelial cells by expression of CD31 or CD141. In infarcted hearts, abundant FAP+ cells were primarily located in fibrotic areas and surrounding arterioles. FAP+ cells expressed Thy-1, a marker commonly used for fibroblast identification, in varying degrees (Fig. 4B): Using quantitative coexpression analysis, the fraction of Thy-1+ expression in FAP+ cells varied from 16 to 63% among patients (Supplementary Fig. C.4), demonstrating that FAP is expressed both in Thy-1+ and Thy-1− fibroblast populations. Only a minor part of interstitial spindle-shaped FAP+ cells were mature myofibroblasts as identified by co-expression of SMA. Of note, expression of FAP in inflammatory cells, endothelial cells or vascular smooth muscle cells was rare (Supplementary data Fig. C.3B). These results show that the FAP+ cell population in hearts of patients with ischemic cardiomyopathy is heterogeneous, consisting of both Thy-1+ and Thy-1− spindle-shaped fibroblasts and to a minor part of SMA+ myofibroblasts. 3.4. FAP expression in cardiac fibroblasts is stimulated by TGFβ1 via canonical SMAD2/3 signaling We then investigated the signaling pathways leading to expression of FAP in cardiac fibroblasts. TGFβ1, a known mediator of wound healing
Fig. 3. Characterization of the FAP+ cell population after MI. (A) Quantitative co-localization analysis showed a marked increase of cells within the border zone after MI. A major part of these cells consisted of activated P4H+ fibroblasts, especially at 7 days after MI. The majority of FAP+ cells co-expressed P4H on days 7 and 28 after MI, indicating an activated fibroblastic phenotype. A significant amount of FAP+/P4H− cells was only present at day 3 after MI, primarily comprising P4H− fibroblasts. FAP+/P4H− non-fibroblasts were detected very rarely at all time points. (B) The FAP+ cell population consisted largely of P4H+ proto-myofibroblasts (37 ± 3%, 61 ± 4% and 65 ± 3%) and to a lesser extent of SMA+ differentiated myofibroblasts (24 ± 1%, 30 ± 7%, 17 ± 6% at 3, 7 and 28 days after MI, respectively). Sham: n = 4, MI 3 days: n = 3, MI 7 days: n = 3, MI 28 days: n = 4; *: p b 0.05 vs. Sham.
198
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
and tissue fibrosis after MI, stimulated FAP concentration in fibroblasts in a time- and dose-dependent manner (Fig. 5A, B). Next, we investigated potential signaling pathways mediating TGFβ1-induced FAP expression. The increase in FAP expression by treatment with TGFβ1 was abolished using SB431542, a potent inhibitor of TGFβ-receptor I signaling (Fig. 5C). Treatment with SB431542 resulted in the inhibition of SMAD2 and SMAD3 phosphorylation, indicating activation of FAP by the canonical SMAD pathway (Fig. 5D, E). Based on this result, we used siRNA to knock down expression of SMAD2 and SMAD3. Treatment of SMAD2- and SMAD3-deficient fibroblasts with 2 ng/ml TGFβ1 resulted in markedly reduced levels of phosphorylated forms of SMAD2 and SMAD3, together with a blunted increase of FAP in both cases (Fig. 5F–H). These results demonstrated that induction of FAP by TGFβ1 is dependent on phosphorylation of SMAD2 and SMAD3 via activation of the TGFβ-receptor I. In contrast, inhibition of the ERK/MAPK pathway with U0126 did not alter FAP expression after treatment with TGFβ1 (Fig. 5C–E). 3.5. FAP exerts gelatinolytic activity and mediates migration in cardiac fibroblasts Because FAP possesses gelatinolytic activity, we analyzed its ability to degrade gelatin by zymography. Cardiac fibroblasts exhibited a marked gelatinolytic activity at molecular weight 170 kDa, corresponding to the dimerized and enzymatically active form of FAP. Since FAP is a serine-protease, this gelatinolytic activity persisted when matrixmetalloproteinase activity was blocked by addition of EDTA. As expected, treatment of cardiac fibroblasts with 2 ng/ml TGFβ1 resulted in increased gelatinolytic activity (Fig. 6A). A second gelatinolytic band was identified as matrix metalloproteinase 2 (MMP-2), a well-known gelatinase of the heart. In comparison, gelatinolytic activity exerted by MMP-2 was higher than of FAP. Next, we wanted to gain insight into the functional role of FAP, and used cardiac fibroblasts with siRNA-mediated knockdown for further experiments (Fig. 6B). Migratory capacity of FAP-deficient fibroblasts was significantly reduced by −30% compared to control cells in a transwell migration assay (Fig. 6C). Complete inhibition of all serine proteases by the broad-band serine-protease inhibitor AEBSF resulted in further reduction of migratory capacity by −69% compared to control cells, supporting an essential role of FAP for fibroblast migration (Fig. 6C). Importantly, proliferation was unaffected by FAP deficiency (Fig. 6D). Finally, gelatin zymography showed loss of FAP-mediated gelatinolytic activity in FAPdeficient cardiac fibroblasts. Together, our results support a role of FAP in extracellular matrix degradation and fibroblast migration after MI. 4. Discussion The invasion of fibroblasts into the injured and surviving myocardium after MI is a hallmark during infarct healing. Still, the mechanisms of fibroblast activation and tissue repair are not fully understood. In this study, we identified the expression of FAP in activated fibroblasts during infarct healing after MI, and showed a role of FAP in fibroblast migration and matrix degradation. 4.1. FAP identifies activated fibroblasts after MI Until now, no marker exclusively identifies fibroblasts, and therefore characterization is established by a combination of protein markers and morphological characteristics [4,23,24].
199
In our study, we used P4H as marker of activated fibroblasts in rat hearts. P4H has been shown to be enriched in α-1 type I collagenproducing cardiac fibroblasts, but not in endothelial cells [25]. In a detailed in vitro analysis, P4H was expressed in dermal fibroblasts, fibrocytes and macrophages, but not monocytes [26]. To ascertain that we primarily identified fibroblasts in our study, we showed coexpression of FAP+ cells with the fibroblast markers P4H, SMA and vimentin in rat hearts, and used Thy-1 as additional maker to identify fibroblast subpopulations in human hearts [22,24]. Importantly, we excluded relevant co-expression of FAP in monocytes, macrophages, endothelial cells and vascular smooth muscle cells. We detected robust FAP expression in fibroblasts within the granulation and scar tissue of healing infarcts after MI, while it was rarely expressed in healthy hearts. Since nearly all FAP expressing cells showed staining for P4H in rat hearts at days 7 and 28 after MI, we conclude that FAP identifies activated, collagen-synthesizing fibroblasts in the scar. The fibroblast phenotype is further supported by demonstrating that the majority of FAP+ cells in human hearts with ischemic cardiomyopathy have a spindle-shaped morphology, show overlap with Thy-1+ fibroblast and SMA+ myofibroblast populations, and do not express endothelial or inflammatory cell markers. In fact, only a negligible part of FAP+ cells were endothelial cells, inflammatory cells or vascular smooth muscle cells in both rat and human heart tissues. Because of the rare occurrence of these non-fibroblast FAP+ cells, we argue that FAP+/P4H− and FAP+/Thy − 1− cells are mostly fibroblasts in an early activation state. This is supported by our data in rat hearts showing a significantly elevated level of FAP+/P4H− cells only at 3 days after MI, which is declining on days 7 and 28 after MI. Since FAP was only expressed at sites of active tissue remodeling, we argue that FAP expression identifies activated, but not resting fibroblasts. Thy-1+ cells were recently shown to identify fibroblasts with increased expression of extracellular-matrix associated genes in the healthy adult mouse heart [24]. However, fibroblasts are a heterogeneous population possessing a large diversity in activity, function and origin, in part depending on their activation state [23,24]. In this regard, Thy-1+ and Thy-1− populations have been described to co-exist [27], and these subsets show functional heterogeneity upon stimulation with TGFβ [28]. In fact, loss of Thy-1 has been associated with progressing fibrosis and fibroblast activation in a mouse model of bleomycin induced lung fibrosis [29], illustrating the dynamic alteration of a fibroblast marker in fibrotic disease. In our study, the expression of FAP was not restricted to differentiated myofibroblasts, and FAP was expressed in P4H+ and P4H− cells in rat hearts as well as Thy-1+ and Thy-1− cells in human hearts after MI. Therefore, expression of FAP might be indicating a pro-fibrotic phenotype in a wide range of fibroblast and fibroblast-like subpopulations including early proto-myofibroblasts and differentiated myofibroblasts. Considering various origins and heterogeneity of fibroblasts on one hand and vast expression of FAP on the other hand, we suggest that FAP is a marker for fibroblast activation, irrespective of their origin and differentiation stage. 4.2. FAP is activated in cardiac fibroblasts by TGFβ1 Activation of fibroblasts by TGFβ1 is a crucial step of wound healing early after MI by enabling the pro-fibrotic phenotype, leading to the formation of a strong extracellular matrix and scar [30]. While cardiomyocyte-specific inhibition of TGFβ signaling improved cardiac remodeling [31], broad inhibition of TGFβ signaling resulted in
Fig. 4. FAP is expressed in human hearts after acute MI. (A) Quantitative analysis by immunohistochemistry showed elevated FAP-expression in hearts of patients with acute MI and severely reduced LV-function (n = 5) as compared to non-failing donor hearts (control, n = 4, P b 0.05). In infarcted hearts, FAP+ cells were surrounding vessels (black arrows) and dispersed in fibrotic tissue (white arrows). In control hearts, FAP+ cells were at times found in fibrotic foci (asterisk). FAP: brown. Scale bars = 100 μM. (B) Heterogeneity of Thy-1expression in FAP+ fibroblasts after acute MI. In infarcted human hearts, areas with abundant predominantly FAP+/Thy-1− cells (left panel, asterisk) or predominantly FAP+/Thy-1+ cells (right panel, white arrows) existed. Left panel: A cluster of FAP−/Thy-1+ cells is shown as reference (black arrow). FAP: red; Thy-1: green; DAPI: blue. Scale bars: 10 μm. Image processing included changes in brightness, contrast and tonal range, and was applied equally across the entire image. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
200
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fig. 5. TGFβ1 induces FAP protein expression in adult human cardiac fibroblasts. (A, B) FAP-expression is stimulated by TGFβ1 in a time- (A) and dose-dependent manner (B). A slight increase in expression of FAP was evident over time in cardiac fibroblasts after N48 h of culture. (C–H) TGFβ1-induced FAP expression is mediated by the canonical TGFβ signaling pathway: Inhibition of TGFβ type I receptor using SB431542 blunted TGFβ1-mediated increase of FAP, while blocking the MAP kinase pathway by U0126 did not alter FAP levels (C). SB431542 efficiently inhibited SMAD2 and SMAD3 phosphorylation (D, E). Moreover, knockdown of SMAD2 and SMAD3 by siRNA in cardiac fibroblasts prevented formation of phospho-SMAD2 (F) and phospho-SMAD3 (G) and blunted increase of FAP expression (H) after stimulation with TGFβ1 in CF. A: * = P b 0.05 vs. 6 h, #: P b 0.05 vs. control; n = 3–4 each time point. B: * = P b 0.05 vs. no TGFβ1; n = 3–6 each time point. C n = 15. D–G n = 3. H n = 6.
increased LV dimensions and higher mortality after MI [32], supporting beneficial effects of TGFβ on non-myocytes, including fibroblasts, after MI. In our study, we show that FAP is highly expressed in fibroblasts within the borderzone after permanent LAD ligation, with a peak at 7 days after MI and decline thereafter. This upregulation of FAP might be mediated at least in part by TGFβ1 in vivo: Temporal and regional expression changes of inflammatory, angiogenic, differentiation and homing related genes after MI have been demonstrated previously. Using laser microdissection to identify infarct scar and borderzone, permanent LAD ligation in mice resulted in increased expression of various genes including TGFß1 within the first hours, with a peak between 2 and 7 days after MI and decrease thereafter [33]. Of note, TGFβ1 protein
expression was highest in the borderzone in this study. This corresponds to another report using a rat model of MI, demonstrating that maximum expression of TGFβ1 mRNA and TGFβ1 protein was detected within days 1 to 7 after MI, together with a peak of procollagen α1(I) at 3 to 7 days after MI [34]. Together, these previous reports provide evidence of TGFβ1 expression in the border zone within the first days after MI in mice and rats, which corresponds to our findings of FAP expression after MI in rats and humans. TGFβ1 exerts its effects on extracellular matrix production mainly by activation of proteases via the canonical SMAD2/3 cascade [30]. Indeed, our data shows that FAP is induced in cardiac fibroblasts by TGFβ1 via TGFβ receptor I, SMAD2 and SMAD3, suggesting that FAP contributes to TGFβ1-mediated effects in wound healing after MI.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
201
Fig. 6. FAP exerts gelatinolytic activity and promotes migration in adult human cardiac fibroblasts. (A) Gelatin zymography demonstrated gelatinolytic activity of FAP derived from cardiac fibroblasts, which was further increased after treatment with TGFβ1. FAP gelatinolytic activity persisted after blocking matrix metalloproteinase activity with EDTA. (B) Efficient knockdown of FAP in cardiac fibroblasts using FAP siRNA as compared to scrambled control siRNA was verified by Western blot and ELISA. (C) Cell migration was reduced in FAP-deficient cells, and even further reduced by adding serine protease blocker AEBSF. (D) Control and FAP-deficient fibroblasts showed no difference in proliferation index. (E) Gelatinolytic activity of FAP is blunted in FAP-deficient cells. A: n = 2. B: n = 6. C: n = 3–9. D: n = 12. E: n = 5.
202
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
4.3. FAP exerts gelatinolytic activity and mediates fibroblast migration Our study demonstrated gelatinolytic activity of cardiac fibroblastderived FAP. Proteases are an essential part for the homeostasis of cardiac extracellular matrix and various proteases have been shown to be of crucial importance for healing after MI. Matrix metalloproteinases including MMP2, MMP9, MMP14 and MMP28 have been extensively studied and shown to affect myocardial post-MI remodeling by matrix degradation as well as proteolytic activation of cytokines [35–38]. Similar effects were shown for serine proteases in post-MI remodeling: Heymans et al. described altered scar formation and impaired remodeling in animals with deficiency of the serine proteases tissue-type plasminogen activator and urokinase-type plasminogen activator [39]. Cell surface proteolysis represents an important mechanism for cell migration. In fact, we found impaired fibroblast migration in FAP deficient cells. This is in concordance with previous reports demonstrating a role for migration in endothelial cells [40], embryonic fibroblasts [41], hepatic stellate cells [42] and bone marrow mesenchymal stem cells [43]. Together, we show that cardiac fibroblast-derived FAP has gelatinolytic activity and mediates fibroblast migration, which may contribute to extracellular matrix scar formation and ultimately remodeling after MI. Apart from its gelatinolytic activity, other enzymatic substrates of FAP such as α2-antiplasmin, neuropeptide Y, B-type natriuretic peptide, substance P and peptide YY are known so far, and other yet unknown substrates probably exist [44]. Also, gelatinolytic activity of FAP was much less as compared to MMP2, suggesting that FAP has other physiologic roles in post-MI remodeling. Whether FAP itself is a therapeutic target remains unknown, and further studies on the effects of FAP on wound healing and cardiac remodeling after MI are warranted. Apart from its potential pathophysiological roles, FAP might be suitable as a specific marker for activated fibroblasts in cardiac imaging. Molecular imaging of myocardial fibrosis has been attempted using various targets, including collagen fibers, myofibroblasts, matricellular proteins and matrix metalloproteinases [45]. Activated fibroblasts are the major source of extracellular matrix after MI, and myocardial fibrosis is a major determinant of left ventricular dysfunction. The development of FAP-specific imaging probes for noninvasive assessment of activated fibroblasts could help in evaluating the individual extent and progression of fibrosis, and therefore aid in antifibrotic treatments. In this regard, first studies using fluorochrome-tagged FAP substrates for in vivo imaging of tumor fibroblasts have been reported in small animal models [46], and a FAP antibody was used to visualize arthritic joints in murine experimental arthritis by radionuclide imaging [47].
4.4. Summary In this study, we characterized for the first time the expression and phenotype of FAP+ cells in the healing wound after MI. The identification of FAP in activated, but not resting fibroblasts in rat and human tissues after MI will help to further delineate the role of fibroblasts in inflammation and wound healing, and holds promise for further evaluation as target for drug treatments and molecular imaging to better understand and improve cardiac remodeling after MI. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yjmcc.2015.08.016.
Acknowledgments We gratefully acknowledge Annette Berbner, Annemieke Klan and Silke Pretzer for expert technical assistance. We thank Dr. Länger for providing archival myocardial tissue samples for immunohistochemistry. This work was initially supported by grants from the Interdisciplinary Clinical Research Center Würzburg (to J.T., J.B.: IZKF E-140).
Disclosures None. References [1] D. Fraccarollo, P. Galuppo, J. Bauersachs, Novel therapeutic approaches to postinfarction remodelling, Cardiovasc. Res. 94 (2012) 293–303. [2] D. Fraccarollo, P. Galuppo, S. Motschenbacher, H. Ruetten, A. Schafer, J. Bauersachs, Soluble guanylyl cyclase activation improves progressive cardiac remodeling and failure after myocardial infarction. Cardioprotection over ACE inhibition, Basic Res. Cardiol. 109 (2014) 421. [3] S.W. van den Borne, J. Diez, W.M. Blankesteijn, J. Verjans, L. Hofstra, J. Narula, Myocardial remodeling after infarction: the role of myofibroblasts, Nat. Rev. Cardiol. 7 (2010) 30–37. [4] E.C. Goldsmith, A.D. Bradshaw, M.R. Zile, F.G. Spinale, Myocardial fibroblast-matrix interactions and potential therapeutic targets, J. Mol. Cell. Cardiol. 70 (2014) 92–99. [5] J.J. Tomasek, G. Gabbiani, B. Hinz, C. Chaponnier, R.A. Brown, Myofibroblasts and mechano-regulation of connective tissue remodelling, Nat. Rev. Mol. Cell Biol. 3 (2002) 349–363. [6] A. Leask, Potential therapeutic targets for cardiac fibrosis: TGFbeta, angiotensin, endothelin, CCN2, and PDGF, partners in fibroblast activation, Circ. Res. 106 (2010) 1675–1680. [7] T. Thum, C. Gross, J. Fiedler, T. Fischer, S. Kissler, M. Bussen, et al., MicroRNA-21 contributes to myocardial disease by stimulating MAP kinase signalling in fibroblasts, Nature 456 (2008) 980–984. [8] J.E. Park, M.C. Lenter, R.N. Zimmermann, P. Garin-Chesa, L.J. Old, W.J. Rettig, Fibroblast activation protein, a dual specificity serine protease expressed in reactive human tumor stromal fibroblasts, J. Biol. Chem. 274 (1999) 36505–36512. [9] P. Garin-Chesa, L.J. Old, W.J. Rettig, Cell surface glycoprotein of reactive stromal fibroblasts as a potential antibody target in human epithelial cancers, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 7235–7239. [10] H. Dolznig, N. Schweifer, C. Puri, N. Kraut, W.J. Rettig, D. Kerjaschki, et al., Characterization of cancer stroma markers: in silico analysis of an mRNA expression database for fibroblast activation protein and endosialin, Cancer Immun. 5 (2005) 10. [11] J. Niedermeyer, M. Kriz, F. Hilberg, P. Garin-Chesa, U. Bamberger, M.C. Lenter, et al., Targeted disruption of mouse fibroblast activation protein, Mol. Cell. Biol. 20 (2000) 1089–1094. [12] M. Kraman, P.J. Bambrough, J.N. Arnold, E.W. Roberts, L. Magiera, J.O. Jones, et al., Suppression of antitumor immunity by stromal cells expressing fibroblast activation protein-alpha, Science 330 (2010) 827–830. [13] P.S. Acharya, A. Zukas, V. Chandan, A.L. Katzenstein, E. Pure, Fibroblast activation protein: a serine protease expressed at the remodeling interface in idiopathic pulmonary fibrosis, Hum. Pathol. 37 (2006) 352–360. [14] M.T. Levy, G.W. McCaughan, G. Marinos, M.D. Gorrell, Intrahepatic expression of the hepatic stellate cell marker fibroblast activation protein correlates with the degree of fibrosis in hepatitis C virus infection, Liver 22 (2002) 93–101. [15] S. Bauer, M.C. Jendro, A. Wadle, S. Kleber, F. Stenner, R. Dinser, et al., Fibroblast activation protein is expressed by rheumatoid myofibroblast-like synoviocytes, Arthritis Res. Ther. 8 (2006) R171. [16] D. Fraccarollo, P. Galuppo, S. Schraut, S. Kneitz, N. van Rooijen, G. Ertl, et al., Immediate mineralocorticoid receptor blockade improves myocardial infarct healing by modulation of the inflammatory response, Hypertension 51 (2008) 905–914. [17] J. Tillmanns, M. Rota, T. Hosoda, Y. Misao, G. Esposito, A. Gonzalez, et al., Formation of large coronary arteries by cardiac progenitor cells, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 1668–1673. [18] J. Tillmanns, M. Schneider, D. Fraccarollo, J.D. Schmitto, F. Langer, D. Richter, et al., PET imaging of cardiac wound healing using a novel [68Ga]-labeled NGR probe in rat myocardial infarction, Mol. Imaging Biol. 17 (2015) 76–86. [19] J. Tillmanns, C. Widera, Y. Habbaba, P. Galuppo, T. Kempf, K.C. Wollert, et al., Circulating concentrations of fibroblast activation protein alpha in apparently healthy individuals and patients with acute coronary syndrome as assessed by sandwich ELISA, Int. J. Cardiol. 168 (2013) 3926–3931. [20] D. Fraccarollo, J. Bauersachs, M. Kellner, P. Galuppo, G. Ertl, Cardioprotection by long-term ET(A) receptor blockade and ACE inhibition in rats with congestive heart failure: mono-versus combination therapy, Cardiovasc. Res. 54 (2002) 85–94. [21] K. Aertgeerts, I. Levin, L. Shi, G.P. Snell, A. Jennings, G.S. Prasad, et al., Structural and kinetic analysis of the substrate specificity of human fibroblast activation protein alpha, J. Biol. Chem. 280 (2005) 19441–19444. [22] F. Hudon-David, F. Bouzeghrane, P. Couture, G. Thibault, Thy-1 expression by cardiac fibroblasts: lack of association with myofibroblast contractile markers, J. Mol. Cell. Cardiol. 42 (2007) 991–1000. [23] G. Krenning, E.M. Zeisberg, R. Kalluri, The origin of fibroblasts and mechanism of cardiac fibrosis, J. Cell. Physiol. 225 (2010) 631–637. [24] S.R. Ali, S. Ranjbarvaziri, M. Talkhabi, P. Zhao, A. Subat, A. Hojjat, et al., Developmental heterogeneity of cardiac fibroblasts does not predict pathological proliferation and activation, Circ. Res. 115 (2014) 625–635. [25] T. Moore-Morris, N. Guimaraes-Camboa, I. Banerjee, A.C. Zambon, T. Kisseleva, A. Velayoudon, et al., Resident fibroblast lineages mediate pressure overload-induced cardiac fibrosis, J. Clin. Invest. 124 (2014) 2921–2934. [26] D. Pilling, T. Fan, D. Huang, B. Kaul, R.H. Gomer, Identification of markers that distinguish monocyte-derived fibrocytes from monocytes, macrophages, and fibroblasts, PLoS One 4 (2009), e7475.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203 [27] L. Koumas, A.E. King, H.O. Critchley, R.W. Kelly, R.P. Phipps, Fibroblast heterogeneity: existence of functionally distinct Thy 1(+) and Thy 1(−) human female reproductive tract fibroblasts, Am. J. Pathol. 159 (2001) 925–935. [28] L. Koumas, T.J. Smith, S. Feldon, N. Blumberg, R.P. Phipps, Thy-1 expression in human fibroblast subsets defines myofibroblastic or lipofibroblastic phenotypes, Am. J. Pathol. 163 (2003) 1291–1300. [29] J.S. Hagood, P. Prabhakaran, P. Kumbla, L. Salazar, M.W. MacEwen, T.H. Barker, et al., Loss of fibroblast Thy-1 expression correlates with lung fibrogenesis, Am. J. Pathol. 167 (2005) 365–379. [30] M. Bujak, N.G. Frangogiannis, The role of TGF-beta signaling in myocardial infarction and cardiac remodeling, Cardiovasc. Res. 74 (2007) 184–195. [31] P.P. Rainer, S. Hao, D. Vanhoutte, D.I. Lee, N. Koitabashi, J.D. Molkentin, et al., Cardiomyocyte-specific transforming growth factor beta suppression blocks neutrophil infiltration, augments multiple cytoprotective cascades, and reduces early mortality after myocardial infarction, Circ. Res. 114 (2014) 1246–1257. [32] S. Frantz, K. Hu, A. Adamek, J. Wolf, A. Sallam, S. Maier, et al., Transforming growth factor beta inhibition increases mortality and left ventricular dilatation after myocardial infarction, Basic Res. Cardiol. 103 (2008) 485–492. [33] S. Vandervelde, M.J. van Luyn, M.H. Rozenbaum, A.H. Petersen, R.A. Tio, M.C. Harmsen, Stem cell-related cardiac gene expression early after murine myocardial infarction, Cardiovasc. Res. 73 (2007) 783–793. [34] R.G. Dean, L.C. Balding, R. Candido, W.C. Burns, Z. Cao, S.M. Twigg, et al., Connective tissue growth factor and cardiac fibrosis after myocardial infarction, J. Histochem. Cytochem. 53 (2005) 1245–1256. [35] A. Ducharme, S. Frantz, M. Aikawa, E. Rabkin, M. Lindsey, L.E. Rohde, et al., Targeted deletion of matrix metalloproteinase-9 attenuates left ventricular enlargement and collagen accumulation after experimental myocardial infarction, J. Clin. Invest. 106 (2000) 55–62. [36] S. Hayashidani, H. Tsutsui, M. Ikeuchi, T. Shiomi, H. Matsusaka, T. Kubota, et al., Targeted deletion of MMP-2 attenuates early LV rupture and late remodeling after experimental myocardial infarction, Am. J. Physiol. Heart Circ. Physiol. 285 (2003) H1229–H1235. [37] F.G. Spinale, R. Mukherjee, J.A. Zavadzkas, C.N. Koval, S. Bouges, R.E. Stroud, et al., Cardiac restricted overexpression of membrane type-1 matrix metalloproteinase causes adverse myocardial remodeling following myocardial infarction, J. Biol. Chem. 285 (2010) 30316–30327.
203
[38] Y. Ma, G.V. Halade, J. Zhang, T.A. Ramirez, D. Levin, A. Voorhees, et al., Matrix metalloproteinase-28 deletion exacerbates cardiac dysfunction and rupture after myocardial infarction in mice by inhibiting M2 macrophage activation, Circ. Res. 112 (2013) 675–688. [39] S. Heymans, A. Luttun, D. Nuyens, G. Theilmeier, E. Creemers, L. Moons, et al., Inhibition of plasminogen activators or matrix metalloproteinases prevents cardiac rupture but impairs therapeutic angiogenesis and causes cardiac failure, Nat. Med. 5 (1999) 1135–1142. [40] G. Ghersi, Q. Zhao, M. Salamone, Y. Yeh, S. Zucker, W.T. Chen, The protease complex consisting of dipeptidyl peptidase IV and seprase plays a role in the migration and invasion of human endothelial cells in collagenous matrices, Cancer Res. 66 (2006) 4652–4661. [41] G. Ghersi, H. Dong, L.A. Goldstein, Y. Yeh, L. Hakkinen, H.S. Larjava, et al., Regulation of fibroblast migration on collagenous matrix by a cell surface peptidase complex, J. Biol. Chem. 277 (2002) 29231–29241. [42] X.M. Wang, D.M. Yu, G.W. McCaughan, M.D. Gorrell, Fibroblast activation protein increases apoptosis, cell adhesion, and migration by the LX-2 human stellate cell line, Hepatology 42 (2005) 935–945. [43] K.M. Chung, S.C. Hsu, Y.R. Chu, M.Y. Lin, W.T. Jiaang, R.H. Chen, et al., Fibroblast activation protein (FAP) is essential for the migration of bone marrow mesenchymal stem cells through RhoA activation, PLoS One 9 (2014), e88772. [44] F.M. Keane, N.A. Nadvi, T.W. Yao, M.D. Gorrell, Neuropeptide Y, B-type natriuretic peptide, substance P and peptide YY are novel substrates of fibroblast activation protein-alpha, FEBS J. 278 (2011) 1316–1332. [45] H.J. de Haas, E. Arbustini, V. Fuster, C.M. Kramer, J. Narula, Molecular imaging of the cardiac extracellular matrix, Circ. Res. 114 (2014) 903–915. [46] R. Ruger, F.L. Tansi, M. Rabenhold, F. Steiniger, R.E. Kontermann, A. Fahr, et al., In vivo near-infrared fluorescence imaging of FAP-expressing tumors with activatable FAP-targeted, single-chain Fv-immunoliposomes, J. Control. Release 186 (2014) 1–10. [47] P. Laverman, T. van der Geest, S.Y. Terry, D. Gerrits, B. Walgreen, M.M. Helsen, et al., Immuno-PET and immuno-SPECT of rheumatoid arthritis with radiolabeled antifibroblast activation protein antibody correlates with severity of arthritis, J. Nucl. Med. 56 (2015) 778–783.
Contents lists available at ScienceDirect
Journal of Molecular and Cellular Cardiology journal homepage: www.elsevier.com/locate/yjmcc
Fibroblast activation protein alpha expression identifies activated fibroblasts after myocardial infarction☆ Jochen Tillmanns a,⁎, Daniel Hoffmann b,1, Yasmin Habbaba a,2, Jan D. Schmitto c, Daniel Sedding a, Daniela Fraccarollo a, Paolo Galuppo a, Johann Bauersachs a a b c
Department of Cardiology and Angiology, Hannover Medical School, 30625 Hannover, Germany Department of Medicine, University Hospital Wurzburg, 97080 Wurzburg, Germany Department of Cardiothoracic Surgery, Hannover Medical School, 30625 Hannover, Germany
a r t i c l e
i n f o
Article history: Received 16 April 2015 Received in revised form 3 August 2015 Accepted 20 August 2015 Available online 28 August 2015 Keywords: Fibroblast activation protein α Inflammation Myofibroblast Wound healing Heart Myocardial infarction
a b s t r a c t Introduction: Fibroblast activation protein α (FAP) is a membrane-bound serine protease expressed by activated fibroblasts during wound healing in the skin. Expression of FAP after myocardial infarction (MI) and potential effects on cardiac wound healing are largely unknown. Methods: MI was induced in rats and FAP expression was analyzed at 3, 7 and 28 days post-MI by microarray, Western blot and immunohistochemistry. In human hearts after MI, a FAP+ fibroblast population was identified, and characterized by immunohistochemistry for prolyl-4-hydroxylase β, α-smooth muscle actin, Thy-1 and vimentin. Signaling pathways leading to FAP expression were studied in human cardiac fibroblasts by Western blot and ELISA using TGFβ1, TGF-beta type I-receptor (TGFbR1)-inhibitor SB431542 or the MAPK-inhibitor U0126 as well as siRNA targeting SMAD2 and SMAD3. Finally, fibroblasts were assayed for FAP-dependent migration (modified Boyden-chamber), proliferation (BrdU-assay) and gelatinolytic activity by gelatin zymography. Results: In rats, FAP expression was increased after MI especially in the peri-infarct area peaking at 7 days post-MI. Co-localization analysis identified the majority of FAP+ cells as activated proto-myofibroblasts and myofibroblasts. Concordantly, FAP+ fibroblasts were abundant in ischemic tissue of human hearts after MI, but not in healthy control hearts. In vitro, FAP was induced by TGFβ1 via the canonical SMAD2/SMAD3 pathway. Depletion of FAP in fibroblasts reduced migratory capacity, while proliferation was not affected. Gelatin zymography revealed gelatinase activity by fibroblast-derived FAP. Conclusion: In this study, we show for the first time the expression of FAP in activated fibroblasts after MI and its activation by TGFβ1. Effects of FAP on fibroblast migration and gelatinolytic activity indicate a potential role in cardiac wound healing and remodeling. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Cardiac fibroblasts are central mediators of the reparative response after myocardial infarction (MI) [1,2]. Following MI, fibroblasts are activated, resulting in dynamic phenotypic changes and differentiation into collagen-secreting proto-myofibroblasts, mediated by various Abbreviations: FAP, fibroblast activation protein α; APCE, antiplasmin-cleaving enzyme; rh, recombinant human; DPPIV, dipeptidyl peptidase IV; MI, myocardial infarction; LV, left ventricular; CF, cardiac fibroblast. ☆ The authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors declare no conflict of interest. ⁎ Corresponding author at: Department of Cardiology and Angiology, Hannover Medical School, Carl-Neuberg-Str. 1, 30625 Hannover, Germany. E-mail address: [email protected] (J. Tillmanns). 1 Present address: Department of Trauma and Reconstructive Surgery, University Hospital, 37099 Goettingen, Germany. 2 Present address: Department of Psychiatry and Psychotherapy, University Hospital Wurzburg, 97080 Wurzburg, Germany.
http://dx.doi.org/10.1016/j.yjmcc.2015.08.016 0022-2828/© 2015 Elsevier Ltd. All rights reserved.
factors including mechanical stress and TGFβ-signaling [3,4]. Protomyofibroblasts express β- and γ-cytoplasmic actins, and can further differentiate into mature myofibroblasts, characterized by the expression of α-smooth muscle actin (SMA) [3,5]. Myofibroblasts migrate into the injured tissue and secrete increased amounts of growth factors, cytokines, as well as peri-cellular proteases and their inhibitors, thereby regulating the composition of the extracellular matrix and promoting healing of the infarct scar [6]. However, persistence of myofibroblasts within the surviving myocardium contributes to detrimental cardiac fibrosis, leading to systolic and diastolic heart failure by decreased contraction and increased cardiac muscle stiffness [3]. Because of their central role in cardiac remodeling, fibroblasts are under investigation as therapeutic targets in ischemic cardiomyopathy [7]. Beneficial effects of fibroblasts in early infarct healing and detrimental effects of excessive amounts of fibroblasts in uninjured tissue require careful timing for safe and effective anti-fibrotic treatments. Therefore, a better understanding of the tempro-spatial activation and function of fibroblasts within the heart after MI is needed.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fibroblast activation protein α (FAP, also known as Seprase) is a type II integral membrane glycoprotein with dipeptidyl-peptidase and type I collagenase activity [8]. FAP is transiently expressed in some fetal mesenchymal tissues, but is absent or expressed at low levels in most adult tissues under baseline conditions [9,10]. Elevated production of FAP is associated with active tissue remodeling in mesenchymal tissues during embryogenesis [11], cancer [9,12], idiopathic pulmonary fibrosis [13], liver cirrhosis [14] and rheumatoid arthritis [15]. Of note, FAPexpressing fibroblasts have been detected in healing wounds [9], and its expression level correlates with fibrosis severity in liver cirrhosis [14]. Since FAP is expressed at sites of active tissue remodeling, we hypothesized that it is expressed in activated cardiac fibroblasts, and contributes to healing and matrix turnover of the infarcted tissue after MI. 2. Materials and methods Additional materials and methods are presented as Supplementary online material. 2.1. Experimental myocardial infarction in rats Myocardial infarction (MI) was induced in male Wistar rats weighing 200–250 g, as described previously [16,17]. Briefly, under 1.5–2% isoflurane anesthesia (induction with 5% isoflurane), the thorax was opened and the proximal left anterior descending coronary artery was occluded using a 5–0 suture. Sham-operated control rats underwent the same surgical procedure except that the suture around the coronary artery was not tied. Animals were kept warm with a heating pad. Depth of anesthesia was tested using the pedal withdrawal reflex. Analgesia was maintained using buprenorphine (0.05 mg/kg BW i.p.). Before and after surgery, animals were housed in the animal facility and monitored daily for activity and signs of pain. Animals were euthanized and hearts removed at time points 3 days (n = 7), 7 days (n = 7) and 28 days after MI (n = 8), and from sham-operated animals (n = 6). Animal studies were conducted in accordance with the principles and procedures outlined in the Guide for the Care and Use of Laboratory Animals and were approved by the local government (Regierung Unterfranken AZ: K 55.2-2531.01-6409). 2.2. Immunohistochemistry of rat and human myocardial tissue Frozen sections from rat hearts were stained with antibodies against FAP, α-smooth muscle actin (SMA), Ki67, prolyl-4-hydroxylase β (P4H), CD31, CD68, CD11b/c, and vimentin, and visualized by fluorescent labeling as described previously [18] and in Appendix B. Quantitative analysis of staining for FAP, SMA and P4H was analyzed in rat hearts at 3 days (n = 3), 7 days (n = 3) and 28 days (n = 4) after MI, and in sham-operated animals (n = 4). Pictures of the infarct borderzone area were taken at 63× magnification using a Zeiss Axiovert microscope with Axiovison 4.0 software (Carl Zeiss MicroImaging). Counting was performed by two independent individuals using Image Pro Plus software (Media Cybernetics, Bethesda, USA), and results were averaged. FAP expression was evaluated in paraffin-embedded formalin-fixed tissue from the left ventricular apex obtained from discarded tissue of patients receiving a left ventricular assist device (LVAD) due to acute MI (n = 5) as described previously [18] and in Appendix B. Biopsies obtained from non-failing donor hearts served as controls (n = 4). The number of FAP+ cells was quantified in each sample using large images taken at 10 × magnification. To minimize sampling error, quantified areas ranged from 2.8 to 9.8 mm2, depending on the number of positive cells. The study was approved by the ethics committee of Hannover Medical School and conforms to the ethical guidelines of the 1975 Declaration of Helsinki.
195
2.3. In vitro studies with cardiac fibroblasts Adult human cardiac fibroblasts (CFs) were obtained commercially (Cell Applications, no. 306-05a) and cultured according to the manufacturer instructions. FAP concentrations in cell lysates were quantified using a commercial ELISA (R&D Systems) according to manufacturer instructions and as described previously [19] and in Appendix B. For FAP knockdown experiments, CF were transfected with siRNA against FAP, SMAD2 or SMAD3 (50 nM, Ambion) using oligofectamine or lipofectamine RNAiMAX reagent (Invitrogen). The efficiency of FAP silencing was documented 4 days later. Fibroblasts transfected with scrambled siRNA (Ambion) were used as control in all studies. Before the experiments, cells were starved in serum-free medium consisting of DMEM high glucose (Lonza) supplemented with 10 mM HEPES and 0.2% bovine serum albumin (Sigma-Aldrich) for 24 h. Cells were studied in passages 3 to 6. For stimulation experiments, fibroblasts at 70% confluence were supplemented with fresh serum-free medium containing various concentrations of recombinant human TGFβ1 (R&D Systems) and cultured for 48 h to 72 h. In selected experiments, the TGFβ-receptor-inhibitor SB431542 (10 μM, Sigma-Aldrich) or the MAP kinase inhibitor U0126 (10 μM, Sigma-Aldrich) was added to the culture prior to the addition of TGFβ1. Cultured fibroblasts were analyzed for proliferation and migration as described in Appendix B. 2.4. Gelatin zymography Fibroblasts were homogenized in ice-cold cacodylic acid buffer, mixed with loading buffer and electrophoresis was performed under nonreducing conditions on a 10% SDS polyacrylamide gel containing 1 mg/ml of gelatin as described previously [20]. Zymography was also performed with EDTA to inhibit Ca++-dependent matrix-metalloproteinase activity. FAP gelatinolytic activity resulted in a band at 170 kDa, corresponding to the enzymatically active FAP dimer [21]. Gelatinolytic bands were quantified by densitometry using NIH ImageJ software. 2.5. Statistics Data are expressed as mean ± SE. We used the unpaired Student's t test for analysis of differences between two groups, and one-way ANOVA with multiple comparisons for analysis of differences between more than two groups. A value of P less than 0.05 was considered statistically significant. Statistical analysis was performed with Prism 5 (GraphPad). 3. Results 3.1. Expression levels of FAP are increased after MI Analysis of gene expression by microarray was performed in the ischemic tissue of rat hearts explanted 3 and 7 days after MI. From a total of 31,099 genes, 1287 genes (4%) were induced N2 fold at 3 days after MI, and 878 genes (3%) were induced N 2 fold at 7 days after MI,
Fig. 1. A marked increase of FAP protein expression was present in the left ventricular ischemic tissue and scar until day 28 after MI. FAP expression after MI was maximal in scar and border zone at 7 days after MI, and persisted within the infarct scar until at least 28 days after MI. In contrast, FAP expression was low within non-infarcted interventricular septum of the same animals. MI 3 days, n = 2; MI 7 days, n = 2; MI 28 days, n = 2.
196
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fig. 2. (A) Immunohistological analysis revealed abundant FAP+ fibroblasts infiltrating the infarcted myocardium especially in the border zone. In hearts of sham-operated animals, FAP+ cells were scarce. Abundant co-localization of FAP and P4H was present, indicating increased collagen synthesis of FAP+ cells. (B) FAP+ fibroblasts have an activated fibroblast phenotype, as indicated by co-localization with P4H and vimentin. A part of FAP+ cells also co-localized with SMA, identifying them as differentiated myofibroblasts. FAP: red, P4H, Vimentin, SMA: green, DAPI: blue. Scale bars: A: 10 μm, B: 100 μm. Image processing included changes in brightness, contrast and tonal range, and was applied equally across the entire image. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
compared to expression levels in hearts of sham-operated animals. Thirty six of the induced genes were classified as proteolytic, including FAP (Supplementary data Table A.1). FAP gene expression was increased 2.0 ± 0.1 fold at 3 days and 2.8 ± 0.0 fold at 7 days after MI, respectively. These results were confirmed on the protein level using Western blot analysis of rat hearts at 3, 7 or 28 days after MI (Fig. 1).
197
myofibroblast phenotypes: We analyzed the frequency of SMA+ expression as a marker indicating differentiation of fibroblasts into mature myofibroblasts. The majority of FAP+ cells showed strong co-expression of P4H at days 7 and 28 after MI, while a minor part of FAP+ fibroblasts co-expressed SMA, indicating that the FAP+ fibroblast subpopulation consisted predominantly of proto-myofibroblasts and to a minor part of myofibroblasts (Fig. 3B).
3.2. FAP is expressed in activated fibroblasts after MI Morphometric analysis demonstrated induction of large myocardial infarctions in rat hearts. Infarct sizes were 44 ± 3%, 52 ± 3% and 44 ± 5% on days 3, 7 and 28 after MI, respectively (Supplementary data Fig. C.1). To characterize the expression of FAP in cardiac cells, we performed multi-color fluorescence immunohistochemistry analyses in infarcted rat hearts. On days 3 and 7 after MI, FAP was increasingly expressed in cells located within the border zone of the infarcted area. While there were few FAP+ cells in the necrotic area on day 3 and day 7 after MI, FAP+ cells were frequently detected clustered within the scar tissue 28 days after MI (Fig. 2A). Only few FAP+ cells were found in the remote area and interventricular septum. In myocardium of sham-operated animals, FAP+ cells were only rarely found within connective tissue. We then aimed to further characterize the phenotype of FAP+ cells. We used prolyl-4-hydroxylase β (P4H) as marker for activated, collagen-synthesizing fibroblasts, and α-smooth muscle actin (SMA) as marker to identify differentiated, mature myofibroblasts. The fibroblast-phenotype was further supported by expression of vimentin. After MI, FAP+ cells were indeed identified as activated fibroblasts, as evidenced by co-expression of FAP with P4H, SMA and vimentin (Fig. 2B). FAP+ activated fibroblasts were actively replicating, indicated by expression of Ki67 (Supplementary data Fig. C.2). Next, we used P4H and FAP co-expression analysis to quantify the fraction of activated, collagen-synthesizing FAP+ cells after MI. The density of activated P4H+ fibroblasts peaked on day 7 and declined thereafter, and the density of FAP+ cells showed a similar time course (Fig. 3A). On days 7 and 28 after MI, the majority of FAP+ cells co-expressed P4H while there were few FAP+/P4H− cells, indicating an activated fibroblastic phenotype of FAP+ cells at these time points. At day 3 after MI, a significant number of FAP+ cells did not express P4H. However, using co-expression analysis we found that FAP was only rarely expressed in non-fibroblast cells such as inflammatory cells, endothelial cells or vascular smooth muscle cells (Supplementary data Fig. C.3A), indicating that FAP+/P4H− cells are fibroblasts in an early activation state with low collagen production. This was supported by further analyses to establish whether FAP+ fibroblasts exhibited proto-myofibroblast or
3.3. FAP is expressed in human hearts after acute MI To investigate a potential role for FAP in human hearts after MI, we examined heart tissue sections from individuals with severe heart failure due to acute MI. As reference, tissue from donor organs for heart transplantation was used. Patient characteristics are detailed in Supplementary data Table A.4. Indeed, immunohistochemical staining revealed accumulation of spindle-shaped, FAP expressing cells in areas of acute injury after MI, whereas FAP staining was rarely detected within normal myocardium (Fig. 4A). To establish the phenotype of FAP+ cells after MI in human hearts, we characterized FAP+ cells in human tissues by multi-color fluorescence immunohistochemistry. We identified fibroblasts by spindle-shaped morphology, expression of Thy-1 (CD90) [22] and α-smooth muscle actin, inflammatory cells by expression of CD68, and endothelial cells by expression of CD31 or CD141. In infarcted hearts, abundant FAP+ cells were primarily located in fibrotic areas and surrounding arterioles. FAP+ cells expressed Thy-1, a marker commonly used for fibroblast identification, in varying degrees (Fig. 4B): Using quantitative coexpression analysis, the fraction of Thy-1+ expression in FAP+ cells varied from 16 to 63% among patients (Supplementary Fig. C.4), demonstrating that FAP is expressed both in Thy-1+ and Thy-1− fibroblast populations. Only a minor part of interstitial spindle-shaped FAP+ cells were mature myofibroblasts as identified by co-expression of SMA. Of note, expression of FAP in inflammatory cells, endothelial cells or vascular smooth muscle cells was rare (Supplementary data Fig. C.3B). These results show that the FAP+ cell population in hearts of patients with ischemic cardiomyopathy is heterogeneous, consisting of both Thy-1+ and Thy-1− spindle-shaped fibroblasts and to a minor part of SMA+ myofibroblasts. 3.4. FAP expression in cardiac fibroblasts is stimulated by TGFβ1 via canonical SMAD2/3 signaling We then investigated the signaling pathways leading to expression of FAP in cardiac fibroblasts. TGFβ1, a known mediator of wound healing
Fig. 3. Characterization of the FAP+ cell population after MI. (A) Quantitative co-localization analysis showed a marked increase of cells within the border zone after MI. A major part of these cells consisted of activated P4H+ fibroblasts, especially at 7 days after MI. The majority of FAP+ cells co-expressed P4H on days 7 and 28 after MI, indicating an activated fibroblastic phenotype. A significant amount of FAP+/P4H− cells was only present at day 3 after MI, primarily comprising P4H− fibroblasts. FAP+/P4H− non-fibroblasts were detected very rarely at all time points. (B) The FAP+ cell population consisted largely of P4H+ proto-myofibroblasts (37 ± 3%, 61 ± 4% and 65 ± 3%) and to a lesser extent of SMA+ differentiated myofibroblasts (24 ± 1%, 30 ± 7%, 17 ± 6% at 3, 7 and 28 days after MI, respectively). Sham: n = 4, MI 3 days: n = 3, MI 7 days: n = 3, MI 28 days: n = 4; *: p b 0.05 vs. Sham.
198
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
and tissue fibrosis after MI, stimulated FAP concentration in fibroblasts in a time- and dose-dependent manner (Fig. 5A, B). Next, we investigated potential signaling pathways mediating TGFβ1-induced FAP expression. The increase in FAP expression by treatment with TGFβ1 was abolished using SB431542, a potent inhibitor of TGFβ-receptor I signaling (Fig. 5C). Treatment with SB431542 resulted in the inhibition of SMAD2 and SMAD3 phosphorylation, indicating activation of FAP by the canonical SMAD pathway (Fig. 5D, E). Based on this result, we used siRNA to knock down expression of SMAD2 and SMAD3. Treatment of SMAD2- and SMAD3-deficient fibroblasts with 2 ng/ml TGFβ1 resulted in markedly reduced levels of phosphorylated forms of SMAD2 and SMAD3, together with a blunted increase of FAP in both cases (Fig. 5F–H). These results demonstrated that induction of FAP by TGFβ1 is dependent on phosphorylation of SMAD2 and SMAD3 via activation of the TGFβ-receptor I. In contrast, inhibition of the ERK/MAPK pathway with U0126 did not alter FAP expression after treatment with TGFβ1 (Fig. 5C–E). 3.5. FAP exerts gelatinolytic activity and mediates migration in cardiac fibroblasts Because FAP possesses gelatinolytic activity, we analyzed its ability to degrade gelatin by zymography. Cardiac fibroblasts exhibited a marked gelatinolytic activity at molecular weight 170 kDa, corresponding to the dimerized and enzymatically active form of FAP. Since FAP is a serine-protease, this gelatinolytic activity persisted when matrixmetalloproteinase activity was blocked by addition of EDTA. As expected, treatment of cardiac fibroblasts with 2 ng/ml TGFβ1 resulted in increased gelatinolytic activity (Fig. 6A). A second gelatinolytic band was identified as matrix metalloproteinase 2 (MMP-2), a well-known gelatinase of the heart. In comparison, gelatinolytic activity exerted by MMP-2 was higher than of FAP. Next, we wanted to gain insight into the functional role of FAP, and used cardiac fibroblasts with siRNA-mediated knockdown for further experiments (Fig. 6B). Migratory capacity of FAP-deficient fibroblasts was significantly reduced by −30% compared to control cells in a transwell migration assay (Fig. 6C). Complete inhibition of all serine proteases by the broad-band serine-protease inhibitor AEBSF resulted in further reduction of migratory capacity by −69% compared to control cells, supporting an essential role of FAP for fibroblast migration (Fig. 6C). Importantly, proliferation was unaffected by FAP deficiency (Fig. 6D). Finally, gelatin zymography showed loss of FAP-mediated gelatinolytic activity in FAPdeficient cardiac fibroblasts. Together, our results support a role of FAP in extracellular matrix degradation and fibroblast migration after MI. 4. Discussion The invasion of fibroblasts into the injured and surviving myocardium after MI is a hallmark during infarct healing. Still, the mechanisms of fibroblast activation and tissue repair are not fully understood. In this study, we identified the expression of FAP in activated fibroblasts during infarct healing after MI, and showed a role of FAP in fibroblast migration and matrix degradation. 4.1. FAP identifies activated fibroblasts after MI Until now, no marker exclusively identifies fibroblasts, and therefore characterization is established by a combination of protein markers and morphological characteristics [4,23,24].
199
In our study, we used P4H as marker of activated fibroblasts in rat hearts. P4H has been shown to be enriched in α-1 type I collagenproducing cardiac fibroblasts, but not in endothelial cells [25]. In a detailed in vitro analysis, P4H was expressed in dermal fibroblasts, fibrocytes and macrophages, but not monocytes [26]. To ascertain that we primarily identified fibroblasts in our study, we showed coexpression of FAP+ cells with the fibroblast markers P4H, SMA and vimentin in rat hearts, and used Thy-1 as additional maker to identify fibroblast subpopulations in human hearts [22,24]. Importantly, we excluded relevant co-expression of FAP in monocytes, macrophages, endothelial cells and vascular smooth muscle cells. We detected robust FAP expression in fibroblasts within the granulation and scar tissue of healing infarcts after MI, while it was rarely expressed in healthy hearts. Since nearly all FAP expressing cells showed staining for P4H in rat hearts at days 7 and 28 after MI, we conclude that FAP identifies activated, collagen-synthesizing fibroblasts in the scar. The fibroblast phenotype is further supported by demonstrating that the majority of FAP+ cells in human hearts with ischemic cardiomyopathy have a spindle-shaped morphology, show overlap with Thy-1+ fibroblast and SMA+ myofibroblast populations, and do not express endothelial or inflammatory cell markers. In fact, only a negligible part of FAP+ cells were endothelial cells, inflammatory cells or vascular smooth muscle cells in both rat and human heart tissues. Because of the rare occurrence of these non-fibroblast FAP+ cells, we argue that FAP+/P4H− and FAP+/Thy − 1− cells are mostly fibroblasts in an early activation state. This is supported by our data in rat hearts showing a significantly elevated level of FAP+/P4H− cells only at 3 days after MI, which is declining on days 7 and 28 after MI. Since FAP was only expressed at sites of active tissue remodeling, we argue that FAP expression identifies activated, but not resting fibroblasts. Thy-1+ cells were recently shown to identify fibroblasts with increased expression of extracellular-matrix associated genes in the healthy adult mouse heart [24]. However, fibroblasts are a heterogeneous population possessing a large diversity in activity, function and origin, in part depending on their activation state [23,24]. In this regard, Thy-1+ and Thy-1− populations have been described to co-exist [27], and these subsets show functional heterogeneity upon stimulation with TGFβ [28]. In fact, loss of Thy-1 has been associated with progressing fibrosis and fibroblast activation in a mouse model of bleomycin induced lung fibrosis [29], illustrating the dynamic alteration of a fibroblast marker in fibrotic disease. In our study, the expression of FAP was not restricted to differentiated myofibroblasts, and FAP was expressed in P4H+ and P4H− cells in rat hearts as well as Thy-1+ and Thy-1− cells in human hearts after MI. Therefore, expression of FAP might be indicating a pro-fibrotic phenotype in a wide range of fibroblast and fibroblast-like subpopulations including early proto-myofibroblasts and differentiated myofibroblasts. Considering various origins and heterogeneity of fibroblasts on one hand and vast expression of FAP on the other hand, we suggest that FAP is a marker for fibroblast activation, irrespective of their origin and differentiation stage. 4.2. FAP is activated in cardiac fibroblasts by TGFβ1 Activation of fibroblasts by TGFβ1 is a crucial step of wound healing early after MI by enabling the pro-fibrotic phenotype, leading to the formation of a strong extracellular matrix and scar [30]. While cardiomyocyte-specific inhibition of TGFβ signaling improved cardiac remodeling [31], broad inhibition of TGFβ signaling resulted in
Fig. 4. FAP is expressed in human hearts after acute MI. (A) Quantitative analysis by immunohistochemistry showed elevated FAP-expression in hearts of patients with acute MI and severely reduced LV-function (n = 5) as compared to non-failing donor hearts (control, n = 4, P b 0.05). In infarcted hearts, FAP+ cells were surrounding vessels (black arrows) and dispersed in fibrotic tissue (white arrows). In control hearts, FAP+ cells were at times found in fibrotic foci (asterisk). FAP: brown. Scale bars = 100 μM. (B) Heterogeneity of Thy-1expression in FAP+ fibroblasts after acute MI. In infarcted human hearts, areas with abundant predominantly FAP+/Thy-1− cells (left panel, asterisk) or predominantly FAP+/Thy-1+ cells (right panel, white arrows) existed. Left panel: A cluster of FAP−/Thy-1+ cells is shown as reference (black arrow). FAP: red; Thy-1: green; DAPI: blue. Scale bars: 10 μm. Image processing included changes in brightness, contrast and tonal range, and was applied equally across the entire image. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
200
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
Fig. 5. TGFβ1 induces FAP protein expression in adult human cardiac fibroblasts. (A, B) FAP-expression is stimulated by TGFβ1 in a time- (A) and dose-dependent manner (B). A slight increase in expression of FAP was evident over time in cardiac fibroblasts after N48 h of culture. (C–H) TGFβ1-induced FAP expression is mediated by the canonical TGFβ signaling pathway: Inhibition of TGFβ type I receptor using SB431542 blunted TGFβ1-mediated increase of FAP, while blocking the MAP kinase pathway by U0126 did not alter FAP levels (C). SB431542 efficiently inhibited SMAD2 and SMAD3 phosphorylation (D, E). Moreover, knockdown of SMAD2 and SMAD3 by siRNA in cardiac fibroblasts prevented formation of phospho-SMAD2 (F) and phospho-SMAD3 (G) and blunted increase of FAP expression (H) after stimulation with TGFβ1 in CF. A: * = P b 0.05 vs. 6 h, #: P b 0.05 vs. control; n = 3–4 each time point. B: * = P b 0.05 vs. no TGFβ1; n = 3–6 each time point. C n = 15. D–G n = 3. H n = 6.
increased LV dimensions and higher mortality after MI [32], supporting beneficial effects of TGFβ on non-myocytes, including fibroblasts, after MI. In our study, we show that FAP is highly expressed in fibroblasts within the borderzone after permanent LAD ligation, with a peak at 7 days after MI and decline thereafter. This upregulation of FAP might be mediated at least in part by TGFβ1 in vivo: Temporal and regional expression changes of inflammatory, angiogenic, differentiation and homing related genes after MI have been demonstrated previously. Using laser microdissection to identify infarct scar and borderzone, permanent LAD ligation in mice resulted in increased expression of various genes including TGFß1 within the first hours, with a peak between 2 and 7 days after MI and decrease thereafter [33]. Of note, TGFβ1 protein
expression was highest in the borderzone in this study. This corresponds to another report using a rat model of MI, demonstrating that maximum expression of TGFβ1 mRNA and TGFβ1 protein was detected within days 1 to 7 after MI, together with a peak of procollagen α1(I) at 3 to 7 days after MI [34]. Together, these previous reports provide evidence of TGFβ1 expression in the border zone within the first days after MI in mice and rats, which corresponds to our findings of FAP expression after MI in rats and humans. TGFβ1 exerts its effects on extracellular matrix production mainly by activation of proteases via the canonical SMAD2/3 cascade [30]. Indeed, our data shows that FAP is induced in cardiac fibroblasts by TGFβ1 via TGFβ receptor I, SMAD2 and SMAD3, suggesting that FAP contributes to TGFβ1-mediated effects in wound healing after MI.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
201
Fig. 6. FAP exerts gelatinolytic activity and promotes migration in adult human cardiac fibroblasts. (A) Gelatin zymography demonstrated gelatinolytic activity of FAP derived from cardiac fibroblasts, which was further increased after treatment with TGFβ1. FAP gelatinolytic activity persisted after blocking matrix metalloproteinase activity with EDTA. (B) Efficient knockdown of FAP in cardiac fibroblasts using FAP siRNA as compared to scrambled control siRNA was verified by Western blot and ELISA. (C) Cell migration was reduced in FAP-deficient cells, and even further reduced by adding serine protease blocker AEBSF. (D) Control and FAP-deficient fibroblasts showed no difference in proliferation index. (E) Gelatinolytic activity of FAP is blunted in FAP-deficient cells. A: n = 2. B: n = 6. C: n = 3–9. D: n = 12. E: n = 5.
202
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203
4.3. FAP exerts gelatinolytic activity and mediates fibroblast migration Our study demonstrated gelatinolytic activity of cardiac fibroblastderived FAP. Proteases are an essential part for the homeostasis of cardiac extracellular matrix and various proteases have been shown to be of crucial importance for healing after MI. Matrix metalloproteinases including MMP2, MMP9, MMP14 and MMP28 have been extensively studied and shown to affect myocardial post-MI remodeling by matrix degradation as well as proteolytic activation of cytokines [35–38]. Similar effects were shown for serine proteases in post-MI remodeling: Heymans et al. described altered scar formation and impaired remodeling in animals with deficiency of the serine proteases tissue-type plasminogen activator and urokinase-type plasminogen activator [39]. Cell surface proteolysis represents an important mechanism for cell migration. In fact, we found impaired fibroblast migration in FAP deficient cells. This is in concordance with previous reports demonstrating a role for migration in endothelial cells [40], embryonic fibroblasts [41], hepatic stellate cells [42] and bone marrow mesenchymal stem cells [43]. Together, we show that cardiac fibroblast-derived FAP has gelatinolytic activity and mediates fibroblast migration, which may contribute to extracellular matrix scar formation and ultimately remodeling after MI. Apart from its gelatinolytic activity, other enzymatic substrates of FAP such as α2-antiplasmin, neuropeptide Y, B-type natriuretic peptide, substance P and peptide YY are known so far, and other yet unknown substrates probably exist [44]. Also, gelatinolytic activity of FAP was much less as compared to MMP2, suggesting that FAP has other physiologic roles in post-MI remodeling. Whether FAP itself is a therapeutic target remains unknown, and further studies on the effects of FAP on wound healing and cardiac remodeling after MI are warranted. Apart from its potential pathophysiological roles, FAP might be suitable as a specific marker for activated fibroblasts in cardiac imaging. Molecular imaging of myocardial fibrosis has been attempted using various targets, including collagen fibers, myofibroblasts, matricellular proteins and matrix metalloproteinases [45]. Activated fibroblasts are the major source of extracellular matrix after MI, and myocardial fibrosis is a major determinant of left ventricular dysfunction. The development of FAP-specific imaging probes for noninvasive assessment of activated fibroblasts could help in evaluating the individual extent and progression of fibrosis, and therefore aid in antifibrotic treatments. In this regard, first studies using fluorochrome-tagged FAP substrates for in vivo imaging of tumor fibroblasts have been reported in small animal models [46], and a FAP antibody was used to visualize arthritic joints in murine experimental arthritis by radionuclide imaging [47].
4.4. Summary In this study, we characterized for the first time the expression and phenotype of FAP+ cells in the healing wound after MI. The identification of FAP in activated, but not resting fibroblasts in rat and human tissues after MI will help to further delineate the role of fibroblasts in inflammation and wound healing, and holds promise for further evaluation as target for drug treatments and molecular imaging to better understand and improve cardiac remodeling after MI. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.yjmcc.2015.08.016.
Acknowledgments We gratefully acknowledge Annette Berbner, Annemieke Klan and Silke Pretzer for expert technical assistance. We thank Dr. Länger for providing archival myocardial tissue samples for immunohistochemistry. This work was initially supported by grants from the Interdisciplinary Clinical Research Center Würzburg (to J.T., J.B.: IZKF E-140).
Disclosures None. References [1] D. Fraccarollo, P. Galuppo, J. Bauersachs, Novel therapeutic approaches to postinfarction remodelling, Cardiovasc. Res. 94 (2012) 293–303. [2] D. Fraccarollo, P. Galuppo, S. Motschenbacher, H. Ruetten, A. Schafer, J. Bauersachs, Soluble guanylyl cyclase activation improves progressive cardiac remodeling and failure after myocardial infarction. Cardioprotection over ACE inhibition, Basic Res. Cardiol. 109 (2014) 421. [3] S.W. van den Borne, J. Diez, W.M. Blankesteijn, J. Verjans, L. Hofstra, J. Narula, Myocardial remodeling after infarction: the role of myofibroblasts, Nat. Rev. Cardiol. 7 (2010) 30–37. [4] E.C. Goldsmith, A.D. Bradshaw, M.R. Zile, F.G. Spinale, Myocardial fibroblast-matrix interactions and potential therapeutic targets, J. Mol. Cell. Cardiol. 70 (2014) 92–99. [5] J.J. Tomasek, G. Gabbiani, B. Hinz, C. Chaponnier, R.A. Brown, Myofibroblasts and mechano-regulation of connective tissue remodelling, Nat. Rev. Mol. Cell Biol. 3 (2002) 349–363. [6] A. Leask, Potential therapeutic targets for cardiac fibrosis: TGFbeta, angiotensin, endothelin, CCN2, and PDGF, partners in fibroblast activation, Circ. Res. 106 (2010) 1675–1680. [7] T. Thum, C. Gross, J. Fiedler, T. Fischer, S. Kissler, M. Bussen, et al., MicroRNA-21 contributes to myocardial disease by stimulating MAP kinase signalling in fibroblasts, Nature 456 (2008) 980–984. [8] J.E. Park, M.C. Lenter, R.N. Zimmermann, P. Garin-Chesa, L.J. Old, W.J. Rettig, Fibroblast activation protein, a dual specificity serine protease expressed in reactive human tumor stromal fibroblasts, J. Biol. Chem. 274 (1999) 36505–36512. [9] P. Garin-Chesa, L.J. Old, W.J. Rettig, Cell surface glycoprotein of reactive stromal fibroblasts as a potential antibody target in human epithelial cancers, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 7235–7239. [10] H. Dolznig, N. Schweifer, C. Puri, N. Kraut, W.J. Rettig, D. Kerjaschki, et al., Characterization of cancer stroma markers: in silico analysis of an mRNA expression database for fibroblast activation protein and endosialin, Cancer Immun. 5 (2005) 10. [11] J. Niedermeyer, M. Kriz, F. Hilberg, P. Garin-Chesa, U. Bamberger, M.C. Lenter, et al., Targeted disruption of mouse fibroblast activation protein, Mol. Cell. Biol. 20 (2000) 1089–1094. [12] M. Kraman, P.J. Bambrough, J.N. Arnold, E.W. Roberts, L. Magiera, J.O. Jones, et al., Suppression of antitumor immunity by stromal cells expressing fibroblast activation protein-alpha, Science 330 (2010) 827–830. [13] P.S. Acharya, A. Zukas, V. Chandan, A.L. Katzenstein, E. Pure, Fibroblast activation protein: a serine protease expressed at the remodeling interface in idiopathic pulmonary fibrosis, Hum. Pathol. 37 (2006) 352–360. [14] M.T. Levy, G.W. McCaughan, G. Marinos, M.D. Gorrell, Intrahepatic expression of the hepatic stellate cell marker fibroblast activation protein correlates with the degree of fibrosis in hepatitis C virus infection, Liver 22 (2002) 93–101. [15] S. Bauer, M.C. Jendro, A. Wadle, S. Kleber, F. Stenner, R. Dinser, et al., Fibroblast activation protein is expressed by rheumatoid myofibroblast-like synoviocytes, Arthritis Res. Ther. 8 (2006) R171. [16] D. Fraccarollo, P. Galuppo, S. Schraut, S. Kneitz, N. van Rooijen, G. Ertl, et al., Immediate mineralocorticoid receptor blockade improves myocardial infarct healing by modulation of the inflammatory response, Hypertension 51 (2008) 905–914. [17] J. Tillmanns, M. Rota, T. Hosoda, Y. Misao, G. Esposito, A. Gonzalez, et al., Formation of large coronary arteries by cardiac progenitor cells, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 1668–1673. [18] J. Tillmanns, M. Schneider, D. Fraccarollo, J.D. Schmitto, F. Langer, D. Richter, et al., PET imaging of cardiac wound healing using a novel [68Ga]-labeled NGR probe in rat myocardial infarction, Mol. Imaging Biol. 17 (2015) 76–86. [19] J. Tillmanns, C. Widera, Y. Habbaba, P. Galuppo, T. Kempf, K.C. Wollert, et al., Circulating concentrations of fibroblast activation protein alpha in apparently healthy individuals and patients with acute coronary syndrome as assessed by sandwich ELISA, Int. J. Cardiol. 168 (2013) 3926–3931. [20] D. Fraccarollo, J. Bauersachs, M. Kellner, P. Galuppo, G. Ertl, Cardioprotection by long-term ET(A) receptor blockade and ACE inhibition in rats with congestive heart failure: mono-versus combination therapy, Cardiovasc. Res. 54 (2002) 85–94. [21] K. Aertgeerts, I. Levin, L. Shi, G.P. Snell, A. Jennings, G.S. Prasad, et al., Structural and kinetic analysis of the substrate specificity of human fibroblast activation protein alpha, J. Biol. Chem. 280 (2005) 19441–19444. [22] F. Hudon-David, F. Bouzeghrane, P. Couture, G. Thibault, Thy-1 expression by cardiac fibroblasts: lack of association with myofibroblast contractile markers, J. Mol. Cell. Cardiol. 42 (2007) 991–1000. [23] G. Krenning, E.M. Zeisberg, R. Kalluri, The origin of fibroblasts and mechanism of cardiac fibrosis, J. Cell. Physiol. 225 (2010) 631–637. [24] S.R. Ali, S. Ranjbarvaziri, M. Talkhabi, P. Zhao, A. Subat, A. Hojjat, et al., Developmental heterogeneity of cardiac fibroblasts does not predict pathological proliferation and activation, Circ. Res. 115 (2014) 625–635. [25] T. Moore-Morris, N. Guimaraes-Camboa, I. Banerjee, A.C. Zambon, T. Kisseleva, A. Velayoudon, et al., Resident fibroblast lineages mediate pressure overload-induced cardiac fibrosis, J. Clin. Invest. 124 (2014) 2921–2934. [26] D. Pilling, T. Fan, D. Huang, B. Kaul, R.H. Gomer, Identification of markers that distinguish monocyte-derived fibrocytes from monocytes, macrophages, and fibroblasts, PLoS One 4 (2009), e7475.
J. Tillmanns et al. / Journal of Molecular and Cellular Cardiology 87 (2015) 194–203 [27] L. Koumas, A.E. King, H.O. Critchley, R.W. Kelly, R.P. Phipps, Fibroblast heterogeneity: existence of functionally distinct Thy 1(+) and Thy 1(−) human female reproductive tract fibroblasts, Am. J. Pathol. 159 (2001) 925–935. [28] L. Koumas, T.J. Smith, S. Feldon, N. Blumberg, R.P. Phipps, Thy-1 expression in human fibroblast subsets defines myofibroblastic or lipofibroblastic phenotypes, Am. J. Pathol. 163 (2003) 1291–1300. [29] J.S. Hagood, P. Prabhakaran, P. Kumbla, L. Salazar, M.W. MacEwen, T.H. Barker, et al., Loss of fibroblast Thy-1 expression correlates with lung fibrogenesis, Am. J. Pathol. 167 (2005) 365–379. [30] M. Bujak, N.G. Frangogiannis, The role of TGF-beta signaling in myocardial infarction and cardiac remodeling, Cardiovasc. Res. 74 (2007) 184–195. [31] P.P. Rainer, S. Hao, D. Vanhoutte, D.I. Lee, N. Koitabashi, J.D. Molkentin, et al., Cardiomyocyte-specific transforming growth factor beta suppression blocks neutrophil infiltration, augments multiple cytoprotective cascades, and reduces early mortality after myocardial infarction, Circ. Res. 114 (2014) 1246–1257. [32] S. Frantz, K. Hu, A. Adamek, J. Wolf, A. Sallam, S. Maier, et al., Transforming growth factor beta inhibition increases mortality and left ventricular dilatation after myocardial infarction, Basic Res. Cardiol. 103 (2008) 485–492. [33] S. Vandervelde, M.J. van Luyn, M.H. Rozenbaum, A.H. Petersen, R.A. Tio, M.C. Harmsen, Stem cell-related cardiac gene expression early after murine myocardial infarction, Cardiovasc. Res. 73 (2007) 783–793. [34] R.G. Dean, L.C. Balding, R. Candido, W.C. Burns, Z. Cao, S.M. Twigg, et al., Connective tissue growth factor and cardiac fibrosis after myocardial infarction, J. Histochem. Cytochem. 53 (2005) 1245–1256. [35] A. Ducharme, S. Frantz, M. Aikawa, E. Rabkin, M. Lindsey, L.E. Rohde, et al., Targeted deletion of matrix metalloproteinase-9 attenuates left ventricular enlargement and collagen accumulation after experimental myocardial infarction, J. Clin. Invest. 106 (2000) 55–62. [36] S. Hayashidani, H. Tsutsui, M. Ikeuchi, T. Shiomi, H. Matsusaka, T. Kubota, et al., Targeted deletion of MMP-2 attenuates early LV rupture and late remodeling after experimental myocardial infarction, Am. J. Physiol. Heart Circ. Physiol. 285 (2003) H1229–H1235. [37] F.G. Spinale, R. Mukherjee, J.A. Zavadzkas, C.N. Koval, S. Bouges, R.E. Stroud, et al., Cardiac restricted overexpression of membrane type-1 matrix metalloproteinase causes adverse myocardial remodeling following myocardial infarction, J. Biol. Chem. 285 (2010) 30316–30327.
203
[38] Y. Ma, G.V. Halade, J. Zhang, T.A. Ramirez, D. Levin, A. Voorhees, et al., Matrix metalloproteinase-28 deletion exacerbates cardiac dysfunction and rupture after myocardial infarction in mice by inhibiting M2 macrophage activation, Circ. Res. 112 (2013) 675–688. [39] S. Heymans, A. Luttun, D. Nuyens, G. Theilmeier, E. Creemers, L. Moons, et al., Inhibition of plasminogen activators or matrix metalloproteinases prevents cardiac rupture but impairs therapeutic angiogenesis and causes cardiac failure, Nat. Med. 5 (1999) 1135–1142. [40] G. Ghersi, Q. Zhao, M. Salamone, Y. Yeh, S. Zucker, W.T. Chen, The protease complex consisting of dipeptidyl peptidase IV and seprase plays a role in the migration and invasion of human endothelial cells in collagenous matrices, Cancer Res. 66 (2006) 4652–4661. [41] G. Ghersi, H. Dong, L.A. Goldstein, Y. Yeh, L. Hakkinen, H.S. Larjava, et al., Regulation of fibroblast migration on collagenous matrix by a cell surface peptidase complex, J. Biol. Chem. 277 (2002) 29231–29241. [42] X.M. Wang, D.M. Yu, G.W. McCaughan, M.D. Gorrell, Fibroblast activation protein increases apoptosis, cell adhesion, and migration by the LX-2 human stellate cell line, Hepatology 42 (2005) 935–945. [43] K.M. Chung, S.C. Hsu, Y.R. Chu, M.Y. Lin, W.T. Jiaang, R.H. Chen, et al., Fibroblast activation protein (FAP) is essential for the migration of bone marrow mesenchymal stem cells through RhoA activation, PLoS One 9 (2014), e88772. [44] F.M. Keane, N.A. Nadvi, T.W. Yao, M.D. Gorrell, Neuropeptide Y, B-type natriuretic peptide, substance P and peptide YY are novel substrates of fibroblast activation protein-alpha, FEBS J. 278 (2011) 1316–1332. [45] H.J. de Haas, E. Arbustini, V. Fuster, C.M. Kramer, J. Narula, Molecular imaging of the cardiac extracellular matrix, Circ. Res. 114 (2014) 903–915. [46] R. Ruger, F.L. Tansi, M. Rabenhold, F. Steiniger, R.E. Kontermann, A. Fahr, et al., In vivo near-infrared fluorescence imaging of FAP-expressing tumors with activatable FAP-targeted, single-chain Fv-immunoliposomes, J. Control. Release 186 (2014) 1–10. [47] P. Laverman, T. van der Geest, S.Y. Terry, D. Gerrits, B. Walgreen, M.M. Helsen, et al., Immuno-PET and immuno-SPECT of rheumatoid arthritis with radiolabeled antifibroblast activation protein antibody correlates with severity of arthritis, J. Nucl. Med. 56 (2015) 778–783.