Vascular Injury Induces Expression of Periostin
http://www.100md.com
《动脉硬化血栓血管生物学》
From the Center for Molecular Medicine, Maine Medical Center Research Institute, Scarborough, Me.
Correspondence to Dr Volkhard Lindner, Center for Molecular Medicine, Maine Medical Center Research Institute, Scarborough, ME 04074. E-mail lindnv@mmc.org
Abstract
Objective— Periostin mRNA is among the most strongly upregulated transcripts in rat carotid arteries after balloon injury. The goal of the present study was to gain insight into the significance of periostin in the vasculature.
Methods and Results— Periostin expression after injury was localized to smooth muscle cells of the neointima and the adventitia. The expression of periostin in smooth muscle cells in vitro was not regulated by cytokines such as fibroblast growth factor-2 (FGF-2). In contrast, stimulation of MC3T3-E1 osteoblastic cells, NIH3T3 fibroblasts, or mesenchymal C3H10T1/2 cells with FGF-2 reduced periostin mRNA levels to <5% of controls, whereas conversely bone morphogenetic protein-2 (BMP-2) increased periostin mRNA levels. Periostin expression was induced and maintained during retinoic acid-induced smooth muscle cell differentiation in A404 cells. In addition, overexpression of periostin in C3H10T1/2 cells caused an increase in cell migration that could be blocked with an anti-periostin antibody.
Conclusions— Periostin expression is associated with smooth muscle cell differentiation in vitro and promotes cell migration. Unlike other mesenchymally derived cell lines, periostin expression is not regulated by FGF-2 in smooth muscle cells. This distinction may be useful in discriminating smooth muscle and fibroblast lineages.
Periostin is an abundant sequence induced on vascular injury arteries. We find that periostin is associated with smooth muscle cell differentiation and that it functions as a migratory stimulus. In addition, inability to downregulate periostin expression with FGF-2 may be a useful criterion to discriminate between smooth muscle and fibroblast lineages.
Key Words: smooth muscle ? adventitia ? myofibroblast ? intimal hyperplasia
Introduction
Tissue remodeling after injury involves the coordinate regulation of numerous genes, including those involved with extracellular matrix (ECM) synthesis and turnover, cell adhesion, and cell migration. Major components of the ECM include structural proteins and a number of soluble ECM-associated proteins. The soluble ECM-associated proteins include SPARC, lumican, and the fasciclin domain-containing proteins ?ig-H3 and periostin. Periostin, originally named osteoblast-specific factor-2 (Genbank D13664) was identified as a soluble ECM protein that was expressed in bone and, to a lesser extent, lung.1 The periostin mRNA is alternatively spliced, resulting in polypeptides that differ in the length of their C-terminal domains and in their tissue-specific patterns of expression.2
Periostin is a 90-kDa heparin-binding N-glycosylated protein that was proposed to associate with bone ECM, and it is highly homologous to ?ig-h3, previously known as the 68-kDa transforming growth factor (TGF)-?1-inducible protein. Both periostin and ?ig-h3 contain 4 tandem fasciclin (Fas1) domains homologous to the insect protein fasciclin.3 In the case of ?ig-h3, the second and fourth fasciclin domains interact with the 3?1 integrin, and all 4 fasciclin domains can interact with the v?5 integrin, thus mediating adhesion.4 Periostin also mediates cell adhesion by binding to v?3 and v?5, and increases cell motility.5 The localization of recombinant periostin with v?5 integrin at sites of focal adhesion5 suggests the contribution of periostin to cell adhesion and motility. Despite these recent observations of periostin interacting with cell surface receptors, its functional roles are not clear.
More recently, periostin expression was found in embryonic hearts by embryonic day E10.5, and increased to a plateau by E14. Periostin mRNA expression was localized to the endocardial cushions that ultimately divide the primitive heart into the mature 4-chambered heart.6 In addition, periostin expression is increased in pulmonary aortic smooth muscle cells (SMCs)in response to hypoxia.7
The goal of the present study was to examine the role of periostin in SMCs. Our data indicate that periostin is associated with SMC differentiation and functions as a migratory stimulus in vitro. After balloon catheter injury, periostin expression coincides with the proliferative phase of SMCs in the neointima.
Methods
Animals
Sprague-Dawley rats were used and the left carotid artery and the aorta were denuded with a 2-French balloon catheter as described.8
Subtractive Hybridization and Cloning of the Periostin cDNA
Suppressive subtractive hybridization was performed with a kit (polymerase chain reaction Select cDNA Subtraction Kit; Clontech), following the manufacturer’s instructions. cDNA was prepared from 2 μg of mRNA extracted from normal rat carotid arteries and aortae, as well as 8-day balloon-injured vessels. For the purpose of isolating sequences overexpressed in injured arteries, cDNA from injured vessels was used as "tester" and cDNA from normal vessels as "driver" cDNA. Partial sequences of 250 clones were obtained by automated sequencing DNA sequence analysis and their identities were determined by searching Genbank databases, including nonredundant and expressed sequence tag databases.
Northern Blotting and cDNA Probes
Total RNA isolated from normal carotid arteries, 8-day balloon-injured carotid arteries, and indicated cell lines were examined by Northern blot analysis as described.9 Blots were hybridized simultaneously using a rat periostin cDNA and a mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) cDNA labeled with -[32P]-dCTP. The intensities of the periostin band and the GAPDH band were quantified by phosphorimaging. The ratio of the signal intensities for periostin/GAPDH before growth factor addition (time point, 0 hours) were set to 1.0 (100%). Identical blots were probed in duplicate with similar results.
In Situ Hybridization
[35S]-UTP labeled RNA probes corresponding to the sense and antisense strand of the coding region of rat periostin were prepared and in situ hybridization was performed on paraffin sections as described.9
Cells and Culture
A404 cells (kindly provided by Dr Gary Owens, University of Virginia School of Medicine) were maintained as described.10 Differentiation of A404 cells was initiated with 1 μmol/L all-trans retinoic acid (RA) for 3 days. After withdrawal of RA after 3 days, selection for SMCs was achieved by addition of 0.5 μg/mL puromycin (Clontech) to the medium for either 3 or 5 days.
Primary rat aortic SMCs (RASMC), MC3T3-E1, C3H10T1/2, NIH3T3, rat PAC1 smooth muscle, ROS17/2.8 osteosarcoma, BOSC23,11 and human aortic SMC (AoSMC) (Clonetics) cell lines were used. Primers for reverse-transcriptase polymerase chain reaction were: GAPDH (5'GGAGATTGTTGCCATCAACGA3', 5'GAAGACACCAGTAGACTCCACGACA3'), achaete-scute homolog-1 (MASH-1, 5'CAAGTTGGTCAACCTGGGTTTTG3', 5'CACTAAAGATGCAGGATCTGCTG3'), periostin (5'AAAGTAAAAGTTGGCCTTAGCGACC3', 5'CAGAAGCTCCCTTTCTTCGCTAGT3'), and smooth muscle myosin heavy chain (SM-MHC, 5'CAGTTGGACACTATGTCAGGGAAA3', 5'ATGGAGACAAATGCTAATCAGCC3').
For stimulation of cells with cytokines, confluent cells were fed with fresh medium containing 10% fetal bovine serum, and 2 days later cells were stimulated with one of the following: fibroblast growth factor-2 (FGF-2) (10 ng/mL), PDGF-AB (10 ng/mL), BMP-2 (25 ng/mL), or TGF-? (10 ng/mL) (all from R&D Systems).
Retrovirus Production and Isolation
BOSC23 cells (Clontech) were transiently transfected with either control pWzl12 or pWzl bearing periostin cDNA. Immunofluorescence analysis was used to confirm the efficiency of retroviral gene expression, which was >95% (data not shown).
Western Blot Analysis
Rabbit anti-periostin antibody (kindly provided by Dr Roger Markwald, Medical University of South Carolina) was used at a 1:1000 dilution. FGF-2 expression levels in the various cell lines were determined by immunoblotting with a mouse anti-FGF-2 antibody13 after SDS-PAGE of 80 μg cell lysate under reducing conditions. Quantification was performed by densitometry, and the relative amounts of FGF-2 present were expressed as a fraction of FGF-2 detected in RASMC (value set at 1.0).
Migration Assay
Migration of cells to 3% serum was assayed in a modified Boyden chamber as described, in 3 separate experiments.14 In some cases, rabbit anti-periostin antibody or control rabbit anti-HA antibody were included at 0.5 μg/mL. The data are means±SEM of 6 wells per experimental condition. Student t test was used to compare the means between the 2 groups, and P0.05 was considered significant.
Results
Periostin Expression Is Increased After Arterial Injury
Suppressive subtractive hybridization was conducted on cDNA derived from 8-day balloon-injured rat carotid arteries and control vessels. A survey of 250 clones obtained using this method identified 3 of them as periostin cDNA fragments (Table). Other identified sequences were various collagens, smoothelin, osteopontin, and SPARC, as well as genes that have been described in other tissues such as connective tissue growth factor (CCN2), hevin/SC1, and WISP2 (CCN5).15 These data are consistent with the view that vascular responses to injury include the regulation of groups of genes that regulate the ECM environment.
Frequency of Selected Upregulated Sequences Among 250 Identified Clones From Balloon-Injured Rat Carotid Arteries
Northern blot analysis performed on RNA isolated from normal and 8-day balloon-injured rat carotid arteries demonstrated increased levels of periostin mRNA in injured vessels and only low levels of expression in normal arteries (Figure 1A). This differential expression pattern was also verified at the protein level using an anti-periostin antibody (Figure 1B). This antibody recognized a major band (molecular weight 90 kDa) and a slightly smaller and less intense band, which could reflect a difference in glycosylation. Among a variety of tissues from adult rats, only lung and, to a lesser extent, uterus revealed significant levels of periostin mRNA (Figure 1A). To identify the cells expressing periostin mRNA within the vessel wall after injury, we performed in situ hybridization with an antisense periostin riboprobe. Low levels of periostin expression were detectable in normal arteries (Figure 2A). Eight days after balloon injury, high levels of periostin mRNA expression were observed in the adventitia, where adventitial fibroblasts are undergoing rapid proliferation at this time16 (Figure 2B). Two weeks after injury, a SMC-rich neointimal lesion has developed and cells throughout the neointima revealed abundant expression of periostin mRNA (Figure 2C). At 4 weeks after injury, only very few intimal SMCs on the luminal surface still expressed periostin mRNA (Figure 2D). En face preparations did not show detectable levels of periostin expression in either normal endothelium (Figure 2E) or wounded endothelium (not shown). The induction of periostin in the adventitia prompted us to examine whether this was a general response of connective tissue to injury. Dermal fibroblasts at sites of full-thickness skin incisional wounding also exhibited abundant expression of periostin mRNA (Figure 2F).
Figure 1. Periostin transcripts are dynamically regulated after vascular injury. A, Northern blot analysis of periostin mRNA expression in vivo. Compared with uninjured rat carotid arteries, increased periostin mRNA levels were detected in arteries 8 days after balloon catheter denudation. In organs from normal adult rats, significant periostin mRNA levels were detectable in the lung and at lower levels also in uterine tissue. B, Western blot analysis of total protein from uninjured and injured arteries 8 days after balloon catheter denudation demonstrates increased periostin protein levels in injured vessels.
Figure 2. Periostin is expressed by adventitial cells and neointimal cells. Photomicrographs of in situ hybridization performed with [35S]-UTP-labeled periostin antisense (A to F) and sense (G) riboprobes. A, In the uninjured rat carotid artery, low levels of periostin expression were detectable in the media (m). B, Eight days after balloon injury, increased periostin expression is detectable in the outer layers of the adventitia (a), where fibroblast proliferation and matrix production are prominent at this time. C, Two weeks after injury, the newly formed neointima (i) reveals abundant expression levels of periostin, whereas (D) at 4 weeks after denudation only scattered cells on the surface of the intimal lesion still express this mRNA. E, An en face preparation of a normal aorta shows only background levels of hybridization in endothelial cells. F, During wound healing, 8 days after a full-thickness skin incision reveals abundant periostin mRNA expression in dermal fibroblasts (d) but not in the keratinocytes of the epidermis (e). G, A section of a carotid artery 2 weeks after balloon injury was hybridized with a periostin sense probe for evaluation of nonspecific background hybridization. All images are seen under dark-field illumination with silver grains appearing as white specks. Nuclei were counterstained with hematoxylin. Original magnification was 200x except for (F), which was 50x.
Regulation of Periostin Expression in SMC and Other Cell Lines
In response to injury, SMCs in the vessel wall are exposed to a number of growth factors and cytokines released from activated platelets, inflammatory cells, and damaged vascular cells.17 For the purpose of identifying growth factors that might be involved in the regulation of periostin expression, we examined the steady-state levels of periostin mRNA after growth factor administration in rat and human primary SMC, as well as in the PAC1 SMC line and NIH3T3 fibroblasts. Because a number of other genes normally expressed in bone are also expressed in the injured arterial wall,18,19 we examined the growth factor-dependent modulation of periostin expression also in the MC3T3-E1 osteoblast cell line, as well as in the osteosarcoma cell line ROS17/2.8. We also included the oligopotent mesenchymal C3H10T1/2 cells in the analysis because they can be differentiated toward smooth muscle. C3H10T1/2, NIH3T3, and MC3T3-E1 cells all responded to FGF-2 with a profound downregulation of periostin (5%, 7%, and 2% of control levels, respectively; Figure 3A). RASMC and human AoSMC (not shown), as well as the PAC1 and ROS17/2.8 cells (not shown), however, demonstrated no significant change in periostin mRNA levels in response to FGF-2 (Figure 3A). BMP-2 caused periostin mRNA levels to increase 2- to 3-fold in C3H10T1/2, NIH3T3, and MC3T3-E1 cells, but this effect was not observed in SMCs or ROS17/2.8 cells. These cell-specific differences with regard to FGF-2-dependent downregulation of periostin prompted us to examine their endogenous FGF-2 levels by immunoblotting because high levels of endogenous FGF-2 could limit the cells ability to respond to exogenous FGF-2. RASMC and AoSMC had by far the highest levels of FGF-2 of all cell lines examined (Figure 3B). For example, FGF-2 levels of RASMC were 11-fold higher than those detected in NIH3T3 cells. The 18-kDa form, as well as higher molecular weight forms, of FGF-2 (21/22kDa in rat RASMC, 24-kDa in human AoSMC) were detected (Figure 3B). These data indicate periostin is a FGF-responsive gene in certain cell types with low endogenous FGF-2 levels.
Figure 3. Periostin expression is regulated by growth factors. A, The time course of periostin mRNA expression in response to BMP-2 (25 ng/mL) and FGF-2 (10 ng/mL) was examined by Northern blotting in various cell lines at the indicated time points. Blots were hybridized simultaneously with a GAPDH cDNA and the quantification of expression was performed with a phosphoimager. Periostin expression was normalized to GAPDH mRNA levels and the fold difference from control levels at 0 hours is shown at the bottom of each lane. B, Immunoblot analysis of cell lysates with anti-FGF-2 antibody; 80 μg protein were loaded in each lane. The amounts of FGF-2 present in the sample relative to RASMC are shown on the bottom of each lane. Note high levels of 18-kDa and high-molecular-weight FGF-2 forms in RASMC and AoSMC.
Periostin and SMC Differentiation
The A404 cell line is a P19-derived line that differentiates into several lineages including SMCs after RA treatment.20 A404 cells bear a stably transfected copy of a smooth muscle -actin promoter-driven puromycin resistance cassette. This cassette permits the elimination of nonSMC lineages in favor of smooth muscle -actin-expressing cells. Reverse-transcription polymerase chain reaction expression analysis of A404 cells that have been induced to differentiate with RA indicates that periostin mRNA is upregulated by day 3 of RA treatment, as shown in Figure 4. RA induced coordinate upregulation of the definitive smooth muscle marker SM-MHC, indicating that a true SMC lineage is present in the differentiating cell population. The neurogenic differentiation marker MASH-121 is also induced by RA as expected, indicating the presence of neural lineage cells. However, unlike periostin and SM-MHC, MASH-1 is eliminated after continued presence of puromycin, which eliminates nonsmooth muscle -actin-expressing cell lineages. These results support the view that periostin expression is responsive to RA in a pluripotent SMC precursor population and that periostin persists in RA-differentiated A404-derived SMC.
Figure 4. Smooth muscle cell differentiation induces periostin expression. Reverse-transcriptase polymerase chain reaction of A404 cells induced to differentiate into smooth muscle cells with all-trans retinoic acid (RA, 1 μmol/L). mRNA and cDNA were prepared from untreated A404 cells (no RA, lane 1) from cells with RA treatment for 3 days (3d RA, lane 2), from cells with RA treatment for 3 days followed by selection with puromycin for an additional 3 days (3d RA plus 3d puro, lane 3), and from cells with RA treatment for 3 days followed by selection with puromycin for an additional 5 days (3d RA plus 5d puro, lane 4). Primers for periostin, SM-MHC, MASH-1, and GAPDH were used for reverse-transcription polymerase chain reaction. In the absence of cDNA addition, no bands were amplified (no RT, lane 5).
Perostin and Cell Migration
C3H10T1/2 were chosen for migration experiments because we were able to achieve high levels of periostin overexpression with retroviral expression systems. Cells were transduced with retroviral constructs of a vector control or a full-length periostin expression construct. Immunofluorescence of retrovirally transduced C3H10T1/2 cells, using anti-periostin antibody, showed that essentially 100% of the infected cell population was expressing periostin (Figure 5B) as compared with empty virus control-infected cells (Figure 5A).
Figure 5. Periostin expression enhances cell migration. Immunofluorescence of C3H10T1/2 cells infected with (A) insertless pWzl retrovirus or (B) virus bearing periostin cDNA using an anti-periostin antibody. C, A representative migration assay with control (10T1/2-vector) and periostin-expressing C3H10T1/2 cells (10T1/2-periostin) is shown. The migration assay was performed in the absence of added antibodies (No Ab) or in the presence of either rabbit anti-HA antibody (Anti-HA) or rabbit anti-periostin antibody (Anti-Peri) using 3% serum as the migration stimulus. Bovine serum albumin (200 μg/mL) was used as the control for migration in the absence of stimulus.
Migration assays of cells to a stimulus with 3% serum were performed in a modified Boyden chamber as described.14 Periostin-overexpressing C3H10T1/2 cells showed a significantly enhanced migratory response to serum (Figure 5C). As shown in Figure 5C, periostin-dependent enhancement of cell migration was unchanged in the presence of a rabbit anti-HA antibody, used here as an isotype-matched irrelevant antibody. Periostin-overexpressing cells, but not control cells, showed a significant reduction of cell migration after treatment with the anti-periostin antibody (Figure 5C). These data indicate that periostin overexpression stimulates cell migration in C3H10T1/2 cells and that this stimulation is caused by increased periostin expressed outside the cell.
Discussion
Periostin was originally identified as a Fas1 domain-containing osteoblast-specific factor. Periostin has been suggested to function as a cell adhesion molecule for preosteoblasts and to participate in osteoblast recruitment, attachment, and spreading.2 Since then, several studies have demonstrated that periostin may have a role in diverse tissues because periostin was found to be expressed in the developing mouse embryonic and fetal heart.6 Furthermore, periostin expression was observed in skeletal muscle after injury, where it revealed coordinated expression with other ECM constituents including the collagen genes.22 The present study demonstrates that periostin is also part of the coordinated transcriptional response of the vessel wall to injury, where both neointimal SMCs and adventitial myofibroblasts reveal a dramatic but transient increase in periostin expression. This increase in expression is paralleled by a fibrotic process with abundant deposition of collagen type I and type III in the adventitia as well as in the neointima.23 Together with collagen genes, periostin was among the most abundant mRNAs associated with vascular injury. It was interesting to note that a number of other ECM-associated molecules were also identified as injury-induced sequences. These included osteopontin, which had previously been reported in injured arteries,24 and several members of the CCN family of proteins. The latter includes connective tissue growth factor (CCN2), NOV (CCN3), and WISP-2 (CCN5), and in an earlier study we had also detected CYR61 (CCN1) in injured arteries.25 All of these CCN family members promote adhesion for a variety of cell types.26 Several of the injury-induced matrix molecules found in our study have recently been shown to play roles in collagen matrix assembly. This applies to osteopontin,27 SPARC,28 and Hevin/SC1,29 and this property would be consistent with the extensive deposition of collagenous matrix in the adventitia and neointima on injury.
The function of periostin is poorly understood; however, the Fas1 domains of periostin and related proteins such as ?ig-h3 have been shown to interact with integrins and to modulate cell adhesion and migration.4,5 Other Fas1 domain-containing proteins include stabilin-1 and stabilin-2, and these also play a role in cell-matrix interactions and adhesion.30 These diverse Fas-1 domain proteins seem to share a common function as proteins that participate in the processes of adhesion, migration, or protein turnover in the ECM and the data presented in this study show that periostin also promotes cell migration.
On balloon catheter injury, SMCs are induced to proliferate and this change from a previously quiescent to a highly proliferative state is accompanied by a decrease in expression of certain SMC markers such as smooth muscle -actin and smoothelin.16,31 In the adventitia, however, TGF-? signaling after arterial injury is responsible for the induction of smooth muscle -actin, causing fibroblasts to adopt the myofibroblast phenotype.23 The levels of periostin expression in the injured arterial wall appear to correlate with the highly proliferative state of neointimal SMCs and adventitial cells (Figure 2B and 2C), and this indicates an association of elevated periostin expression with the proliferative state of those cells. This observation is further substantiated by the fact that little periostin is expressed in nonproliferating SMC found in the normal artery wall (Figure 1) and in arteries at 4 weeks after injury when SMC proliferation is restricted to cells on the surface of the neointima (Figure 2D). SMCs in culture more closely represent the proliferative phenotype seen after arterial injury and the expression of periostin by primary SMC cultures is consistent with this notion. We have previously shown that arterial injury causes the release of endogenous FGF-2, which then induces proliferation of previously quiescent SMCs.17 We have also demonstrated that FGF-2 expression is increased in proliferating SMC in vivo,32 whereas others have shown that FGF-2 is required for SMC proliferation and neointimal lesion formation.33,34 In addition, the proliferative SMC phenotype of the injured artery is less susceptible to proliferative effects of FGF-2.35 Collectively, these data indicate that proliferation of SMC both in vitro and in vivo is, to a large extent, driven by FGF-2, and the high levels of FGF-2 found in primary SMC (Figure 3B) support this concept. This autocrine stimulation of SMC by endogenous FGF-2 may also contribute to the ability of SMC to grow under low serum conditions, and this is in contrast to fibroblast cell lines that exhibit very low thymidine incorporation under the same conditions (data not shown). The highest endogenous FGF-2 levels were observed in primary SMC (RASMC and AoSMC), and this may at least in part explain why these cells did not respond to exogenous FGF-2 with downregulation of periostin mRNA expression. Furthermore, our data suggest that periostin expression in primary SMC and PAC1 SMC does not inversely correlate with FGF signaling. Thus, one characteristic that sets SMCs apart from other mesenchyme-derived cell lines, such as NIH3T3, C3H10T1/2, and MC3T3-E1 cells, is their inability to respond to exogenous FGF-2 with downregulation of periostin mRNA. It should be emphasized that the lack of downregulation of periostin mRNA by FGF-2 in SMC cannot be explained by the absence of FGF receptors, because we and others have shown that proliferating SMC express elevated levels of FGF-R1.32 FGF functions generally as an inhibitor of differentiation, and this is true for neuronal differentiation, myogenic differentiation, as well as osteoblast differentiation, including osteoblast differentiation of MC3T3-E1 cells.36 The presence of periostin expression in unstimulated or BMP-2-stimulated C3H10T1/2, NIH3T3, and MC3T3-E1 cells may reflect a more differentiated phenotype than the corresponding FGF-2-stimulated cultures.
RA functions as a differentiation signal for a variety of cell types including the undifferentiated P19 embryonal carcinoma cells.37 The A404 cells20 used here are derived from P19 cells. Stimulation of these pluripotent cells with RA causes some of them to differentiate into smooth muscle and acquire a periostin-expressing phenotype. Our findings related to periostin expression in smooth muscle is summarized as follows: (1) development of the SMC phenotype from highly undifferentiated cells such as P19 cells correlates with the acquisition of periostin expression; (2) SMC respond to a proliferative stimulus (as seen after arterial injury or in vitro) with an increase in periostin expression; and (3) periostin mRNA levels in SMC are not responsive to stimulation with FGF-2. The inability to downregulate periostin expression with FGF-2 may be a useful criterion to discriminate between smooth muscle and fibroblast lineages. Furthermore, periostin promotes cell migration.
Acknowledgments
This study was supported by National Institutes of Health (NIH) grants HL69182 (V.L.) and NIH grant P20 15555 (V.L., R.E.F., C.P.H.V.) from the COBRE program of the National Center for Research Resources.
Received March 11, 2004; accepted October 20, 2004.
References
Takeshita S, Kikuno R, Tezuka K, Amann E. Osteoblast-specific factor 2: cloning of a putative bone adhesion protein with homology with the insect protein fasciclin I. Biochem J. 1993; 294 (Pt 1): 271–278.
Horiuchi K, Amizuka N, Takeshita S, Takamatsu H, Katsuura M, Ozawa H, Toyama Y, Bonewald LF, Kudo A. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J Bone Miner Res. 1999; 14: 1239–1249.
Sugiura T, Takamatsu H, Kudo A, Amann E. Expression and characterization of murine osteoblast-specific factor 2 (OSF-2) in a baculovirus expression system. Protein Expr Purif. 1995; 6: 305–311.
Kim JE, Jeong HW, Nam JO, Lee BH, Choi JY, Park RW, Park JY, Kim IS. Identification of motifs in the fasciclin domains of the transforming growth factor-beta-induced matrix protein betaig-h3 that interact with the alphavbeta5 integrin. J Biol Chem. 2002; 277: 46159–46165.
Gillan L, Matei D, Fishman DA, Gerbin CS, Karlan BY, Chang DD. Periostin secreted by epithelial ovarian carcinoma is a ligand for alpha(V)beta(3) and alpha(V)beta(5) integrins and promotes cell motility. Cancer Res. 2002; 62: 5358–5364.
Kruzynska-Frejtag A, Machnicki M, Rogers R, Markwald RR, Conway SJ. Periostin (an osteoblast-specific factor) is expressed within the embryonic mouse heart during valve formation. Mech Dev. 2001; 103: 183–188.
Li P, Oparil S, Feng W, Chen YF. Hypoxia-responsive growth factors upregulate periostin and osteopontin expression via distinct pathways in rat pulmonary arterial smooth muscle cells. J Appl Physiol. 2004; 97: 1550–1558.
Clowes AW, Reidy MA, Clowes MM. Kinetics of cellular proliferation after arterial injury. I. Smooth muscle growth in the absence of endothelium. Lab Invest. 1983; 49: 327–333.
Landry DB, Couper LL, Bryant SR, Lindner V. Activation of the NF-kappa B and I kappa B system in smooth muscle cells after rat arterial injury. Induction of vascular cell adhesion molecule-1 and monocyte chemoattractant protein-1. Am J Pathol. 1997; 151: 1085–1095.
Manabe I, Owens GK. Recruitment of serum response factor and hyperacetylation of histones at smooth muscle-specific regulatory regions during differentiation of a novel P19-derived in vitro smooth muscle differentiation system. Circ Res. 2001; 88: 1127–1134.
Pear WS, Nolan GP, Scott ML, Baltimore D. Production of high-titer helper-free retroviruses by transient transfection. Proc Natl Acad Sci U S A. 1993; 90: 8392–8396.
Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997; 88: 593–602.
Bryant SR, Bjercke RJ, Erichsen DA, Rege A, Lindner V. Vascular remodeling in response to altered blood flow is mediated by fibroblast growth factor-2. Circ Res. 1999; 84: 323–328.
Conley BA, Koleva R, Smith JD, Kacer D, Zhang D, Bernabeu C, Vary CP. Endoglin controls cell migration and composition of focal adhesions: function of the cytosolic domain. J Biol Chem. 2004; 279: 27440–27449.
Tai JT, Brooks EE, Liang S, Somogyi R, Rosete JD, Lawn RM, Shiffman D. Determination of temporal expression patterns for multiple genes in the rat carotid artery injury model. Arterioscler Thromb Vasc Biol. 2000; 20: 2184–2191.
Ryan ST, Koteliansky VE, Gotwals PJ, Lindner V. Transforming growth factor-beta-dependent events in vascular remodeling following arterial injury. J Vasc Res. 2003; 40: 37–46.
Lindner V, Olson NE, Clowes AW, Reidy MA. Inhibition of smooth muscle cell proliferation in injured rat arteries. Interaction of heparin with basic fibroblast growth factor. J Clin Invest. 1992; 90: 2044–2049.
Tintut Y, Alfonso Z, Saini T, Radcliff K, Watson K, Bostrom K, Demer LL. Multilineage potential of cells from the artery wall. Circulation. 2003; 108: 2505–2510.
Bostrom K, Demer LL. Regulatory mechanisms in vascular calcification. Crit Rev Eukaryot Gene Expr. 2000; 10: 151–158.
Blank RS, Swartz EA, Thompson MM, Olson EN, Owens GK. A retinoic acid-induced clonal cell line derived from multipotential P19 embryonal carcinoma cells expresses smooth muscle characteristics. Circ Res. 1995; 76: 742–749.
Itoh F, Nakane T, Chiba S. Gene expression of MASH-1, MATH-1, neuroD and NSCL-2, basic helix-loop-helix proteins, during neural differentiation in P19 embryonal carcinoma cells. Tohoku J Exp Med. 1997; 182: 327–336.
Goetsch SC, Hawke TJ, Gallardo TD, Richardson JA, Garry DJ. Transcriptional profiling and regulation of the extracellular matrix during muscle regeneration. Physiol Genomics. 2003; 14: 261–271.
Smith JD, Bryant SR, Couper LL, Vary CP, Gotwals PJ, Koteliansky VE, Lindner V. Soluble transforming growth factor-beta type II receptor inhibits negative remodeling, fibroblast transdifferentiation, and intimal lesion formation but not endothelial growth. Circ Res. 1999; 84: 1212–1222.
Giachelli CM, Bae N, Almeida M, Denhardt DT, Alpers CE, Schwartz SM. Osteopontin is elevated during neointima formation in rat arteries and is a novel component of human atherosclerotic plaques. J Clin Invest. 1993; 92: 1686–1696.
Grzeszkiewicz TM, Lindner V, Chen N, Lam SC, Lau LF. The angiogenic factor cysteine-rich 61 (CYR61, CCN1) supports vascular smooth muscle cell adhesion and stimulates chemotaxis through integrin alpha(6)beta(1) and cell surface heparan sulfate proteoglycans. Endocrinology. 2002; 143: 1441–1450.
Ellis PD, Chen Q, Barker PJ, Metcalfe JC, Kemp PR. Nov gene encodes adhesion factor for vascular smooth muscle cells and is dynamically regulated in response to vascular injury. Arterioscler Thromb Vasc Biol. 2000; 20: 1912–1919.
Liaw L, Birk DE, Ballas CB, Whitsitt JS, Davidson JM, Hogan BL. Altered wound healing in mice lacking a functional osteopontin gene (spp1). J Clin Invest. 1998; 101: 1468–1478.
Puolakkainen P, Bradshaw AD, Kyriakides TR, Reed M, Brekken R, Wight T, Bornstein P, Ratner B, Sage EH. Compromised production of extracellular matrix in mice lacking secreted protein, acidic and rich in cysteine (SPARC) leads to a reduced foreign body reaction to implanted biomaterials. Am J Pathol. 2003; 162: 627–635.
Hambrock HO, Nitsche DP, Hansen U, Bruckner P, Paulsson M, Maurer P, Hartmann U. SC1/hevin. An extracellular calcium-modulated protein that binds collagen I. J Biol Chem. 2003; 278: 11351–11358.
Politz O, Gratchev A, McCourt PA, Schledzewski K, Guillot P, Johansson S, Svineng G, Franke P, Kannicht C, Kzhyshkowska J, Longati P, Velten FW, Goerdt S. Stabilin-1 and -2 constitute a novel family of fasciclin-like hyaluronan receptor homologues. Biochem J. 2002; 362: 155–164.
Bar H, Wende P, Watson L, Denger S, van Eys G, Kreuzer J, Jahn L. Smoothelin is an indicator of reversible phenotype modulation of smooth muscle cells in balloon-injured rat carotid arteries. Basic Res Cardiol. 2002; 97: 9–16.
Lindner V, Reidy MA. Expression of basic fibroblast growth factor and its receptor by smooth muscle cells and endothelium in injured rat arteries. An en face study. Circ Res. 1993; 73: 589–595.
Hanna AK, Fox JC, Neschis DG, Safford SD, Swain JL, Golden MA. Antisense basic fibroblast growth factor gene transfer reduces neointimal thickening after arterial injury. J Vasc Surg. 1997; 25: 320–325.
Miyamoto T, Leconte I, Swain JL, Fox JC. Autocrine FGF signaling is required for vascular smooth muscle cell survival in vitro. J Cell Physiol. 1998; 177: 58–67.
Olson NE, Kozlowski J, Reidy MA. Proliferation of intimal smooth muscle cells. Attenuation of basic fibroblast growth factor 2-stimulated proliferation is associated with increased expression of cell cycle inhibitors. J Biol Chem. 2000; 275: 11270–11277.
Diez del Corral R, Storey KG. Opposing FGF and retinoid pathways: a signalling switch that controls differentiation and patterning onset in the extending vertebrate body axis. Bioessays. 2004; 26: 857–869.
Linney E. Retinoic acid receptors: transcription factors modulating gene regulation, development, and differentiation. Curr Top Dev Biol. 1992; 27: 309–350.(Volkhard Lindner; Qiaozen)
Correspondence to Dr Volkhard Lindner, Center for Molecular Medicine, Maine Medical Center Research Institute, Scarborough, ME 04074. E-mail lindnv@mmc.org
Abstract
Objective— Periostin mRNA is among the most strongly upregulated transcripts in rat carotid arteries after balloon injury. The goal of the present study was to gain insight into the significance of periostin in the vasculature.
Methods and Results— Periostin expression after injury was localized to smooth muscle cells of the neointima and the adventitia. The expression of periostin in smooth muscle cells in vitro was not regulated by cytokines such as fibroblast growth factor-2 (FGF-2). In contrast, stimulation of MC3T3-E1 osteoblastic cells, NIH3T3 fibroblasts, or mesenchymal C3H10T1/2 cells with FGF-2 reduced periostin mRNA levels to <5% of controls, whereas conversely bone morphogenetic protein-2 (BMP-2) increased periostin mRNA levels. Periostin expression was induced and maintained during retinoic acid-induced smooth muscle cell differentiation in A404 cells. In addition, overexpression of periostin in C3H10T1/2 cells caused an increase in cell migration that could be blocked with an anti-periostin antibody.
Conclusions— Periostin expression is associated with smooth muscle cell differentiation in vitro and promotes cell migration. Unlike other mesenchymally derived cell lines, periostin expression is not regulated by FGF-2 in smooth muscle cells. This distinction may be useful in discriminating smooth muscle and fibroblast lineages.
Periostin is an abundant sequence induced on vascular injury arteries. We find that periostin is associated with smooth muscle cell differentiation and that it functions as a migratory stimulus. In addition, inability to downregulate periostin expression with FGF-2 may be a useful criterion to discriminate between smooth muscle and fibroblast lineages.
Key Words: smooth muscle ? adventitia ? myofibroblast ? intimal hyperplasia
Introduction
Tissue remodeling after injury involves the coordinate regulation of numerous genes, including those involved with extracellular matrix (ECM) synthesis and turnover, cell adhesion, and cell migration. Major components of the ECM include structural proteins and a number of soluble ECM-associated proteins. The soluble ECM-associated proteins include SPARC, lumican, and the fasciclin domain-containing proteins ?ig-H3 and periostin. Periostin, originally named osteoblast-specific factor-2 (Genbank D13664) was identified as a soluble ECM protein that was expressed in bone and, to a lesser extent, lung.1 The periostin mRNA is alternatively spliced, resulting in polypeptides that differ in the length of their C-terminal domains and in their tissue-specific patterns of expression.2
Periostin is a 90-kDa heparin-binding N-glycosylated protein that was proposed to associate with bone ECM, and it is highly homologous to ?ig-h3, previously known as the 68-kDa transforming growth factor (TGF)-?1-inducible protein. Both periostin and ?ig-h3 contain 4 tandem fasciclin (Fas1) domains homologous to the insect protein fasciclin.3 In the case of ?ig-h3, the second and fourth fasciclin domains interact with the 3?1 integrin, and all 4 fasciclin domains can interact with the v?5 integrin, thus mediating adhesion.4 Periostin also mediates cell adhesion by binding to v?3 and v?5, and increases cell motility.5 The localization of recombinant periostin with v?5 integrin at sites of focal adhesion5 suggests the contribution of periostin to cell adhesion and motility. Despite these recent observations of periostin interacting with cell surface receptors, its functional roles are not clear.
More recently, periostin expression was found in embryonic hearts by embryonic day E10.5, and increased to a plateau by E14. Periostin mRNA expression was localized to the endocardial cushions that ultimately divide the primitive heart into the mature 4-chambered heart.6 In addition, periostin expression is increased in pulmonary aortic smooth muscle cells (SMCs)in response to hypoxia.7
The goal of the present study was to examine the role of periostin in SMCs. Our data indicate that periostin is associated with SMC differentiation and functions as a migratory stimulus in vitro. After balloon catheter injury, periostin expression coincides with the proliferative phase of SMCs in the neointima.
Methods
Animals
Sprague-Dawley rats were used and the left carotid artery and the aorta were denuded with a 2-French balloon catheter as described.8
Subtractive Hybridization and Cloning of the Periostin cDNA
Suppressive subtractive hybridization was performed with a kit (polymerase chain reaction Select cDNA Subtraction Kit; Clontech), following the manufacturer’s instructions. cDNA was prepared from 2 μg of mRNA extracted from normal rat carotid arteries and aortae, as well as 8-day balloon-injured vessels. For the purpose of isolating sequences overexpressed in injured arteries, cDNA from injured vessels was used as "tester" and cDNA from normal vessels as "driver" cDNA. Partial sequences of 250 clones were obtained by automated sequencing DNA sequence analysis and their identities were determined by searching Genbank databases, including nonredundant and expressed sequence tag databases.
Northern Blotting and cDNA Probes
Total RNA isolated from normal carotid arteries, 8-day balloon-injured carotid arteries, and indicated cell lines were examined by Northern blot analysis as described.9 Blots were hybridized simultaneously using a rat periostin cDNA and a mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) cDNA labeled with -[32P]-dCTP. The intensities of the periostin band and the GAPDH band were quantified by phosphorimaging. The ratio of the signal intensities for periostin/GAPDH before growth factor addition (time point, 0 hours) were set to 1.0 (100%). Identical blots were probed in duplicate with similar results.
In Situ Hybridization
[35S]-UTP labeled RNA probes corresponding to the sense and antisense strand of the coding region of rat periostin were prepared and in situ hybridization was performed on paraffin sections as described.9
Cells and Culture
A404 cells (kindly provided by Dr Gary Owens, University of Virginia School of Medicine) were maintained as described.10 Differentiation of A404 cells was initiated with 1 μmol/L all-trans retinoic acid (RA) for 3 days. After withdrawal of RA after 3 days, selection for SMCs was achieved by addition of 0.5 μg/mL puromycin (Clontech) to the medium for either 3 or 5 days.
Primary rat aortic SMCs (RASMC), MC3T3-E1, C3H10T1/2, NIH3T3, rat PAC1 smooth muscle, ROS17/2.8 osteosarcoma, BOSC23,11 and human aortic SMC (AoSMC) (Clonetics) cell lines were used. Primers for reverse-transcriptase polymerase chain reaction were: GAPDH (5'GGAGATTGTTGCCATCAACGA3', 5'GAAGACACCAGTAGACTCCACGACA3'), achaete-scute homolog-1 (MASH-1, 5'CAAGTTGGTCAACCTGGGTTTTG3', 5'CACTAAAGATGCAGGATCTGCTG3'), periostin (5'AAAGTAAAAGTTGGCCTTAGCGACC3', 5'CAGAAGCTCCCTTTCTTCGCTAGT3'), and smooth muscle myosin heavy chain (SM-MHC, 5'CAGTTGGACACTATGTCAGGGAAA3', 5'ATGGAGACAAATGCTAATCAGCC3').
For stimulation of cells with cytokines, confluent cells were fed with fresh medium containing 10% fetal bovine serum, and 2 days later cells were stimulated with one of the following: fibroblast growth factor-2 (FGF-2) (10 ng/mL), PDGF-AB (10 ng/mL), BMP-2 (25 ng/mL), or TGF-? (10 ng/mL) (all from R&D Systems).
Retrovirus Production and Isolation
BOSC23 cells (Clontech) were transiently transfected with either control pWzl12 or pWzl bearing periostin cDNA. Immunofluorescence analysis was used to confirm the efficiency of retroviral gene expression, which was >95% (data not shown).
Western Blot Analysis
Rabbit anti-periostin antibody (kindly provided by Dr Roger Markwald, Medical University of South Carolina) was used at a 1:1000 dilution. FGF-2 expression levels in the various cell lines were determined by immunoblotting with a mouse anti-FGF-2 antibody13 after SDS-PAGE of 80 μg cell lysate under reducing conditions. Quantification was performed by densitometry, and the relative amounts of FGF-2 present were expressed as a fraction of FGF-2 detected in RASMC (value set at 1.0).
Migration Assay
Migration of cells to 3% serum was assayed in a modified Boyden chamber as described, in 3 separate experiments.14 In some cases, rabbit anti-periostin antibody or control rabbit anti-HA antibody were included at 0.5 μg/mL. The data are means±SEM of 6 wells per experimental condition. Student t test was used to compare the means between the 2 groups, and P0.05 was considered significant.
Results
Periostin Expression Is Increased After Arterial Injury
Suppressive subtractive hybridization was conducted on cDNA derived from 8-day balloon-injured rat carotid arteries and control vessels. A survey of 250 clones obtained using this method identified 3 of them as periostin cDNA fragments (Table). Other identified sequences were various collagens, smoothelin, osteopontin, and SPARC, as well as genes that have been described in other tissues such as connective tissue growth factor (CCN2), hevin/SC1, and WISP2 (CCN5).15 These data are consistent with the view that vascular responses to injury include the regulation of groups of genes that regulate the ECM environment.
Frequency of Selected Upregulated Sequences Among 250 Identified Clones From Balloon-Injured Rat Carotid Arteries
Northern blot analysis performed on RNA isolated from normal and 8-day balloon-injured rat carotid arteries demonstrated increased levels of periostin mRNA in injured vessels and only low levels of expression in normal arteries (Figure 1A). This differential expression pattern was also verified at the protein level using an anti-periostin antibody (Figure 1B). This antibody recognized a major band (molecular weight 90 kDa) and a slightly smaller and less intense band, which could reflect a difference in glycosylation. Among a variety of tissues from adult rats, only lung and, to a lesser extent, uterus revealed significant levels of periostin mRNA (Figure 1A). To identify the cells expressing periostin mRNA within the vessel wall after injury, we performed in situ hybridization with an antisense periostin riboprobe. Low levels of periostin expression were detectable in normal arteries (Figure 2A). Eight days after balloon injury, high levels of periostin mRNA expression were observed in the adventitia, where adventitial fibroblasts are undergoing rapid proliferation at this time16 (Figure 2B). Two weeks after injury, a SMC-rich neointimal lesion has developed and cells throughout the neointima revealed abundant expression of periostin mRNA (Figure 2C). At 4 weeks after injury, only very few intimal SMCs on the luminal surface still expressed periostin mRNA (Figure 2D). En face preparations did not show detectable levels of periostin expression in either normal endothelium (Figure 2E) or wounded endothelium (not shown). The induction of periostin in the adventitia prompted us to examine whether this was a general response of connective tissue to injury. Dermal fibroblasts at sites of full-thickness skin incisional wounding also exhibited abundant expression of periostin mRNA (Figure 2F).
Figure 1. Periostin transcripts are dynamically regulated after vascular injury. A, Northern blot analysis of periostin mRNA expression in vivo. Compared with uninjured rat carotid arteries, increased periostin mRNA levels were detected in arteries 8 days after balloon catheter denudation. In organs from normal adult rats, significant periostin mRNA levels were detectable in the lung and at lower levels also in uterine tissue. B, Western blot analysis of total protein from uninjured and injured arteries 8 days after balloon catheter denudation demonstrates increased periostin protein levels in injured vessels.
Figure 2. Periostin is expressed by adventitial cells and neointimal cells. Photomicrographs of in situ hybridization performed with [35S]-UTP-labeled periostin antisense (A to F) and sense (G) riboprobes. A, In the uninjured rat carotid artery, low levels of periostin expression were detectable in the media (m). B, Eight days after balloon injury, increased periostin expression is detectable in the outer layers of the adventitia (a), where fibroblast proliferation and matrix production are prominent at this time. C, Two weeks after injury, the newly formed neointima (i) reveals abundant expression levels of periostin, whereas (D) at 4 weeks after denudation only scattered cells on the surface of the intimal lesion still express this mRNA. E, An en face preparation of a normal aorta shows only background levels of hybridization in endothelial cells. F, During wound healing, 8 days after a full-thickness skin incision reveals abundant periostin mRNA expression in dermal fibroblasts (d) but not in the keratinocytes of the epidermis (e). G, A section of a carotid artery 2 weeks after balloon injury was hybridized with a periostin sense probe for evaluation of nonspecific background hybridization. All images are seen under dark-field illumination with silver grains appearing as white specks. Nuclei were counterstained with hematoxylin. Original magnification was 200x except for (F), which was 50x.
Regulation of Periostin Expression in SMC and Other Cell Lines
In response to injury, SMCs in the vessel wall are exposed to a number of growth factors and cytokines released from activated platelets, inflammatory cells, and damaged vascular cells.17 For the purpose of identifying growth factors that might be involved in the regulation of periostin expression, we examined the steady-state levels of periostin mRNA after growth factor administration in rat and human primary SMC, as well as in the PAC1 SMC line and NIH3T3 fibroblasts. Because a number of other genes normally expressed in bone are also expressed in the injured arterial wall,18,19 we examined the growth factor-dependent modulation of periostin expression also in the MC3T3-E1 osteoblast cell line, as well as in the osteosarcoma cell line ROS17/2.8. We also included the oligopotent mesenchymal C3H10T1/2 cells in the analysis because they can be differentiated toward smooth muscle. C3H10T1/2, NIH3T3, and MC3T3-E1 cells all responded to FGF-2 with a profound downregulation of periostin (5%, 7%, and 2% of control levels, respectively; Figure 3A). RASMC and human AoSMC (not shown), as well as the PAC1 and ROS17/2.8 cells (not shown), however, demonstrated no significant change in periostin mRNA levels in response to FGF-2 (Figure 3A). BMP-2 caused periostin mRNA levels to increase 2- to 3-fold in C3H10T1/2, NIH3T3, and MC3T3-E1 cells, but this effect was not observed in SMCs or ROS17/2.8 cells. These cell-specific differences with regard to FGF-2-dependent downregulation of periostin prompted us to examine their endogenous FGF-2 levels by immunoblotting because high levels of endogenous FGF-2 could limit the cells ability to respond to exogenous FGF-2. RASMC and AoSMC had by far the highest levels of FGF-2 of all cell lines examined (Figure 3B). For example, FGF-2 levels of RASMC were 11-fold higher than those detected in NIH3T3 cells. The 18-kDa form, as well as higher molecular weight forms, of FGF-2 (21/22kDa in rat RASMC, 24-kDa in human AoSMC) were detected (Figure 3B). These data indicate periostin is a FGF-responsive gene in certain cell types with low endogenous FGF-2 levels.
Figure 3. Periostin expression is regulated by growth factors. A, The time course of periostin mRNA expression in response to BMP-2 (25 ng/mL) and FGF-2 (10 ng/mL) was examined by Northern blotting in various cell lines at the indicated time points. Blots were hybridized simultaneously with a GAPDH cDNA and the quantification of expression was performed with a phosphoimager. Periostin expression was normalized to GAPDH mRNA levels and the fold difference from control levels at 0 hours is shown at the bottom of each lane. B, Immunoblot analysis of cell lysates with anti-FGF-2 antibody; 80 μg protein were loaded in each lane. The amounts of FGF-2 present in the sample relative to RASMC are shown on the bottom of each lane. Note high levels of 18-kDa and high-molecular-weight FGF-2 forms in RASMC and AoSMC.
Periostin and SMC Differentiation
The A404 cell line is a P19-derived line that differentiates into several lineages including SMCs after RA treatment.20 A404 cells bear a stably transfected copy of a smooth muscle -actin promoter-driven puromycin resistance cassette. This cassette permits the elimination of nonSMC lineages in favor of smooth muscle -actin-expressing cells. Reverse-transcription polymerase chain reaction expression analysis of A404 cells that have been induced to differentiate with RA indicates that periostin mRNA is upregulated by day 3 of RA treatment, as shown in Figure 4. RA induced coordinate upregulation of the definitive smooth muscle marker SM-MHC, indicating that a true SMC lineage is present in the differentiating cell population. The neurogenic differentiation marker MASH-121 is also induced by RA as expected, indicating the presence of neural lineage cells. However, unlike periostin and SM-MHC, MASH-1 is eliminated after continued presence of puromycin, which eliminates nonsmooth muscle -actin-expressing cell lineages. These results support the view that periostin expression is responsive to RA in a pluripotent SMC precursor population and that periostin persists in RA-differentiated A404-derived SMC.
Figure 4. Smooth muscle cell differentiation induces periostin expression. Reverse-transcriptase polymerase chain reaction of A404 cells induced to differentiate into smooth muscle cells with all-trans retinoic acid (RA, 1 μmol/L). mRNA and cDNA were prepared from untreated A404 cells (no RA, lane 1) from cells with RA treatment for 3 days (3d RA, lane 2), from cells with RA treatment for 3 days followed by selection with puromycin for an additional 3 days (3d RA plus 3d puro, lane 3), and from cells with RA treatment for 3 days followed by selection with puromycin for an additional 5 days (3d RA plus 5d puro, lane 4). Primers for periostin, SM-MHC, MASH-1, and GAPDH were used for reverse-transcription polymerase chain reaction. In the absence of cDNA addition, no bands were amplified (no RT, lane 5).
Perostin and Cell Migration
C3H10T1/2 were chosen for migration experiments because we were able to achieve high levels of periostin overexpression with retroviral expression systems. Cells were transduced with retroviral constructs of a vector control or a full-length periostin expression construct. Immunofluorescence of retrovirally transduced C3H10T1/2 cells, using anti-periostin antibody, showed that essentially 100% of the infected cell population was expressing periostin (Figure 5B) as compared with empty virus control-infected cells (Figure 5A).
Figure 5. Periostin expression enhances cell migration. Immunofluorescence of C3H10T1/2 cells infected with (A) insertless pWzl retrovirus or (B) virus bearing periostin cDNA using an anti-periostin antibody. C, A representative migration assay with control (10T1/2-vector) and periostin-expressing C3H10T1/2 cells (10T1/2-periostin) is shown. The migration assay was performed in the absence of added antibodies (No Ab) or in the presence of either rabbit anti-HA antibody (Anti-HA) or rabbit anti-periostin antibody (Anti-Peri) using 3% serum as the migration stimulus. Bovine serum albumin (200 μg/mL) was used as the control for migration in the absence of stimulus.
Migration assays of cells to a stimulus with 3% serum were performed in a modified Boyden chamber as described.14 Periostin-overexpressing C3H10T1/2 cells showed a significantly enhanced migratory response to serum (Figure 5C). As shown in Figure 5C, periostin-dependent enhancement of cell migration was unchanged in the presence of a rabbit anti-HA antibody, used here as an isotype-matched irrelevant antibody. Periostin-overexpressing cells, but not control cells, showed a significant reduction of cell migration after treatment with the anti-periostin antibody (Figure 5C). These data indicate that periostin overexpression stimulates cell migration in C3H10T1/2 cells and that this stimulation is caused by increased periostin expressed outside the cell.
Discussion
Periostin was originally identified as a Fas1 domain-containing osteoblast-specific factor. Periostin has been suggested to function as a cell adhesion molecule for preosteoblasts and to participate in osteoblast recruitment, attachment, and spreading.2 Since then, several studies have demonstrated that periostin may have a role in diverse tissues because periostin was found to be expressed in the developing mouse embryonic and fetal heart.6 Furthermore, periostin expression was observed in skeletal muscle after injury, where it revealed coordinated expression with other ECM constituents including the collagen genes.22 The present study demonstrates that periostin is also part of the coordinated transcriptional response of the vessel wall to injury, where both neointimal SMCs and adventitial myofibroblasts reveal a dramatic but transient increase in periostin expression. This increase in expression is paralleled by a fibrotic process with abundant deposition of collagen type I and type III in the adventitia as well as in the neointima.23 Together with collagen genes, periostin was among the most abundant mRNAs associated with vascular injury. It was interesting to note that a number of other ECM-associated molecules were also identified as injury-induced sequences. These included osteopontin, which had previously been reported in injured arteries,24 and several members of the CCN family of proteins. The latter includes connective tissue growth factor (CCN2), NOV (CCN3), and WISP-2 (CCN5), and in an earlier study we had also detected CYR61 (CCN1) in injured arteries.25 All of these CCN family members promote adhesion for a variety of cell types.26 Several of the injury-induced matrix molecules found in our study have recently been shown to play roles in collagen matrix assembly. This applies to osteopontin,27 SPARC,28 and Hevin/SC1,29 and this property would be consistent with the extensive deposition of collagenous matrix in the adventitia and neointima on injury.
The function of periostin is poorly understood; however, the Fas1 domains of periostin and related proteins such as ?ig-h3 have been shown to interact with integrins and to modulate cell adhesion and migration.4,5 Other Fas1 domain-containing proteins include stabilin-1 and stabilin-2, and these also play a role in cell-matrix interactions and adhesion.30 These diverse Fas-1 domain proteins seem to share a common function as proteins that participate in the processes of adhesion, migration, or protein turnover in the ECM and the data presented in this study show that periostin also promotes cell migration.
On balloon catheter injury, SMCs are induced to proliferate and this change from a previously quiescent to a highly proliferative state is accompanied by a decrease in expression of certain SMC markers such as smooth muscle -actin and smoothelin.16,31 In the adventitia, however, TGF-? signaling after arterial injury is responsible for the induction of smooth muscle -actin, causing fibroblasts to adopt the myofibroblast phenotype.23 The levels of periostin expression in the injured arterial wall appear to correlate with the highly proliferative state of neointimal SMCs and adventitial cells (Figure 2B and 2C), and this indicates an association of elevated periostin expression with the proliferative state of those cells. This observation is further substantiated by the fact that little periostin is expressed in nonproliferating SMC found in the normal artery wall (Figure 1) and in arteries at 4 weeks after injury when SMC proliferation is restricted to cells on the surface of the neointima (Figure 2D). SMCs in culture more closely represent the proliferative phenotype seen after arterial injury and the expression of periostin by primary SMC cultures is consistent with this notion. We have previously shown that arterial injury causes the release of endogenous FGF-2, which then induces proliferation of previously quiescent SMCs.17 We have also demonstrated that FGF-2 expression is increased in proliferating SMC in vivo,32 whereas others have shown that FGF-2 is required for SMC proliferation and neointimal lesion formation.33,34 In addition, the proliferative SMC phenotype of the injured artery is less susceptible to proliferative effects of FGF-2.35 Collectively, these data indicate that proliferation of SMC both in vitro and in vivo is, to a large extent, driven by FGF-2, and the high levels of FGF-2 found in primary SMC (Figure 3B) support this concept. This autocrine stimulation of SMC by endogenous FGF-2 may also contribute to the ability of SMC to grow under low serum conditions, and this is in contrast to fibroblast cell lines that exhibit very low thymidine incorporation under the same conditions (data not shown). The highest endogenous FGF-2 levels were observed in primary SMC (RASMC and AoSMC), and this may at least in part explain why these cells did not respond to exogenous FGF-2 with downregulation of periostin mRNA expression. Furthermore, our data suggest that periostin expression in primary SMC and PAC1 SMC does not inversely correlate with FGF signaling. Thus, one characteristic that sets SMCs apart from other mesenchyme-derived cell lines, such as NIH3T3, C3H10T1/2, and MC3T3-E1 cells, is their inability to respond to exogenous FGF-2 with downregulation of periostin mRNA. It should be emphasized that the lack of downregulation of periostin mRNA by FGF-2 in SMC cannot be explained by the absence of FGF receptors, because we and others have shown that proliferating SMC express elevated levels of FGF-R1.32 FGF functions generally as an inhibitor of differentiation, and this is true for neuronal differentiation, myogenic differentiation, as well as osteoblast differentiation, including osteoblast differentiation of MC3T3-E1 cells.36 The presence of periostin expression in unstimulated or BMP-2-stimulated C3H10T1/2, NIH3T3, and MC3T3-E1 cells may reflect a more differentiated phenotype than the corresponding FGF-2-stimulated cultures.
RA functions as a differentiation signal for a variety of cell types including the undifferentiated P19 embryonal carcinoma cells.37 The A404 cells20 used here are derived from P19 cells. Stimulation of these pluripotent cells with RA causes some of them to differentiate into smooth muscle and acquire a periostin-expressing phenotype. Our findings related to periostin expression in smooth muscle is summarized as follows: (1) development of the SMC phenotype from highly undifferentiated cells such as P19 cells correlates with the acquisition of periostin expression; (2) SMC respond to a proliferative stimulus (as seen after arterial injury or in vitro) with an increase in periostin expression; and (3) periostin mRNA levels in SMC are not responsive to stimulation with FGF-2. The inability to downregulate periostin expression with FGF-2 may be a useful criterion to discriminate between smooth muscle and fibroblast lineages. Furthermore, periostin promotes cell migration.
Acknowledgments
This study was supported by National Institutes of Health (NIH) grants HL69182 (V.L.) and NIH grant P20 15555 (V.L., R.E.F., C.P.H.V.) from the COBRE program of the National Center for Research Resources.
Received March 11, 2004; accepted October 20, 2004.
References
Takeshita S, Kikuno R, Tezuka K, Amann E. Osteoblast-specific factor 2: cloning of a putative bone adhesion protein with homology with the insect protein fasciclin I. Biochem J. 1993; 294 (Pt 1): 271–278.
Horiuchi K, Amizuka N, Takeshita S, Takamatsu H, Katsuura M, Ozawa H, Toyama Y, Bonewald LF, Kudo A. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J Bone Miner Res. 1999; 14: 1239–1249.
Sugiura T, Takamatsu H, Kudo A, Amann E. Expression and characterization of murine osteoblast-specific factor 2 (OSF-2) in a baculovirus expression system. Protein Expr Purif. 1995; 6: 305–311.
Kim JE, Jeong HW, Nam JO, Lee BH, Choi JY, Park RW, Park JY, Kim IS. Identification of motifs in the fasciclin domains of the transforming growth factor-beta-induced matrix protein betaig-h3 that interact with the alphavbeta5 integrin. J Biol Chem. 2002; 277: 46159–46165.
Gillan L, Matei D, Fishman DA, Gerbin CS, Karlan BY, Chang DD. Periostin secreted by epithelial ovarian carcinoma is a ligand for alpha(V)beta(3) and alpha(V)beta(5) integrins and promotes cell motility. Cancer Res. 2002; 62: 5358–5364.
Kruzynska-Frejtag A, Machnicki M, Rogers R, Markwald RR, Conway SJ. Periostin (an osteoblast-specific factor) is expressed within the embryonic mouse heart during valve formation. Mech Dev. 2001; 103: 183–188.
Li P, Oparil S, Feng W, Chen YF. Hypoxia-responsive growth factors upregulate periostin and osteopontin expression via distinct pathways in rat pulmonary arterial smooth muscle cells. J Appl Physiol. 2004; 97: 1550–1558.
Clowes AW, Reidy MA, Clowes MM. Kinetics of cellular proliferation after arterial injury. I. Smooth muscle growth in the absence of endothelium. Lab Invest. 1983; 49: 327–333.
Landry DB, Couper LL, Bryant SR, Lindner V. Activation of the NF-kappa B and I kappa B system in smooth muscle cells after rat arterial injury. Induction of vascular cell adhesion molecule-1 and monocyte chemoattractant protein-1. Am J Pathol. 1997; 151: 1085–1095.
Manabe I, Owens GK. Recruitment of serum response factor and hyperacetylation of histones at smooth muscle-specific regulatory regions during differentiation of a novel P19-derived in vitro smooth muscle differentiation system. Circ Res. 2001; 88: 1127–1134.
Pear WS, Nolan GP, Scott ML, Baltimore D. Production of high-titer helper-free retroviruses by transient transfection. Proc Natl Acad Sci U S A. 1993; 90: 8392–8396.
Serrano M, Lin AW, McCurrach ME, Beach D, Lowe SW. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell. 1997; 88: 593–602.
Bryant SR, Bjercke RJ, Erichsen DA, Rege A, Lindner V. Vascular remodeling in response to altered blood flow is mediated by fibroblast growth factor-2. Circ Res. 1999; 84: 323–328.
Conley BA, Koleva R, Smith JD, Kacer D, Zhang D, Bernabeu C, Vary CP. Endoglin controls cell migration and composition of focal adhesions: function of the cytosolic domain. J Biol Chem. 2004; 279: 27440–27449.
Tai JT, Brooks EE, Liang S, Somogyi R, Rosete JD, Lawn RM, Shiffman D. Determination of temporal expression patterns for multiple genes in the rat carotid artery injury model. Arterioscler Thromb Vasc Biol. 2000; 20: 2184–2191.
Ryan ST, Koteliansky VE, Gotwals PJ, Lindner V. Transforming growth factor-beta-dependent events in vascular remodeling following arterial injury. J Vasc Res. 2003; 40: 37–46.
Lindner V, Olson NE, Clowes AW, Reidy MA. Inhibition of smooth muscle cell proliferation in injured rat arteries. Interaction of heparin with basic fibroblast growth factor. J Clin Invest. 1992; 90: 2044–2049.
Tintut Y, Alfonso Z, Saini T, Radcliff K, Watson K, Bostrom K, Demer LL. Multilineage potential of cells from the artery wall. Circulation. 2003; 108: 2505–2510.
Bostrom K, Demer LL. Regulatory mechanisms in vascular calcification. Crit Rev Eukaryot Gene Expr. 2000; 10: 151–158.
Blank RS, Swartz EA, Thompson MM, Olson EN, Owens GK. A retinoic acid-induced clonal cell line derived from multipotential P19 embryonal carcinoma cells expresses smooth muscle characteristics. Circ Res. 1995; 76: 742–749.
Itoh F, Nakane T, Chiba S. Gene expression of MASH-1, MATH-1, neuroD and NSCL-2, basic helix-loop-helix proteins, during neural differentiation in P19 embryonal carcinoma cells. Tohoku J Exp Med. 1997; 182: 327–336.
Goetsch SC, Hawke TJ, Gallardo TD, Richardson JA, Garry DJ. Transcriptional profiling and regulation of the extracellular matrix during muscle regeneration. Physiol Genomics. 2003; 14: 261–271.
Smith JD, Bryant SR, Couper LL, Vary CP, Gotwals PJ, Koteliansky VE, Lindner V. Soluble transforming growth factor-beta type II receptor inhibits negative remodeling, fibroblast transdifferentiation, and intimal lesion formation but not endothelial growth. Circ Res. 1999; 84: 1212–1222.
Giachelli CM, Bae N, Almeida M, Denhardt DT, Alpers CE, Schwartz SM. Osteopontin is elevated during neointima formation in rat arteries and is a novel component of human atherosclerotic plaques. J Clin Invest. 1993; 92: 1686–1696.
Grzeszkiewicz TM, Lindner V, Chen N, Lam SC, Lau LF. The angiogenic factor cysteine-rich 61 (CYR61, CCN1) supports vascular smooth muscle cell adhesion and stimulates chemotaxis through integrin alpha(6)beta(1) and cell surface heparan sulfate proteoglycans. Endocrinology. 2002; 143: 1441–1450.
Ellis PD, Chen Q, Barker PJ, Metcalfe JC, Kemp PR. Nov gene encodes adhesion factor for vascular smooth muscle cells and is dynamically regulated in response to vascular injury. Arterioscler Thromb Vasc Biol. 2000; 20: 1912–1919.
Liaw L, Birk DE, Ballas CB, Whitsitt JS, Davidson JM, Hogan BL. Altered wound healing in mice lacking a functional osteopontin gene (spp1). J Clin Invest. 1998; 101: 1468–1478.
Puolakkainen P, Bradshaw AD, Kyriakides TR, Reed M, Brekken R, Wight T, Bornstein P, Ratner B, Sage EH. Compromised production of extracellular matrix in mice lacking secreted protein, acidic and rich in cysteine (SPARC) leads to a reduced foreign body reaction to implanted biomaterials. Am J Pathol. 2003; 162: 627–635.
Hambrock HO, Nitsche DP, Hansen U, Bruckner P, Paulsson M, Maurer P, Hartmann U. SC1/hevin. An extracellular calcium-modulated protein that binds collagen I. J Biol Chem. 2003; 278: 11351–11358.
Politz O, Gratchev A, McCourt PA, Schledzewski K, Guillot P, Johansson S, Svineng G, Franke P, Kannicht C, Kzhyshkowska J, Longati P, Velten FW, Goerdt S. Stabilin-1 and -2 constitute a novel family of fasciclin-like hyaluronan receptor homologues. Biochem J. 2002; 362: 155–164.
Bar H, Wende P, Watson L, Denger S, van Eys G, Kreuzer J, Jahn L. Smoothelin is an indicator of reversible phenotype modulation of smooth muscle cells in balloon-injured rat carotid arteries. Basic Res Cardiol. 2002; 97: 9–16.
Lindner V, Reidy MA. Expression of basic fibroblast growth factor and its receptor by smooth muscle cells and endothelium in injured rat arteries. An en face study. Circ Res. 1993; 73: 589–595.
Hanna AK, Fox JC, Neschis DG, Safford SD, Swain JL, Golden MA. Antisense basic fibroblast growth factor gene transfer reduces neointimal thickening after arterial injury. J Vasc Surg. 1997; 25: 320–325.
Miyamoto T, Leconte I, Swain JL, Fox JC. Autocrine FGF signaling is required for vascular smooth muscle cell survival in vitro. J Cell Physiol. 1998; 177: 58–67.
Olson NE, Kozlowski J, Reidy MA. Proliferation of intimal smooth muscle cells. Attenuation of basic fibroblast growth factor 2-stimulated proliferation is associated with increased expression of cell cycle inhibitors. J Biol Chem. 2000; 275: 11270–11277.
Diez del Corral R, Storey KG. Opposing FGF and retinoid pathways: a signalling switch that controls differentiation and patterning onset in the extending vertebrate body axis. Bioessays. 2004; 26: 857–869.
Linney E. Retinoic acid receptors: transcription factors modulating gene regulation, development, and differentiation. Curr Top Dev Biol. 1992; 27: 309–350.(Volkhard Lindner; Qiaozen)