Background
Idiopathic pulmonary fibrosis (IPF) is a characteristically progressive chronic lung disease, with irreversible lung scarring and the histological features of interstitial fibrosis of lung tissues, and unknown etiology [
1]. IPF is devastating; the median survival of IPF patients is 3 to 5 years. Aberrant activation of the lung epithelium triggers pulmonary fibrosis, producing mediators of fibroblast migration, proliferation, and differentiation into active myofibroblasts. The fibroblasts and myofibroblasts then secrete exaggerated amounts of extracellular matrix (ECM) proteins, which then remodel the lung architecture.
Fibroblast proliferation is a hallmark of IPF, as is the secretion of ECM proteins from fibroblasts [
2]. It is still uncertain whether, in the lung tissues of IPF patients, the progressive proliferation of fibroblasts is programmed in fibroblasts themselves or is in some way influenced by the extracellular milieu. Several trials have been performed to compare gene profiles in fibroblasts derived from IPF patients and normal donors [
3‐
5], showing that gene expression patterns differ in these two types of fibroblasts. Expression of fibrosis-related genes (
IGFBP3,
IGFBP7,
LOX, and
POSTN) chemokines (
CCL2,
CCL8, and
CCL26), and growth factors such as
FGF7 are upregulated in IPF fibroblasts. Several proliferation-related genes such as
WNT5A and
RGCC are sporadically observed in the gene profiles of upregulated genes in IPF fibroblasts. However, we are far from understanding how IPF fibroblasts acquire the ability to progressively proliferate.
Periostin encoded by the
POSTN gene is a matricellular protein of 93.3 kDa in size belonging to the fasciclin family and is involved in the pathogenesis of various inflammatory and fibrotic diseases by accelerating inflammation or fibrosis [
6,
7]. We and others have demonstrated that periostin is highly expressed in the lung tissue of IPF patients [
8‐
12]. It is of note that expression of periostin is significant in fibroblastic foci in which fibrosis is active and that upon stimulation by either IL-4 or IL-13 periostin can be detected in the supernatant of lung fibroblasts, but not of airway epithelial cells [
8‐
10,
13], suggesting that fibroblasts are main sources of periostin in lung. Moreover, we and another group have shown that a genetic deficiency of periostin or the administration of neutralizing antibodies (Abs) against periostin protected mice from bleomycin (BLM)-induced pulmonary fibrosis [
10,
14], suggesting the significance of periostin in generating pulmonary fibrosis. Periostin acts by binding several integrin molecules—α
Vβ
1, α
Vβ
3, α
Vβ
5, α
6β
4, and α
Mβ
2—on cell surfaces [
7]. We have previously shown that periostin derived from fibroblasts acts on fibroblasts by co-operating with inflammatory cytokines such as TNF-α activating NF-κB, followed by inducing pro-inflammatory cytokines or chemokines [
14]. This is one underlying mechanism by which periostin causes pulmonary fibrosis. Moreover, we have recently found that cross-talk between TGF-β, a critical mediator for pulmonary fibrosis, and periostin via α
Vβ
3 integrin is important for generating pulmonary fibrosis [
15]. However, it has remained undetermined whether or how periostin affects the proliferation of fibroblasts.
In this study, we first aimed to identify periostin-dependently expressed genes in lung fibroblasts, comparing the gene profile in periostin-silenced fibroblasts and found that many cell-cycle–related genes are involved in this profile. Accordingly, periostin- or integrin αVβ3-silenced fibroblasts showed slower proliferation. Lung fibroblasts derived from IPF patients also required periostin for maximum proliferation. Moreover, an inhibitor of integrin αVβ3, a periostin receptor, downregulated proliferation along with expression of cell-cycle–related genes in IPF lung fibroblasts as well as in normal lung fibroblasts. These results offer the first formal proof that periostin plays a critical role in the proliferation of lung fibroblasts.
Methods
Cell culture
MRC-5 cells (Riken BioResource Center, Tsukuba, Japan) were maintained as previously described [
16]. NHLFs (normal human lung fibroblasts) were purchased from Lonza (Basel, Switzerland). RNA extracts were applied to quantitative reverse transcription PCR (qRT-PCR). Five clones of lung fibroblasts were cultured from the explanted lungs of IPF patients undergoing lung transplantation and nine clones of lung fibroblasts were also cultured from normal donor lungs that were not used for transplantation [
3,
17].
Knockdown of mRNA by siRNA
siRNA oligonucleotides were purchased from Dharmacon/GE Healthcare (Lafayette, CO, USA). Cells were transfected with ON-TARGET plus siRNA for POSTN, ITGAV, ITGB3, or control at the indicated concentrations and for the indicated times in the presence of RNAiMAX reagent (Thermo Fisher Scientific, Rockford, IL, USA).
DNA microarray analysis
MRC-5 cells were transfected with 10 nM periostin siRNA for 48 h. Total RNA with an RNA integrity number greater than 9.2 was applied to Agilent Expression Array (SurePrint G3 Human GE8x60K v2 Microarray, Takara Bio, Shiga, Japan). The calculated relative signal intensity values were presented on a heat map and subjected to MultiExperiment Viewer (MeV) v4.9 software (Dana-Farber Cancer Institute, Boston, MA, USA). For gene ontology analysis, the Database for Annotation, Visualization, and Integrated Discovery (DAVID) tool (National Cancer Institute, Frederick, MD, USA) was used. This database includes the Gene Ontology Database (
http://geneontology.org/, GSE132917).
Cell proliferation assay
After incubation in serum-free medium for 24 h, cells were treated with control or with periostin siRNA. Cell proliferation was evaluated using a Cell Counting kit-8 (Dojindo, Kumamoto, Japan) or BrdU Cell Proliferation ELISA Kit (Abcam, Cambridge, UK).
Flow cytometry
To analyze cell death, the cells were treated with 50 μg/mL cycloheximide (Wako, Osaka, Japan) or 50 ng/mL TNF-α (PeproTech, Rocky Hill, NJ, USA) or either control siRNA or periostin siRNA for the indicated times. After harvested cells were labeled with annexin V FITC (Apoptosis Detection Kit I, BD Biosciences, Tokyo, Japan) and propidium iodide (PI, Sigma-Aldrich, San Diego, CA, USA), the cells were subjected to flow cytometry analysis using FACSCalibur (BD Biosciences).
For cell cycle analysis, MRC-5 cells and NHLFs were treated with control siRNA or periostin siRNA and then were fixed in ice-cold 70% ethanol for 1 h. After washing, cells were incubated in PI staining buffer (50 μg/mL PI, 0.1 mg/mL RNase A and 0.5% Triton X-100) for 30 min at 37 °C in the dark. After staining with PI (Sigma-Aldrich), the DNA content was analyzed by flow cytometry.
qRT-PCR
qRT-PCR was performed as previously described [
15]. Primers for qRT-PCR are described in Additional file
2: Table S1.
ELISA
ELISA for human periostin was performed using two rat anti-human periostin mAbs, SS18A and SS17B (Shino-test, Tokyo, Japan), as previously reported [
15.
Adhesion assay
MRC-5 cells were transfected with 10 nM periostin siRNA for 72 h. Detached cells in the culture medium were collected to the tube and centrifuged. The cells were resuspended in the culture medium and counted the number of cells.
Recombinant human periostin protein
Recombinant human periostin protein was purchased from R&D systems (Minneapolis, MN, USA). Ten μg/mL of recombinant human periostin protein was added for the cell proliferation assay.
Transient transfection
Overexpression of periostin was performed as previously described [
15].
αVβ3 inhibitor
CP4715, an α
Vβ
3 integrin inhibitor, was prepared as previously described [
18‐
21]. In some experiments, MRC-5 cells and IPF lung fibroblasts were treated with 1 μM CP4715 dissolved in DMSO for 24 h.
Statistical analysis
Data are presented as mean ± SD. Statistical analyses were performed using the Prism 5.0 software (GraphPad Software, La Jolla, CA, USA). Significance was assessed using an unpaired or paired Student’s t-test. Values of P < 0.05 were considered statistically significant.
Discussion
Periostin is a matricellular protein that exerts various effects on cells by binding to several integrins on the cell surface [
6]. The ability of periostin to promote cell growth in cancer cells has been well studied; either exposure of periostin, transfection of the periostin gene, or co-existence of periostin-producing cells can enhance proliferation of cancer cells [
28‐
33]. This may be an underlying mechanism explaining why, in cancer, high periostin levels reflect aggressive tumor behavior, advanced stage, and poor prognosis [
34]. Activation of the Erk pathways and the cross-talk with EGF signals have both been proposed as the underlying mechanism of how periostin accelerates proliferation of cancer cells [
31‐
33]. Moreover, it has been reported that periostin induces cell cycle reentry in cardiomyocytes, followed by improving ventricular remodeling and cardiac function after myocardial infarction [
35], although these effects are still controversial [
36]. Our present study shows that periostin is required for maximal proliferation of normal lung fibroblasts and, moreover, that IPF lung fibroblasts retain this activity. We found that neither exposure to periostin nor overexpression of periostin enhances proliferation or expression of cell cycle-related genes in lung fibroblasts (Additional file
1: Figures S1 and S2). These results may suggest that the cell cycle in lung fibroblasts is strictly regulated compared with cancer cells [
37], and that excess amounts of periostin do not add proliferative effects on lung fibroblasts in vitro. It has been reported that the negatively regulatory mechanism of cell cycle is impaired in IPF patients [
38]. In such a situation, stimulation by periostin may enhance cell proliferation of lung fibroblasts. The finding that lung structure is normally maintained in periostin-deficient mice points to a dispensable role for periostin on proliferation of lung fibroblasts at steady state [
14]. However, given the aggressive status of proliferation for lung fibroblasts such as IPF, expression level of periostin may make a difference in expansion of fibroblasts.
It is widely accepted that the mitogen and integrin signals via the PI3K/Akt and Erk pathways are important for the transition of the cell cycle from the G1 to the S phase, the first checkpoint of the cell cycle [
39]. Integrin signals are unique among integrin members [
39,
40] and periostin is a unique ligand for α
Vβ
3 because periostin does not have an RGD sequence like other ligands such as vitronectin, osteopontin, and fibronectin. Nevertheless, the cell cycle analysis in the present study shows that periostin is important for the G1/S transition of the cell cycle as well as other integrin ligands (Fig.
4). In the G1 phase, the cyclin D1/CDK4 complex phosphorylates Rb protein, followed by dissociation of E2F from Rb [
36]. E2F increases transcription of cyclin E followed by formation of the cyclin E/CDK2 complex, which furthers Rb phosphorylation. Our present study shows that periostin is important for the expression of cyclin E2, CDK2, and E2F members, but not of cyclin D1 or CDK4. These results suggest that periostin promotes G1/S transition by enhancing production of the cyclin E/CDK2 complex via E2F members, rather than via the cyclin D/CDK4 complex. Although periostin is important for expression of the G2/M phase–related molecules―cyclin B, CDK1, B-MYB, and FOXM1―periostin silencing is unlikely to cause obvious impairment of the G2/M checkpoint.
It is appreciated that fibroblasts taken from IPF patients and cultured in vitro still retain the characteristics of the fibroblasts in vivo in IPF patients. IPF fibroblasts display enhanced proliferation on polymerized collagen matrices [
41]. Moreover, profiles of expressed genes differ between IPF patients and normal donors [
3‐
5] . These differences include several signature molecules of IPF such as IGFBP-3 and lysyl oxidase [
3]. Lee et al. have reported that periostin expression is enhanced in fibroblasts derived from IPF patients, although expression levels of periostin vary among the clones [
5]. We observed that some clones of IPF lung fibroblasts show high expression of periostin compared to normal lung fibroblasts, whereas there was no statistical significance because of the heterogeneity of IPF lung fibroblasts (Fig.
5). The concept of the heterogeneity of IPF lung fibroblasts is consistent with our previous finding that high expression of periostin is relatively limited to the fibroblastic foci, which are not broadly observed in the lungs of IPF patients [
7]. In spite of the heterogeneity of periostin expression in IPF lung fibroblasts, all IPF lung fibroblasts retain the effects of periostin silencing on cell growth and expression of cell-cycle–related molecules (Fig.
5). These results suggest that neither programming nor the extracellular milieu in IPF affect the signal pathway of periostin for proliferation of lung fibroblasts.
Sadly, the median survival for IPF patients is only 3 to 5 years. Thus far, only two drugs, pirfenidone and nintedanib, have been approved by by FDA to treat IPF, and the efficacy of these drugs is limited. There is still an unmet need to develop a novel and effective therapeutic drug to treat IPF. Given that periostin is a key molecule in the pathogenesis of pulmonary fibrosis, it is a promising therapeutic target for IPF. Building on this concept, we have recently found that cross-talk between TGF-β and periostin is important for the generation of pulmonary fibrosis and that CP4715, a potent inhibitor of integrin α
Vβ
3, improves pulmonary fibrosis in mice by inhibiting TGF-β signaling [
15]. Our present study shows that CP4715 has a potent ability to slow proliferation of IPF fibroblasts, as does periostin silencing, although CP4715 has weaker abilities to downregulate cell-cycle–related genes than periostin silencing (Fig.
6). These results give us a basis for applying inhibitors of the periostin/integrin α
Vβ
3 interaction to IPF patients.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.