Background
Bone marrow-derived mesenchymal stem cells are a population of multipotent adult stem cells, distinct from haematopoietic stem cells, that classically can differentiate into mesodermal lineages including osteoblasts, chondrocytes, adipocytes, and cardiomyocytes [
1‐
3]. Numerous reports also suggest differentiation into other, non-mesodermal lineages including neurons [
4,
5], hepatocytes [
6] and lung epithelial cells [
7‐
9].
This evidence provides a strong rational for the potential application of hMSCs in regenerative therapeutic approaches in many diseases including those of the lung where effective treatment options may be limited [
10]. Idiopathic pulmonary fibrosis (IPF) is a chronic, progressive fibrotic lung disorder of unknown aetiology and the most common and lethal form of interstitial lung diseases with a post diagnosis median survival time of 3–5 years irrespective of its treatment status [
11]. Hypothetically, the pathophysiology of IPF is most likely associated with multiple alveolar injuries, failure or delayed alveolar reepithelialisation, abnormal immune responses and subsequent fibrosis [
12‐
14]. Studies involving the bleomycin-induced pulmonary fibrosis mouse model, a widely used animal model of pulmonary fibrosis [
15], demonstrated the migration and homing of endotracheal or systematically transplanted MSCs towards the site of injury and attenuation of pulmonary fibrosis [
16,
17]. However, the magnitude of the amelioration of fibrosis appeared out of proportion to the numbers of engrafted MSCs which had differentiated into alveolar epithelial cells (AECs) indicating the involvement of other mechanisms in this MSC-mediated reparative process. An emerging consensus is that paracrine mechanisms could be associated with MSC-mediated wound repair and tissue regenerative process [
18]. However, the identity of these paracrine factors with a putative role in alveolar injury repair and regeneration is not clear.
A wide range of different growth factors, cytokines and extracellular matrix proteins (ECM) have been identified as constituents of the in vitro cultured MSC secretome [
19,
20]. Many of these secretory proteins are biologically active with anti-inflammatory, anti-fibrotic and immunomodulatory functions [
21,
22]. Previous reports have demonstrated that conditioned media obtained from MSC culture improved cutaneous wound healing [
19,
23] and cardiac repair [
24]. However, data supporting the role of MSC-secreted paracrine factors in the mediation of AEC wound repair is absent. In this study, we have tested hMSC serum-free conditioned media (CM) on AEC wound repair using an in vitro scratch wound repair assay. We demonstrate that hMSC-CM alone increased AEC migration, contained an array of secretory proteins, but had little impact on the rate of wound repair. However, supplementation of hMSC-CM with trace levels of serum (0.2%) significantly increased both migration and wound repair. A selected cohort of hMSC secretory proteins were tested for their effect on repair of AEC and small airway epithelial cells (SAEC) isolated from small airways of distal human lung, and diverse effects on wound healing and cell migration were noted. By developing a direct contact co-culture wound repair system we also demonstrated that when placed in close proximity, hMSCs would migrate into, and repair, AEC wounds in vitro. These findings provide an insight in understanding the cellular and paracrine effects of hMSC in alveolar wound repair and for possible application of hMSC secretory products or their equivalent recombinant proteins as an alternative pharmacoregenerative therapeutic option for pulmonary fibrosis.
Discussion
Human mesenchymal stem cells (hMSC) are in current clinical trials for incurable diseases including osteogenesis imperfecta, graft-versus-host disease, chronic ischemic heart disease, and chronic obstructive pulmonary disease (
http://www.clinicaltrials.gov) and are the focus of many other clinical applications. Studies on animal pulmonary fibrosis models demonstrate that intravenous and endotracheal administration of MSCs attenuate lung injury and fibrosis suggesting a potential clinical application of MSCs for the treatment of IPF [
16,
17,
31]. However, the mechanism of MSC-mediated amelioration of pulmonary fibrosis is not clear and an active participation of MSCs through differentiation into AEC and lung regeneration is under debate [
16,
32]. Bleomycin-induced mouse lung fibrosis models have demonstrated an MSC stimulated reduction in pulmonary fibrosis via inhibition of pro-fibrotic cytokines TNF-α and IL-1 through a paracrine mechanism [
33]. Furthermore, rat models of pulmonary emphysema demonstrated that MSCs reduce AEC apoptosis through paracrine mechanism via up-regulation of anti-apoptotic Bcl-2 gene [
34]. Here we demonstrate that in response to injury, hMSC display site-specific migration into AEC wounds; this observation has also been noted by others in animal lung injury models [
16,
17]. Coupled to this hMSC migration, we also demonstrate that in response to SF-MSC CM, AEC and SAEC migration and wound repair occur with distinct trace serum augmentation requirements. We also provide evidence supporting specific hMSC-secreted paracrine components for effective alveolar and small airway epithelial wound repair.
hMSC secret a myriad of proteins that include growth factors, cytokines and ECM proteins [
19,
20]. Many of these secretory proteins are biologically active with anti-inflammatory, anti-fibrotic and immunomodulatory functions [
21,
22]. Previous and current studies are mostly focused on evaluation of MSC-secreted growth factors and their effects on tissue repair. Here, in line with others, our mass spectrometry-mediated hMSC secretome analyses indicate the presence of a high abundance of ECM/matricellular protein components with diverse biological activities on wound healing and tissue repair [
20,
23]. Fibronectin (a multifunctional glycoprotein), Lumican (collagen-binding keratin sulfate proteoglycan), Periostin (matricellular N-glycoprotein), and IGFBP-7 (IGF-I, -II low affinity binding protein) were identified as major components of SF-MSC CM through high peptide counts and ion score confidence intervals (Table
1). Data supportive of a role for these factors in alveolar wound repair is relatively scarce though implied in previous studies. Fibronectin upregulation is implicated in abdominal wall, corneal, and skin in vivo and ex vivo wound repair and migration where values ranging from 100 ng/ml to 60 μg/ml of Fibronectin have been evaluated [
35‐
39]. In vitro models including corneal epithelial cells, corneal fibroblasts, keratinocyte, dermal fibroblast, nasal airway epithelial cells, gingival fibroblast, and SAEC have further demonstrated that fibronectin stimulated wound repair and cell migration as either soluble factor or substrate component frequently through integrin signaling activation [
40‐
45].
Lumican, a small leucine-rich proteoglycan (SLRP) family member, is a major component of the proteoglycan-based ECM. Lumican displays a heterogeneous, diffuse, staining profile in the alveolar walls and peripheral regions of adult human lung and its sub-epithelial deposition is implicated in airway remodeling and counteracting the severity of asthma [
46,
47]. In vivo and ex vivo studies indicate that Lumican is essential for corneal epithelial cell, skin, and oral mucosa wound repair and promotion of chemotactic migration [
48‐
51]. In vitro migration and wound healing of corneal epithelial cells was also induced by Lumican [
52,
53]. Circulating plasma and cellular Fibronectin concentrations (300 μg/ml and 2.46 μg/ml, respectively) are substantially in excess of those required to induce wound repair and migration in our, and previously described (see above), in vitro models whereas physiological levels of Lumican are yet to be defined [
54]. In support and extension of these previous reports we have now demonstrated that both Fibronectin and Lumican facilitated AEC and SAEC wound repair as soluble factors (with trace serum supplementation for AEC only) and migration of AEC as a substrate component.
Periostin (or OSF-2, Osteoblast Specific Factor-2) is a matricellular protein with a poorly defined role in wound repair and a physiological serum level of approximately 39 ng/ml [
55,
56]. In vivo studies revealed an association with fracture healing, wound-derived blood vessels, acute myocardial infarction response, skin wounds, and ligament repair [
57‐
62]. In addition to those above low levels of Periostin expression are common in normal lung while high levels of Periostin are detected in IPF lungs and patient serum although the role of this protein in the pathogenesis of lung fibrosis has not been clarified [
56]. A relationship between Periostin and wound repair across multiple tissues is immediately apparent though little is known, at this time, of mechanism of action. A solitary report describes exogenous overexpression of Periostin in A549 cells, as used in our study, which enhanced both proliferation and migration in routine culture [
63]. We have herein reported that Periostin has a role in AEC and SAEC wound repair and migration as a soluble factor and identified a putative hMSC source of wound associated Periostin.
IGFBP-7 (also known as IGFBP-rP1) is an IGFBP-related protein (IGFBP-rPs) superfamily member. IGFBP-rPs have less binding affinity to IGF (insulin growth factor) and are involved in diverse biological activities in an IGF-independent manner [
64]. IGFBP-rP2 (CTGF (connective tissue growth factor)), another IGFBP-rP family member, plays a critical role during fibrogenesis in IPF [
65,
66]. However, the role of IGFBP-7, which has a normal serum concentration of 33 ng/ml [
67], in pulmonary fibrosis is unknown. IGFBP-7 is up-regulated in the fibrotic regions of IPF lung tissue as well as in isolated IPF fibroblasts though absent from controls [
68]. Here we have demonstrated that recombinant human IGFBP-7 significantly increased human primary SAEC wound repair when applied in serum-free basal media, whereas AEC were non-responsive to this protein. The mechanism behind these divergent responses remains to be clarified. The development of transferable and accessible primary human AEC cultures (which replicate in vivo characteristics; i.e. monolayer formation) will be necessary before clarification can be achieved. However at this time we cannot preclude the possibility that the distinct response is due tissue source or phenotypic background of the AEC and SAEC used in this study. However this distinction in response provides strong support for broad therapeutic applicability across multiple clinical applications for the hMSC secretome. Characterization of these cell-specific responses will underpin the continuing development of the stem cell-driven regenerative medicine industry.
The protein composition of ECM is variable, tissue specific, and provides essential scaffold and biochemical signals required for cell growth, tissue homeostasis, development and wound repair [
55,
69]. Our in vitro AEC wound repair data encouraged us to investigate the substrate roles of Fibronectin, Lumican and Periostin on AEC migration. The in vitro scratch wound repair system is not suitable to test the effectiveness of individual substrate components as the wounding process would disrupt the protein coating. We developed a novel ‘collagen drop’ cell migration assay to evaluate individual substrate component capacity to support or inhibit migration. Normal basement membrane architecture provides a favourable substrate for AEC migration whereas disrupted alveolar basement membrane and aberrant ECM remodeling play a crucial role in the abrogation of alveolar re-epithelialisation in pulmonary fibrosis [
13]. Data from our ‘collagen drop’ cell migration assay has identified that both Fibronectin and Lumican are supportive of cell migration as substrate component. From consideration of the data obtained from wound repair and collagen drop assays we suggest a putative dual role (as topical and substrate components) of Fibronectin and Lumican, and a topical role of Periostin on alveolar epithelial cell migration and wound repair. Fibronectin, Lumican and Periostin have diverse biological effects on various cell types depending on their interactions with different receptor families, such as integrins, receptor tyrosine kinases (RTKs) and IGF-receptor [
69‐
73]. It can be speculated that an individual candidate protein may stimulate different receptors depending on the mode of application with the potential to trigger different outcomes. Here our two different assay systems illustrated the effects of Periostin on AEC migration operated in an administration-dependent fashion. Further elucidation of hMSC secretory proteins and their interactions with target receptors in alveolar epithelial wound repair will provide a clear understanding in hMSC-mediated alveolar injury repair and regeneration.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
Conceived and designed the study: KMA, MAS, NRF. Performed the experiments: KMA, SS. Analysed the data and prepared results: KMA. Wrote the manuscript: KMA, NRF; MAS. Study supervised: NRF, MAS. All authors read and approved the final manuscript.