Developmental Reprogramming in Mesenchymal Stromal Cells of Human Subjects with Idiopathic Pulmonary Fibrosis.

Cellular plasticity and de-differentiation are hallmarks of tissue/organ regenerative capacity in diverse species. Despite a more restricted capacity for regeneration, humans with age-related chronic diseases, such as cancer and fibrosis, show evidence of a recapitulation of developmental gene programs. We have previously identified a resident population of mesenchymal stromal cells (MSCs) in the terminal airways-alveoli by bronchoalveolar lavage (BAL) of human adult lungs. In this study, we characterized MSCs from BAL of patients with stable and progressive idiopathic pulmonary fibrosis (IPF), defined as <5% and ≥10% decline, respectively, in forced vital capacity over the preceding 6-month period. Gene expression profiles of MSCs from IPF subjects with progressive disease were enriched for genes regulating lung development. Most notably, genes regulating early tissue patterning and branching morphogenesis were differentially regulated. Network interactive modeling of a set of these genes indicated central roles for TGF-β and SHH signaling. Importantly, fibroblast growth factor-10 (FGF-10) was markedly suppressed in IPF subjects with progressive disease, and both TGF-β1 and SHH signaling were identified as critical mediators of this effect in MSCs. These findings support the concept of developmental gene re-activation in IPF, and FGF-10 deficiency as a potentially critical factor in disease progression.

Idiopathic pulmonary fibrosis (IPF) is a chronic fibrotic lung disease characterized by impaired repair/regenerative responses and aberrant tissue remodeling12. It has been proposed that IPF may represent a re-capitulation of developmental programs based on global genomic studies demonstrating that IPF lungs are enriched with genes associated with lung development, e.g. transcription factors that regulate tissue morphogenesis of embryonic lung34; however, cell-specific expression patterns and the interaction of developmental genes in IPF have not been elucidated. IPF is a heterogeneous disease process with variable clinical courses and some patients are relatively stable for long periods, while others progress more rapidly567. Factors governing this heterogeneity in disease progression are not well understood.

During early lung development, signals from the mesenchyme are critical to specification of epithelial cell proliferation and differentiation8910. Interactions and signaling between mesenchymal and epithelial cells are critical for later stages of lung development including branching morphogenesis and alveologenesis11. Lung branching morphogenesis is regulated by coordinated action of fibroblast growth factor (FGF-10), sonic hedgehog (SHH) and bone morphogenetic protein (BMP-4)121314. Homeobox (Hox) genes are master regulators of tissue patterning and organ development. HoxA1 to A5 and HoxB1 to B6 are expressed in the developing lung15. Recently HoxA5 genes have been shown to be key upstream mesenchymal regulators of the Wnt2/2b, one of the main regulators of FGF-10 expression in the lung1617. Mesenchyme homeobox-2 (Meox2) regulates TGF-β signaling18, nuclear factor-kappa B activity19, microRNA-22120, and DNA methylation21, processes known to be relevant to IPF pathogenesis. Although the precise roles of Hox genes in lung development have not been elucidated, they are known to be expressed at early stages, preceding branching morphogenesis. Roles of these molecules have also been reported in the maintenance of adult lung homeostasis and fibrosis2223. FGF-10 is reported to play a major role in alveolar epithelial cell progenitor cell viability242526, and repression of Meox2 is required for TGF-β1 induced myofibroblast differentiation27. Thus, dysregulation of these pathways may negatively affect adult lung injury repair.

The participation and contribution of mesenchymal stromal cells (MSCs) to injury repair processes in adult tissues/organs is well recognized28. We have previously identified a lung-resident population of MSCs isolated from the lower respiratory tract of human subjects via bronchoscopy and broncho-alveolar lavage (BAL)29. BAL-derived MSCs in ex vivo culture lack hematopoietic cell markers (CD14, CD34, and CD45), express CD73, CD90, and CD105, and demonstrate the capacity to differentiate into adipocytes, chondrocytes, and osteocytes. These cells were found to be donor-derived up to >11 years (based on sex-mismatch analyses of lung transplant recipients), suggesting that this MSC population is long-lived and reside locally in the terminal airspaces to regulate injury-repair processes29. We postulated that these BAL-derived MSCs represent a specific subpopulation of mesenchymal cells that are “embryonic remnants” that lie quiescent within the alveolar interstitium and are mobilized into the alveolar space in the context of lung injury repair. In this study, we hypothesized that gene expression patterns in MSCs from human subjects with varying disease activities/phenotypes may provide clues to aberrant developmental re-programming in IPF. Using differential gene expression and network analyses, we identified central roles for transforming growth factor-β1 (TGF-β1) and sonic hedgehog (SHH) pathways in human subjects with progressive disease; additionally, validation studies indicate a convergence of these pathways on the down-regulation of FGF-10, a critical homeostatic growth factor in alveolar epithelial cell survival and maintenance242526.

Results

Gene expression profiling of MSCs in progressive vs. stable IPF

Previous studies from our group demonstrated the presence of tissue-resident MSCs isolated by bronchoscopy and BAL from human subjects29. Gene expression profiles in MSCs from IPF have not been previously characterized. To determine the changes in global mRNA expression during IPF progression, MSCs were isolated from patients with stable IPF (s-IPF) and progressive IPF (p-IPF). s-IPF and p-IPF patients were defined by a decline in forced vital capacity (FVC) < 5% and FVC ≥ 10% respectively over the preceding 6 months (n = 4 in each group; protocol for MSC isolation is shown in Supplementary Figure S1). MSCs were characterized, post-isolation and ex vivo growth, for their ability to form colonies in tissue culture plates and uniform expression of mesenchymal cell phenotype markers, prolyl-4-hydroxylase and vimentin (Supplementary Figure S2); additionally, we confirmed their ability to differentiate into mesodermal lineages (Supplementary Figure S3). No significant difference was observed in osteogenic, adipogenic and chondrogenic differentiation potential of BAL-derived MSCs between s-IPF and p-IPF MSCs. Early passage MSCs were placed in low serum conditions (0.05%) for 24 hours, then total RNA was isolated and subjected to whole genome transcriptome analysis to determine cellular/molecular mechanisms which separate these distinct clinical subphenotypes. Expression values for each gene using a robust multi-array average were calculated after a number of quality control steps and normalization30. As a quality control step, a principal components analysis (PCA) was performed to analyze the segregation of gene expression profiles in each group (Supplementary Figure S4). Based on this PCA, one sample from each group was excluded (sample #3 from s-IPF and sample #8 from p-IPF) resulting in n = 3 in each group for further analysis (GEO repository accession ID # GSE73854).

Statistical analysis was performed fitting a linear model designed for microarray analysis31. 428 probesets were selected based on an unadjusted p-value less than 0.005. The gene expression data was then analyzed using Ingenuity Pathway Analysis (IPA). The “top functions” predicted by this analysis (score >30; number of molecules ≥20) and associated genes, including those which were up-regulated or down-regulated are shown in Supplementary Table S1. Interestingly, many of the differentially expressed genes are known to play central roles in organismal development (cluster #2 and 5). Among these, the major growth factors that regulate branching morphogenesis during embryonic lung development, FGF-10 and BMP-4 were down-regulated while transcription regulators, Meox2 and HoxA2, were up-regulated in p-IPF when compared to s-IPF. A heat map comprising genes in cluster #2 (with addition of HoxA2 from cluster #5) is shown in Fig. 1, and description of these genes including fold change and p-values are provided in Table 1.

Figure 1

Heat map analysis of lung developmental genes.

Heat map representing color coded expression levels of differentially expressed genes in progressive IPF compared to stable IPF (n = 3 in each group; p < 0.005). Up-regulated genes were shown in shades of red whereas down-regulated genes were shown in shades of green.

Table 1

Differentially expressed lung developmental genes in progressive (vs. stable) IPF.

Symbol	Entrez Gene Name	Affymetrix ID	Fold Change (LOG2)	p-value	Location	Type(s)
Down-regulated genes in progressive IPF
BMP4	Bone morphogenetic protein 4	211518_s_at	−2.58	0.001	Extracellular Space	growth factor
COL18A1	Collagen, type XVIII, alpha 1	209082_s_at	−1.58	0.002	Extracellular Space	other
FGF10	Fibroblast growth factor 10	231762_at	−3.23	0.000	Extracellular Space	growth factor
GREM2	Gremlin 2, DAN family BMP antagonist	220794_at	−1.54	0.004	Extracellular Space	other
IGFBP2	Insulin-like growth factor binding protein 2, 36 kDa	202718_at	−2.58	0.000	Extracellular Space	other
MMP11	Matrix metallopeptidase 11	203878_s_at	−2.65	0.001	Extracellular Space	peptidase
MTUS1	Microtubule associated tumor suppressor 1	212095_s_at	−1.56	0.002	Other	other
PERP	PERP, TP53 apoptosis effector	236009_at	−1.55	0.001	Plasma Membrane	other
SHC3	SHC (Src homology 2 domain containing) transforming protein 3	229824_at	−1.75	0.001	Cytoplasm	other
Up-regulated genes in progressive IPF
ANXA3	Annexin A3	209369_at	3.01	0.000	Cytoplasm	enzyme
CNN1	Calponin 1, basic, smooth muscle	203951_at	2.29	0.000	Cytoplasm	other
COL4A1	Collagen, type IV, alpha 1	211981_at	1.65	0.004	Extracellular Space	other
CYR61	Cysteine-rich, angiogenic inducer, 61	210764_s_at	1.41	0.004	Extracellular space	other
F3	Coagulation factor III (thromboplastin, tissue factor)	204363_at	2	0.000	Plasma Membrane	transmembrane receptor
F11R	F11 receptor	223000_s_at	2.71	0.000	Plasma Membrane	other
GATA2	GATA binding protein 2	209710_at	1.16	0.004	Nucleus	transcription regulator
HMMR	Hyaluronan-mediated motility receptor (RHAMM)	207165_at	2.91	0.000	Plasma Membrane	other
ID4	Inhibitor of DNA binding 4, dominant negative helix-loop-helix protein	209291_at	2.65	0.000	Nucleus	transcription regulator
ITGA6	Integrin, alpha 6	201656_at	2.09	0.000	Plasma Membrane	other
MEOX2	Mesenchyme homeobox 2	206201_s_at	2.82	0.001	Nucleus	transcription regulator
RCAN1	Regulator of calcineurin 1	215253_s_at	1.71	0.000	Nucleus	transcription regulator
SEMA3F	Sema domain, immunoglobulin domain (Ig), short basic domain, secreted, (semaphorin) 3 F	206832_s_at	1.2	0.004	Extracellular Space	other
SERPINE1	Serpin peptidase inhibitor, clade E (nexin, plasminogen activator inhibitor type 1), Member 1	1568765_at	2.35	0.000	Extracellular Space	other
HOXA2	Homeobox A2	1557051_s_at	1.73	0.002	Nucleus	transcription regulator

To validate the expression of the developmental genes of interest, we analyzed a new cohort of IPF patients (n = 15), including stable-IPF (s-IPF, n = 7) and progressive-IPF (p-IPF, n = 8). Real-time PCR for the genes of interest (FGF-10, BMP-4, Meox2 and HoxA2) was performed on BAL-MSCs on these patients. The primary finding of a down-regulation of FGF-10 in p-IPF was validated with an even greater fold change (−3.77 Log2-fold decrease, p < 0.04; t-test) (Fig. 2). To determine if the levels of FGF-10 from IPF subjects is diminished relative to control, non-IPF lungs, we assessed the constitutive expression of FGF-10 mRNA in MSCs obtained from surveillance bronchoscopies and BAL from lung transplant recipients without bronchiolitis obliterans or infection and grown ex vivo. We found that steady-state levels of FGF-10 gene expression were >25-fold higher in these non-IPF control subjects (Supplementary Figure S5). BMP-4, Meox2 and HoxA2 showed a trend in the direction of change observed in the original cohort although statistical significance was not achieved (Fig. 2). Together, these data support the concept of the activation of developmental pathways in lung MSCs derived from human subjects with IPF, and a deficiency in MSC expression of FGF-10 in subjects with progressive disease.

Figure 2

Validation of lung developmental genes of interest.

Total RNA was isolated from MSCs from stable IPF (n = 7) and progressive IPF (n = 8), and subjected to real-time PCR analysis for FGF-10, BMP-4, Meox2 and HoxA2. Data were normalized to 18 S rRNA and represented graphically as fold change compared to stable IPF (s-IPF).

Up-regulation of TGF-β and SHH signaling pathways in progressive IPF

IPA is one approach to identifying the roles of differentially regulated genes in biological pathways/processes32. An IPA network analysis of the selected developmental genes, FGF-10, BMP-4, Meox2, and HoxA2, revealed many direct/indirect interactions with other genes within the network that were either up-regulated (red) or down-regulated (green) (Fig. 3A). When “shortest path” analysis between these four genes was performed, two principal biological pathways were uncovered, the canonical TGF-β and SHH signaling pathways (Fig. 3B). These data suggest that both the canonical TGF-β and SHH signaling may participate in developmental programming and contribute to the observed patterns in MSC gene expression.

Figure 3

Gene interaction network of organismal development-related genes.

(A) Ingenuity Pathway Analysis (IPA) was used to generate gene interaction network in progressive IPF. This network contains four developmental genes of interest (FGF-10, BMP-4, Meox2, HOxA2; Table 1). Transcriptional information was projected onto the interaction map such that up-regulated genes are depicted in shades of red and down-regulated genes are in shades of green. (B) Shortest path gene interaction network of growth factors (FGF-10 and BMP-4) and transcription regulators (Meox2 and HoxA2). IPA was used to generate this network using their Path Explorer filter which calculates the shortest path between these genes. Blue lines mark the genes found in the canonical transforming growth factor-β (TGF-β) and sonic hedgehog (SHH) signaling; whereas, orange lines indicate genes found to be involved in lung development.

TGF-β signaling down-regulates FGF-10, BMP-4, Meox2 and HoxA2 in BAL-MSCs

Hyperactivation of TGF-β signaling has been implicated in the pathogenesis of IPF including myofibroblast differentiation and survival33. Therefore, we first determined whether higher levels of TGF-β are associated with myofibroblast differentiation from BAL-derived MSCs. To test this hypothesis, MSCs were obtained from surveillance bronchoscopies and BAL from lung transplant recipients without bronchiolitis obliterans or infection and grown ex vivo. After serum deprived for 24 h, cells were treated with recombinant TGF-β1 (2.5 ng/ml) in vitro for 0 to 48 h and expression of α-smooth muscle actin (α-SMA), a marker for the myofibroblast phenotype, was assessed by western blotting. A time-dependent increase in α-SMA was observed in the MSC following TGF-β1 treatment indicating myofibroblast differentiation (Fig. 4A). High levels of SHH have also been reported in IPF lungs34. To test if SHH, similar to TGF-β1, induces myofibroblast differentiation of BAL-MSCs, MSCs were serum deprived for 24 h and treated in vitro with recombinant SHH (0, 50, 100 and 500 ng/ml) for 48 h and analyzed for α-SMA protein expression. In contrast to TGF-β1, SHH had no effect on α-SMA expression at any of the doses tested (Fig. 4B). These data indicate that TGF-β1, but not the direct actions of SHH, likely contributes to myofibroblast differentiation of MSCs, a key event in fibrogenesis.

Figure 4

Effects of TGF-β1, SHH and SAG on BAL-derived MSCs.

(A,B) Myofibroblast differentiation. MSCs were isolated from surveillance bronchoscopies and BAL from lung transplant recipients without bronchiolitis obliterans or infection. MSCs were seeded in 6-well tissue culture plates and serum deprived for 24 h followed by either TGF-β1 treatment (2.5 ng/ml) or SHH (0, 50, 100, 500 ng/ml) for 48 h. Cell lysates were prepared in RIPA buffer and subjected to SDS-PAGE and western blot analysis for α-SMA; GAPDH antibody was used as loading control. Densitometry was performed to quantitate the ratio of α-SMA and GAPDH and plotted graphically; bar graphs represent mean ± SEM, n = 3; *p < 0.05, compared to vehicle treated control. Full-length western blots are presented in Supplementary Figure S6. (C–F) RNA expression of developmental genes of interest. Total RNA was isolated from MSCs 48 h post-treatment with either TGF-β1 or SHH or combination and subjected to real-time PCR analysis. Data were normalized to 18 S rRNA and relative mRNA expressions are represented graphically as fold change compared to control. Data represents mean ± SEM; n = 3 (each analyzed in triplicate); *p < 0.01; **p < 0.001; ***p < 0.0001. Individual colored marker represents average relative mRNA expression of single lung transplant recipient. (G) The smoothened agonist, SAG, downregulates FGF-10. MSCs were treated with vehicle and SAG (100 ng/ml) for 48 h; total RNA was extracted and subjected to real-time PCR analysis. Data were normalized to 18 S rRNA and relative mRNA expressions represented graphically as fold change compared to control. Data represents mean ± SEM; n = 3, *p < 0.01.

Next, we sought to determine whether changes in the pattern of developmental gene expression between s-IPF and p-IPF could be attributed to hyperactive TGF-β or SHH signaling in IPF. MSCs obtained from healthy transplant recipients (n = 3, each analyzed in triplicate) were treated with TGF-β1 (2.5 ng/ml), SHH (500 ng/ml) or in combination for 48 h following 24 h of serum deprivation and assessed gene expression of FGF-10, BMP-4, Meox2 and HoxA2 by real-time PCR. Both TGF-β1 and combination treatment of TGF-β1 with SHH resulted in significant down-regulation of FGF-10, BMP-4, Meox2 and HoxA2 mRNA expression, while SHH by itself did not alter the expression of these genes (Fig. 4C–F). Since exogenous addition of recombinant SHH may fail to bind/activate PTCH1 to de-repress smoothened (SMO; SHH co-receptor), essential for SHH signaling35, we tested the ability of a SMO agonist (cell-permeable smoothened agonist, SAG (100 ng/ml; Calbiochem). SAG induced a marked down-regulation in FGF-10 expression in normal human MSCs (Fig. 4G; similar results were obtained in MSCs from a second human subject). SAG had no effect on the expression of BMP-4, Meox2, HoxA2 and α-SMA (data not shown). This data suggests high levels of TGF-β and SHH signaling in IPF may account for the observed deficiency of FGF-10 in human subjects with progressive disease.

FGF-10 expression is reduced in fibroblastic foci of IPF lungs

Based on the finding that FGF-10 expression is decreased in MSCs derived from the BAL of human subjects with progressive IPF, we evaluated the localization of this protein in lungs of control and IPF subjects. Tissue sections obtained from failed donor lungs (controls, n = 4) and IPF subjects (explants from lung transplantation, n = 4) were immunostained for FGF-10 expression. α-SMA immunostaining was also performed to identify areas of active fibrosis. While robust expression of α-SMA was observed within fibroblastic foci in IPF lungs, these foci were largely devoid of FGF-10 expression, despite the expression of this growth factor in interstitial mesenchymal cells in regions with less fibrotic remodeling. We were unable to detect FGF-10 immunostaining in normal lungs and, as expected, only airway/vascular smooth muscle cells stained positive for α-SMA (Fig. 5). IPF lung tissues were also co-immunostained for vimentin, a mesenchymal cell maker, and FGF-10 to determine identity of the FGF-10 producing cells. Immunofluorescence based co-localization demonstrated that almost all FGF-10 positive cells expressed vimentin, while a number of vimentin positive cells failed to express FGF-10 (Supplementary Figure S7). This suggests that FGF-10 expressing cells represent a subset of mesenchymal cells in IPF lung tissues.

Figure 5

Immunohistochemical localization of α-SMA and FGF-10 in IPF and normal lungs.

Six micron (6 μ) sections were obtained from normal (n = 4, failed donor lung) and IPF (n = 4; explant during lung transplantation). Immunohistochemical staining showing expression of the myofibroblast marker, α-SMA, in fibroblastic foci of IPF lung tissues; note that α-SMA staining in normal lung is restricted to large airways and blood vessels (arrows). FGF-10 immunostaining was not detected in the fibroblastic foci although expression of this growth factor was observed in interstitial mesenchymal cells in regions with less fibrotic remodeling (arrows). Higher magnification images of specific regions (boxes) are shown in middle panels.

Discussion

Effective lung regeneration following injury requires reactivation of developmental programs which involves the crosstalk between the alveolar epithelium and underlying mesenchymal cells136. In response to lung injury, stromal fibroblasts/myofibroblasts deposit extracellular matrix (ECM) proteins, mainly fibrillar collagens and fibronectin, to form a provisional matrix that allows for alveolar epithelial cell proliferation and differentiation to regenerate damaged epithelium. Chronic injury and aging may exhaust mechanisms by which the mesenchyme and epithelium regenerate functional alveoli and lead, instead, to aberrant mesenchymal activation characterized by myofibroblast accumulation and excessive ECM deposition. The roles of alveolar mesenchymal cells in lung injury repair are not well understood, and likely represent a heterogeneous population37. We have previously identified lung-resident MSCs that can be isolated and analyzed by bronchoscopy and BAL29.

In the current study, we identified an interesting pattern of gene expression in BAL-derived MSCs from patients with IPF, when differentially analyzed based on a clinical definition of disease progression [forced vital capacity (FVC) decline of ≥10% over a 6-month period] vs. stability (FVC < 5% over a 6-month period). The robustness of this definition of disease progression has been validated by 3 independent research groups that showed that a decline in FVC of ≥10% over a 6–12 month period is predictive of decreased survival in IPF subjects567. For this study, we relied on early passage (P1–4) MSCs ex vivo to maintain relative purity of the mesenchymal cell population (rather than P0 MSCs that are less pure; no specific cell surface marker of MSCs has been reported). Although gene expression patterns are influenced by ex vivo cell culture conditions, these cells also maintain a stable and heritable pattern of gene expression, likely via cell autonomous epigenetic mechanisms3839. Transcriptomic analyses on MSCs isolated from a discovery cohort of IPF patients with progressive vs. stable disease (n = 4 in each group, and n = 3 per group after PCA) revealed enrichment for genes involved in organismal development. Among these genes, differential expression of genes regulating branching morphogenesis (FGF-10, BMP-4) and the homeobox genes, Meox2 and HoxA2, were identified. Interactome analysis of these four genes by shortest pathway network algorithm indicated their association with canonical TGF-β1 and SHH signaling. We explored whether TGF-β1 and SHH signaling in MSCs were sufficient to explain the observed patterns of developmental gene expression that discriminated progressive from stable IPF. TGF-β1 suppressed FGF-10, BMP4, Meox2 and HoxA2 gene expressions in MSCs. While recombinant SHH had no effect on these genes of interest, treatment of MSCs with an agonist of smoothened (SMO; SHH co-receptor) resulted in inhibition of FGF-10 gene expression without significant effects on BMP4, Meox2 and HoxA2. We re-analyzed these 4 candidate genes in a replication cohort of 15 IPF subjects (8-progressive and 7-stable) and observed an even greater decrease in FGF-10 (−3.77 Log2-fold change; p < 0.04) in IPF subjects with progressive disease; the other developmental genes of interest showed similar trends to the discovery cohort, although statistical significance was not achieved. Non-IPF control MSCs obtained from surveillance bronchoscopies and BAL from lung transplant recipients without bronchiolitis obliterans or infection showed significantly higher FGF-10 expression compared to IPF MSCs. This data supports the concept that there may be a loss of a protective effect of FGF-10 in progressive IPF.

Mesenchyme-derived FGF-10 is critical for maintaining lung epithelial progenitor cells during early lung development40. FGF-10 knockout mice fail to develop lungs41. FGF-10 expression is also required for lung epithelial regeneration following injury in adults, and suppression of FGF-10 has been shown to compromise regenerative capacity after naphthalene injury42. FGF-10 acts as a protective and therapeutic agent against bleomycin-induced pulmonary fibrosis24 and attenuates H2O2-induced alveolar epithelial cell DNA damage26. FGF-10 overexpression during bleomycin-induced lung injury has been shown to have a protective effect on epithelial progenitor cells by inhibiting TGF-β124. Thus, TGF-β1 and FGF-10 reciprocally and negatively regulate each other to influence the outcome of the repair and regenerative response. Absence of FGF-10 immunostaining in the fibroblastic foci (myofibroblasts) in IPF lungs substantiates TGF-β1-stimulated suppression of FGF-10 gene expression in our in vitro studies. A recent report also demonstrated that intra-tracheal delivery of FGF-10 protected from LPS-induced acute lung injury in rats via mobilizing MSCs in the BAL43. It remains to be determined whether the mechanism of protection observed in this study is solely related to MSC mobilization or whether this protection may involve mesenchymal-to-epithelial cell signaling. Our study focused on human subjects with progressive (vs. stable) fibrotic lung disease, supporting a role for FGF-10 in disease progression rather than disease initiation or susceptibility.

SHH has been shown to inhibit FGF-10 gene expression in embryonic lung tissue354445. SHH was also reported to stimulate proliferation and myofibroblast differentiation of human lung fibroblasts3446. High levels of SHH is also reported in IPF lungs3447. SHH expression is also required during embryogenesis and tissue repair4849. Recently, SHH was shown to be required for maintenance of a quiescent mesenchyme in the adult lung50. During epithelial injury, SHH levels decline leading to a proliferative repair response by mesenchymal cells. SHH levels promptly returns to baseline post injury restoring mesenchymal quiescence50. In our study, while exogenous treatment with recombinant SHH failed to suppress FGF-10 expression in MSCs, activation of this pathway by SAG (a smoothened agonist) markedly down-regulated FGF-10 in lung MSCs. The failure of recombinant SHH to mediate signaling may be related to its inability to bind/activate PTCH1 and de-repress SMO, essential for SHH signaling35. Together, these data support a role for heightened SHH/smoothened signaling, in concert with TGF-β1, in the downregulation of FGF-10 gene expression observed in MSCs of human IPF subjects with progressive disease.

BMP-4 is principally expressed by the lung epithelial cells and is weakly expressed by the mesenchyme during lung development51. Unlike other BMPs, BMP-4 expression remains restricted to the distal epithelial cells and regulates branching morphogenesis along with FGF-10 and SHH4852. FGF-10 is known to stimulate BMP-4 expression in the epithelium during branching and in return is inhibited by SHH and BMP-4 at high concentrations, thus controlling branch outgrowth53. TGF-β1 has been reported to inhibit BMP-4 signaling in pulmonary artery smooth muscle cells54. Similar to this latter report, we observed a significant down-regulation of BMP-4 gene expression by TGF-β1; the precise role for mesenchyme-derived BMP-4 in wound healing and repair requires further investigation.

TGF-β1’s role in fibrotic disorders is well recognized, not only in the lung but in almost all organs studied55. Alveolar epithelial cells are thought to be the primary source of TGF-β1 in IPF lungs5657. TGF-β1 is known to regulate embryonic development, cellular growth/differentiation, host defense and tissue repair58. Our findings showed significant downregulation of FGF-10 gene expression in MSCs in response to TGF-β1 and SHH signaling. The sustained up-regulation of TGF-β1 and SHH, with concomitant deficiency of FGF-10, may prevent restoration of normal epithelial homeostasis.

In summary, this study demonstrates lung-resident MSCs harbor gene expression patterns that recapitulate developmental reprogramming in an adult, age-related fibrotic lung disease, IPF. This is the first study to demonstrate a deficiency in MSC expression of FGF-10 in progressive IPF, a specific clinical endotype. This is consistent with a protective role of FGF-10 in animal models of lung injury and fibrosis24404359. Further studies are required to determine the impact of the FGF-10 deficiency on epithelial cell fate and function. FGF-10low population may be a biomarker of “pathologic” MSCs and that loss of FGF-10 alone is unlikely to explain disease progression. The studies reported here also suggest that gene expression patterns of alveolar MSCs may be useful in prognostication of disease progression in IPF, and potentially other chronic lung diseases; future studies will determine if such approaches can be exploited to develop personalized therapies that target altered developmental pathways.

Methods

Isolation of alveolar mesenchymal cells

BAL fluid was collected from patient with stable and progressive IPF following patient informed consent and methods approved by the Institutional Review Board (IRB) of the University of Michigan (UM) and University of Alabama at Birmingham (UAB). All methods involving human participants were performed in accordance with the relevant guidelines and regulations. The cellular fraction of the BAL fluid is pelleted by centrifugation and plated on tissue culture plates at a density of 5 × 106 cells/100 mm dish. The media is changed daily for the first two days and thrice weekly, thereafter. This resulted in the removal of terminally differentiated, apoptotic cells and/or non-adherent cells, mainly inflammatory cell populations (macrophages). By 10–14 days in cell culture, a number of colony forming units of MSCs (CFU-MSCs) are seen on Giemsa staining. Serial dilutions allowed an estimate of the number of clonally expanding cells in the original BAL; for example, in the patient represented, 4 colonies were visualized at 1:1000 dilutions, indicating ~4000 cells in the original 5 × 106 cells (0.08%). MSCs, at passage 1, show uniform staining for prolyl-4-hydroxylase and vimentin, supporting mesenchymal cell phenotype (Supplementary Figure S2). All BAL-MSCs used in this study were between passages 1–4.

Mesodermal lineage differentiation Assay

For osteogenic differentiation of BAL-derived MSCs, cells were seeded at a concentration of 104 cells/cm2 in 24-well culture dishes and allowed to grow for 24 h. The growth medium was replaced with medium containing 10–8 M dexamethasone, 0.2 mM ascorbate phosphate, and 10 mM β-glycerophosphate in α-MEM (Invitrogen; Thermo Fisher Scientific, Waltham, MA) with 10% FBS. Medium was changed every 2 days. After 14 days cells were washed, fixed with 10% formalin and stained with Alizarin red stain to visualize calcium deposition. For adipogenic differentiation of MSCs, cells were seeded at a density of 104 cells/cm2 in 24-well culture dishes and allowed to grow for 24 h. The MSCs were then exposed to StemPro adipocyte differentiation medium (Thermo Fisher Scientific). Fresh medium was added to the MSCs every 2 days. After 10 days the MSCs were fixed and stained with Oil Red-O stain (SIGMA) to visualize lipid accumulation in the MSCs. Adipocyte nuclei were counterstained with hematoxyllin. For chondrogenesis assay, 3 × 105 MSCs were pelleted in 15 ml conical bottom tubes and cell pellet was exposed to StemPro chondrocyte differentiation medium (Thermo Fisher Scientific) for 14 days. Medium was changed every 2 days. The cell pellets were stained with Safranin-O aqueous staining solution (EMD Millipore, Billerica, MA) for cellular proteoglycans specific to cartilage.

Affymetrix Gene Chip Analysis

MSCs were grown up to 80–90% confluence; and serum deprived for 24 h at passage 2, which represents the “pooled” colonies. Total RNA was isolated using Qiagen RNeasy Mini Kit (Valencia, CA) and subjected to whole genome transcriptomal analysis. RNA isolates (n = 4 in each group) were hybridized on Affymetrix U133A microarray chips with 22976 probe-pairs and statistical analyses were performed at the University of Michigan Microarray Core Facility.

Systems Biology analysis

For generating networks, a data set containing gene identifiers and corresponding expression values was uploaded into the Ingenuity Pathway Analysis software (Ingenuity Systems, www.qiagen.com/ingenuity). Each identifier was mapped to its corresponding object in Ingenuity’s Knowledge Base. A fold change cutoff of ±2 and a q-value cutoff of 0.05 were set to identify molecules whose expression was significantly differentially regulated. These molecules, called Network Eligible molecules, were overlaid onto a global molecular network developed from information contained in Ingenuity’s Knowledge Base. Networks of Network Eligible Molecules were then algorithmically generated based on their connectivity. The Functional Analysis identified the biological functions and/or diseases that were most significant to the entire data set. Molecules from the dataset that met the fold change cutoff of ± 2 and a q-value cutoff of 0.05 and were associated with biological functions and/or diseases in Ingenuity’s Knowledge Base were considered for the analysis. Right-tailed Fisher’s exact test was used to calculate a p-value determining the probability that each biological function and/or disease assigned to that data set is due to chance alone.

Ex vivo myofibroblast differentiation

BAL-derived MSCs (P3; 100,000 cells) were plated in each well of 6-well tissue culture dishes in DMEM supplemented with 4.5 g/L glucose, 10% FBS, 1 mM L-glutamine and penicillin-streptomycin. The next day the cells were serum starved for 24 h and subjected to either vehicle or recombinant TGF-β1 (R&D Systems, Minneapolis, MN) or recombinant SHH (R&D Systems) for 48 h to test for myofibroblast differentiation. Proteins were isolated from MSCs post-treatment and subjected to western blotting for α-SMA expression.

Real-Time PCR

Total RNA was isolated from cells using the RNeasy Mini Kit (Qiagen) and reverse transcribed using iScript Reverse Transcription SuperMix for real-time PCR (Bio-Rad, Hercules, CA). Real-time PCR reactions were performed using SYBR Green PCR Master Mix (Life Technologies, Grand Island, NY) and gene specific primer pairs for FGF-10, BMP-4, Meox2, HoxA2, and 18 S rRNA (Integrated DNA Technologies, Coralville, IA; for primer sequences, see Supplementary Table S2). Reactions were carried out for 40 cycles (95 °C for 15 sec, 60 °C for 1 min for FGF-10, BMP-4 and HoxA2; 95 °C for 15 sec, 66 °C for 1 min for Meox2) in a StepOnePlus Real-Time PCR System (Life Technologies). Real-time PCR data is expressed for each target gene normalized to endogenous 18 S, as 2−ΔΔCt and relative mRNA expression is represented graphically as fold change compared to control.

Western Blotting

Protein lysates were collected from each time point using RIPA cell lysis buffer supplemented with sodium orthovanadate (SIGMA-Aldrich, St Louis, MO) and protease inhibitor cocktail (EMD Millipore). Protein concentrations were determined using Pierce BCA micro protein assay kit (Thermo Fisher Scientific) and a microplate reader (BioTek Instruments, Winooski, VT). The lysates were then denatured and reduced using 4x Laemmli sample buffer and 10x reducing agent (Life Technologies) at 70 °C for 10 minutes. The proteins were separated on a 4–20% Tris-glycine Bio-Rad Criterion precast gradient gel and subjected to western blotting. Briefly, proteins were transferred to a nitrocellulose membrane using a Bio-Rad transfer apparatus. Membranes were blocked with Superblock T20 (Thermo Fisher Scientific) and probed with antibodies against human α-SMA (American Research Product Inc; Waltham, MA). GAPDH (Cell Signaling, Danvers, MA; dilution 1:1000) was used as a loading control. Signals were detected by probing with anti-mouse secondary antibodies labeled with horse radish peroxidase (Pierce) and immunoblots were developed with Luminata Crescendo Western Blot HRP substrate (EMD Millipore). Blots were imaged in Amersham Imager 600 (GE Health Care, Pittsburg, PA). Density of the protein bands were measured using ImageQuant TL software (GE). Data presented as the mean of three experiments. The error bar represents the standard error of mean.

Immunostaining of lung tissue sections

Human lung tissues were obtained from patients with confirmed diagnosis of IPF60, under protocol approved by the IRB of UM and UAB. Informed consents were obtained from all individuals enrolled through the Airway Tissue Procurement Program at UM and UAB. Normal and IPF lung tissues were fixed in formaldehyde, dehydrated and paraffin embedded using standard protocol. Six micron thick sections were cut and mounted on glass slides and immunostained for human α-SMA (mouse monoclonal; American Research Products; 1:500), FGF-10 (rabbit polyclonal; EMD Millipore; 1:100), and vimentin (mouse monoclonal; SIGMA-Aldrich; 1:100). Briefly, tissue sections were deparaffinized in xylene and hydrated through ethanol series and water. Antigen retrieval was performed using citrate buffer at pH 6.0 in a 95 °C water bath followed by quenching of endogenous peroxidases using 3% hydrogen peroxide. Tissue sections were blocked using 5% normal goat (immunohistochemistry) or donkey serum (immunofluorescence) and were then incubated in primary antibodies overnight at 4 °C. No primary antibody and appropriate IgG isotype controls were utilized to determine specificity of staining. Secondary detection was performed using the Dako Envision Dual Link System (Carpinteria, CA) for α-SMA and FGF-10 (IHC). For immunofluorescence detection of vimentin and FGF-10, anti-mouse/anti-rabbit Alexa Fluor 594/488 (Molecular Probes, Thermo Fisher) secondary antibodies were used. No primary antibody and appropriate IgG isotype controls were utilized to determine specificity of staining. Colorimetric detection was achieved using the DAB/H2O2 kit from the Vector Laboratories (Burlingame, CA). Nuclei were counterstained with hematoxylin (Vector Labs) for IHC and Hoechst dye (Molecular Probes) for immunofluorescence.

Statistical Analyses

Statistical analyses and graphical presentations were prepared in GraphPad PRISM version 6.0 software. Western blotting and real-time PCR data obtained in this study were analyzed by unpaired two-tailed Student’s t-test for pairwise comparisons. If more than two groups were present, data were analyzed by ANOVA followed by Tukey’s post hoc test for multiple pairwise comparisons. Results were considered significant if p < 0.05.

Additional Information

How to cite this article: Chanda, D. et al. Developmental Reprogramming in Mesenchymal Stromal Cells of Human Subjects with Idiopathic Pulmonary Fibrosis. Sci. Rep. 6, 37445; doi: 10.1038/srep37445 (2016).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.