Abstract

Background COPD is an incurable disease and a leading cause of death worldwide. In mice, fibroblast growth factor (FGF)10 is essential for lung morphogenesis, and in humans, polymorphisms in the human FGF10 gene correlate with an increased susceptibility to develop COPD.

Methods We analysed FGF10 signalling in human lung sections and isolated cells from healthy donor, smoker and COPD lungs. The development of emphysema and PH was investigated in Fgf10+/− and Fgfr2b+/− (FGF receptor 2b) mice upon chronic exposure to cigarette smoke. In addition, we overexpressed FGF10 in mice following elastase- or cigarette smoke-induced emphysema and pulmonary hypertension (PH).

Results We found impaired FGF10 expression in human lung alveolar walls and in primary interstitial COPD lung fibroblasts. In contrast, FGF10 expression was increased in large pulmonary vessels in COPD lungs. Consequently, we identified impaired FGF10 signalling in alveolar walls as an integral part of the pathomechanism that leads to emphysema and PH development: mice with impaired FGF10 signalling (Fgf10+/− and Fgfr2b+/−) spontaneously developed lung emphysema, PH and other typical pathomechanistic features that generally arise in response to cigarette smoke exposure.

Conclusion In a therapeutic approach, FGF10 overexpression successfully restored lung alveolar and vascular structure in mice with established cigarette smoke- and elastase-induced emphysema and PH. FGF10 treatment triggered an initial increase in the number of alveolar type 2 cells that gradually returned to the basal level when the FGF10-mediated repair process progressed. Therefore, the application of recombinant FGF10 or stimulation of the downstream signalling cascade might represent a novel therapeutic strategy in the future.

Tweetable abstract

Impaired FGF10 signalling is an integral part of the pathomechanism that leads to development of cigarette smoke-induced emphysema and PH. Restoration of FGF10 signalling reverses established cigarette smoke- and elastase-induced emphysema and PH in mice. https://bit.ly/3PgqrBw

Introduction

COPD is a leading cause of morbidity and mortality worldwide, characterised by persistent airflow limitation due to chronic bronchitis and emphysema. The main risk factor contributing to COPD development is the inhalation of noxious particles or gases, particularly cigarette smoke and air pollution [1–3]. Structural remodelling of the pulmonary vasculature, which often accompanies COPD, leads to the development of pulmonary hypertension (PH) [4–7].

Emerging literature reports dysregulation of developmental pathways in lung emphysema, suggesting the importance of homeostatic mechanisms for lung structural integrity in adult life [3, 8–10]. Indeed, lung development, repair and homeostasis share several mechanisms and signalling pathways, including fibroblast growth factors (FGF) [3, 9–11]. However, complex interactions between master regulators of development and homeostasis in emphysema are still poorly understood [3].

Secreted, paracrine-acting FGFs regulate fundamental cellular processes such as proliferation, survival, migration, differentiation and metabolism [12]. FGF10 is indispensable for branching morphogenesis during embryonic lung development [13]. Mesenchyme-derived FGF10 predominantly binds to and activates its cognate receptor, FGFR2b, expressed by epithelial cells. Post-natally, FGF10/FGFR2b signalling regulates various aspects of stem/progenitor cell homeostasis, and activation in the lung promotes wound healing and prevents oxidant- and stretch-induced DNA damage in alveolar epithelial cells [14–19]. FGF10 haploinsufficiency in humans has been linked to impaired lung function and single nuclear polymorphisms in the region of the FGF10 locus were identified in COPD patients [11, 20, 21]. Fgf10 mRNA expression was decreased in alveolar walls and increased in pulmonary vessels in cigarette smoke-exposed mice [6]. Selective inhibition of inducible nitric oxide synthase with l-N6-(1-iminoethyl)lysine (l-NIL) successfully reversed established cigarette smoke-induced emphysema and PH [6]. Of interest, Fgf10 upregulation positively correlated with l-NIL-mediated disease reversion [6]. However, the underlying mechanism remains largely elusive. Against this background, we aimed to investigate the role of FGF10 in the development and in treatment of lung emphysema and PH (supplementary figure S1).

Material and methodsHuman lung tissue samples

Human lung tissue samples were obtained from the DZL Biobank, of individuals with and without COPD who had undergone lung transplantation or from healthy lung-transplantation donors. The pathologist examined all lungs and determined smoking status of the donors. Human serum was collected from donors and COPD patients with variable disease severity and PH status, following routine protocols. The ethics committee of the Justus-Liebig-University School of Medicine approved the studies (AZ 58/15; AZ 17/18).

Animal studies and experimental design

In the loss-of-function experimental setup (shown in supplementary figure S1a), mice that were haploinsufficient for either Fgf10 or Fgfr2b (Fgf10+/− or Fgfr2b+/−) were exposed to room air or cigarette smoke for 3 or 8 months (n=12 per group). Littermate mice with both functional Fgf10 and Fgfr2b alleles were used as wild-type controls (n=12 per group). Furthermore, we overexpressed FGF10 in mice after cigarette smoke- or elastase-induced emphysema and PH (supplementary figure S1b). FGF10 overexpression was induced using doxycycline, using the Tet-ON system in transgenic B6-Gt(ROSA)26Sortm1.1(rtTA,EGFP)Nagy Tg(tetO-Fgf10)1Jaw/SpdlJ mice (Jackson Laboratories, reference 025671). All animal lines were cross-bred to C57Bl/6J background. After 8 months of room air/cigarette smoke exposure, or 1 month after intratracheal saline/elastase instillation, animals were fed ad libitum with feed containing doxycycline (600 mg·kg−1; Altromin Spezialfutter, Lage, Germany) for an additional 1, 5 or 12 weeks (n=12 animals per each group). The doxycycline-containing food was given for a week, followed by a week of “break”, when mice were fed regular feed [15, 22]. In this way, animals could ingest high doses of doxycycline, but the common side-effects of long-term antibiotic treatment, or massive Fgf10 overexpression, were minimised. Animals in control groups received feed lacking doxycycline. In previously published work, we demonstrated that doxycycline, given in such a way, does not affect cigarette smoke-induced emphysema and PH [22]. Before sacrificing, emphysema and PH phenotype in experimental mice was assessed in vivo by microcomputed tomography, echocardiography, lung function and haemodynamic measurements. In the case of in vivo measurements, n-numbers for final analysis vary due to technical reasons (e.g. animal death during measurements, incorrect placement of the haemodynamic catheter or the echocardiographic transducer or unsuccessful lung fixation). All animal experiments were approved by the local governmental authorities (Regierungspräsidium Gießen) in accordance with the German animal welfare law and the European legislation for the protection of animals used for scientific purposes (2010/63/EU).

Information about all other methods used in this study can be found in the supplementary material.

ResultsFGF10 signalling is impaired in the lungs of individuals with COPD

FGF10 expression was significantly downregulated in alveolar walls and upregulated in vessels of lungs explanted from human subjects with end-stage emphysema (figure 1a). Upregulation of FGF10 was limited to the smooth muscle cell layer of the muscularised proximal pulmonary vessels, whereas no change could be observed in smaller distal vessels (supplementary figure S2a). FGF10 levels in the serum of COPD patients were lower compared to healthy donors and were independent of the Global Initiative for Chronic Obstructive Lung Disease stage or diagnosed pulmonary hypertension (supplementary figure S2b). Slight downregulation of FGF10 in alveolar walls of smokers without emphysema was accompanied by increased expression of FGFR2 (figure 1b), which might explain the absence of emphysema. Moreover, FGF10 levels in alveolar septa negatively correlated with emphysema severity in COPD lungs (supplementary figure S2c).

FIGURE 1

Fibroblast growth factor (FGF)10 signalling in human donor and COPD lungs. a, b) Representative images and quantification of a) FGF10 and b) FGFR2 staining in human lung sections from healthy donors (D; n=10), smokers without emphysema (S; n=8) and smokers with COPD (C; n=18). Asterisks (*) indicate pulmonary vessels. c, d) FGF10 expression quantified by Western blotting in lung interstitial fibroblasts isolated from c) healthy donor lungs (n=6) and COPD lungs (n=6); d) healthy donor (n=6) and COPD (n=6) lungs in vitro treated with cigarette smoke extract (CSE). Protein expression in Western blot was quantified by band densitometry analysis and standardised to β-actin; relative expression was calculated by standardising each value to the mean value of the corresponding control group. Immunohistochemistry: red: FGF10/FGFR2; blue: haematoxylin. In the quantification panels each dot represents a measurement obtained from an individual COPD or healthy donor. The mean value for each group is represented by a horizontal line (±sd). Statistical analysis: a, b) one-way ANOVA (Dunnett's multiple-to-one comparison); c) unpaired and d) paired t-test; p-value for each comparison is given in the graphs. Bold type represents statistical significance.

Next, we investigated FGF10 expression in vitro in interstitial fibroblasts isolated from lungs explanted from healthy donors and people with end-stage emphysema. Interestingly, FGF10 expression remained impaired in COPD fibroblasts compared to donors, despite the in vitro propagation of the cells (figure 1c). Furthermore, cigarette smoke extract (CSE) treatment in vitro decreased FGF10 expression in interstitial lung fibroblasts isolated from healthy donors and from individuals with COPD (figure 1d).

Mice with impaired FGF10 signalling spontaneously develop PH and emphysema

To decipher the role of the FGF10/FGFR2b signalling axis during emphysema development, we exposed Fgf10+/− and Fgfr2b+/− mice along with wild-type littermates to cigarette smoke for 3 or 8 months (supplementary figure S3a). FGF10 was downregulated in lung homogenates from cigarette smoke-exposed mice. FGF10 expression in lungs of room air-exposed Fgf10+/− mice was at a similar level as in cigarette smoke-exposed wild-type animals and, interestingly, did not further decrease upon cigarette smoke exposure (supplementary figure S3b–d).

At the 3-month time point, wild-type and Fgf10+/− had a similar increase in right ventricular systolic pressure (RVSP) upon cigarette smoke exposure (figure 2a). Fgfr2b+/−mice had increased RVSP compared to the corresponding wild-type controls, which could be due to a slight increase in systemic arterial pressure (supplementary figure S4a). Our transgenic animals developed PH with ageing (8-month time point). The pressure measurements were consistent with the heart function data obtained by echocardiography (decreased tricuspid annular plane systolic excursion (TAPSE)) and right ventricular (RV) hypertrophy (figure 2b and c). Notably, we observed lower vessel density in room air-exposed Fgf10+/− and Fgfr2b+/− mice (figure 2d). Vascular pruning is indeed described in smokers’ lungs [23], whereas FGF10 is known to stimulate vasculogenesis and angiogenesis [24]. Cigarette smoke exposure could thus lead to vascular pruning, remodelling and PH via FGF10 downregulation. Moreover, cigarette smoke-exposed mice and mice with impaired FGF10 signalling exhibited pronounced vascular remodelling, measured as an increase in muscularisation of usually nonmuscularised or slightly muscularised pulmonary vessels confirming haemodynamic data (figure 2e, supplementary figure S4b). These results support the hypothesis that FGF10 signalling in the pulmonary vasculature plays an important protective role rather than contributing to pulmonary vascular remodelling [25].

FIGURE 2

Mice with impaired fibroblast growth factor (FGF)10 signalling spontaneously develop pulmonary hypertension. Pulmonary hypertension assessed by measurement of a) right ventricular systolic pressure (RVSP) (n=9–12) and b) echocardiographic tricuspid annular plane systolic excursion (TAPSE) quantification (n=10–12). c) Right ventricular (RV) hypertrophy quantified by the weight ratio between the RV and left ventricle with septum (LV+S) (n=10–12). No significant changes in LV+S mass were detected between analysed groups. d) Pulmonary vasculature visualised ex vivo by microcomputed tomography in lungs perfused with radio-dense Microfil, representative scans of lungs from wild-type (Wt), Fgf10+/− or Fgfr2b+/− mice exposed to room air (RA) for 3 months. e) Representative images of pulmonary vessels (brown: von Willebrand factor; purple: α-smooth muscle actin; green: nuclei) and quantification of average vessel muscularisation in lungs of experimental animals (n=5–7). Analysis was performed in pulmonary vessels with a diameter of 20–70 µm and results are shown as percentage of vessel circumference positive for α-smooth muscle actin staining. In the quantification panels each dot represents a measurement obtained from one individual experimental animal. The mean value for each group is represented by a horizontal line (±sd). Statistical analysis: two-way ANOVA; p-values for each comparison and interaction are given in the graphs. CS: cigarette smoke. Bold type represents statistical significance.

As reported previously, 3 months of cigarette smoke exposure in mice is sufficient to induce significant alterations in the pulmonary vasculature, while 8 months is needed for emphysema development [5, 6, 22, 26]. Fgf10+/− and Fgfr2b+/− mice developed emphysema much earlier than their wild-type littermates: after only 3 months of cigarette smoke exposure. Intriguingly, at the 8-month time point, room air-exposed Fgf10+/− and Fgfr2b+/− mice developed emphysema spontaneously, seen as significantly increased airspace percentage (figure 3a and b), increased mean linear intercept (MLI), decreased alveolar wall thickness and increased air-to-tissue volume ratio, compared to the room air-exposed wild-type controls (supplementary figure S5a–c). These findings were also confirmed using design-based stereology (figure 3c). Furthermore, mice with impaired FGF10 signalling showed significantly higher cellular senescence and spontaneously developed other typical pathomechanistic features that generally arise in response to cigarette smoke exposure (supplementary material, supplementary figures S5e–g and S6–8).

FIGURE 3

Mice with impaired fibroblast growth factor (FGF)10 signalling spontaneously develop pulmonary emphysema. a) Representative images of haematoxylin and eosin-stained lung sections of wild-type (Wt), Fgf10+/− or Fgfr2b+/− mice exposed to cigarette smoke (CS) or room air (RA) for 3 or 8 months. b, c) Pulmonary emphysema was quantified by determination of b) airspace percentage quantified using histological analysis (n=6–8) and c) alveoli number from left lung lobes using design-based stereology (n=5–6). d) In vivo matrix metalloproteinases (MMP) activity (MMPsense probe representing activity of MMP 2, 3, 7, 9, 12 and 13) or apoptosis (Annexin-Vivo probe) assessed using fluorescence molecular tomography imaging in experimental animals (n=4–6). In the quantification panels each dot represents a measurement obtained from one individual experimental animal. The mean value for each group is represented by a horizontal line (±sd). Statistical analysis: a–c) two-way ANOVA; d) paired two-way ANOVA; p-values for each comparison and interaction are given in the graphs. Bold type represents statistical significance.

In vivo fluorescence molecular tomography–computed tomography measurements revealed increased matrix metalloproteinase (MMP) activity and apoptosis (figure 3d, supplementary figure S9a) in cigarette smoke-exposed wild-type animals and room air-exposed mice with impaired FGF10 signalling (Fgf10+/− and Fgfr2b+/−) mice at the 3- and 8-month time points with spontaneous emphysema. FGF10 expression in fibroblasts is hypothesised to affect adjacent alveolar epithelial type 2 (AT2) cells during emphysema formation. Indeed, in vitro pre-treatment with recombinant FGF10 was able to prevent CSE-induced apoptosis and decreased viability in isolated AT2 cells (supplementary figure S9b).

FGF10 overexpression reverses cigarette smoke-induced PH and emphysema

We then tested whether FGF10 overexpression is sufficient to drive lung repair and reverse established cigarette smoke-induced emphysema and PH (supplementary figure S10a). We confirmed FGF10 overexpression via Western blotting and immunostaining in the lungs of animals fed with doxycycline (supplementary figure S10b and c). Between the examined groups, we could not detect changes in FGFR2 expression levels (supplementary figure S10d). Cigarette smoke-exposed mice developed PH and emphysema that remained stable for up to 3 months upon cigarette smoke cessation. FGF10 overexpression for 5 weeks was sufficient to completely reverse the cigarette smoke-induced increase of RVSP and RV hypertrophy and the decrease of RV function (figure 4a–c, supplementary figure S11a and b). Reversal of PH in FGF10-treated animals was associated with the reduction of cigarette smoke-induced remodelling (muscularisation) of small pulmonary vessels (figure 4d, supplementary figure S11c).

FIGURE 4

Fibroblast growth factor (FGF)10 overexpression reverses cigarette smoke (CS)-induced pulmonary hypertension in mice. a) Right ventricular systolic pressure (RVSP) (n=6–12), b) tricuspid annular plane systolic excursion (TAPSE) (n=5–12) and c) weight ratio of right ventricle (RV) and left ventricle plus septum (LV+S) (n=6–12), measured in CS- or room air (RA)-exposed mice that were subsequently fed with standard feed (control) or feed containing doxycycline (FGF10) for 1, 5 or 12 weeks. No significant changes in LV+S mass were detected between analysed groups. d) Representative images of pulmonary vessels (brown: von Willebrand factor; purple: α-smooth muscle actin; green: nuclei) and quantification of average vessel muscularisation in lungs of experimental animals (n=5–6). Analysis was performed in pulmonary vessels with a diameter of 20–70 µm and results are shown as percentage of vessel circumference positive for α-smooth muscle actin staining. In the quantification panels each dot represents a measurement obtained from one individual experimental animal. The mean value for each group is represented by a horizontal line (±sd). Statistical analysis: two-way ANOVA; p-values for each comparison and interaction are given in the graph. Bold type represents statistical significance.

Intriguingly, FGF10 overexpression successfully reversed established cigarette smoke-induced emphysema as demonstrated by the increase of alveolar numbers, decrease in MLI (figure 5a and b), decrease in airspace percentage and increase in alveolar wall thickness (supplementary figure S11d). The number of proliferating (Ki67+) cells was increased in mouse lungs upon FGF10 treatment (supplementary figure S11e). AT2 cell number in the distal lung parenchyma and gene expression of epithelial progenitor markers (Scgb1a1 and Sftpc) in bronchi were significantly increased in cigarette smoke-exposed mice upon FGF10 treatment for 1 week (figure 5c and d). The AT2 cell counts returned to the physiological level (figure 5c), while the area covered by AT1 cells increased (supplementary figure S11f) as lung structure was restored. Our results thus indicate that emphysema reversion could be mediated via the FGF10 effect on epithelial progenitor cells. Moreover, in the lung sections of cigarette smoke-exposed mice, we observed a significantly decreased von Willebrand factor immunostaining compared to room air-exposed controls, indicating loss of vessels. The cigarette smoke-induced decrease in von Willebrand factor staining was reversed by FGF10 overexpression (figure 5e). Additionally, FGF10 overexpression affected the composition of the extracellular matrix and reversed cigarette smoke-induced collagen accumulation in alveolar septa (supplementary figure S12).

FIGURE 5

Fibroblast growth factor (FGF)10 overexpression reverses cigarette smoke (CS)-induced pulmonary emphysema in mice. a) Representative images of haematoxylin and eosin stained lung sections from the respective experimental groups. b) Pulmonary emphysema was histologically quantified in CS- or room air (RA)-exposed mice that were subsequently fed with standard feed (control) or feed containing doxycycline (FGF10) for 1, 5 or 12 weeks, shown as alveoli number (calculated by design-based stereology, n=4–6) or mean linear intercept (MLI) (n=5–7). c) Immunofluorescence staining for alveolar epithelial type 2 cells (AT2, pro-surfactant protein C (pro-SPC)) and type 1 cells (receptor for advanced glycation end products (RAGE)) in lungs of experimental animals. Quantification shows percentage of AT2 of total lung cells in distal lung parenchyma (n=3). No significant changes in nuclei count were detected between analysed groups. d) mRNA expression of epithelial progenitor markers (secretoglobin family 1A member 1 (Scgb1a1) and surfactant protein C (Sftpc)) in laser microdissected bronchi of experimental mice (n=5). mRNA was quantified using reverse transcriptase quantitative PCR and expression of the gene of interest related to the expression of B2m used as a reference gene. e) Immunofluorescence staining and fluorescence intensity quantification of von Willebrand factor (vWF)/4′,6-diamidino-2-phenylindole (DAPI) in lungs from CS-treated mice with or without doxycycline-induced FGF10 overexpression for 12 weeks (n=6). In the quantification panels each dot represents a measurement obtained from one individual experimental animal. The mean value for each group is represented by a horizontal line (±sd). Statistical analysis: two-way ANOVA; p-values for each comparison and interaction are given in the graphs. Bold type represents statistical significance.

Fgf10+/− mice have a similar gene expression pattern to cigarette smoke-exposed wild-type mice

To decipher the underlying mechanisms of emphysema and PH development, we analysed gene expression profiles in wild-type and Fgf10+/− animals at a 3-month time point. We used laser-assisted microdissection to separately collect RNA from alveolar walls and pulmonary vessels of the experimental mouse lungs. Gene expression analysis was then performed using microarrays. We observed a strong correlation between general gene expression patterns of wild-type cigarette smoke and Fgf10+/− room air, compared with wild-type room air control (supplementary figure S13). Further analysis revealed 263 genes in the alveolar walls and 289 genes in pulmonary vessels that are commonly regulated between wild-type cigarette smoke and Fgf10+/− room air, compared with wild-type room air control (figure 6a). Moreover, in both compartments, these genes were regulated in the same direction (figure 6b) and, largely, not further changed upon cigarette smoke exposure in Fgf10+/− mice (supplementary figure S13a–d). Of importance, these findings indicate that the transcriptomic alterations occurring in response to cigarette smoke exposure are already “saturated” in the context of Fgf10 loss of function and that altered FGF10/FGFR2b seems to be central to the pathogenesis of cigarette smoke-induced PH and emphysema.

FIGURE 6

Cigarette smoke (CS)- and fibroblast growth factor (FGF)10-related molecular pathways involved in the development or reversion of emphysema and pulmonary hypertension. a) Venn diagrams were used to show the number of dysregulated genes during disease development or therapeutic intervention by FGF10 overexpression in the alveolar wall or the pulmonary vasculature. The overlap regions show genes that are in common for the compared groups. Genes that are commonly regulated (between wild-type (Wt) CS versus Wt room air (RA) and Fgf10+/− RA versus Wt RA during disease development; between control (Ctrl) CS versus Ctrl RA and FGF10 overexpression (ovxp) CS versus Ctrl CS during the therapeutic intervention) were further investigated. b) Diagrams showing the common genes with the direction of gene expression changes. Between the two compared groups of genes, positive or negative correlation implies the similar or opposite direction of fold changes, respectively. c) Genes that are regulated in Wt CS versus Wt RA and Fgf10+/− RA versus Wt RA during the disease development are investigated using functional protein association networks (STRING-DB). The analysis shows only genes whose products are connected in the alveolar wall compartment or in pulmonary vessels. Genes that play a role in the Wnt signalling pathway and are regulated during the disease development in the alveolar wall compartment are marked in blue. A cluster related to tyrosine kinase signalling in the pulmonary vasculature is marked in orange. Nos3: (endothelial) nitric oxide synthase 3; Ctsk: cathepsin K; Bok: Bcl-2-related ovarian killer; Tox2: TOX high mobility group box family member 2; Wnt5a: Wnt family member 5A; Ror1: receptor tyrosine kinase-like orphan receptor 1; Actr3: actin-related protein 3; Braf: serine/threonine-protein kinase B rapidly accelerated fibrosarcoma; Prkca: protein kinase Cα; Kit: tyrosine kinase receptor KIT; Vefga: vascular endothelial growth factor A; Cyfip2: cytoplasmic FMR1 interacting protein 2; Nhlrc2: NHL repeat containing 2; Pdgfa: platelet-derived growth factor subunit α; LFC: log fold change.

Genes involved in the FGF10-mediated reversal of cigarette smoke-induced emphysema and PH were analysed at the earliest time point: 1 week of FGF10 overexpression after 8 months of cigarette smoke exposure. In the alveolar wall compartment, we found 209 genes that were affected by both cigarette smoke exposure and subsequent FGF10 overexpression. In the pulmonary vasculature, 153 genes were commonly regulated (figure 6a). The vast majority of genes in the alveolar wall compartment were inversely regulated upon FGF10 overexpression compared to cigarette smoke exposure (figure 6b, supplementary figure S13e–h). Furthermore, endothelial nitric oxide synthase (Nos3) was downregulated in alveolar walls upon cigarette smoke exposure and upregulated with FGF10 overexpression. This suggests that the endothelium in the alveolar walls could be affected by cigarette smoke exposure and by FGF10 overexpression, as previously shown in figure 5e.

However, in the pulmonary vasculature, we could distinguish two populations of genes (figure 6b). One group of genes was inversely regulated by FGF10 expression compared to the effect of cigarette smoke exposure. These genes could be involved in the reverse remodelling process and resolution of cigarette smoke-induced PH. We found transcription factor TOX high mobility group box family member 2 (Tox2) that could be of interest for the mechanism of PH reversion. Tox2 was the most prominently downregulated gene in the pulmonary vasculature upon cigarette smoke exposure and markedly upregulated with FGF10 overexpression. Interestingly, single-nucleotide polymorphisms in the human TOX2 gene were associated with increased diastolic blood pressure in systemic circulation [27]. In pulmonary vasculature, the other group of genes was regulated in the same direction by cigarette smoke exposure and FGF10 overexpression. This finding is not surprising, considering that cigarette smoke exposure also upregulates FGF10 in the pulmonary vascular compartment.

The gene regulation pattern and the genes with the most robust regulation are depicted in heat maps (supplementary figure S14). In order to identify pathways and test if the regulated genes align in specific signalling pathways, we performed functional protein association network analysis (STRING-DB) [28]. Interestingly, the targets were rather scattered in several clusters, comprising a few targets each and we could not find enrichment in any specific signalling pathway (figure 6c, supplementary figure S15a and b). It thus seems that FGF10 deficiency and cigarette smoke exposure trigger many targets rather than one distinct signalling pathway. However, using such analysis, we could identify clusters related to Wnt signalling in the alveolar wall compartment and tyrosine kinases in the vascular compartment. Along these lines, we found increased Akt phosphorylation and decreased expression of β-catenin in lung homogenates from room air-exposed Fgf10+/− and cigarette smoke-exposed wild-type mice (supplementary figure S15c and d).