Abstract
FOXF1, a member of the forkhead box family of transcription factors, has been previously shown to be critical for lung development, homeostasis, and injury responses. However, the role of FOXF1 in lung regeneration is unknown. Herein, we performed partial pneumonectomy, a model of lung regeneration, in mice lacking one Foxf1 allele in endothelial cells (PDGFb-iCre/Foxf1 fl/+ mice). Endothelial cell proliferation was significantly reduced in regenerating lungs from mice deficient for endothelial Foxf1. Decreased endothelial proliferation was associated with delayed lung regeneration as shown by reduced respiratory volume in Foxf1-deficient lungs. FACS-sorted endothelial cells isolated from regenerating PDGFb-iCre/Foxf1 fl/+ and control lungs were used for RNAseq analysis to identify FOXF1 target genes. Foxf1 deficiency altered expression of numerous genes including those regulating extracellular matrix remodeling (Timp3, Adamts9) and cell cycle progression (Cdkn1a, Cdkn2b, Cenpj, Tubb4a), which are critical for lung regeneration. Deletion of Foxf1 increased Timp3 mRNA and protein, decreasing MMP14 activity in regenerating lungs. ChIPseq analysis for FOXF1 and histone methylation marks identified DNA regulatory regions within the Cd44, Cdkn1a, and Cdkn2b genes, indicating they are direct FOXF1 targets. Thus FOXF1 stimulates lung regeneration following partial pneumonectomy via direct transcriptional regulation of genes critical for extracellular matrix remodeling and cell cycle progression.
Similar content being viewed by others
Introduction
Lung regeneration plays an important role in lung repair after injury. It is reliant upon proliferation of multiple cell types in the lung, including endothelium, epithelium, and fibroblasts, as well as remodeling of the extracellular matrix. Lung regeneration following injury progresses via an initial inflammatory response during which macrophages clear the tissue of cellular debris. This process continues through cellular proliferation when existing cells and progenitors act to repopulate cells lost during injury, followed by tissue maturation in which newly formed cells achieve a differentiated phenotype1. Signaling pathways critical for lung regeneration include FGF, EGF, WNT, and NOTCH. In addition, HDACs, miRNAs, ELASTIN, and MMP14 have been shown to regulate lung regeneration2, 3. Partial pneumonectomy (PNX) has been used as a therapeutic and investigational tool for several decades. Following PNX the remaining lung increases in size to compensate for loss of volume and respiratory capacity. Much has been learned about the triggers and mechanisms regulating pulmonary regeneration. However, the role of endothelial cells in post-PNX lung growth remains incompletely characterized. Endothelial cells are essential for alveolar structure as well as to mediate gas exchange between the alveoli and circulatory system. Two divergent processes have been proposed to stimulate post-PNX lung growth. Lung volume may be recuperated by increasing the size of individual alveoli, leading to a simpler lung structure while increasing gas exchange surface area4. The alternative process is to increase the number of alveoli, which maintains a consistent pulmonary morphology while expanding lung volume and surface area for gas exchange4. Both processes of post-PNX lung growth involve a period of rapid cell proliferation including endothelial and epithelial cells as both are essential for formation of alveoli to restore respiratory capacity after resection. While endothelial cells stimulate epithelial proliferation and induce lung repair and regeneration via MMP14-mediated release of EGF ligands3, the transcriptional mechanisms regulating endothelial cells during lung regeneration remain largely unknown.
Foxf1 transcription factor (previously known as HFH-8 or Freac1) is known to be an essential mediator of lung development and embryonic angiogenesis. Foxf1 −/− mice are embryonic lethal due to abnormal development of yolk sac, vasculature, and allantois5,6,7,8. Haploinsufficiency for Foxf1 causes severe lung malformations and inhibits development of pulmonary capillaries during embryonic and early postnatal periods7. Heterozygous deletions and point mutations in the FOXF1 gene locus were found in more than 40% of patients with alveolar-capillary dysplasia with misalignment of pulmonary veins (ACD/MPV)9. Conditional deletion of Foxf1 from endothelial cells is sufficient to cause embryonic lethality due to severely underdeveloped vasculature in the yolk sac and placenta as well as the loss of VEGF receptors, FLK1 and FLT1 from the surface of endothelial cells10. In adult mice, endothelial-specific deletion of both Foxf1 alleles results in uniform lethality due to pulmonary hemorrhage, massive inflammatory cell infiltration, and pulmonary edema, resulting from disruption of adherens junctions and loss of endothelial barrier function11. Deletion of only one Foxf1 allele from endothelial cells conferred greater susceptibility to butylated hydroxytoluene- (BHT) or lipopolysaccharide- (LPS) induced lung injury11. FOXF1 regulates expression of genes critical for maintenance of the endothelial barrier, such as VE-cadherin (Cdh5) and sphingosine 1-phosphate receptor 1 (S1pr1)11. Foxf1 +/− mice have also been shown to be more susceptible to chemically induced lung12, 13 and liver injury14. While these studies demonstrate a critical role for FOXF1 in lung injury and inflammation, molecular mechanisms regulated by FOXF1 during proliferative responses induced by PNX remain uncharacterized.
In the present study, a PNX model was used to investigate the role of FOXF1 in lung regeneration. Since deletion of both Foxf1 alleles in endothelial cells is lethal11, a tamoxifen-inducible Cre (PDGFb-iCre) was used to cause Foxf1 haploinsufficiency in endothelial cells prior to PNX surgery. PDGFb-iCre/Foxf1 fl/+ mice exhibited delayed lung regeneration associated with diminished proliferative capacity of endothelial cells. Our study shows that FOXF1 transcriptionally regulates Timp3, Cdkn1a, Cdkn2b, and Cd44, all of which are critical for extracellular matrix remodeling and endothelial proliferation. This study demonstrates that FOXF1 induces lung regeneration and reestablishment of normal alveolar structure after PNX.
Methods
Ethics statement
All animal experiments and protocols used in the study were approved by and carried out in accordance with the guidelines of the Cincinnati Children’s Research Foundation Institutional Animal Care and Use Committee (IACUC) and the NIH IACUC Guidebook. All experiments were covered under our animal protocol (IACUC2016-0038). The Cincinnati Children’s Research Foundation Institutional Animal Care and Use Committee is an AAALAC and NIH accredited institution (NIH Insurance #8310801).
Generation of mice with conditional deletion of Foxf1 from endothelial cells
Mice carrying the Foxf1-floxed allele, in which LoxP sites flank the first exon of the Foxf1 gene encoding the DNA binding domain, have been described previously10, 11. Foxf1-floxed mice were bred with PDGFb-iCre mice, a tamoxifen inducible Cre that targets endothelial cells11, 15. Mice heterozygous for the Foxf1-floxed allele (PDGFb-iCre/Foxf1 fl/+, abbreviated PDGFb-Foxf1 fl/+) were utilized in this study (Fig. 1A). No differences were observed between Foxf1 fl/+ and PDGFb-iCre/Foxf1 fl/+ mice in the absence of tamoxifen, nor did tamoxifen administration affect Foxf1 fl/+ mice lacking Cre.
Partial pneumonectomy (PNX) surgery
Pneumonectomy surgery was performed as previously described16. Briefly, under isoflurane anesthesia, 8–12 week old control or PDGFb-iCre/Foxf1 fl/+ mice were intubated to maintain pulmonary pressure. The left lung lobe was exposed via surgical incision into the intercostal space. The left bronchiole and supporting vasculature were ligated by surgical clamp, the left lung lobe was excised, and the thoracic cavity closed with suture. Negative pressure was returned to the thoracic cavity following wound closure by drawing air out using a 3 cc syringe with needle placed into the thoracic cavity prior to closure. Sham operated mice were treated the same as surgical mice without removal of the lung lobe, they still were intubated and the thoracic cavity breached via incision into the intercostal space. Following surgery mice were given Buprenex for pain management and monitored for adverse side effects. Tamoxifen (3 mg/day, Sigma) was administered by oral gavage the day prior to and the morning of PNX surgery.
Evaluation of lung function, real-time qRT-PCR, Western blot, and immunostaining
Evaluation of lung volume and function was performed on days 2, 4, 7, 14, and 42 post-PNX. The middle right lobe was collected for RNA and protein, while the remaining right lobes, cranial, caudal and accessory, were imbedded for immunostaining. Alterations in lung volume and function following PNX were assessed using the Flexivent small animal ventilator (SCIREQ) as previously described17. Protein extracts were used for either Western blot analysis as previously described18,19,20 or gel zymography. RNA was isolated according to manufacturer’s protocol and reverse transcribed using Applied Biosystems high capacity reverse transcription kit according to manufacturer’s protocol as previously described20,21,22,23. Analysis of changes in gene expression was performed using a StepONE qRT-PCR machine (Applied Biosystems) and inventoried Taqman probes (Applied Biosystems, Supplemental Table 1) as previously described20,21,22,23. Lungs were gravity inflated with 4% paraformaldehyde prior to paraffin imbedding. Immunohistochemistry and immunofluorescence were performed on 5 μm paraffin sections and stained as previously described20,21,22, 24,25,26,27,28. Antibodies used in this study were FOXF1 (ref. 10), Ki-67 (1:200 (immunohistochemistry, IHC), 1:200 (immunofluorescence, IF);Thermo Scientific), PCNA (1:2000 (IHC), 1:500 (IF); Roche Diagnostics), PH3 (1:1000; Santa Cruz Biotechnology), TIMP3 (1:750;Santa Cruz Biotech), SOX17 (1:250; 7 Hills Bioreagents), SPC (1:500; 7 Hills Bioreagents), PECAM-1 (1:1000 (WB); Abcam), ENDOMUCIN (1:100; Abcam), BrdU (1:500 (IF); Iowa Hybridoma Bank), Cleaved Caspase 3 (1:1000; R&D Systems), CDKN1A (P21Waf1/Cip1)(1:1000; Cell Signaling Technolgy), CDKN2B (P15INK4b)(1:1000; Aviva Systems Biology), MMP14 (1:200; Abcam), β-ACTIN (1:6000; Santa Cruz Biotechnology). Images were collected using a Zeiss Axioplan 2 microscope and AxioVision software.
ChIPseq and RNAseq
Formaldehyde cross-linked chromatin was sonicated using Covaris S220 to 200–300 bp and ChIP was performed using SX-8G IP-STAR robot (Diagenode) using the AF4798 antibody. ChIP-Seq library was prepared using ChIPmentation procedure and libraries were sequenced using Illumina HiSeq. 2500 at CCHMC sequencing core. Data analysis was performed using the BioWardrobe platform29. Briefly, ChIP-Seq reads were aligned by Bowtie to the mouse genome (mm10); only unique reads with no more than one mismatch were kept. Reads were extended to estimated fragment length, normalized to total mapped read number and displayed as coverage on a mirror of the University of California Santa Cruz (UCSC) genome browser. MACS230 was used to identify islands of enrichment. FOXF1 sequence logos were identified with MEME-ChIP31. FOXF1 ChIPseq data was aligned with previously published ChIPseq data for histone methylation (GEO Accession GSE31039) using the BioWardrobe platform.
RNAseq analysis was performed on FACS-sorted endothelial cells (CD45−CD31+CD326−) from control and PDGFb-iCre/Foxf1 fl/+ lungs 4 days after PNX surgery. Changes in gene expression were determined using DESeq and analyzed as previously described11. Heat map was generated from differentially expressed genes using JMP Genomics 6.0.
ChIPseq and RNAseq data are available as GEO accession GSE100149.
Analysis of lung air space
Changes in lung morphology were evaluated using the mean intercept method of evaluating respiratory versus air space within the lung. H&E stained 20x fields of lungs from sham operated and PNX mice during the first 2 weeks of recovery were analyzed. Analysis was performed by placing a lattice over the image and counting whether intersections landed on respiratory space or air space. Data is represented as percent of air space. 15 random fields from 4 mice per group were analyzed.
Analysis of ENDOMUCIN staining
Changes in endothelial cell coverage in the lung were evaluated by measuring fluorescence using ImageJ software. ENDOMUCIN-stained 20x fields of lungs from sham operated and PNX mice at four and seven days of recovery were analyzed. Analysis was performed by measuring fluorescent intensity. Data is represented as mean fluorescent intensity. 15 random fields from 4 mice per group were analyzed.
Fluorescence assisted cell sorting (FACS)
Isolation of endothelial and epithelial cell populations by FACS was performed as previously described10, 11, 32, 33. Hematopoetic cells were eliminated based on CD45 expression. Endothelial cells were then defined from CD45− cells as having high CD31 expression and lacking CD326 expression (CD45−CD31+CD326−). Epithelial cells were defined from CD45− cells as having high CD326 expression and lacking CD31 expression (CD45−CD31−CD326+). Antibodies used were CD45 (eBioscience, clone 30-F11), CD31 (eBioscience, clone 390), and CD326 (eBioscience, clone G8.8). Cells were separated using Five-laser FACSAria II, BD Biosciences. Purified cells were used for qRT-PCR and RNAseq.
FACS isolated endothelial cells (CD45−CD31+CD326−) were analyzed for DNA content using DRAQ5, a cell permeable dye. Percentages of cells in G0/G1 and S/G2/M phases of the cell cycle were measured. 3 mice were used per group.
Zymography
Zymography was performed on whole lung protein samples collected on day 7 post-PNX. Samples were homogenized under non-reducing conditions and protein concentration determined by Applied Biosystems DC protein assay according to manufacturer’s instructions. Ten μg of total protein were loaded per sample and run on zymography gel according to manufacturer’s protocol (10% gelatin gel, NOVEX). Gel images were captured on a Canon scanner (Canoscan LiDE 90) and band intensity determined by ImageJ software.
Statistical analysis
Statistical analysis was performed using Excel software. Student’s t-test analysis was used to determine significance which was set at p ≤ 0.05. All data are represented as mean ± standard error of mean (SEM).
Results
Foxf1 deficiency in endothelial cells delays lung regeneration after partial pneumonectomy
To determine the role of FOXF1 in endothelial cells during lung regeneration, we performed partial pneumonectomy (PNX) surgery on mice with endothelial-specific Foxf1 deficiency. Since deletion of both Foxf1 alleles from endothelial cells is lethal11, mice containing a deletion of one Foxf1 allele (PDGFb-iCre/Foxf1 fl/+) were used in the present study (Fig. 1A). Tamoxifen was administered prior to PNX surgery to activate Cre. After PNX, Foxf1 mRNA expression was significantly decreased in both PDGFb-iCre/Foxf1 fl/+ and control mice (Fig. 1B), Foxf1 mRNA levels were significantly lower in PDGFb-iCre/Foxf1 fl/+ mice compared to control mice as determined by qRT-PCR analysis (Fig. 1B). Consistent with decreased Foxf1 mRNA, FOXF1 staining was decreased in PDGFb-iCre/Foxf1 fl/+ lungs compared to controls (Fig. 1C). Evaluation of FOXF1 distribution in sham-operated and regenerating lungs showed that FOXF1 is widely distributed in the alveolar region but can only be detected in a subset of endothelial cells within pulmonary arteries and veins (Supplemental Fig. 1). Measurement of lung function showed PNX caused a 40% decrease in lung volume in control and PDGFb-iCre/Foxf1 fl/+ mice 2 days after surgery, consistent with removal of the left lung lobe. Lung volume increased in mice from both genotypes from day 2 following PNX, ultimately returning to sham levels by 42 days (Fig. 1D). Regeneration was delayed in PDGFb-iCre/Foxf1 fl/+ mice compared to controls, as shown by significantly smaller lung volumes in these mice from days 4 through 14 (Fig. 1D). Interestingly, despite delayed regeneration Foxf1-deficient lungs had normal size of alveoli (Supplemental Fig. 2A,B) and lung function was not affected as shown by measurements of lung compliance, elastance, and resistance (Fig. 1E,F, Supplemental Fig. 2C). Consistent with these findings lung collagen content and distribution was unaltered as shown by qRT-PCR analysis of Col1a1 and Col3a1 (Supplemental Fig. 2E) as well as Masson’s trichrome stain (Supplemental Fig. 2F). Tamoxifen administration did not affect lung regeneration as Foxf1 fl/+ and PDGFb-iCre/Foxf1 fl/+ mice in the absence of tamoxifen showed normal kinetics of lung volume recovery (Supplemental Fig. 2D). Thus, Foxf1 expression in endothelial cells promotes lung regeneration following PNX.
ChIPseq analysis identifies FOXF1 target genes in endothelial cells
To determine molecular mechanisms whereby FOXF1 stimulates lung regeneration, ChIPseq analysis was performed in the MFLM-91U pulmonary endothelial cell line. ChIPSeq data analysis identified 16024 FOXF1 binding sites. MEME-ChIP31 analysis of the sequence under the peaks identified a TAAACA motif (Fig. 2D) which is similar to the motif previously reported for FOXF134 (Fig. 2D), 8916 genes were located in the vicinity of FOXF1 binding sites. Of these genes, FOXF1 bound to known promoter regions for 2196 genes. Gene ontology analysis performed using DAVID35 and ToppFun36 softwares showed that cell cycle regulating genes were the major targets of FOXF1 transcription factor binding (Fig. 2A,B). The main biological processes affected by FOXF1 target genes included cell cycle regulation, cell division, mitotic cell division, regulation of transcription, chromosome organization, and DNA repair (Fig. 2A,B). Other gene categories associated with FOXF1 binding include: transcription cofactor activity, transcription factor binding, and regulation of gene expression. ChIPseq demonstrated that FOXF1 bound to chromatin at Flt1, Kdr, Notch2, Itgb3, Pdgfb, Pecam1, S1pr1, and Tek gene loci, all of which have been previously shown to be direct FOXF1 target genes10, 11, 37 (Supplemental Table 2). Among the novel FOXF1 targets identified were centromere protein J (Cenpj) and tubulin beta-4A (Tubb4a), both of which regulate mitotic cell division, mitotic spindle formation, and chromatin separation38,39,40 (Supplemental Table 2, Fig. 2C). Histone deacetylase (Hdac) 5 and 7 gene loci were also bound by FOXF1 (Supplemental Table 2, Fig. 2C, Supplemental Fig. 3). HDAC proteins regulate histone tertiary structure which is important in both gene transcription and cell division41. FOXF1 also bound to promoter regions of Cdkn1a, Cdkn2b, Ccnd3, and Cd44, genes critical for endothelial proliferation42, 43 (Supplemental Table 2, Supplemental Fig. 3). Other FOXF1 target genes identified by ChIPseq include: Adamts9, Itgb4, Ptgs1, Ptgs2, and Spry4 (Supplemental Table 2, Supplemental Fig. 3).
Identification of FOXF1 target genes in regenerating lungs
Next we determined whether Foxf1-deficiency changed expression of these genes in vivo. FACS was used to purify populations of endothelial (CD45−CD31+CD326−) and epithelial (CD45−CD31−CD326+) cells from regenerating lungs (Fig. 3A). Purity of populations isolated by FACS was confirmed by the presence of Pecam1 in endothelial cells and Sftpc in epithelial cells as determined by qRT-PCR (Supplemental Fig. 4A,B). Foxf1 was expressed in endothelial cells but not epithelial cells (Fig. 3B), which is consistent with previous studies10, 11. A 50% decrease in Foxf1 mRNA was found in PDGFb-iCre/Foxf1 fl/+ endothelial cells compared to endothelial cells from control mice (Fig. 3B), consistent with deletion of one Foxf1 allele. To identify transcripts regulated by FOXF1, FACS-sorted endothelial cells were analyzed by genome-wide RNAseq analysis. Significant changes in expression for over a thousand genes were found in endothelial cells from PDGFb-iCre/Foxf1 fl/+ lungs 4 days after PNX (Fig. 3C). Consistent with ChIPseq data (Fig. 2A,B), mRNA expression of multiple cell cycle regulators was altered. These genes include: Cenpj, Tubb4a, Adamts9, Hdac5, Hdac7, Itgb3, Itgb4, Ptgs1, Ptgs2, and Spry4 (Table 1). Cd44, which promotes endothelial cell proliferation, was one of the genes with decreased expression in Foxf1-deficient endothelial cells (Table 1). Among the genes with increased mRNA levels in PDGFb-iCre/Foxf1 fl/+ mice were the cell cycle inhibitory proteins Cdkn1a, Cdkn2b, and Spry4 (Table 1) all of which inhibit cellular proliferation in vitro and in vivo 44. Expression of Timp3 and Adamts9, both critical for ECM remodeling, was also increased in PDGFb-iCre/Foxf1 fl/+ endothelial cells (Table 1). qRT-PCR confirmed altered expression of these genes in Foxf1-deficient endothelial cells (Fig. 3E). qRT-PCR analysis also showed decreased expression of Ccnd3 in Foxf1-deficient endothelial cells, supporting decreased cell cycle progression after PNX. Consistent with mRNA changes, Western blot analysis showed that protein levels of CDKN1A (P21Cip1) and CDKN2B (P15Ink4b) were increased in lungs from PDGFb-iCre/Foxf1 fl/+ mice at base line and 3 days after PNX (Fig. 3D), consistent with decreased proliferation at these time points. Levels of CDKN1A and CDKN2B increase in control lungs 7 days after PNX as proliferation starts to wane. Thus FOXF1 regulates multiple endothelial genes critical for lung regeneration, ECM remodeling, and cell cycle progression.
Decreased proliferation in PDGFb-iCre/Foxf1fl /+ mice after PNX
Since FOXF1 binds to and regulates expression of cell cycle regulatory genes, we examined cellular proliferation in Foxf1-deficient lungs in a PNX model of lung regeneration. Consistent with previous reports3, 4 we observed increased proliferation after PNX in control mice. However, the number of Ki-67-positive cells was significantly lower in PDGFb-iCre/Foxf1 fl/+ mice compared to controls (Fig. 4A,B). In agreement with Ki-67 staining, immunostaining for proliferating cell nuclear antigen (PCNA) or phospho-Histone 3 (PH3) showed significantly less cell proliferation in PDGFb-iCre/Foxf1 fl/+ mice compared to control mice (Fig. 4C,D). Furthermore, the number of cells undergoing DNA replication after PNX was decreased in lungs of PDGFb-iCre/Foxf1 fl/+ mice as determined by BrdU incorporation (Fig. 5A,B). As PDGFb-iCre specifically targets endothelial cells, immunostaining for Ki-67 and SOX17, a marker of endothelial cells45, was performed to visualize endothelial cells undergoing the cell cycle. In agreement with previous studies3, PNX surgery increased the number of proliferating endothelial cells in both mouse lines (Fig. 5C,D). Endothelial cell proliferation was significantly decreased in regenerating lungs of PDGFb-iCre/Foxf1 fl/+ mice compared to control (Fig. 5C,D). Consistent with decreased endothelial cell proliferation, Pecam-1 mRNA and protein were reduced in PDGFb-iCre/Foxf1 fl/+ lungs as determined by qRT-PCR and Western blot, respectively (Fig. 5E,F). ENDOMUCIN staining and quantification (Supplemental Fig. 5A,B) showed similar expression and patterning during the regenerative phase of post-PNX growth, consistent with maintenance of lung morphology between control and PDGFb-iCre/Foxf1 fl/+ mice during lung regeneration. Although proliferation of alveolar type II epithelial cells also increased after PNX, it was largely unchanged between control and PDGFb-iCre/Foxf1 fl/+ lungs (Supplemental Fig. 5C,D). Staining for activated caspase 3 showed no difference between control and PDGFb-iCre/Foxf1 fl/+ lungs (Supplemental Fig. 6A), suggesting apoptosis does not contribute to delayed lung regeneration in PDGFb-iCre/Foxf1 fl/+ mice. DNA content was analyzed in FACS-sorted endothelial cells (CD45−CD31+CD326−) by DRAQ5 at 19 days post-PNX to evaluate endothelial proliferation during the late phase of regeneration. This analysis showed that a small number of endothelial cells still proliferate at this late stage of regeneration, with decreased numbers of endothelial cells undergoing S, G2, and M phases of the cell cycle in PDGFb-iCre/Foxf1 fl/+ lungs (Supplemental Fig. 6B,C). Therefore, deletion of one Foxf1 allele from endothelial cells was sufficient to delay endothelial proliferation and lung regeneration after PNX.
Since analysis of FOXF1 binding to chromatin indicated cell cycle regulation is the main gene ontology signature for FOXF1, we further investigated three genes critical for endothelial proliferation, Cdkn1a, Cdkn2b, and Cd44. These genes were analyzed to determine whether their respective FOXF1 binding regions had positive or negative histone methylation marks as determined by ChIPseq. FOXF1 bound to Cd44 in the promoter region (−494/−1014 bp) and within the final exon (+6058/+5857 bp) (Fig. 6, Supplemental Table 2). FOXF1 binding regions of Cd44 had tri- and mono-methylation of histone 3 lysine 4 (H3K4me1 and H3K4me3), both of which promote transcriptional activation46,47,48,49 (Fig. 6), suggesting FOXF1 stimulates Cd44 transcription. These results are consistent with decreased Cd44 mRNA levels in Foxf1-deficient endothelial cells (Fig. 3E). In contrast, FOXF1 binding regions for Cdkn1a (−151/+394 bp and +1447/+1627 bp) and Cdkn2b (+897/+1073 bp and +1866/+2032 bp) often aligned with H3K4me1 and H3K4me3 but were also positive for H3K27 tri-methylation (H3K27me3) which causes transcriptional silencing50 (Fig. 6). Consistent with the presence of H3K27me3 gene silencing marks, mRNA and protein for Cdkn1a (P21Cip1) and Cdkn2b (P15Ink4b) were increased in Foxf1-deficient lungs (Fig. 3D,E). Thus, FOXF1 binds to both enhancer and repressor regions to differentially regulate gene expression in regenerating endothelial cells.
Increased TIMP3 expression, decreased MMP14 activity in PDGFb-iCre/Foxf1fl /+ lungs
Rearrangement of the extracellular matrix (ECM) by MMPs is a major component of alveologenesis and it was previously shown that MMP14 induces lung regeneration after PNX3. While neither Mmp14 mRNA nor protein was altered (Supplemental Fig. 7), mRNA expression of Timp3, an inhibitor of MMP1451, was increased in FACS-sorted PDGFb-iCre/Foxf1 fl/+ lung endothelial cells as shown by RNAseq (Fig. 3C) and qRT-PCR (Fig. 7A). FOXF1 directly bound to Timp3 DNA regulatory regions located in the promoter and first intron (Fig. 7E and Supplemental Table 2), indicating that Timp3 is a direct transcriptional target of FOXF1. Consistent with this hypothesis, TIMP3 immunostaining was increased and more TIMP3-positive cells were observed in regenerating lungs from PDGFb-iCre/Foxf1 fl/+ mice compared to controls (Fig. 7B). Furthermore, zymography was used to demonstrate that gelatinase activity of MMP14, a functional readout of TIMP3 activation51, was significantly decreased in regenerating lungs from PDGFb-iCre/Foxf1 fl/+ mice (Fig. 7C,D). Thus, reduced MMP14 activation can contribute to delayed lung regeneration in PDGFb-iCre/Foxf1 fl/+ mice. Altogether, FOXF1 stimulates lung regeneration by regulating transcription of endothelial genes critical for ECM remodeling and cell cycle progression (Fig. 7F).
Discussion
Lung injury or lung reduction surgery diminishes the capacity for gas exchange within the lung, causing hypoxia within the tissues of the body which can lead to organ failure. Due to the profound need for efficient lung function, pathways through which lung regeneration can be expedited could dramatically help overall survival. In this study, we found that FOXF1, a transcription factor expressed in endothelial cells, is critical for lung regeneration following PNX. Mice with endothelial-specific deletion of one Foxf1 allele have slower rates of lung regeneration compared to control mice. Utilizing a combination of RNAseq and ChIPseq analyses novel transcriptional targets of FOXF1 where identified in this study. Our data suggests FOXF1 stimulates lung regeneration by regulating endothelial genes critical for ECM remodeling (Timp3, Adamts9), cell cycle regulation (Cdkn1a, Cdkn2b), and cytokinesis (Tubb4a, Cenpj, Hdac5, and Hdac7). DNA regulatory regions of these genes possess high-affinity binding sites for FOX proteins that were occupied by FOXF1 bound to chromatin. The functional importance of these FOXF1-binding regions was supported by epigenetic marks such as single and triple methylation of lysine 4 of histone 3 which mark enhancer regions46,47,48,49 as well as triple methylation of lysine 27 of histone 3 which marks regions of gene repression50. Thus, FOXF1 is an important regulator of lung regeneration after PNX.
Our data are consistent with published studies demonstrating that mice haploinsufficient for the global Foxf1-null mutation (Foxf1 +/−) exhibited decreased rates of liver regeneration following carbon tertrachloride administration14 and increased susceptibility to butylated hydroxytoluene (BHT)-induced lung injury13. Interestingly, Foxf1 +/− mice have a variety of developmental defects including fusion of lung lobes and vessels7, 52, 53 as well as esophageal7 and gastrointestinal abnormalities54. Therefore, these mice are not practical for use in PNX regeneration studies. Herein, we utilized an inducible mouse model to delete one Foxf1 allele in an endothelial restricted manner in adult mice. This model enabled discovery of novel functions of FOXF1 as a transcriptional driver of lung regeneration. PDGFb-iCre/Foxf1 fl/+ mice had slower rates of regeneration up to two weeks after PNX, but Foxf1-deficient lungs did ultimately regenerate back to pre-PNX volumes by 6 weeks. Nineteen days after PNX, a small percentage of endothelial cells were still proliferating in control and PDGFb-iCre/Foxf1 fl/+ lungs. This low level of proliferation likely continues in PDGFb-iCre/Foxf1 fl/+ mice until the stresses that stimulate lung regeneration are quenched, which occurs quicker in control lungs where endothelial cells proliferate faster. Previous research has shown that pulmonary microvascular endothelial cells stimulate lung regeneration via VEGFR2 and FGFR13; however, expression of these receptors was not altered in the current study, suggesting they do not play a role in decreased endothelial cell proliferation in PDGFb-iCre/Foxf1 fl/+ lungs. It was further shown that MMP14 activity was essential for post-PNX growth as administration of anti-MMP14 monoclonal antibodies blocked lung regeneration3. While Mmp14 expression was not altered in the current study, regenerating PDGFb-iCre/Foxf1 fl/+ lungs had increased mRNA and protein for Timp3, an inhibitor of MMP14 activity. Consistent with increased TIMP3 immunostaining, MMP14 activity was decreased in Foxf1-deficient lungs, suggesting reduced MMP14 activity can contribute to delayed lung regeneration in PDGFb-iCre/Foxf1 fl/+ mice. Interestingly, mRNA expression of Adamts9, which regulates ECM remodeling similar to TIMP355, 56, was also increased in Foxf1-deficient lungs. Since both Adamts9 and Timp3 are direct FOXF1 targets, FOXF1 may induce lung regeneration through transcriptional repression of these genes.
Previous studies have shown decreased cell proliferation in Foxf1-deficient embryos which resulted in decreased angiogenesis and smaller lung size10. Foxf1 depletion in rhabdomyosarcoma cells increased levels of the CDK inhibitor Cdkn1a (P21Cip1), whereas Foxf1-overexpression reduced Cdkn1a levels57. ChIP analysis showed FOXF1 bound directly to the Cdkn1a promoter, contributing to its transcriptional repression57. Interestingly, in addition to the FOXF1 binding site identified in rhabdomyosarcoma cells by Milewski et al.57, the current study identified an additional high affinity FOXF1 binding region located within the first exon of the Cdkn1a gene. Since an endothelial cell line was used for ChIPseq in our studies, FOXF1 may regulate Cdkn1a through distinct regulatory regions in a tissue or cell-type specific manner. It is also possible that FOXF1 requires cofactors to efficiently bind chromatin and that expression of these cofactors in different cell-types could mediate the ability of FOXF1 to bind distinct DNA regulatory regions. The presence or absence of chromatin binding cofactors may determine whether FOXF1 functions as a transcriptional activator or repressor as both capacities were observed in this study. An important contribution of the current study is the use of a genome-wide approach to identify novel FOXF1 targets in endothelial cells. Analysis of ChIPseq data for which FOXF1 bound to the promoter regions showed a high predilection for genes affecting the cell cycle. Among these genes are positive and negative regulators of the cell cycle, genes essential for formation of the mitotic spindle and cytokinesis, and genes known to promote cell cycle induction. CYCLIN D3 promotes cell cycle progression via interaction with CDK4/658. This study shows that FOXF1 binds Cyclin D3 DNA regulatory regions and decreased expression could contribute to delayed lung regeneration in PDGFb-iCre/Foxf1 fl/+ mice. Our findings are consistent with increased expression of Cdkn1a and Cdkn2b (P15Ink4b) in Foxf1-deficient lungs, both of which inhibit cell cycle progression via the CDK2/4/6 pathway59, 60. Thus decreased FOXF1 in endothelial cells could result in decreased CDK activity, which in turn inhibits endothelial proliferation and delays lung regeneration after PNX.
In summary, this study shows FOXF1 induces lung regeneration via transcriptional regulation of endothelial genes critical for ECM remodeling, cell cycle progression, and cytokinesis. Therefore, small molecule compounds capable of promoting Foxf1 transcription or protecting FOXF1 protein from degradation could provide a useful therapeutic for a variety of pulmonary diseases associated with delayed lung regeneration and repair.
References
Forbes, S. J. & Rosenthal, N. Preparing the ground for tissue regeneration: from mechanism to therapy. Nat Med 20, 857–869, doi:10.1038/nm.3653 (2014).
Akram, K. M., Patel, N., Spiteri, M. A. & Forsyth, N. R. Lung Regeneration: Endogenous and Exogenous Stem Cell Mediated Therapeutic Approaches. Int J Mol Sci 17, doi:10.3390/ijms17010128 (2016).
Ding, B. S. et al. Endothelial-derived angiocrine signals induce and sustain regenerative lung alveolarization. Cell 147, 539–553, doi:10.1016/j.cell.2011.10.003 (2011).
Thane, K., Ingenito, E. P. & Hoffman, A. M. Lung regeneration and translational implications of the postpneumonectomy model. Transl Res 163, 363–376, doi:10.1016/j.trsl.2013.11.010 (2014).
Aitola, M., Carlsson, P., Mahlapuu, M., Enerback, S. & Pelto-Huikko, M. Forkhead transcription factor FoxF2 is expressed in mesodermal tissues involved in epithelio-mesenchymal interactions. Developmental dynamics: an official publication of the American Association of Anatomists 218, 136–149, doi:10.1002/(SICI)1097-0177 (2000).
Kalinichenko, V. V., Lim, L., Shin, B. & Costa, R. H. Differential expression of forkhead box transcription factors following butylated hydroxytoluene lung injury. Am J Physiol Lung Cell Mol Physiol 280, L695–704 (2001).
Mahlapuu, M., Enerback, S. & Carlsson, P. Haploinsufficiency of the forkhead gene Foxf1, a target for sonic hedgehog signaling, causes lung and foregut malformations. Development 128, 2397–2406 (2001).
Mahlapuu, M., Ormestad, M., Enerback, S. & Carlsson, P. The forkhead transcription factor Foxf1 is required for differentiation of extra-embryonic and lateral plate mesoderm. Development 128, 155–166 (2001).
Stankiewicz, P. et al. Genomic and genic deletions of the FOX gene cluster on 16q24.1 and inactivating mutations of FOXF1 cause alveolar capillary dysplasia and other malformations. American journal of human genetics 84, 780–791, doi:10.1016/j.ajhg.2009.05.005 (2009).
Ren, X. et al. FOXF1 transcription factor is required for formation of embryonic vasculature by regulating VEGF signaling in endothelial cells. Circ Res 115, 709–720, doi:10.1161/CIRCRESAHA.115.304382 (2014).
Cai, Y. et al. FOXF1 maintains endothelial barrier function and prevents edema after lung injury. Sci Signal 9, ra40, doi:10.1126/scisignal.aad1899 (2016).
Kalin, T. V. et al. Pulmonary mastocytosis and enhanced lung inflammation in mice heterozygous null for the Foxf1 gene. American journal of respiratory cell and molecular biology 39, 390–399, doi:10.1165/rcmb.2008-0044OC (2008).
Kalinichenko, V. V. et al. Wild-type levels of the mouse Forkhead Box f1 gene are essential for lung repair. American journal of physiology. Lung cellular and molecular physiology 282, L1253–1265, doi:10.1152/ajplung.00463.2001 (2002).
Kalinichenko, V. V. et al. Foxf1+/− mice exhibit defective stellate cell activation and abnormal liver regeneration following CCl4 injury. Hepatology 37, 107–117, doi:10.1053/jhep.2003.50005 (2003).
Claxton, S. et al. Efficient, inducible Cre-recombinase activation in vascular endothelium. Genesis 46, 74–80, doi:10.1002/dvg.20367 (2008).
Le Cras, T. D., Fernandez, L. G., Pastura, P. A. & Laubach, V. E. Vascular growth and remodeling in compensatory lung growth following right lobectomy. J Appl Physiol (1985) 98, 1140–1148, doi:10.1152/japplphysiol.00479.2004 (2005).
Xia, H. et al. Foxm1 regulates resolution of hyperoxic lung injury in newborns. American journal of respiratory cell and molecular biology 52, 611–621, doi:10.1165/rcmb.2014-0091OC (2015).
Bolte, C., Newman, G. & Schultz Jel, J. Kappa and delta opioid receptor signaling is augmented in the failing heart. J Mol Cell Cardiol 47, 493–503, doi:10.1016/j.yjmcc.2009.06.016 (2009).
Bolte, C., Newman, G. & Schultz Jel, J. Hypertensive state, independent of hypertrophy, exhibits an attenuated decrease in systolic function on cardiac kappa-opioid receptor stimulation. Am J Physiol Heart Circ Physiol 296, H967–975, doi:10.1152/ajpheart.00909.2008 (2009).
Bolte, C. et al. Expression of Foxm1 transcription factor in cardiomyocytes is required for myocardial development. PloS one 6, e22217, doi:10.1371/journal.pone.0022217 (2011).
Bolte, C. et al. Forkhead box F2 regulation of platelet-derived growth factor and myocardin/serum response factor signaling is essential for intestinal development. J Biol Chem 290, 7563–7575, doi:10.1074/jbc.M114.609487 (2015).
Bolte, C. et al. Postnatal ablation of Foxm1 from cardiomyocytes causes late onset cardiac hypertrophy and fibrosis without exacerbating pressure overload-induced cardiac remodeling. PLoS One 7, e48713, doi:10.1371/journal.pone.0048713 (2012).
Wang, I. C. et al. Increased expression of FoxM1 transcription factor in respiratory epithelium inhibits lung sacculation and causes Clara cell hyperplasia. Dev Biol 347, 301–314, doi:10.1016/j.ydbio.2010.08.027 (2010).
Kalinichenko, V. V., Gusarova, G. A., Shin, B. & Costa, R. H. The forkhead box F1 transcription factor is expressed in brain and head mesenchyme during mouse embryonic development. Gene Expr Patterns 3, 153–158 (2003).
Kim, I. M. et al. The forkhead box m1 transcription factor is essential for embryonic development of pulmonary vasculature. J Biol Chem 280, 22278–22286, doi:10.1074/jbc.M500936200 (2005).
Ustiyan, V. et al. Forkhead box M1 transcriptional factor is required for smooth muscle cells during embryonic development of blood vessels and esophagus. Dev Biol 336, 266–279, doi:10.1016/j.ydbio.2009.10.007 (2009).
Ustiyan, V. et al. Foxm1 transcription factor is critical for proliferation and differentiation of Clara cells during development of conducting airways. Dev Biol 370, 198–212, doi:10.1016/j.ydbio.2012.07.028 (2012).
Wang, I. C. et al. Foxm1 mediates cross talk between Kras/mitogen-activated protein kinase and canonical Wnt pathways during development of respiratory epithelium. Mol Cell Biol 32, 3838–3850, doi:10.1128/MCB.00355-12 (2012).
Kartashov, A. V. & Barski, A. BioWardrobe: an integrated platform for analysis of epigenomics and transcriptomics data. Genome Biol 16, 158, doi:10.1186/s13059-015-0720-3 (2015).
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9, R137, doi:10.1186/gb-2008-9-9-r137 (2008).
Ma, W., Noble, W. S. & Bailey, T. L. Motif-based analysis of large nucleotide data sets using MEME-ChIP. Nat Protoc 9, 1428–1450, doi:10.1038/nprot.2014.083 (2014).
Ren, X. et al. FOXM1 promotes allergen-induced goblet cell metaplasia and pulmonary inflammation. Mol Cell Biol 33, 371–386, doi:10.1128/MCB.00934-12 (2013).
Ren, X. et al. Forkhead box M1 transcription factor is required for macrophage recruitment during liver repair. Mol Cell Biol 30, 5381–5393, doi:10.1128/MCB.00876-10 (2010).
Peterson, R. S. et al. The winged helix transcriptional activator HFH-8 is expressed in the mesoderm of the primitive streak stage of mouse embryos and its cellular derivatives. Mechanisms of development 69, 53–69 (1997).
Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4, 44–57, doi:10.1038/nprot.2008.211 (2009).
Chen, J., Bardes, E. E., Aronow, B. J. & Jegga, A. G. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res 37, W305–311, doi:10.1093/nar/gkp427 (2009).
Malin, D. et al. Forkhead box F1 is essential for migration of mesenchymal cells and directly induces integrin-beta3 expression. Molecular and cellular biology 27, 2486–2498, doi:10.1128/MCB.01736-06 (2007).
Al-Dosari, M. S., Shaheen, R., Colak, D. & Alkuraya, F. S. Novel CENPJ mutation causes Seckel syndrome. J Med Genet 47, 411–414, doi:10.1136/jmg.2009.076646 (2010).
Bond, J. et al. A centrosomal mechanism involving CDK5RAP2 and CENPJ controls brain size. Nat Genet 37, 353–355, doi:10.1038/ng1539 (2005).
Cushion, T. D. et al. De novo mutations in the beta-tubulin gene TUBB2A cause simplified gyral patterning and infantile-onset epilepsy. Am J Hum Genet 94, 634–641, doi:10.1016/j.ajhg.2014.03.009 (2014).
Lakshmaiah, K. C., Jacob, L. A., Aparna, S., Lokanatha, D. & Saldanha, S. C. Epigenetic therapy of cancer with histone deacetylase inhibitors. J Cancer Res Ther 10, 469–478, doi:10.4103/0973-1482.137937 (2014).
Griffioen, A. W. et al. CD44 is involved in tumor angiogenesis; an activation antigen on human endothelial cells. Blood 90, 1150–1159 (1997).
Trochon, V. et al. Evidence of involvement of CD44 in endothelial cell proliferation, migration and angiogenesis in vitro. Int J Cancer 66, 664–668, doi:10.1002/(SICI)1097-0215 (1996).
Perl, A. K., Hokuto, I., Impagnatiello, M. A., Christofori, G. & Whitsett, J. A. Temporal effects of Sprouty on lung morphogenesis. Dev Biol 258, 154–168 (2003).
Lee, S. et al. Deficiency of endothelium-specific transcription factor Sox17 induces intracranial aneurysm. Circulation 131, 995–1005, doi:10.1161/CIRCULATIONAHA.114.012568 (2015).
Barski, A. et al. High-resolution profiling of histone methylations in the human genome. Cell 129, 823–837, doi:10.1016/j.cell.2007.05.009 (2007).
Nakamura, T. et al. ALL-1 is a histone methyltransferase that assembles a supercomplex of proteins involved in transcriptional regulation. Mol Cell 10, 1119–1128 (2002).
Sedkov, Y. et al. Methylation at lysine 4 of histone H3 in ecdysone-dependent development of Drosophila. Nature 426, 78–83, doi:10.1038/nature02080 (2003).
Wang, H. et al. Purification and functional characterization of a histone H3-lysine 4-specific methyltransferase. Mol Cell 8, 1207–1217 (2001).
Cao, R. et al. Role of histone H3 lysine 27 methylation in Polycomb-group silencing. Science 298, 1039–1043, doi:10.1126/science.1076997 (2002).
Will, H., Atkinson, S. J., Butler, G. S., Smith, B. & Murphy, G. The soluble catalytic domain of membrane type 1 matrix metalloproteinase cleaves the propeptide of progelatinase A and initiates autoproteolytic activation. Regulation by TIMP-2 and TIMP-3. J Biol Chem 271, 17119–17123 (1996).
Kalinichenko, V. V. et al. Defects in pulmonary vasculature and perinatal lung hemorrhage in mice heterozygous null for the Forkhead Box f1 transcription factor. Developmental biology 235, 489–506, doi:10.1006/dbio.2001.0322 (2001).
Lim, L., Kalinichenko, V. V., Whitsett, J. A. & Costa, R. H. Fusion of lung lobes and vessels in mouse embryos heterozygous for the forkhead box f1 targeted allele. Am J Physiol Lung Cell Mol Physiol 282, L1012–1022, doi:10.1152/ajplung.00371.2001 (2002).
Ormestad, M. et al. Foxf1 and Foxf2 control murine gut development by limiting mesenchymal Wnt signaling and promoting extracellular matrix production. Development 133, 833–843, doi:10.1242/dev.02252 (2006).
Du, W. et al. ADAMTS9 is a functional tumor suppressor through inhibiting AKT/mTOR pathway and associated with poor survival in gastric cancer. Oncogene 32, 3319–3328, doi:10.1038/onc.2012.359 (2013).
Jungers, K. A., Le Goff, C., Somerville, R. P. & Apte, S. S. Adamts9 is widely expressed during mouse embryo development. Gene Expr Patterns 5, 609–617, doi:10.1016/j.modgep.2005.03.004 (2005).
Milewski, D. et al. FoxF1 and FoxF2 transcription factors synergistically promote rhabdomyosarcoma carcinogenesis by repressing transcription of p21Cip1 CDK inhibitor. Oncogene, doi:10.1038/onc.2016.254 (2016).
Han, L. P., Fu, T., Lin, Y., Miao, J. L. & Jiang, Q. F. MicroRNA-138 negatively regulates non-small cell lung cancer cells through the interaction with cyclin D3. Tumour Biol, doi:10.1007/s13277-015-3757-8 (2015).
Chen, Y. H. et al. High glucose decreases endothelial cell proliferation via the extracellular signal regulated kinase/p15(INK4b) pathway. Arch Biochem Biophys 465, 164–171, doi:10.1016/j.abb.2007.05.010 (2007).
Staversky, R. J. et al. Normal remodeling of the oxygen-injured lung requires the cyclin-dependent kinase inhibitor p21(Cip1/WAF1/Sdi1). Am J Pathol 161, 1383–1393, doi:10.1016/S0002-9440(10)64414-8 (2002).
Acknowledgements
We would like to thank the following: Yufang Zheng for generation of PDGFb-iCre/Foxf1 fl/+ mice and animal husbandry, Yuqi Cai for assistance with FACS sorting, Arun Pradhan for assistance with Western blot, Tatiana Tomley for assistance with immunostaining, and Logan Fulford for RNAseq analysis. We would like to acknowledge the following funding sources: NIH Grants HL84151 (V.V.K.), HL123490 (V.V.K.) and HL132849 (T.V.K.)
Author information
Authors and Affiliations
Contributions
Conception and design: C.B. and V.V.K.; analysis and interpretation: C.B., H.M.F., X.R., J.S., A.B., T.V.K., V.V.K.; drafting the manuscript for important intellectual content: C.B. and V.V.K.
Corresponding authors
Ethics declarations
Competing Interests
The authors declare that they have no competing interests.
Additional information
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Bolte, C., Flood, H.M., Ren, X. et al. FOXF1 transcription factor promotes lung regeneration after partial pneumonectomy. Sci Rep 7, 10690 (2017). https://doi.org/10.1038/s41598-017-11175-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-017-11175-3
- Springer Nature Limited
This article is cited by
-
Identification of endothelial and mesenchymal FOXF1 enhancers involved in alveolar capillary dysplasia
Nature Communications (2024)
-
FOXF1 promotes tumor vessel normalization and prevents lung cancer progression through FZD4
EMBO Molecular Medicine (2024)
-
Lung endothelial cells regulate pulmonary fibrosis through FOXF1/R-Ras signaling
Nature Communications (2023)
-
Transcriptome profile of spleen tissues from locally-adapted Kenyan pigs (Sus scrofa) experimentally infected with three varying doses of a highly virulent African swine fever virus genotype IX isolate: Ken12/busia.1 (ken-1033)
BMC Genomics (2022)
-
Endothelial progenitor cells stimulate neonatal lung angiogenesis through FOXF1-mediated activation of BMP9/ACVRL1 signaling
Nature Communications (2022)