The Forkhead Box (Fox) family of transcription factors plays important roles in regulating expression of genes involved in cellular proliferation and differentiation. In a previous study, we showed that newbornfoxf1(+/−) mice with diminished Foxf1 levels exhibited abnormal formation of pulmonary alveoli and capillaries and died postnatally. Interestingly, surviving newborn foxf1(+/−) mice exhibited increased pulmonary Foxf1 levels and normal adult lung morphology, suggesting that wild-type Foxf1 levels are required for lung development and function. The present study was conducted to determine whether adult foxf1(+/−) mice were able to undergo lung repair similar to that observed in wild-type mice. We demonstrated that adult foxf1(+/−) mice died from severe lung hemorrhage after butylated hydroxytoluene (BHT) lung injury and that this phenotype was associated with a 10-fold decrease in pulmonary Foxf1 expression and increased alveolar endothelial cell apoptosis that disrupted capillary integrity. Furthermore, BHT-induced lung hemorrhage of adult foxf1(+/−) mice was associated with a drastic reduction in expression of the Flk-1, bone morphogenetic protein-4, surfactant protein B, platelet endothelial cell adhesion molecule, and vascular endothelial cadherin genes, whereas the expression of these genes was either transiently diminished or increased in wild-type lungs after BHT injury. Because these proteins are critical for lung morphogenesis and endothelial homeostasis, their decreased mRNA levels are likely contributing to BHT-induced lung hemorrhage in foxf1(+/−) mice. Collectively, our data suggest that sustained expression of Foxf1 is essential for normal lung repair and endothelial cell survival in response to pulmonary cell injury.
- winged helix deoxyribonucleic acid-binding domain
- bone morphogenetic protein-4
- surfactant protein B
- platelet endothelial cell adhesion molecule-1
- vascular endothelial cadherin
- butylated hydroxytoluene lung injury
- endothelial cells
the forkhead box (Fox) family of transcription factors (20) shares homology in the winged helix DNA-binding domain (6), and its members play important roles in regulating transcription of genes involved in cellular proliferation, differentiation, and metabolic homeostasis (9, 10, 19, 24, 25,37, 42, 48). One of these family members, Foxf1 (also known as HFH-8 or Freac-1), initiates expression during gastrulation in a subset of mesodermal cells arising from the primitive streak region that contributes to the extraembryonic mesoderm and lateral mesoderm (38). Consistent with this early expression pattern,foxf1(−/−) embryos die in utero from defects in extraembryonic and lateral mesoderm differentiation (29). During organogenesis, high levels of Foxf1 expression persist in the mesenchyme of the respiratory and gastrointestinal tracts. Foxf1 RNA is expressed at mesenchymal-epithelial interfaces involved in lung and gut morphogenesis (30, 38). In the adult mouse, Foxf1 RNA is detected in smooth muscle layers of pulmonary bronchioles, lamina propria of the stomach and the intestine, and in alveolar endothelial cells (22, 38).
A study conducted in our laboratory demonstrated that ∼55% of newborn foxf1(+/−) mice died of severe lung hemorrhage (21), and a study conducted in another laboratory confirmed perinatal lethality offoxf1(+/−) mice using a different genetic background (28). The lethal phenotype was associated withfoxf1(+/−) lungs displaying an 80% reduction in Foxf1 mRNA levels compared with wild-type lungs [designated as lowfoxf1(+/−) mice; see Ref. 21]. Lowfoxf1(+/−) newborn mice exhibited defects in formation of pulmonary capillaries and alveoli with disruption of the mesenchymal-epithelial cell interfaces and increased apoptosis in the alveolar and bronchiolar regions of the lung parenchyma (21). Furthermore, lung hemorrhage was associated with reduced expression of the vascular endothelial growth factor (VEGF) receptor 2 (Flk-1), platelet endothelial cell adhesion molecule-1 (PECAM-1), surfactant protein B (SP-B), and bone morphogenetic protein-4 (BMP-4) as well as the lung Kruppel-like factor and T-box (Tbx2-Tbx5) transcription factors. Interestingly, expression of these genes was unchanged in 40% of the newborn foxf1(+/−) mice that possessed wild-type pulmonary levels of Foxf1 mRNA [high foxf1(+/−) mice] but exhibited diminished alveolar septation without pulmonary hemorrhage (21). Moreover, the high foxf1(+/−) mice had normal life spans and adult lung morphology, suggesting that compensation for developmental defects in septation had occurred.
Butylated hydroxytoluene (BHT)-mediated lung injury is characterized initially by extensive damage to the distal lung epithelial and endothelial cells and subsequently by cellular proliferation between 3 and 7 days after BHT administration (1, 31). The cellular repair process after BHT lung injury is associated with a transient 65% reduction in pulmonary Foxf1 mRNA levels between 4 and 6 days after BHT injury and a return to normal levels by day 8(22). In this study, we used BHT lung injury to examine whether adult foxf1(+/−) mice are capable of normal lung repair in response to cellular damage. Within 7 days after BHT lung injury, all foxf1(+/−) mice died from severe lung hemorrhage. BHT-mediated induction of foxf1(+/−) pulmonary hemorrhage was associated with a 10-fold decrease in pulmonary Foxf1 expression and increased apoptosis of alveolar endothelial cells that were positive for Foxf1, PECAM-1, and CD34 expression. Furthermore, the severe foxf1(+/−) pulmonary phenotype after BHT injury was associated with reduced expression of Flk-1, vascular endothelial (VE) cadherin, BMP-4, SP-B, and PECAM-1, all of which are required for lung morphogenesis and endothelial cell homeostasis.
MATERIALS AND METHODS
foxf1(+/−) mice and BHT treatment.
The generation of mice heterozygous for a targeted deletion of the Foxf1 locus has been described previously, and foxf1(+/−) mice were maintained in the 129/black Swiss mouse background (21). The winged helix DNA-binding domain of Foxf1 was replaced by an in-frame insertion of a nuclear localizing β-galactosidase (β-Gal) gene, disrupting function of the mouse gene in vivo (21). Expression of the β-Gal gene was therefore under the control of Foxf1 regulatory sequences, thus allowing the use of β-Gal enzyme staining for visualizing the expression pattern of Foxf1. Tail tissue samples were used to prepare genomic DNA for genotyping of the newborn mice by PCR analysis, as described previously (21).
BHT (3,5-di-t-butyl-4-hydroxytoluene; Sigma, St. Louis, MO) was dissolved in corn oil (Mazola) at a concentration 30 mg/ml, and a single intraperitoneal injection of BHT (300 mg/kg body wt) was given to foxf1(+/−) mice or their wild-type littermates (22). To determine statistical significance of any observed differences, we used three mice per time point after BHT administration, which included 2, 3, 4, 6, and 8 days. The mice were killed by CO2 asphyxiation, and lung tissue was weighed and used to prepare total RNA, or lungs were inflated with 4% paraformaldehyde (PFA), fixed overnight in 4% PFA at 4°C, and then paraffin embedded as described previously (21, 22).
Immunohistochemical staining, TdT-UTP nick end-labeling apoptosis assay and transmission electron microscopy.
Sections of paraffin-embedded lung tissue were stained with hematoxylin and eosin or used for immunohistochemical staining with the following rat monoclonal antibodies: PECAM-1 (clone MEC 13.3) and CD34 (clone RAM34) both from Pharmingen (San Diego, CA). Antibody-antigen complexes were detected using biotinylated horse anti-rat antibody and avidin-alkaline phosphatase complex with 5-bromo-4-chloro-3-indolyl phosphate (BCIP)/nitro blue tetrazolium (NBT) substrate and then were counterstained with nuclear fast red (Vector Laboratories, Burlingame, CA) as described previously (21). We also used mouse monoclonal antibodies against α-smooth muscle actin (clone 1A4; Sigma) and proliferation cell nuclear antigen (PCNA, clone PC10; Roche Molecular Biochemicals, Indianapolis, IN) and detected the antibody-antigen complex using horse anti-mouse antibody conjugated with alkaline phosphatase (Vector Laboratories), as described previously (21). To measure apoptosis in wild-type and foxf1(+/−) mice at various intervals after BHT lung injury (0, 4, and 6 days), TdT-UTP nick end-labeling (TUNEL) assay was performed using an in situ cell death detection kit from Roche Diagnostic according to the manufacturer's recommendations. Apoptotic cells were visualized using BCIP/NBT substrate for alkaline phosphatase and then were counterstained with nuclear fast red (21). We counted apoptotic or PCNA-positive cells from five ×200 lung viewing fields for each of three mice to determine the mean and used this to calculate the mean number of apoptotic cells (±SD) for three mice.
Adult lung tissue was isolated from foxf1(+/−) or wild-type mice at 4 days after BHT lung injury, PFA fixed, and prepared for transmission electron microscopy (TEM), as described previously (39). Lung tissue sections were photographed using a Jeol (Peabody, MA) JEM 1210 transmission electron microscope at 80 or 60 kV on electron microscope film (ESTAR thick base; Kodak, Rochester, NY) and printed on photographic paper.
β-Gal enzyme staining.
Staining for expression of β-Gal was performed according to Clevidence et al. (7), with a few modifications. Briefly, dissected lungs were fixed for 1 h in PBS (pH 7.8) containing 2% formaldehyde-0.2% glutaraldehyde, 0.02% Nonidet P-40 (NP-40), and 0.01% sodium deoxycholate. Lung tissue was then stained for β-Gal enzyme activity in a PBS solution containing 5 mM potassium ferrocyanide, 5 mM ferricyanide, 2 mM magnesium chloride, 0.02% NP-40, 0.01% sodium deoxycholate, and 1 mg/ml 5-bromo-4-chloro-3-indolyl-β-d-galactoside for 4–6 h at 37°C. After β-Gal staining, samples were rinsed three times in PBS, postfixed overnight in 4% PFA, and then paraffin embedded (21). For colocalization studies, β-Gal-stained sections were treated with xylene to remove paraffin wax, rehydrated, and used for TUNEL assay to measure apoptosis.
RNA extraction and RNase protection assay.
Total mouse lung RNA was prepared by an acid guanidium-thiocyanate-phenol-chloroform extraction method using RNA-STAT-60 (Tel-Test “B,” Friendswood, TX). RNase protection assay was performed with [32P]UTP-labeled antisense RNA synthesized from plasmid templates with the appropriate RNA polymerase, as described previously (8). RNA probe hybridization, RNase One (Promega, Madison, WI) digestion, electrophoresis of RNA-protected fragments, and autoradiography were performed as described previously (22, 39, 41, 49). Quantitation of expression levels was determined from scanned X-ray films by using the BioMax 1D program (Kodak). The cyclophilin hybridization signal was used for normalization control between different lung RNA samples. Synthesis of antisense mouse Foxf1, Flk-1, PECAM-1, VEGF, BMP-4, FoxM1B, cyclophilin, and rat SP-B RNA probes was described previously (21, 22, 41). Antisense RNA probes for VE cadherin, heme oxygenase 1, and vimentin were generated from Atlas cDNA plasmids purchased from Clontech (Palo Alto, CA). The expression of bcl-w and interleukin (IL)-6 genes was determined using mouse Apo-2 and mCK-2 templates from Pharmingen. The Foxf2 RNase protection probe was generated by PCR amplification of mouse genomic DNA using the following primers: 5′-gcggaattccctgacctcaagcagccg and 5′-gcgggattccagccttggcgctcttta. The resulting Foxf2 genomic PCR fragment was cloned into BlueScript plasmid (Stratagene) as anEcoRI and BamHI fragment, and T3 RNA polymerase was used to synthesize Foxf2 antisense RNA probe from aClaI-digested template.
BHT lung injury induces pulmonary hemorrhage in foxf1(+/−) mice.
Hematoxylin and eosin staining of adult lung sections showed that thefoxf1(+/−) pulmonary morphology was indistinguishable from the wild type (Fig. 1, A andB). We therefore used BHT lung injury to examine whether adult foxf1(+/−) mice displayed a normal lung repair process. Adult foxf1(+/−) mice and their wild-type littermates were injected with BHT, and their lungs were harvested at intervals after BHT injury. Although wild-type mice exhibited only a single mortality after BHT lung injury, none of the 16foxf1(+/−) mice survived >7 days after BHT administration (Fig. 2 A). By 3 days after BHT injury, the lung weight-to-body weight ratio was significantly increased in foxf1(+/−) mice compared with wild-type controls (Fig. 2 B), suggesting an accumulation of extracellular fluid in foxf1(+/−) lung tissue. At 4 days post-BHT lung injury, both wild-type and foxf1(+/−) lungs displayed visible lung damage with alveolar infiltration of inflammatory cells, but hemorrhage was seen only in lungs fromfoxf1(+/−) mice (Fig. 1, C–E). By day 6 after BHT injury, histological improvement was noted in the lungs of wild-type mice (Fig. 1 F), whereasfoxf1(+/−) lungs exhibited severe pulmonary hemorrhage (Fig. 1, G and H). These results suggest that thefoxf1(+/−) mice died from severe pulmonary hemorrhage induced by BHT injury.
foxf1(+/−) pulmonary hemorrhage is associated with significant decreases in pulmonary expression of Foxf1 mRNA.
We recently reported that a subset of newborn foxf1(+/−) mice exhibited lethal pulmonary hemorrhage, which correlated with diminished pulmonary expression of Foxf1 mRNA (21). The surviving foxf1(+/−) mice displayed wild-type pulmonary levels of Foxf1 expression (Fig. 3,A–C), suggesting that normal Foxf1 levels are necessary for appropriate lung morphogenesis and function (21). To determine whether BHT-induced lung hemorrhage was associated with decreased Foxf1 expression, RNase protection assays were performed to measure pulmonary Foxf1 mRNA levels. At 2 days after BHT lung injury,foxf1(+/−) mice displayed a more precipitous 70% decline in pulmonary Foxf1 mRNA than the 20% reduction noted in wild-type lungs (Fig. 3, A–C). Both foxf1(+/−) and wild-type mice displayed a similar reduction in pulmonary Foxf1 mRNA between days 3 and 4 post-BHT lung injury (Fig.3, A–C). Although wild-type mice exhibited increased pulmonary expression of Foxf1 by day 6 after BHT lung injury, foxf1(+/−) lungs exhibited further decreases in Foxf1 mRNA (Fig. 3, B and C). This 90% reduction in pulmonary Foxf1 mRNA was associated with severe lung hemorrhage and mortality in foxf1(+/−) mice at 6 days after BHT lung injury.
Because of the similarities between Foxf1 and Foxf2 (Freac-2, Lun) proteins in expression pattern and the winged helix DNA-binding sequence (2, 16, 17, 30, 33, 38), we next examined Foxf2 mRNA levels in BHT-injured lungs (Fig. 3 D). Wild-type lungs exhibited a transient 60% reduction in Foxf2 mRNA between 2 and 4 days after BHT injury, and its expression recovered by day 6(Fig. 3, A and D). As was the case with Foxf1 expression, BHT-injured foxf1(+/−) lungs exhibited a further decrease in Foxf2 mRNA at day 6 (Fig. 3,A–D). This result suggests that, infoxf1(+/−) lungs, BHT did not induce Foxf2 mRNA to compensate for the precipitous decline in Foxf1 expression.
Because the targeted Foxf1 allele possesses an in-frame insertion of the nuclear localizing β-Gal gene, staining for β-Gal enzyme activity allows the identification of Foxf1-expressing cells (21,22). We next stained foxf1(+/−) lungs for β-Gal enzyme activity at different intervals after BHT injury to determine whether the lung damage caused a reduction in the number of Foxf1-expressing cells (Fig. 4,A–D). In untreated foxf1(+/−) lung tissue, β-Gal-positive cells represent ∼52% of the total cells in the lung parenchyma (Fig. 3 E). However, the foxf1(+/−) alveolar region exhibited a 50% decrease in β-Gal-positive cells by 6 days after BHT treatment (Figs. 3 E and 4,A–D). In contrast, no decrease in β-Gal-positive peribronchiolar smooth muscle cells was noted after BHT lung injury (Fig. 4, E–F). Consistent with this result, BHT-injured foxf1(+/−) lungs display unchanged expression of α-smooth muscle actin in the peribronchiolar smooth muscle cells (Fig. 4, I–J), suggesting that these Foxf1-expressing cells were not severely affected by BHT lung injury.
BHT-injured foxf1(+/−) lungs exhibit an increase in apoptosis and persistent proliferation.
Because BHT injury of foxf1(+/−) lungs caused a decline in alveolar Foxf1-expressing endothelial cells, we next used TUNEL assay to examine whether the foxf1(+/−) lungs displayed an increase in apoptosis. Although BHT injury induced visible apoptosis in wild-type lungs (Figs.5 A and6, A, C, andE), a fivefold increase in the number of apoptotic cells was observed throughout the alveolar region of foxf1(+/−) lungs (Figs. 5 A and 6, B, D, andF). Colocalization studies showed that the majority of BHT-induced apoptotic cells were the β-Gal-positive (Foxf1-expressing) endothelial cells of the pulmonary alveolar region (Fig. 4, K–L). These results suggest that increased alveolar endothelial cell apoptosis may contribute to the observed lung hemorrhage in BHT-treated foxf1(+/−) mice. However, the TUNEL assay cannot rule out the possibility that necrosis also occurred in BHT-injured foxf1(+/−) lungs and contributed to the foxf1(+/−) phenotype. Immunohistochemical staining with PCNA antibody demonstrated that wild-type and foxf1(+/−) lungs exhibited equivalent PCNA staining at 4 days post-BHT treatment (Fig. 5 B). Although PCNA staining had subsided in wild-type lungs by day 6 after BHT injury, proliferation remained elevated in foxf1(+/−) lungs (Fig. 5 B), perhaps in response to increased alveolar apoptosis (Fig. 5 A). In support of this concept, BHT-injured foxf1(+/−) lungs also displayed a 60–80% reduction in the expression of the bcl-2 cell survival family member bcl-w (Fig. 5 C). These results correspond with recently reported data describing a correlation between apoptosis and differential regulation of bcl-2 family genes after hyperoxia-induced lung injury (13, 35). Normal induction of the proliferation-specific FoxM1B transcription factor was found in thefoxf1(+/−) lungs after BHT injury (22, 41,47), supporting the notion that the global proliferative response was not impaired in foxf1(+/−) lungs (Fig.5 C). These results suggest that the decline in alveolar Foxf1-expressing cells in BHT-injured foxf1(+/−) lungs was the result of increased apoptosis and not diminished cellular proliferation.
We next used TEM to examine the interface between the alveolar endothelial cells and type I epithelial cells in wild-type andfoxf1(+/−) lungs at 4 days after BHT lung injury (Fig.7). Although BHT injury of both wild-type and foxf1(+/−) lungs caused edema between the two cell types (Fig. 7, A and B), onlyfoxf1(+/−) lungs displayed extensive apoptosis of alveolar endothelial cells and disruption of the capillary wall (Fig.7, C and D). These TEM studies confirm that BHT injury of foxf1(+/−) lungs leads to increased apoptosis of alveolar endothelial cells, causing capillary damage. These results suggest that Foxf1 plays an important role in mediating lung morphogenesis and alveolar endothelial cell survival during cellular repair in response to pulmonary injury.
Decreased alveolar PECAM-1 and CD34 staining of endothelial cells in BHT-injured foxf1(+/−) lungs.
Previous studies demonstrated that Foxf1 expression colocalizes with PECAM-1-positive alveolar endothelial cells (22). We therefore used immunohistochemical staining to assess the pulmonary expression of PECAM-1 in wild-type andfoxf1(+/−) lungs after BHT injury. Before BHT injury, equivalent alveolar PECAM-1 staining was found infoxf1(+/−) and wild-type mice (Fig.8, A and B). Although PECAM-1 staining did not change in BHT-injured wild-type lung (Fig. 8, C, E, and G), drastic reductions in alveolar PECAM-1 staining were observed withfoxf1(+/−) lungs at 4 and 6 days after BHT injury (Fig. 8,D, F, and H). Consistent with this finding, we observed a diminished number of CD34-positive alveolar endothelial cells in foxf1(+/−) lungs at 6 days after BHT lung injury (Fig. 8, I–J). In contrast, large pulmonary vessels, which do not express Foxf1, displayed normal PECAM-1 staining in response to BHT injury (Fig. 8, D–E), suggesting that the decline in PECAM-1 expression was restricted to alveolar capillaries and arterioles. Moreover, RNase protection assays demonstrated that, before BHT injury, foxf1(+/−) lungs displayed a twofold increase in PECAM-1 mRNA compared with wild-type lungs (Fig. 9). Although BHT injury offoxf1(+/−) lungs caused an 80% reduction in PECAM-1 expression at 6 days, only minor decreases were observed in wild-type lungs (Fig. 9). Furthermore, PECAM-1 has been implicated in vascular permeability (34) and in protection of endothelial cells against apoptosis (11, 18), suggesting that decreased PECAM-1 expression in foxf1(+/−) lungs may contribute to an increase in BHT-mediated apoptosis of alveolar endothelial cells (Figs. 5 and 6).
BHT-induced pulmonary hemorrhage of foxf1(+/−) mice is associated with diminished pulmonary expression of Flk-1, BMP-4, VE cadherin, and SP-B.
To identify genes whose altered expression was associated with lung hemorrhage after BHT injury, RNase protection assays were performed in duplicate. Examination of Flk-1, BMP-4, PECAM-1, and VE cadherin mRNA levels in adult foxf1(+/−) and wild-type lungs before BHT injury demonstrated that their expression was increased in heterozygous lungs (Fig. 9). Because these proteins are critical for pulmonary mesenchymal cell homeostasis and function (5, 11, 12, 18, 27, 34), their compensatory increased expression may play a role in survival of foxf1(+/−) mice. Furthermore, we found that VEGF expression was not significantly affected in either wild-type or foxf1(+/−) mice after BHT injury, but an 18-fold decline in its receptor Flk-1 mRNA was observed by day 6 in foxf1(+/−) lungs (Fig. 9). Because Flk-1 mediates endothelial cell proliferation and survival (5), its reduced expression in foxf1(+/−) lungs after BHT injury correlates with a diminished number of PECAM-1-positive endothelial cells. Interestingly, BHT injury of wild-type lungs caused increased expression of BMP-4, which is consistent with their roles in mediating lung branching morphogenesis and proliferation of mesenchymal cells that may be involved in the repair process (4, 45). In contrast, expression of BMP-4 mRNA was decreased in foxf1(+/−) lungs at day 6after BHT injury, and their reduced levels are associated with defective lung injury repair. Furthermore, foxf1(+/−) lungs displayed a sudden reduction in VE cadherin and SP-B mRNA levels at 6 days after BHT injury (Fig. 9), a finding that may explain increased permeability in the foxf1(+/−) lungs (12,40). The foxf1(+/−) lungs also exhibited a drastic increase in IL-6 at day 6 after BHT injury, perhaps consistent with the role of IL-6 in protection from hyperoxic lung injury (44). Interestingly, wild-type andfoxf1(+/−) lungs displayed similar induced expression of heme oxygenase-1 (Fig. 9), an enzyme that is implicated in protection of endothelial cells against acute lung injury (36, 46). Finally, we did not observe significant changes in mRNA levels of vimentin (Fig. 9) and lung Kruppel-like factor (data not shown), the latter of which is a transcription factor required for normal lung development (43).
In a previous study, we reported that ∼55% of newbornfoxf1(+/−) mice died of severe lung hemorrhage and that the severity of their pulmonary abnormalities correlated with an 80% reduction in Foxf1 mRNA levels [designated as lowfoxf1(+/−); see Ref. 21]. Defects in formation of lung alveoli and capillaries were observed in lowfoxf1(+/−) newborn mice, suggesting that wild-type levels of Foxf1 are required for normal lung development (21). The pulmonary hemorrhage was associated with reduced expression of SP-B, VEGF receptor 2 (Flk-1), BMP-4, and lung Kruppel-like factor and T-box transcription factors, which are critical for lung morphogenesis and function. Interestingly, wild-type pulmonary levels of these regulatory genes and Foxf1 mRNA were observed in a subset of newbornfoxf1(+/−) mice that did not exhibit lung hemorrhage [designated as high foxf1(+/−) mice]. Although highfoxf1(+/−) mice had normal life spans and their lungs appear morphologically normal, the fact that BHT injury was sufficient to induce a lethal pulmonary hemorrhage suggests thatfoxf1(+/−) lungs had limited capacity to respond to cellular damage. Consistent with a critical role of wild-type Foxf1 levels for lung morphogenesis (21), BHT-inducedfoxf1(+/−) lung hemorrhage was associated with a 10-fold decline in pulmonary Foxf1 mRNA without compensatory increases in expression of a related family member, Foxf2. The BHT-induced decrease in pulmonary Foxf1 levels in foxf1(+/−) mice was associated with significant decreases in expression of Flk-1, PECAM-1, VE cadherin, BMP-4, SP-B, and bcl-w genes, all of which are critical for lung morphogenesis and alveolar endothelial cell survival required for normal lung repair. These results suggest that Foxf1 plays an important role in regulating the transcriptional network of genes involved in mediating lung morphogenesis and repair in response to pulmonary injury.
We used TEM to confirm that BHT-injured foxf1(+/−) lungs possessed apoptotic alveolar endothelial cells and had damage to capillaries, both of which contribute to the severe pulmonary hemorrhage. We found that BHT injury of foxf1(+/−) lungs caused alveolar endothelial cells to express undetectable levels of PECAM-1 protein, which is required for protection against apoptosis (11, 18), and increased vascular permeability (34). Our results suggest that diminished PECAM-1 expression may contribute to elevated apoptosis of alveolar endothelial cells in BHT-injured foxf1(+/−) lungs. Interestingly, we found that BHT lung injury did not cause apoptosis of Foxf1-expressing bronchiolar smooth muscle cells, suggesting that apoptosis of this cell type does not contribute to pulmonary hemorrhage. This result is in contrast to the newborn lowfoxf1(+/−) mice, which displayed increased apoptosis of the smooth muscle cells and disruption of its interface with the bronchiolar epithelial cell layer (21). This discrepancy in damage of the smooth muscle cell layer offoxf1(+/−) lungs may reflect the fact that BHT injury is restricted to the pulmonary epithelial and endothelial cells. We also found that PCNA staining and expression of the proliferation-specific FoxM1B gene remained elevated in foxf1(+/−) lungs at 6 days after BHT injury, suggesting that sustained proliferation occurs in response to an increase in alveolar apoptosis. Together, our data suggest that elevated apoptosis of alveolar endothelial cells in BHT-injured foxf1(+/−) lungs causes increased capillary permeability, contributing to a severe lung hemorrhage phenotype.
Interestingly, adult foxf1(+/−) lungs express wild-type levels of Foxf1 mRNA and exhibit increased expression of Flk-1, BMP-4, PECAM-1, and VE cadherin. Because these proteins are critical for pulmonary mesenchymal cell homeostasis and function (4, 5, 11,12, 18, 34, 45), their elevated expression may compensate for initial defects in pulmonary septation and may contribute to survival of Foxf1 heterozygous mice. BHT-induced lung hemorrhage of adultfoxf1(+/−) mice was associated with a drastic reduction in expression of the Flk-1, PECAM-1, and VE cadherin genes, whereas the expression of these genes was only slightly diminished in wild-type lungs after BHT injury. Furthermore, their large reduction infoxf1(+/−) lungs at 6 days after BHT injury (Fig. 9) cannot be solely explained by the loss of endothelial cells because the injured foxf1(+/−) lungs exhibited only a 50% reduction in β-Gal-positive alveolar endothelial cells (Fig. 3 E). Interestingly, potential Foxf1-binding sites (38) were found in distinct enhancer regions that are required for correct endothelial expression of the mouse Flk-1 (CATTGTTTATGgA) and VE cadherin (−635 bp, CAGTATTTGTAAA) promoters in transgenic mice (14, 23) as well as in the human PECAM-1 promoter (−48 bp, GAGTGTTTACTCt) region (3). The finding that potential Foxf1-binding sites are found in the functional Flk-1, VE cadherin, and PECAM-1 promoter regions suggests that Foxf1 may be directly regulating their expression. These proteins are also known to be involved in endothelial cell survival, vascular morphogenesis, and maintenance of capillary integrity (5, 11, 12, 15, 18, 34), suggesting that their decreased expression is likely contributing to lung hemorrhage. Recent studies have implicated endothelial cells as critical for the morphogenesis of liver and pancreas, suggesting that they secrete growth factors or provide cell-cell contacts essential for organ development (26, 32). It is tempting to speculate that the reduction in alveolar endothelial cells after BHT injury offoxf1(+/−) lungs may also contribute to the loss of growth factors critical for lung morphogenesis and repair in response to cellular damage.
We also found that BHT injury of wild-type lungs caused increased levels of BMP-4, which is consistent with their roles in mediating lung development (4, 45), and their increased expression is recapitulated during cellular repair in response to lung injury. In contrast, BHT-injured foxf1(+/−) lungs displayed an opposite response with decreased pulmonary expression of BMP-4 mRNA, which was associated with defective lung injury repair and subsequent development of pulmonary hemorrhage. Our current BHT injury studies and previous studies in newborn foxf1(+/−) mice (21) demonstrate that diminished pulmonary Foxf1 levels correlate with reduced expression of BMP-4. The finding that the BMP-4 promoter contains binding sites for Foxf1 (38) and significant reduction in mesodermal expression of BMP-4 was found infoxf1(−/−) embryos provides further support for this concept (29). Moreover, although expression of SP-B was elevated during the first 4 days after BHT injury, its mRNA levels decreased in foxf1(+/−) mice at day 6, coinciding with lung hemorrhage (Fig. 9). Exposure ofSP-B(+/−) mice to hyperoxia caused increased severity of pulmonary edema and hemorrhage (40), indicating that diminished SP-B levels compromise its protective role in the lung. These findings are consistent with the concept that wild-type levels of Foxf1 are required to regulate genes necessary for pulmonary morphogenesis and endothelial cell homeostasis during cellular proliferation in response to lung injury.
In summary, we demonstrated that adult foxf1(+/−) mice were highly susceptible to BHT-induced lung injury. BHT injury caused lethal pulmonary hemorrhage in adult foxf1(+/−) mice and was associated with a 10-fold decrease in pulmonary Foxf1 expression and an increase in apoptosis of alveolar endothelial cells. Furthermore, BHT injury of foxf1(+/−) lungs caused an aberrant reduction in levels of Flk-1, PECAM-1, VE cadherin, BMP-4, SP-B, and bcl-w mRNA, which are critical for cell survival and lung morphogenesis during repair after pulmonary injury. Collectively, our data suggest that Foxf1 is a potential target for therapeutic intervention during a variety of pulmonary diseases involving defective lung repair.
We thank Fengli Guo, Ana Bursick, Mara Sullivan, and Jean Clark for excellent technical assistance and Pradip Raychaudhuri for critically reviewing the manuscript.
This work was supported by National Institutes of Health Grants HL-62446 (R. H. Costa), HL-56387 (J. A. Whitsett), HL-41496 (J. A. Whitsett), and CA-76541 (D. B. Stolz).
Address for reprint requests and other correspondence: R. H. Costa, Dept. of Molecular Genetics (M/C 669), Univ. of Illinois at Chicago, College of Medicine, 900 S. Ashland Ave., Rm. 2220 MBRB, Chicago, IL 60607-7170 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
First published January 11, 2002;10.1152/ajplung.00463.2001
- Copyright © 2002 the American Physiological Society