Published ahead of print on February 15, 2007, doi:10.1164/rccm.200601-050OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200601-050OC
Decreased Lung Fibroblast Growth Factor 18 and Elastin in Human Congenital Diaphragmatic Hernia and Animal Models1 INSERM, U841, Institut Mondor de Médecine Moléculaire, and 2 Faculté de Médecine, IFR10, Université Paris 12, Créteil, France; 3 Maternité and 4 Service de Fœtopathologie, Faculté de Médecine, AP-HP, Université Paris-Descartes, Hôpital Necker-Enfants Malades, Paris, France; and 5 Division of Neonatology, Department of Pediatrics, University of Alberta, Edmonton, Alberta, Canada Correspondence and requests for reprints should be addressed to Jacques Bourbon, Ph.D., Inserm U841, Université Paris 12, Faculté de Médecine, 8 rue du Général Sarrail, 94010 Créteil, France. E-mail: jacques.bourbon{at}creteil.inserm.fr
Rationale: Lung hypoplasia in congenital diaphragmatic hernia (CDH) seems to involve impaired alveolar septation. We hypothesized that disturbed deposition of elastin and expression of fibroblast growth factor 18 (FGF18), an elastogenesis stimulus, occurs in CDH. Objectives: To document FGF18 and elastin in human CDH and ovine surgical and rat nitrofen models and to use models to evaluate the benefit of treatments. Methods: Human CDH and control lungs were collected post mortem. Diaphragmatic hernia was created in sheep at 85 days; fetal lungs were collected at 139 days (term = 145 days). Pregnant rats received nitrofen at 12 days; fetal lungs were collected at 21 days (term = 22 days). Some of the sheep fetuses with hernia underwent tracheal occlusion (TO); some of the nitrofen-treated pregnant rats received vitamin A. Both treatments are known to promote lung growth. Measurements and Main Results: Coincidental with the onset of secondary septation, FGF18 protein increased threefold in control human lungs, which failed to occur in CDH. FGF18 labeling was found in interstitial cells of septa. Elastin staining demonstrated poor septation and markedly decreased elastin density in CDH lungs. Consistently, lung FGF18 transcripts were diminished 60 and 83% by CDH in sheep and rats, respectively, and elastin density and expression were diminished. TO and vitamin A restored FGF18 and elastin expression in sheep and rats, respectively. TO restored elastin density. Conclusions: Impaired septation in CDH is associated with decreased FGF18 expression and elastic fiber deposition. Simultaneous correction of FGF18 and elastin defects by TO and vitamin A suggests that defective elastogenesis may result, at least partly, from FGF18 deficiency.
Key Words: alveolarization • tracheal occlusion • nitrofen • vitamin A
Congenital diaphragmatic hernia (CDH) is a developmental abnormality that is associated with high mortality and morbidity because of respiratory insufficiency due to lung hypoplasia and pulmonary hypertension (1). The incidence of CDH is about 1 out of 3,000 live births. Despite changing concepts and methodology in treatment (2), the mortality rate remains high (3, 4). CDH lungs present fewer and smaller airspaces, reduced radial alveolar count, and thicker alveolar septa (14). This results in part from early impairment of airway branching (6), each bronchiolar end giving rise to a limited number of saccules. Changes in key control factors involved in branching morphogenesis have been consistently reported. The sonic hedgehog system was lowered at early stages in CDH and peaked later in humans and in a rat model of CDH induced by the herbicide nitrofen (7). Fibroblast growth factor (FGF) 10 was decreased in the nitrofen model (8). The expression of FGF7, known to control alveolar epithelial cell proliferation and differentiation, was decreased in the nitrofen model (8) and in a model of surgically induced CDH in sheep (9). Morphogenesis of distal lung, including alveolar septation, also seems to be impaired in CDH. This disorder is a common feature of hypoplastic lungs. Thus, previous studies in human pulmonary hypoplasia of various origins, including hydrops fetalis, renal anomalies, oligohydramnios, and CDH, have established retarded acinar complexity and maturation (10, 11). Elastic fiber deposition, which is essential to build alveolar walls (16–18), was reported to be disturbed in human lung hypoplasia in association with oligohydramnios (15–17) or CDH (17) and experimentally in drainage-induced lung hypoplasia in fetal sheep (18). Moreover, in the ovine model, alveolar hypoplasia occurred in the absence of reduction in bronchiolar generations due to late creation of hernia (19), and discontinuous, uncondensed elastin aggregates have been described in alveolar septa (20). Decreased elastin expression with less elastin deposition and disorganized distribution have also been reported in the nitrofen model (21). Compared with branching morphogenesis, less is known about mechanisms that control saccular and alveolar development. FGF18 is believed to play important role. Thus, lung-targeted FGF18 overexpression inhibited distal lung development (22), whereas FGF18-null mouse fetuses displayed smaller distal airspaces and thickened septa (23). FGF18 expression markedly increases coincidently with postnatal formation of secondary septa in the rat (24), and FGF18 enhances proliferation and elastogenesis in myofibroblasts (24), the source of septal elastin. Moreover, FGFR3 is a high-affinity receptor for FGF18 (25), and alveolarization is abolished in mice devoid of FGFR3 and FGFR4 (26). FGF18 has not been documented in the developing human lung in normal or pathologic conditions. The potential benefit of two treatments aiming to restore lung development in CDH, tracheal occlusion (TO) and vitamin A administration, have been investigated in the surgical ovine model and in the rat nitrofen model, respectively. In sheep fetuses with hernia and lung hypoplasia induced by drainage, TO restored lung growth, increased gas exchange surface area (27–29), and ameliorated respiratory function at delivery (30, 31). This treatment is under trial in human fetuses with CDH (32). In the nitrofen model, vitamin A decreased the incidence and severity of CDH, enhanced lung growth, and restored lung maturation (33, 34). The first objective of the present study was to investigate whether CDH affected FGF18 expression in the developing human lung and in models. Elastic fiber deposits and elastin expression were studied in parallel to further document qualitative and quantitative changes. The second objective consisted of using the sheep and rat models of CDH to evaluate the effects of TO and vitamin A treatment, respectively, on pulmonary FGF18 expression and elastin deposition. Some of the results of these studies have been previously reported in the form of an abstract (35).
Human Lung Tissue Human lung samples were collected during the autopsy after medical terminations of pregnancy in bad-prognosis fetuses or after death postdelivery. Parents were informed about the procedure and issues of post mortem study, and signed consent was obtained for all included patients. The study received approval from the local Ethics Committee. Detailed clinical data are depicted in Table 1.
Sheep Model of CDH and TO Surgical procedures have been extensively described elsewhere (36). Biological samples were collected from the same animals as in previous reports (20, 27).
Nitrofen Exposure in Rats
Histochemical Elastin Staining and Quantification
Immunohistochemical FGF18 Analyses
RNA Extraction
Determination of Ovine Partial cDNA Sequence for FGF18
Northern Blot Analysis
Reverse Transcription and Real-Time Quantitative Polymerase Chain Reaction
Western Blot Analysis Membranes were exposed to goat anti-rhFGF18 antibody (R&D Systems, Lille, France) diluted 1:500, washed in Tween-20 Tris-buffered Saline (TTBS), and exposed to horseradish peroxidase–conjugated donkey anti-goat IgG antibody (Santa Cruz Biotechnology, Santa Cruz, CA) and incubated in enhanced chemiluminescence (ECL) reagent (Amersham Biosciences) before exposure to Kodak BioMax MS film. Signals were quantified by densitometry (NIH image).
Statistical Analysis
FGF18 in Human Lungs with CDH FGF18 protein was studied in human lung samples by Western blot analysis. Preservation of RNA was not constant enough for us to perform a study at the pretranslational level. Because FGF18 had never been documented in developing human lung, a first step consisted in studying changes in FGF18 level during the course of intrauterine development. The FGF18 proportion increased with progressing pregnancy between 14 and 37 weeks (fetal age) with a particularly marked rise around 28 weeks (Figure 1A). Densitometry analysis of data corrected for variations of protein loading (Figure 1B) indicated significant positive exponential correlation with time (r = 0.865; p < 0.001). The highest amount, observed between 32 and 37 weeks (i.e., in early alveolar stage), reached about 20 times those observed at 14 to 16 weeks (pseudoglandular stage) and about 10 times those observed at 19 to 21 weeks (canalicular stage). When densitometric values were gathered into two groups corresponding to presaccular stages (  26 wk; n = 6) and saccular-alveolar stages (age  27 wk; n = 8), the mean FGF18 amount increased from 165 ± 69 arbitrary units (a.u.) in the former to 782 ± 138.5 a.u. in the latter (p < 0.01 by t test). Six pairs of CDH and age-matched control lungs ranging from 27 to 37 weeks (saccular to alveolar stages) were then studied comparatively (Figure 2A). All CDH lungs ipsilateral to hernia displayed lower FGF18 levels than their respective age-matched controls, and FGF18 failed to increase in CDH samples over the period, whereas control values increased by about three times (Figure 2B).
Elastin in Human Lungs with CDH Elastin staining was performed in five pairs of control and CDH lungs ranging from 27 to 37 weeks. In control fetuses, lung parenchyma matured homogeneously during the period, with thinning of septal walls, increased proportion of airspaces, and a surge of secondary septa (Figures 3A and 3C). Airspaces displayed regular distribution. Elastin staining demonstrated bundles beneath the surface of walls and punctate, dense deposits at the tips of growing septa that considerably increased in number with maturation (Figures 3A' and 3C'). Consistent with previous reports, CDH lungs displayed thicker walls and denser tissue (Figures 3B and 3D). Elastin staining also demonstrated mostly bundles in thick septa with an extreme paucity of tip deposits (Figures 3B' and 3D'), which illustrates deficient secondary septation. Lung vessels strongly stained for elastin and presented no noticeable difference between CDH and control lungs. Quantitative evaluation of parenchymal elastic fiber deposits was achieved by determination of the proportion of tissue surface area occupied by elastin patches, excluding vessel and airway elastin (Table 3). The elastic fiber density was low at 27 and 29 weeks, with no difference between CDH and control lungs. It was considerably higher at 31, 33, and 37 weeks in control subjects but failed to increase in CDH lungs, which displayed much lower values than their respective age-matched control subjects.
FGF18 in Sheep Lungs with Surgical Diaphragmatic Hernia and Surgical Diaphragmatic Hernia plus TO The experimental groups were surgical diaphragmatic hernia (sDH) created at 85 days (n = 6; term = 145 days), sDH + TO (n = 4), and control group (n = 5), all retrieved at 139 days of gestation. The expression level of FGF18 in sheep lungs was determined at pre- and post-translational levels. Because the ovine FGF18 cDNA sequence had not been reported previously, a first step consisted of determining it after RT-PCR amplification. Sheep lung mRNAs were retrotranscribed and amplified using two oligonucleotide primers chosen in the human and mouse/rat cDNA sequences in regions proximate to 5' and 3' ends and conserved among species described thus far. This allowed amplification of a 531-bp product representing a partial sequence of ovine FGF18 corresponding to positions 49 through 579 of the 621-nucleotide coding sequences of human, mouse, and rat FGF18 (Figure 4A). Its predicted amino-acid sequence shared 99% homology with human, mouse, and rat FGF18 as shown by sequence alignment (Figure 4B). The ovine cDNA sequence then served to design oligonucleotide primers for use in real-time RT-PCR to evaluate FGF18 expression level among the various experimental groups. We found a 60% reduction of FGF18 transcript in the lung ipsilateral to hernia in the sDH group compared with the control group (Figure 5A). There was a trend toward a decrease of FGF18 mRNA level in contralateral lung as well, but the difference was not statistically significant. FGF18 mRNA level was increased to twice the control level in ipsilateral and contralateral lungs in the sDH+TO group (Figure 5A). FGF18 was also appraised at the posttranslational level through Western blot analysis in ipsilateral lung. FGF-18 antibody raised against the human peptide recognized ovine FGF18 at the same apparent molecular weight (Figure 5B). Densitometric analysis normalized for gel loading indicated values of 193.4 and 596.5, 23.3 and 70.1, and 1,381.0 and 327.5 a.u. in control subjects, sDH, and sDH+TO, respectively. Consistent with mRNA findings, FGF18 protein was therefore decreased to very low level in sDH compared with control subjects and reestablished in sDH+TO. Because FGFR3-FGFR4 double null mutation abolished alveolar septation in the mouse (26) and because FGFR3 is a putative receptor of FGF18 (25), expression of the transcript of FGFR3 was studied in parallel. By contrast with FGF18, FGFR3 expression was unaffected by sDH, and although it tended to increase after TO, the difference was not statistically significant (Figure 5C).
Elastin in Sheep Lungs with sDH and sDH+TO Elastin staining on sections from the lung ipsilateral to hernia showed changes in elastin deposition similar to those in human CDH. Control lungs displayed regularly distributed airspaces with thin parenchymal tissue and dense focal elastic fiber deposits localized primary at the tips of newly forming septa and lining alveolar walls (Figures 6A and 6A'). All sDH lungs appeared immature with thicker septa. Some presented regularly distributed airspaces with evenly distributed elastin fibers, but, similar to human CDH lungs, secondary crests with typical punctate elastin deposit at the apex were rare (Figure 6B). Other subjects displayed zones of dilated airspaces among large zones of dense and thick parenchyma, with elastin deposits being confined to dilated airspaces and extremely scarce in dense areas (Figure 6C). The paucity of elastin staining in dense areas and its presence at the tip of septa in dilated ones indicate that the presence of dilated and nondilated zones is not an artifactual consequence of absence of pressure fixation, but rather is a feature of these lungs. In fetuses with sDH+TO, alveolarization seemed to be restored as indicated by the reappearance of regular elastin lining and punctate elastin deposits located at the tip of crests, which presented an appearance similar to that observed in controls (Figures 6D and 6D'). Elastin in walls of blood vessels and airways did not seem to be altered in the sDH and sDH+TO groups. Quantification of parenchymal elastic fiber density indicated dramatically reduced amounts in sDH lungs that were only one seventh those in control lungs; TO restored the proportion of elastin deposits to a level not significantly different from that in control lungs (Table 4). In addition, tropoelastin mRNA level was decreased about half in ipsilateral lung in the sDH group but was unchanged in contralateral lung, and TO enhanced the level to twice that in controls in ipsilateral and contralateral lungs (Figure 7), thus abolishing the DH effect.
Expression of Tropoelastin and FGF18 in Rat Lungs with Induced CDH and CDH + Vitamin A Nitrofen treatment resulted in lung hypoplasia in all fetuses and was associated with a 60 to 70% incidence of right-sided CDH. Lung wet weights were 111.1 ± 1.0, 85.5 ± 1.9, and 67.7 ± 1.6 mg in control fetuses (given olive oil, the vehicle of nitrofen), nitrofen-treated fetuses without CDH, and nitrofen-treated fetuses with CDH, respectively (p < 0.001 vs. controls for both nitrofen-treated groups). Vitamin A increased lung growth without fully restoring the control level in nitrofen-treated fetuses (97.8 ± 3.0 mg without CDH and 76.0 ± 1.9 mg with CDH, p < 0.05 as compared with nitrofen-treated fetuses without vitamin A). Tropoelastin and FGF18 mRNAs were deeply decreased to about 13 and 17% of the control levels, respectively, in fetuses with nitrofen-induced CDH as compared with controls (Figure 8). Similar decreases were observed in nitrofen-treated fetuses without CDH (not shown), suggesting an association with nitrofen-induced lung hypoplasia in the presence or absence of hernia. Vitamin A administration 2 days after nitrofen treatment prevented the drop of tropoelastin and FGF18 mRNAs that displayed levels not significantly different from those in controls (Figure 8).
FGF18 Immunolocalization in Lung Tissues This investigation was performed in control lung tissues of the three species to define cell localization of FGF18. Figure 9 shows results obtained with the same anti-FGF18 antibody in distal lung tissue. Immunoreactivity appeared in parenchymal cells as a dotted labeling. Figure 9A depicts immunofluorescence micrograph of 37-week fetal human lung tissue, indicating the presence of FGF18 in septal cells. Epithelial airway cells were slightly labeled, whereas smooth muscle cells of small airways or arteries were negative. In bronchial cartilaginous plates, chondrocytes, which are known to express FGF18 (37), were also labeled (not shown). In 139-day fetal sheep lung tissue (Figure 9C), labeling was similarly found in the septa. Weak labeling was present in airway epithelial cells (not shown) but was absent from airway and vascular smooth muscle cells. In the rat, we studied lung tissue on postnatal Day 4 (Figure 9E), when FGF18 expression has been reported to be elevated (24). FGF18 immunoreactivity was present in cells in the thickness of walls of primary and secondary septa (left and right inserts in Figure 9E, respectively). Distribution of FGF18 labeling seemed to be similar to that of elastin or  -smooth muscle actin (Figure 9G).
We report that decreases of FGF18 expression and of elastic fiber deposition in alveolar septa characterize CDH in humans. The presence of similar features in two animal models allowed us to use the latter to evaluate the effects of potential therapeutic approaches on these abnormalities. Tracheal occlusion in the surgical model and vitamin A treatment in the nitrofen model allowed nearly normal features to be recovered. The work supports further studies to determine the potential benefit of these treatments in promoting lung growth and maturation in the presence of diaphragmatic hernia.
Limitations of the Study
Lung Expression of FGF18 Is Deficient in Diaphragmatic Hernia The expression pattern of FGF18 in the developing human lung had not been examined previously. A first step therefore consisted of studying FGF18 expression in the course of fetal lung development. A major observation of the present study was the marked increase in FGF18 protein, starting from 27 to 28 weeks of gestation and reaching an elevated level at 36 to 37 weeks. Although the precise time when secondary alveolar septation begins in humans is a matter of debate, in part because of the difficulty in defining an alveolus in microscopic sections (12), it is generally accepted that the process starts between 30 and 36 weeks of gestation. Therefore, FGF18 increases in human lung coincidently with starting secondary septation. This finding, consistent with studies in rodents, reinforces the assumption of involvement of FGF18 in alveologenesis. Moreover, FGF18 localization in interstitial cells of alveolar walls, presumably myofibroblasts, at the sites and time of alveolarization and elastin deposition strongly argues in favor of FGF18 involvement in the process. Previous investigations had consistently indicated that FGF18 expression was located principally in interstitial tissue of distal lung areas in the mouse (22). Furthermore, FGF18 absence from perivascular or periairway wall tissue where elastin is also abundant is in favor of specific involvement for alveolarization at this stage of lung development. In a second step, human CDH lungs were compared with age-matched control lungs. We found a low FGF18 protein level in CDH lungs, with failure to increase in late pregnancy. Consistently, lowered FGF18 expression was found also in the ovine sDH model and in the nitrofen model in rats. The absence of change in FGFR3 expression in sDH indicates impairment of signaling at the ligand level, not at the receptor level. Secondary septation was begun in humans and advanced in sheep at stages when FGF18 was determined in CDH. Impaired expression is therefore likely to be related to the impairment of this process. In the rat, the process of secondary septation was not initiated at the stage when the study was performed. Nevertheless, elastin and FGF18 transcripts were decreased in the nitrofen model, which suggests disturbance in the prenatal formation of saccular walls. In rat lung, FGF18 presents two developmental peaks: a twofold prenatal increase between Fetal Days 19 and 21 and a sevenfold increase between postnatal Days 2 and 3, separated by a transient fall at the time of birth (24). The changes reported here correspond to the inhibition of the first rise that may therefore be related to the building of saccular walls.
Elastic Fiber Density Is Diminished in Diaphragmatic Hernia Decreased density of elastin fibers in the distal parenchyma of CDH lungs is a novel finding of the present investigation. It indicates that defective septation results from deficient elastin deposition. Decreased tropoelastin transcripts in the fetal sheep model indicated that impairment occurs at the pretranslational level. Reduced lung expansion induced by lung fluid drainage has been reported to decrease tropoelastin mRNA 2.5-fold (38). Contradictory observations have been reported in the nitrofen model with decreased (Reference 21 and present data) or increased (39) tropoelastin transcripts. The reason for the discrepancy between studies is unclear but may be due to methodologic differences (21). Moreover, reduced levels of tropoelastin transcripts in the nitrofen model were corroborated by reduction of desmosine content (indicative of cross-linked elastin) in the lung ipsilateral to hernia (21). Although no quantitative evaluation of elastin synthesis was performed in oligohydramnios, the absence of elastin deposits in alveolar septa reported at the ultrastuctural level in hypoplastic lungs with this disease (16) also suggests defective synthesis. Impaired elastogenesis in pulmonary septa therefore presents as a common feature in underexpanded hypoplastic lungs, whatever the leading cause, and may therefore represent a direct consequence of insufficient lung–tissue tension. Although the present investigation does not demonstrate a causal relationship between FGF18 changes and impairment or restoration of alveolarization, developmental lung disturbances in prenatal mice lacking FGF18 (23) and the coordinated effects of FGF18 upon various proteins involved in elastogenesis by neonatal rat lung myofibroblasts (24) strongly argue in favor of such a link. The presence of FGF18 transcripts (22) and immunoreactive FGF18 (present data) in distal lung parenchyma reinforces this hypothesis.
Treatments Restore FGF18 Expression and Elastin Deposition in CDH Models The second therapeutic approach, which consisted of the use of vitamin A in the nitrofen model, is based on the importance of retinoids, including vitamin A and its active metabolites, in the alveolarization process (47). Several studies support the hypothesis that abnormalities within the retinoid-signaling pathway contribute to the etiology of CDH (48). Moreover, decreased plasma retinol and retinol-binding protein levels have been reported in human newborns with CDH (49). Restoration of FGF18 and tropoelastin expressions by vitamin A in the lung of nitrofen-treated rat fetuses is in keeping with our previous finding that FGF18 and tropoelastin expression were up-regulated subsequently to vitamin A administration to normal rat neonates (24). This suggests that the beneficial effect of vitamin A for pulmonary hypoplasia (33) might have resulted, among possible changes in other growth factors, from promotion of FGF18 expression. However, in this model, lung hypoplasia and immaturity also occur in pups that do not develop CDH. Although less marked, the morphology of lungs of fetuses without hernia was reported to be similar to that of CDH lungs, including a paucity of elastic fiber deposits in septa (21). In agreement with this observation, we found reduced FGF18 and tropoelastin expression in nitrofen-treated fetuses devoid of hernia. Disorders could therefore result from pulmonary toxic effect(s) of nitrofen independently of CDH. Thyroid transcription factor-1, which is essential to lung morphogenesis, was down-regulated by nitrofen in fetal rat lungs independently of the presence of CDH (50) and in a time- and dose-dependent manner in cultured lung epithelial H-441 cells (51). Nitrofen is also believed to interfere with vitamin A signaling (52, 53). Therefore, it cannot be excluded that vitamin A supplementation counteracted some pulmonary effects of nitrofen, including FGF18 and elastin changes, that were not direct consequences of hernia. Considering the use of vitamin A as a possible treatment of lung abnormalities in CDH therefore requires further evaluation of its benefits and safety. In conclusion, changes in FGF18 induced by CDH and treatments are novel and significant findings in the present study. Simultaneous correction of FGF18 and elastin defects by TO and vitamin A suggests that disordered alveolarization may result, at least in part, from FGF18 deficiency.
Supported by a Legs Poix Grant from the Chancellerie des Universités de Paris. This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org Originally Published in Press as DOI: 10.1164/rccm.200601-050OC on February 15, 2007 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form January 13, 2006; accepted in final form February 12, 2007
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
2, vascular endothelial growth factor, insulin-like growth factor-1, and insulin-like growth factor-2 (
