INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Congenital diaphragmatic hernia (CDH), the incidence of which is 1 in 3,000 births (1), has dramatic consequences for fetal lung development. It leads to severe lung hypoplasia as a result of ascension of abdominal viscera into the rib cage and subsequent lung compression. Despite progress in pre- and postnatal management, the mortality rate remains 60% (2). The possibility of managing these fetuses with intrauterine surgery has recently been explored, but technical difficulties and tocolysis problems have been responsible for the limited success in humans (3, 4). An alternative approach has been based on the observation that tracheal occlusion increases lung growth in utero, apparently as a consequence of lung fluid retention. The fetal lung epithelium effectively secretes fluid which is released into the amniotic space. This fluid exerts a set pressure in the future air spaces, due to the resistance of upper airways. This pressure appears to represent an important determinant of growth and development of the fetal lung (5). Experimental data (6) from studies using animal models of diaphragmatic hernia (DH) have shown that fetal tracheal ligation forces viscera to reenter the abdominal cavity and allows the fetal lung to grow in models of CDH. Hedrick and coworkers (6) called this technique the PLUG method for "plug the lung until it grows." Whereas reversible occlusion is an absolute requirement in human fetuses, reversible approach has not been previously attempted in these models. Furthermore, no comprehensive study of lung growth and maturation has been made in lungs that have initially experienced hypoplasia as a consequence of DH, and subsequently expansion, due to tracheal occlusion. Despite this lack of experimental evaluation in animal models, some clinical attempts have been made with poor success (9). We have recently developed a fetal lamb model of DH with subsequent tracheal occlusion with the aid of a releasable-extractible latex balloon (10). The aim of the present study was to further document pulmonary growth and to evaluate fetal lung maturity in this model.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Design

Three groups of lamb fetuses (full term = 145 d) were studied. The first group, designated the diaphragmatic hernia (DH) group (n = 5), was obtained by creating DH in utero at 85 d of gestation. The second group, designated the DH + tracheal occlusion (TO) group (n = 5), was obtained by creating DH at 85 d as in the first group, then TO at 120 d by the introduction of a balloon via a noninvasive endoscopic technique. The third group, designated the control group (n = 6), consisted of unoperated fetuses, derived from twins of DH or DH + TO fetuses. Two fetuses had a spontaneous closure of DH and were used as sham-operated controls (shamDH). In three fetuses with DH + TO, the balloon deflated, no fluid retention occurred and no sign of tracheal lesion was observed; these fetuses were used as controls for TO procedure (shamTO). All fetuses were retrieved by cesarean section at 139 d of gestation and killed for histological and biochemical lung analyses.

Fetal Surgical Procedures

At 85 d of gestation, maternal and fetal anesthesia was obtained with thiopental, 10 mg/kg, and maintained with 1% halothane in O₂/N₂O (50:50 vol/vol). The left-sided DH was created at 85 d of gestation via hysterotomy, using a modification of the fetal sheep model of DH previously described by Soper and coworkers (11).

At 120 d of gestation, the ewes underwent a second laparotomy under the same anesthetic procedure. Fetal tracheal occlusion was performed using a double-channel endoscope (Karl Storz, Tuttlingen, Switzerland) allowing irrigation and insertion of a latex detachable balloon with its catheter (Nycomed Laboratory, Paris, France). The technique of hernia creation and placement of the plug are described in detail elsewhere (10).

Delivery and Lung Preparation

At 139 d of gestation all fetuses were retrieved by cesarean section and killed by immediate injection of a lethal dose of thiopental + KCl in the umbilical vein before sectioning the umbilical cord, to prevent air breathing. Each lamb was then weighed and the chest opened through a median sternotomy. The lungs were excised and weighed. Tissue samples from the same lobe of each lung were immediately frozen in liquid nitrogen for biochemical analyses. Samples of all lobes were fixed in 4% paraformaldehyde for immunocytochemical and histological analyses.

Histological Studies

Specimens were fixed for 24 h, embedded in paraffin, sectioned, and stained with hematoxylin and eosin. The radial alveolar count was estimated according to Emery and Mithal (12) and the architecture of the pulmonary acinus was examined.

Biochemical Analyses

Samples were homogenized in water. DNA, proteins, glycogen, and phospholipids were quantified from aliquot parts of the same homogenate. DNA content was determined by the diphenylamine method of Burton (13). Protein content was assayed by the method of Bradford (14). Glycogen content was determined by the method of Chan and Exton (15). Lipids were extracted from homogenates by chloroform:methanol, 2:1 (vol/vol). Individual phospholipids were separated by thin layer chromatography according to Serrano de la Cruz and coworkers (16) and disaturated phosphatidylcholine (DSPC) was separated by osmium tetroxide and thin layer chromatography according to Patterson and coworkers (17).

Surfactant (lamellar body fraction) was extracted from homogenized tissue by Frosolono's method as modified by Farrell and coworkers (18). Individual phospholipids were determined in surfactant by the same methods as for whole tissue extracts. For tissue and surfactant phospholipids, total lung content was calculated from the average concentration value of right and left lung, then referred to fetal body weight.

Fluorescent Immunohistochemical Study

An antisurfactant protein B (SP-B) antiserum was used to label lung type II cells (19) (or type II pneumocytes), the surfactant-producing cells. The antiserum, prepared by immunization of rabbits against purified bovine SP-B (20), was kindly provided by Dr. J. A. Whitsett (Cincinnati, OH).

After being fixed in 4% paraformaldehyde for 3 h, samples were incubated in 1 M sucrose overnight, frozen and freeze-cut at 5 µm, then stored at 20° C until labeling. Sections were saturated with phosphate-buffered saline (PBS) containing 1% bovine serum albumin and incubated with the SP-B antibody (1/1,000) for 2 h. A 45-min incubation was performed with a Texas red-conjugated anti-rabbit IgG antibody (Amersham France SA, Les Ulis, France). To determine total cell number, nuclei were labeled with the bisbenzimide dye Hoechst 33258 (Sigma, l'Isle d'Abeau, Chesnes, France) at 0.25 µg/ml for 30 min. Each incubation was followed by rinsing with PBS. The sections were examined with a Nikon fluorescence microscope. Countings were performed on photographs of different fields of each sample excluding the large bronchial areas. Up to 3,000 cells have been counted for each sample. Results are expressed as percent of SP-B immunoreactive cells per total lung parenchymal cells. SP-B is produced both by alveolar type II cells and bronchiolar Clara cells (21). Since only parenchymal cells were analyzed, the number of SP-B immunoreactive cells represents principally type II cells.

Statistical Analyses

Data are expressed as mean ± SEM. The groups were compared using unpaired t test with a Statview statistical package (Abacus Concepts, Berkeley, CA). p Values of less than 0.05 were considered statistically significant. In the figures, two groups that are statistically different have no letter in common; two groups that are not statistically different have a letter in common.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung Weight, Proteins, DNA, and Glycogen

Twenty-six ewes underwent surgery. We obtained 21 living fetuses: five DH, five DH + TO, two shamDH, three shamTO, and six control fetuses. In all animals with DH, spleen, stomach, and small bowel were present in the left chest and the lungs were reduced in size. In all fetuses with DH + TO, herniated viscera were reduced from the chest and the lungs were filled with liquid, but their gross anatomy appeared normal. Lung size and lung weight to body weight ratio have been reported previously (10). In brief, as compared with control fetuses, lung weight was unchanged in shamDH, diminished in the DH and shamTO groups, and considerably increased in the DH + TO group but the latter increase was due at least in part to the fluid retained in the occluded lungs. In order to further evaluate lung growth, lung protein and DNA concentrations were determined. In shamDH, proteins and DNA (27.3 ± 1.6 µg/ml and 4.8 ± 0.3 µg/ml, respectively) were identical to those in the control group (Figure 1A, B). Both were increased in the DH group (Figure 1A, B) and in the shamTO group (50.4 ± 4.0 µg/mg and 7.7 ± 0.4 µg/mg, respectively, no significant difference with DH), but decreased in the DH + TO group (Figure 1A, B). The protein to DNA ratio was however significantly higher (+57%) in the DH + TO group than in the control group (Figure 1C). Total DNA content of the whole organ was significantly decreased in DH (399 ± 56 mg) as compared with control fetuses (626 ± 56 mg, p < 0.05); it tended to be increased in DH + TO (489 ± 89 mg, not significant) as compared with DH alone. Lung glycogen concentration was changed neither in the DH group nor in the DH + TO group as compared with the control group (not shown).

View larger version (31K): [in this window] [in a new window]   Figure 1.   Mean values ± SEM for: (A) lung proteins (left: concentration in µg/mg wet lung tissue; right: total content in g/kg body weight), (B) lung DNA (left: concentration in µg/mg wet lung tissue; right, total content in g/kg body weight), and (C ) protein to DNA ratio in control, diaphragmatic hernia (DH) alone, and DH + tracheal occlusion (TO) groups. Two groups that are statistically different have no letter in common. Two groups that are not statistically different have a letter in common. Both protein and DNA content were increased in the DH group and decreased in the DH + TO group, but the protein to DNA ratio was changed in the DH + TO group only.

Morphology

The analysis of the hematoxylin and eosin stained sections showed that the normal lungs at 139 d of gestation were characterized by future air spaces lined by a thin epithelium and subdivided by wall projections (Figure 2A). ShamDH fetuses were not different from control fetuses (not shown). The DH lungs showed thickened alveolar walls, an epithelium comprised of cuboidal cells, an increased amount of interstitial tissue, and numerous capillary vessels (Figure 2B). In the DH + TO lungs, future alveolar walls were thin with very little interstitial tissue (Figure 2C).

View larger version (77K): [in this window] [in a new window]   Figure 2.   Fetal lamb lung morphology at 139 d of gestation in the three experimental groups (hematoxylin-eosin stained sections). (A) Control lung showing thin alveolar septa and minimal amount of interstitial tissue. (B) DH lung showing increased interstitial tissue and decreased alveolar airspace. (C ) DH + TO lung showing an aspect similar to that of normal lungs with dilatation of the airspaces by pulmonary fluid. (Original magnification: ×250.)

No difference was found for the radial alveolar count (RAC) between control lungs, shamDH lungs, and the DH + TO lungs, but all these groups had a significantly higher RAC than the DH and shamTO groups (Figure 3).

View larger version (50K): [in this window] [in a new window]   Figure 3.   Mean values ± SEM for radial alveolar count (RAC) in the three experimental groups. Two groups that are statistically different have no letter in common. Two groups that are not statistically different have a letter in common.

Tissue and Surfactant Phospholipids

Phospholipids were analyzed both in whole lung tissue and in isolated surfactant. Concentrations in both lungs were found to be significantly different neither for tissue phospholipids nor for surfactant phospholipids. Considering whole lung tissue, the distribution of the various tissue phospholipids was not considerably changed by treatments, except for a reduction from 54 to 46% of phosphatidylcholine (PC) which was reflected by a parallel increase in various other phospholipids in the DH + TO group. The amount of PC referred to fetal body weight of the shamDH fetuses (199.4 ± 39.4 µmol/kg BW) was in the range of control group values. By contrast, it was markedly reduced in the DH group (Figure 4) and in the shamTO group (116.1 ± 13.0 µmol/kg BW). The value in the DH + TO group was intermediary between the control and DH groups, suggesting partial compensation, with regard to DSPC, whereas a significant decrease as compared with control fetuses was found in the shamTO group (34.7 ± 3.4 µmol/ kg BW, minus 60%) as well as in the DH and DH + TO groups (minus 44% and minus 38%, respectively; Figure 4).

View larger version (24K): [in this window] [in a new window]   Figure 4.   Mean values ± SEM for total phosphatidylcholine (PC) and disaturated phosphatidylcholine (DSPC) in whole lung tissue in the three experimental groups (µmol/kg body weight). Two groups that are statistically different have no letter in common. Two groups that are not statistically different have a letter in common. DSPC was significantly diminished both in the DH and DH + TO groups.

The pool size of surfactant, as indicated by the amount of total phospholipids of isolated fraction, was not changed in shamDH (198.2 ± 10.5 µmol/kg BW), reduced in DH (Figure 5) and in shamTO (100.5 ± 47.2 µmol/kg BW), and further reduced in DH + TO (Figure 5) as compared with the control group (310.9 ± 91.3 µmol/kg BW). PC and DSPC exhibited the same changes (Figure 5). In the DH + TO group, the total amount of surfactant DSPC was only 29.5 ± 5.6 µmol per lung versus 170.1 ± 46.1 µmol in the control group, 200.3 ± 22.5 µmol in shamDH, 98.3 ± 15.2 µmol in DH, and 80.7 ± 16.1 µmol in shamTO. In other words, surfactant DSPC in the DH + TO group represented only 17.5% of the normal amount versus 58% in the DH group.

View larger version (16K): [in this window] [in a new window]   Figure 5.   Mean values ± SEM for total phospholipids (TPL), total phosphatidylcholine (PC), and disaturated phosphatidylcholine (DSPC) of isolated surfactant fraction in the three experimental groups (µmol/kg body weight). Two groups that are statistically different have no letter in common. Two groups that are not statistically different have a letter in common. Surfactant phospholipids were decreased in the DH group and further decreased in the DH + TO group.

Immunohistochemistry

SP-B immunoreactive cells displayed a bright fluorescence of the cytoplasm after incubation with the anti-SP-B antibody. No labeling was observed after incubation with nonimmune serum (Figure 6A). The normal lungs showed an epithelial lining of the alveoli that consisted of cuboidal cells with fluorescence of the entire cytoplasm interspersed with dark areas probably corresponding to mature type I cells (19) (Figure 6B). In the DH group, despite the compaction of the lungs, the alveoli were also lined with cuboidal cells, but fewer dark zones were present. When referred to total number of parenchymal cells as determined through labeling of nuclei, the number of SP-B-positive cells was higher in DH (27.0 ± 0.2%, Figure 6C) and reduced in DH + TO (7.8 ± 3.6%, Figure 6D) when compared with control fetuses (15.1 ± 0.1%, Figure 6B).

View larger version (104K): [in this window] [in a new window]   Figure 6.   Immunofluorescent detection of SP-B as a marker of type II cells in fetal lamb lung at 139 d of gestation (original magnification: ×100). (A) Negative control lung incubated with nonimmune serum. No fluorescence is visible. (B) Control lung showing alveoli lined with some immunoreactive cells regularly interspersed with dark zones. (C  ) DH lung showing numerous immunoreactive cells. Absence of staining around large vessels and airways (a, pulmonary artery branch; b, bronchiolus) corresponds to interstitial tissue. (D) DH + TO lung: few immunoreactive cells are scattered along the distended alveoli.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aim of this study was to analyze growth and maturation of fetal lungs after occlusion of the trachea in a model of DH. No such evaluation had been carried out in those previous studies in which tracheal ligation had been performed in the presence of experimental DH (6, 8, 22). The present study is therefore the first attempt to document histological and biochemical features of the DH lung after occluding the trachea endoscopically. In addition, we have shown that the endoscopic technique is easier to perform than invasive surgical methods of in utero DH correction, and may avoid the preterm labor complications (10).

For none of the considered parameters (lung weight, lung proteins and DNA, morphological criteria, tissue and surfactant phospholipids), the sham-operated controls (complete surgical procedure with spontaneous closure of diaphragmatic hernia) were found to be different from unoperated control fetuses. These findings confirm the conclusion of Harrison and associates (23) that unoperated twin fetuses are suitable controls in experimental studies of diaphragmatic hernia. Furthermore, the shamTO fetuses which experienced the surgical procedure of balloon insertion, but in which tracheal occlusion failed, were found to be identical to fetuses in the DH group that did not experience this procedure. This clearly shows that changes in the DH + TO group as compared with the DH group are entirely due to tracheal closure and fluid retention.

Lung growth study evidenced hypoplasia in the DH group, consistent with the findings of numerous previous studies (24, 25). The PLUG technique after creation of a DH, is responsible for a considerable enlargement of lung size with a lung weight/body weight ratio higher than normal for the DH + TO group. These lungs, however, were filled with pulmonary fluid. Lung dry weight could not be determined owing to the necessity of drying whole lung to avoid fluid leakage, which would not have been compatible with sectioning of the tissue for the various immunochemical and biochemical determinations. In order to determine whether pulmonary growth was effective and was not only a reflection of expansion, other measurements were necessary. It should be pointed out, however, that in other studies (6, 26) dry lung weight was found to be increased after tracheal ligation.

RAC (which reflects pulmonary growth, especially that of terminal bronchioles) was found to be similar in both the control and the DH + TO groups, whereas it was considerably diminished in the DH group. This indicates that an actual growth and morphogenesis of the tissue was induced by tracheal occlusion. Furthermore, the latter reduced interstitial tissue which is a feature of morphological maturation. The hypoplastic lung growth was therefore partially corrected by tracheal occlusion on a histological point of view. Although alveolar wall thickness has not been measured because arithmetic and harmonic mean septal thickness cannot be measured without inflation fixation (27), it is obvious from morphology and RAC determination that lung growth and maturation in the presence of DH have been improved by tracheal occlusion. It should be pointed out that RAC is reliable in tissue that has not been inflated-fixed (12).

The biochemical study is also very informative in this respect. The total amount of lung DNA was reduced by DH, which means that the number of cells was reduced. The amount of DNA in the DH + TO group is intermediate between the control and DH groups, which suggests that the number of cells increased as compared with the DH group, but without complete compensation to the control level. This partial compensation of lung hypoplasia was illustrated by the abnormally high protein/DNA ratio in the DH + TO group. In other words, the apparently normal amount of protein in the whole organ appears to reflect an increased mean cell protein content. Some deficit in cell number remained in DH + TO fetuses, but this probably represents relative hypertrophy rather than hyperplasia.

Comparison of these results with those of previous studies is difficult, because of differences in experimental design. Di Fiore and colleagues (8) found no difference between the control and DH + tracheal ligation groups for the protein/DNA ratio, but it should be noted that the DH and ligation of the trachea were performed simultaneously. The same is true for the study of Hashim and colleagues (7) who ligated the trachea without creating any diaphragmatic hernia. Possibly, the theory of Liu and coworkers (28, 29) who reported that stretch stimulus applied to a fetal cell with a specific frequency, amplitude, and periodicity is responsible for eliciting a mitogenic response, might be suitable for normal lungs but not to the same extent for damaged tissues of hypoplastic lungs. Tracheal ligation in a model of formerly established diaphragmatic hernia has been performed by Hedrick and coworkers (6). Consistent with our results, they found no difference for the DNA/protein ratio between the DH and DH + TO groups.

The other important question for lung function was the maturational degree of DH + TO lungs, i.e., the presence of an air-blood barrier and the presence of adequate surfactant amount. From the morphological point of view, when comparing the histological sections, DH lungs were found to be immature with an important delay in saccular development. At term, when compared with control lungs, they were made of few alveolar spaces with walls which seemed thicker. These results are similar to those found in previous studies (26, 30). By contrast DH + TO lungs showed similar histological structures to those in control lungs. More significantly, thin walls with large dark areas in the SP-B immunofluorescence study suggest the presence of extended alveolar regions lined with type I pneumocytes. Tracheal occlusion could therefore favor the development of the air-blood barrier. Confirmation of this assumption could only be achieved by ultrastructural observation.

A major criterion of lung maturation is the amount of surfactant. Surfactant is a compound of about 90% lipids and 10% proteins. Among lipids, phospholipids are the major components. We therefore evaluated the amount of surfactant phospholipids and the amount of the principal surface-active phospholipid component, namely DSPC. Phospholipids of whole lung, which gather cell membrane and surfactant phospholipids, were also determined. Our results show a quantitative decrease of tissue PC and DSPC and of surfactant total phospholipids, PC and DSPC, in the DH group. It is generally accepted that DH lungs are surfactant-deficient (25). Although PC and DSPC were increased in whole lung by tracheal occlusion, surfactant levels in the DH + TO lungs were dramatically decreased when compared with the control and even with the DH lungs. Increased whole lung PC and DSPC after tracheal occlusion therefore reflect changes in tissue growth only. Another recent study by O'Toole and coworkers (22) led to a similar conclusion, although both the occlusion technique and surfactant quantification technique (bronchoalveolar lavage approach) were different. The lavage technique quantitates only secreted surfactant, whereas our method allows the total pool of surfactant in the organ to be evaluated. It therefore appears that despite increased PC and DSPC synthesis for lung growth, these changes occur at the expense of surfactant phospholipid synthesis and that surfactant deficits are aggravated by tracheal occlusion.

Surfactant decrease could have resulted either from decreased number of surfactant-producing cells, namely type II cells, or from another unknown mechanism leading to decreased rate of surfactant synthesis. To investigate the cause of surfactant deficiency we tried to identify and count type II cells by immunochemistry. To analyze our results we used Ten Have-Opbroek's histologic criteria: the type II cells are approximately cuboidal cells with a round nucleus and a cytoplasmic staining for surfactant proteins (31). SP-B labeling was therefore used as a feature for characterizing type II cells. Immunofluorescent images of SP-B labeling and total cell labeling allowed us to determine the proportion of SP-B-producing cells. The results indicate that DH lungs are richer in SP-B immunoreactive cells than control lungs, which is consistent with previous assumptions (9, 30, 32). On the contrary, the number of SP-B-positive cells was reduced in the DH + TO group. The reduced surfactant content in the latter therefore appears as a consequence of reduced type II cell number. Although the immunological labeling approach has been used to identify type II cells in previous lung fluid aspiration or tracheal ligation studies (30, 33), no attempt has been made to evaluate their proportion, even when a decreased type II cell number has been assumed (30). In one study reported in abstract form (34), type II cell number and volume have been determined and found to be decreased after tracheal ligation, which is consistent with our findings, but this work did not involve DH. As a whole, tracheal occlusion clearly appears as a cause of a decrease in the number of type II cells and subsequent diminished surfactant production. The cause of type II cell number reduction remains, however, unknown. That forced lung expansion favors their conversion into type I cells is a likely hypothesis that further investigations will explore.

We conclude from this study that occluding the trachea of fetal sheep with DH and hypoplastic lungs might have both beneficial and adverse effects. Although growth and morphological development of the organ are reestablished, at least partially, the method leads to dramatic surfactant deficit likely due to a decreased number of type II pneumocytes. Therefore, trials of reversible tracheal occlusion in human fetuses should be attempted only after careful consideration, until further functional evaluation of the model has been performed and treatments to increase endogenous surfactant or to supplement with exogenous surfactant have been defined.