Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The lung development of fetuses with congenital diaphragmatic hernia (DH) is impaired by the ascension of abdominal viscera into the thorax. Pulmonary hypoplasia is one of the major factors responsible for the 60% mortality rate associated with this disease (1). Lungs from fetuses with a DH have a decreased number of bronchial divisions, leading to a reduction in the number of alveoli (2), and have been described at term as immature with acinar hypoplasia (3). Although in animal models of DH the number of alveolar type II cells is increased as compared with normal lungs (4, 5), the DH lungs seem surfactant-deficient in humans (6, 7) and in various experimental models (8).

In the early 1980s, intrauterine surgery appeared as a potential approach to reduce the hernia before delivery and thus allow lung growth to occur. Technical difficulties and problems of tocolysis have been responsible for the limited success of this approach in humans (12). It has been shown that ligation of the trachea of rabbit fetuses resulted in lung fluid accumulation and alveolar distension (13), and that neonates born with congenital high-airway obstruction present hyperplastic lungs (14, 15). With the aim of inducing fluid retention and lung distension, Hedrick and colleagues (16) applied tracheal ligation to a fetal lamb model of DH (designated the PLUG technique, for "Plug the Lung Until it Grows"); they showed that viscera reentered the abdominal cavity and that lung hypoplasia was compensated. Tracheal occlusion (TO) has been performed by tracheal ligation or, more recently, by endoscopy (17, 18), which reduces the risk of preterm labor. Experimental studies reported that TO on the fetal lung with (16, 19, 20) or without DH (21) enhanced pulmonary arterial and alveolar developmental growth.

This technique at first appeared to reverse the lung hypoplasia associated with DH, but a few studies in animal models with (24, 25) and without DH (26) have now demonstrated that the lungs might display abnormal maturation after TO. These lungs are severely surfactant-deficient (24, 25), and the number of type II cells is decreased (25). Recently, nine human fetuses with severe DH underwent TO in utero (eight with maternal hysterotomy and one by endoscopy) with poor outcome (12, 29).

The aim of the present study was to investigate maturation in the DH and DH-plus-endoscopic TO (DH+TO) models further through ultrastructural examination of the fetal lungs at term. Tissues were examined for cell composition of alveolar epithelium, septal structure and its microvasculature, and the presence of the elastin component.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Experimental Design

Three groups of fetal lambs (Pre-Alp breed) were studied. Extensive fetal surgical procedures are described elsewere (25). In brief, fetuses in the DH group (n = 3) underwent creation of an experimental diaphragmatic hernia on Day 85 of gestation (full term, 145 d). Maternal and fetal anesthesia was achieved with thiopental, 10 mg/kg, and maintained with 1% halothane in O₂-N₂O (50:50, vol/vol). The left-sided DH was created via hysterotomy, using a modification of the technique described by Soper and associates (30). In the DH+TO group (n = 3), creation of a diaphragmatic hernia was performed at 85 d of gestation followed by a tracheal occlusion at 120 d. The ewes underwent a second laparotomy under the same anesthetic procedure. Fetal TO 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 control group (n = 3) consisted of unoperated fetuses of twin pregnancies.

Delivery and Lung Preparation

Fetuses were retrieved by cesarean section at 139 d of gestation. Neonatal breathing was prevented by injection of thiopental (500 mg) and KCl (2 g) into the umbilical vein before sectioning the cord. A median sternotomy was performed, lungs were retrieved, and pulmonary fluid was collected. Samples from the periphery of the inferior left lobe were taken for electron microscopic studies.

Epithelial Cell Study

Three random lung-tissue samples of 1 mm³ were taken from the midsagittal slice of the inferior left lobe of each of the animal lungs. Blocks were fixed in 2.5% glutaraldehyde in 0.1 M phosphate-buffered saline for 16 h at 4°C, and postfixed for 1 h with 2% osmium tetroxide in 0.1 M sodium cacodylate buffer at room temperature. The samples were dehydrated in graded alcohols, transferred to propylene oxide, and embedded in Epon 812. Semithin sections were stained with toluidine blue. Ultrathin sections were cut, mounted on 200-mesh copper grids, stained with uranyl acetate and lead citrate, and examined with a Hitachi HU-12A transmission electron microscope at 50 kV. Cells were counted by direct examination of the whole sections on the microscope screen by one blinded observer. All the alveolar epithelial cells around the air spaces were counted. Bronchiolar epithelial cells, interstitial cells, and endothelial cells were ignored. Cells were identified by their characteristic features and classified into three types; namely, alveolar type II, alveolar type I, and undifferentiated/indeterminate epithelial cells. Only cells showing a nucleus in the plane of the section and lying on basement membrane were taken into account. Areas including large blood vessels and airways were excluded. Approximately 1,000 nucleated cells were counted from each fetus. The number of lamellar bodies in type II cell cross-sections was counted, and the average number per cell profile was compared in the different groups.

Septal Structure

Three different septal components have been studied: the microvasculature of the primary and/or secondary septa, the composition of the air-blood barrier, and the presence of elastin fiber in the alveolar wall. Ultrathin sections were stained with tannic acid/uranyl acetate to highlight the elastic component according to the technique of Kajikawa and coworkers (31). All alveolar septal crests (primary or secondary, depending on the group) were scrutinized for the presence of elastin fibers. The elastic lamina of the blood vessels acted as a built-in positive control for elastin component.

Statistics

A nested design was used with three tissue blocks sampled from each of three animals in each treatment group, making nine tissue blocks in total in each treatment group. Analysis used the PROC MIXED procedure of the SAS for Windows package, version 6.11 (SAS Institute Inc., Cary, NC), which carries out a nested analysis of variance using a mixed model with treatments as fixed effects but tissue blocks and animals as random effects. Linear contrasts were used to compare means between treatment groups. Values were considered significantly different when P < 0.05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Alveolar Epithelial Cells

Three fetal lamb groups were studied: the DH group, the DH+TO group, and the control group.

In each group, epithelial cells were observed on ultrathin sections. Cells were classified into three types: alveolar type II, alveolar type I, and undifferentiated epithelial cells. Approximately cuboidal cells with short apical microvilli and numerous intracellular lamellar bodies (Figure 1A) were identified as alveolar type II cells. Flattened cells with thin attenuated cytoplasmic extensions stretching around the air space away from the main cell body (Figure 1B) were identified as alveolar type I cells. Cuboidal cells with a central nucleus and abundant glycogen were considered undifferentiated cells (Figure 1C). To this group were also arbitrarily ascribed "indeterminate" cells characterized by a few short microvilli, presence of lateral attenuated cytoplasmic extensions, less than two lamellar bodies per cell section, and a more abundant cytoplasm than type I cells (Figure 1D).

View larger version (121K): [in this window] [in a new window]   Figure 1.   Electron micrographs of characteristic features of the various epithelial cell types in the lungs of fetal lambs from the three experimental groups. (A) Typical alveolar type II cell with numerous lamellar bodies in a control lung. (B) Typical alveolar type I cell with attenuated cytoplasm in cell body as well as in the extensions (arrowheads) overlying interstitial layer (il) and a capillary (cap) in a DH+TO lung. (C) Undifferentiated cuboidal cells rich in glycogen (g) in a DH lung. (D) Indeterminate cell in a DH+TO lung with abundant cytoplasm in the core of the cell, rare lamellar bodies, and more or less attenuated cytoplasm in lateral extension. Bars: 2 µm.

To avoid bias secondary to alveolar distension, the results are expressed as a mean percentage for each cell type. The percentages of the three cell types in each group are shown in Figure 2. The type II cell proportion was found to be increased in the DH group (+20%) and decreased in the DH+TO group (19%) compared with the control group. The percentage of type I cells was significantly decreased in the DH group (42%) and increased in the DH+TO group (+13%). Undifferentiated cells were significantly increased in the DH group (+22%) compared with the control group. In the DH+TO group the percentage of undifferentiated cells tended to be increased compared with the control group, but the difference did not reach the P < 0.05 level. Although the third group of cells was composed only of undifferentiated cells in the control group, in the DH+TO group it was made up of two subgroups of cells. Although 24% of those cells were actual undifferentiated cells, the 76% referred to as indeterminate cells presented mixed characteristics of both type II and type I cells. Indeterminate cells were not found in controls and were observed only rarely in the DH group.

View larger version (29K): [in this window] [in a new window]   Figure 2.   Distribution of alveolar epithelial cell types in control, DH, and DH+TO lungs. (A) Type I cells. (B) Type II cells. (C) Undifferentiated cells (white bars) plus indeterminate cells (hatched bars). Significant difference versus control group; *P < 0.05; **P < 0.001.

The difference between the DH and DH+TO groups for the number of lamellar bodies per type II cell profile was not significant, but lamellar bodies were more numerous in the control group than in both of the experimental groups (Figure 3). Secreted lamellar bodies were observed in most of the alveoli in the control and DH groups, but not in the DH+TO group.

View larger version (44K): [in this window] [in a new window]   Figure 3.   Comparison of the number of lamellar bodies (LB) per type II cell profile in control, DH, and DH+TO lungs. Significant difference versus control group for *P < 0.05.

Septal Structure

Septa were examined on semithin (Figure 4) and ultrathin (Figure 5) sections. In the control group (Figure 4C), the pulmonary parenchyma was well developed but presented some degree of structural heterogeneity because of the presence of all the transient stages of septal development. The alveolar primary septa showed a zigzag pattern (32), a single capillary, and a small amount of connective tissue. Secondary septa were present and delimited the primitive air space into alveolar sacs. Some secondary septa presented one capillary layer, a few epithelial cells, and very little connective tissue (Figure 5A). The primary septa also showed crests likely to represent developing secondary septa and containing connective tissue, a single capillary, and a covering epithelium. The alveolar-capillary barrier was well developed and was observed in every alveolus.

View larger version (120K): [in this window] [in a new window]   Figure 4.   Light micrographs of semithin sections of (A) DH lung: note cellularity of walls and paucity of septa; (B) DH+TO lung: air spaces are large, and secondary septa often appear as crests (arrowheads); (C) control lung: numerous well-developed secondary septa delineate air spaces.

View larger version (88K): [in this window] [in a new window]   Figure 5.   Electron micrographs of (A) control lung: septa display classic ultrastructure with alveolar type II cell (II), air-blood barrier and single capillary layer (arrowheads); (B) DH+TO lung: septa display a cellular interstitial layer, secondary septal crests (arrows), and a single capillary layer (arrowhead). Bar: 5 µm.

The pulmonary parenchyma of the fetuses in the DH group displayed a very immature aspect resembling the architecture of the normal late canalicular phase, at a gestational stage when the lung would normally have reached the alveolar phase of development in this species (Figure 4A). Conducting airways occupied a relatively large portion of the lung tissue (not shown). Compared with the alveoli in the control and DH+TO groups, the walls of the primitive air spaces were straight and contained a large amount of connective tissue rich in mesenchymal cells. At the ultrastructural level, a single capillary could be seen alternately to the right or to the left of the connective tissue sheet (not shown). Rare ridges likely to represent septal buds were present. In most of the samples studied, the air- blood barrier was present. This structure did not differ from that of the alveolar-capillary barrier of normal lungs, but its extension was considerably reduced as compared with the control and DH+TO groups because it occurred only in association with the air spaces, which seemed themselves considerably less developed.

Lungs of fetuses in the DH+TO group (Figure 4B) appeared obviously more mature than those of the DH group. True alveoli were present and primary septa were slender, and connective tissue was evidently reduced. At the ultrastructural level (Figure 5B), some secondary septa were seen formed by epithelial cells and a cellular interstitial layer. Capillaries were observed as a single layer. Although varying stages of septal development were seen in this group as in the control group, most of the secondary septa were present as crests. Short, blunt tissue ridges subdivided the peripheral air spaces into shallow alveoli. The air-blood barrier was similar in structure to that in normal lung.

Elastin component was detected in a discontinuous manner at the tip of the rare septal crests seen in the DH group. In the control (not shown) and DH+TO (Figure 6) groups, elastin was seen as condensed aggregates at the tip of the alveolar secondary septal crests and secondary septa. In areas without septa, very small amounts of elastic fibers were scattered in the connective tissue.

View larger version (71K): [in this window] [in a new window]   Figure 6.   Electron micrograph of septa from DH+TO lung shows elastin component as dark aggregates (tannic acid/uranyl acetate staining) at the tips of the alveolar secondary septal crests (arrows). Bar: 5 µm.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

TO has been considered as a potential treatment for fetuses with congenital diaphragmatic hernia (16, 19, 20). The possibility of forcing hypoplastic lungs, consecutively to DH, to enlarge by preventing pulmonary fluid release in the upper airways appeared to be an attractive therapeutic approach. Recently it has been reported that this technique might be beneficial from the point of view of lung growth, allowing the hypoplastic lung to expand and to attain the stage of functional air-blood exchange surface. This approach, however, also seems to have adverse effects. The forced lung expansion has now been shown to induce major surfactant deficiency (24, 25, 33) and reduction of the proportion of type II cells compared with other lung cells (25, 26, 28, 34).

The present study is the first investigation of DH and DH+TO lungs by electron microscopy. Previously, only type II cell examination had been performed after TO without DH (28). Our goal was to document further the stage of cell maturity reached by the alveolar epithelium under both conditions. Type I and type II cells presented normal ultrastructural features in both groups of lambs, which enabled us to determine the type I/type II cell ratio and to compare that with the ratio in age-matched control animals. This ratio, which is about 2 in the control lungs, was reduced 4.5 times in the DH lungs but, conversely, increased 4.5 times in the DH+TO lungs. Because the DH lungs are hypoplastic, the reduction of the ratio in this group could be due to either (1) a decrease in the absolute number of type I cells, (2) an increase in the number of type II cells, or (3) an unequal decrease in the number of both cell types. Balanced hypoplasia with equal reduction of both cell types would not have changed the ratio. Hypoplasia is therefore accompanied by an alteration in the epithelial cell differentiation process. It has previously been shown that lungs of fetuses with DH are morphologically immature (7, 19). The present study confirms the immaturity of DH lung structure at the ultrastructural level with thickening of the lung parenchyma and rare alveolar septa. Moreover, it reveals an increased percentage (29% versus 7%) of undifferentiated alveolar epithelial cells. This increase in the proportion of undifferentiated cells may result from a failure of cells to differentiate or from a delay in their differentiation process. Similarly, the increased proportion of type II cells may also reflect a delay in their conversion into type I cells. The relative increase of type II cells that we observed is in agreement with previous data, but in those studies, either other cell types had not been taken into consideration (5); quantitative data had not been provided (27); or DH had been induced by nitrofen, a drug potentially toxic for the pulmonary parenchyma (4). Despite the increase in the proportion of type II cells, the lamellar body content was found to be reduced on a per-cell basis, which may account for the reduced surfactant content previously observed.

By contrast, in the DH+TO lungs we found an increased type I/type II cell ratio that could be due to either an increased number of type I cells, a decreased number of type II cells, or both. Hyperplasia alone could not be responsible for this increase. It has been suggested by Alcorn and colleagues (27) that TO (through ligation) decreases the number of type II cells, but no quantitative data were provided. We have previously reported a quantitative analysis based on immunohistochemical detection of surfactant protein (SP)-B that showed a decrease in the number of type II cells in the lamb fetus with DH+TO (25), a decrease also shown by in situ hybridization of SP-C transcripts and electron microscopy in lamb fetuses without previously established DH (28, 34). The present work confirms the decrease of type II cell proportion in DH+TO through ultrastructural observation. It also reveals a fact that the previous approaches could not disclose: namely, the appearance of a significant proportion of cells referred to as indeterminate, whose features suggest they could represent a transition form between type II and type I cells. These were not observed in the normal lung.

The mechanism leading to the reduced type II cell proportion remains poorly understood. Several fates, including (1) delayed type II cell maturation, (2) reversion to an undifferentiated state, (3) enhanced conversion into type I cells, (4) damage and destruction of type II cells, or (5) apoptosis, may alternatively or simultaneously account for this phenomenon. Apoptosis can be ruled out because no apoptotic figures were observed in our specimens. The indeterminate cells may represent a transition form between type II and type I cells, as suggested by the start of attenuation of their cytoplasm. Their high proportion in DH+TO fetuses argues for a major role played by conversion of type II or undifferentiated cells into type I cells in the process of increased type I cell proportion. They may result from transdifferentiation of type II into type I cells, but direct differentiation of undifferentiated cells into type I cells may also have occurred as described by Joyce-Brady and Brody (35). Usually, the adaptative response of the alveolar epithelium to environmental stress is replication of type II cells (36), but this is seen in air-filled lungs, not in liquid-filled lungs. One can assume that in order to occupy the extending surface of the growing distended lungs, both type II and undifferentiated cells would form extensions with attenuated cytoplasm. It has recently been shown that the contact of type II cells with endothelial basement membrane induces type I cell differentiation (37). It is possible that forced lung expansion favors type I cell differentiation through increasing epithelial cell contacts with endothelial basement membrane in the stretch-extended septa. As a whole, and whatever the underlying cellular and molecular mechanisms involved, TO seems to induce an overdifferentiation of type I cells. Increased air-blood surface exchange is undoubtedly a sign of enhanced maturity as compared with the DH lung, but it occurs at the expense of type II cell differentiation and surfactant production. Antenatal release from TO may allow redifferentiation of type II cells to occur because, on the one hand, type I cells have been shown to be able to reverse to type II cells in vitro (38), and on the other hand, type II cell number has been reported to normalize 1 wk after release from 2 wk of TO in the fetal lamb (39).

With regard to septal structure, the DH+TO lungs seem to present both mature and immature features. In the lamb, alveolarization is an antenatal event and at birth the lung is fully developed with secondary septa (40). Although more numerous than in the DH lung, most of the secondary septa in the DH+TO group were still crests. Shortness of the septa has also been reported in occluded lungs without previous DH (41). Nevertheless, the elastin fibers that determine the mechanical properties of the lung tissue and have been shown to play a pivotal role in alveolarization process from immature lung saccules (42) were clearly more abundant in the DH+TO lungs than in the DH lungs. The septal structure in the DH+TO lungs seemed normal, with elastin fibers present at the tip of all septa, including septal crests, a location reported for the lamb lung (43) as well as for other species (44, 45). Septation is therefore modified as compared with that in the normal lung, but is progressing. We could not use the criterion of paired capillaries facing each side of the same septum, a characteristic feature of the immature rat (30) and human (44) lung, to evaluate maturation of alveolar septa after tracheal occlusion because pairing of alveolar capillaries has never been described in the lamb fetus, even at early gestational stages (40, 46, 47). In the present study we found a single capillary layer in all three groups, with no sign of pairing of septal capillaries.

In conclusion, the ultrastructural approach showed that TO induced maturational events in the DH lung, which acquired not only a size but also a structure close to that of the normal lung. The process was, however, not complete and included an abnormal increase in the type I cell proportion. We also showed that modifications of mechanical forces applied to the lungs, that is, compression of the lungs or increased intrapulmonary fluid volume, are responsible for alterations in the type II cell differentiation and septal development. These changes in the proportion of the cells may have deleterious consequences that should be further evaluated. Because the ultimate goal of TO studies is to apply this technique to human fetuses with DH, it appears crucial to determine whether changes can be reverted after release from occlusion or compensated by additional treatments.