INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A congenital diaphragmatic hernia (CHD) occurs during the pseudoglandular period of human lung development resulting in lung hypoplasia, and a reduction in the number of conducting airways, alveoli and preacinar pulmonary vessels (1). The work of George and coworkers (4) has also demonstrated impaired development of the acinus and immature lung tissue morphology. The high mortality rate of infants born with CDH is due to a combination of pulmonary hypoplasia and pulmonary hypertension. The pulmonary hypertension has been explained on the basis of a reduction in the total size of the pre-acinar pulmonary vascular bed, a decrease in the number of vessels per unit of lung tissue, and an increase in the muscularization of the pulmonary arterial tree (1). Although it has been suggested that there is a concomitant diminution in the number of air-blood barriers (ABB), this has never been demonstrated. An ABB is composed of capillary endothelium, squamous alveolar epithelium, and a narrow interstitial space, which in its thinnest area is formed exclusively by the fused basement membranes of the two cell types. A deficiency of ABB (i.e., alveolar vascularization) could contribute to pulmonary hypertension and aggrevate the persistent fetal circulation described in these patients. In addition, a deficient capillary surface for gas exchange could explain hypoxemia refractory to medical therapy in some infants.

Our laboratory has developed an early gestational fetal lamb model of CDH (5, 6). Using this model and computerized interactive morphometry, we sought to determine if the lungs demonstrate: (1) a reduced total alveolar and capillary surface area; and (2) a reduction in the number of air blood barriers per unit of alveolar surface density, the capillary load, representing a defect in vascular composition at the acinar level.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Creation of the CDH Model

The lamb model was created as previously described and approved by the Institutional Review Board (6) and complies with all USDA and NIH guidelines for animal care. Briefly, with aseptic surgical technique and the ewe under general anesthesia, a left diaphragmatic hernia was created surgically at 80 d gestation (pseudoglandular period of lung development) on one fetus of a twin gestation. The other fetus served as a control and was not operated upon. Previous work had shown that thoracotomy incision alone, without the creation of a diaphragmatic defect, did not affect lung development (7). The pregnancy was allowed to continue until day 141 (term 140-145 d), when the ewe was fasted for 24 h, and underwent cesarean section with the use of general anesthesia. The fetuses were removed by hysterotomy.

Preparation of the Lung for Study

The fetal head and neck of each animal was exposed with the umbilical cord intact. An endotracheal tube was placed and secured before any spontaneous breaths. Each lamb was killed by a combination of pentobarbital sodium and potassium chloride overdose, delivered, dried, and weighed. The sternal portion of the thoracic wall was removed to expose the lungs. The lungs were fixed by an infusion into the trachea of 1.5% glutaraldehyde in 0.1 M cacodylate buffer at a pressure of 25 cm of water for 10 min. The trachea was ligated and the lungs immersed in the same fixative for 24 h. Glutaraldehyde is used rather than formalin in order to minimize tissue shrinkage. The volumes of the left and right lungs were determined by water displacement. Each lung was then sectioned sagittally from hilum to pleural surface into slices of 0.5-cm thickness and visually inspected. One section of lung tissue, of comparable distance from the hilum, was taken from the lower lobe of the right and left lungs and placed with the hilar surface down. Three horizontal sections of 1-mm thickness, equidistant from the apex of the lobe, were made and each section was cut into numerous 1-mm blocks maintaining their positions within the section. Several blocks were removed from the center of each section, placed in 1.5% osmium tetroxide with S-collidine buffer for 1 h and post-fixed in uranyl acetate in maleate buffer for 45 min. After dehydration in graded ethanol, the tissues were embedded in Epon 812; Epon 812 does not introduce significant dimensional changes in the tissues. Five blocks from each lower lobe were chosen at random and 1-µm sections were cut and stained according to Humphrey and Pittman (8) for light microscopic study and morphometry. Using the same Epon 812 embedded tissues, sections 60-90 nm in thickness were cut with a diamond knife for electron microscopic study and morphometry. The sections were mounted on five copper grids and double stained with uranyl acetate and lead citrate.

Morphometric Analysis

All stereologic measurements were performed by a single investigator in a blinded fashion using computerized interactive morphometry. Light morphometric analysis was performed on an interactive touch screen peripheral overlaid on a monitor displaying the image of the lung tissue from a video camera mounted on a light microscope (9). The projected image is delineated by a rectangular frame generated by the computer, thus demarcating a field for study (test area). When a structure contained in the test area is touched, the touch screen transmits the information on the cartesian coordinates of the points activated to the computer, which then stores the points and echoes by generating a line or point (9).

Using this system on the 1-µm thick embedded tissue sections, the alveolar surface area was determined by linear intersection measurements at a light microscopic magnification of ×40. A grid consisting of a specific number of evenly spaced parallel lines of known length was superimposed onto the slide. This grid delineated the test area where sampling of points occurred. The parallel lines permitted counting of intersections, which were defined as the points at which these lines crossed septal outlines. The number of intersections was counted for each grid. Ten random grids, or test areas, per tissue section were examined. Random selection was performed by moving the stage micrometer a distance of 0.5 mm on the x and y axes of the microscope stage. All lung parenchyma contained within a test area was evaluated by linear intersection counts. Fields devoid of lung parenchyma or containing large airways were not included. Five tissue sections from each of the right and left lobes were examined. The computer kept a tally of the total number of intersections for each tissue section. Using the formula SvA = 2 IL/LT, a total alveolar surface density (SvA) for each test area was determined, where IL is the total number of intersections counted per grid, and LT the length of the individual parallel test line multiplied by the total number of lines per grid (10). The morphometric data for each lung was then averaged. The total alveolar surface area (Sa) of the entire lung was determined using the formula (total lung volume  10%) × SvA = Sa, where 10% is the percentage of the lung composed of large airways.

Using the same lung tissues, sectioned at 60-90 nm and mounted on copper grids, linear intersection measurements were performed on electron microscopic photographs to determine the alveolar capillary surface area. One section per lobe for each specimen was subjected to analysis. At a magnification of ×3,300, 15 electron micrographs from each tissue section were taken of the lung parenchyma. To ensure random selection, the investigator began in the upper left hand corner of the grid and proceeded along the grid in a clockwise manner. Fields devoid of lung parenchyma or containing large airways were not measured. One tissue section from each specimen was examined and photographed in this manner. The electron micrographs were printed at a magnification of ×2.5. Each electron micrograph was viewed through a video camera mounted on an illuminated stand. The image was displayed on a monitor overlaid with an interactive touch sensitive screen as described above (9). Using linear intercepts the alveolar epithelial and alveolar capillary surface area (air-blood barriers) were determined as described above for each specimen. A total of 180 fields were studied on three control and three CDH specimens.

Implementing computerized interactive morphometry at the light microscopic level on the same tissue sections, the internal surface area and capillary load were determined by linear intercept measurements and volumetric point counting at a magnification of ×100 (10). Surface density was calculated using the linear intersection method via the equation shown above. Point counting determines the volume density of a structure of interest (i.e., air blood barriers) contained in the same field of study. For instance the volume density for the air blood barriers contained within a test area is: the number of points within this structure divided by the total number of points within the area of study. The capillary load was determined by dividing the number of air-blood barriers per field by the surface density of that field. Capillary load measurements of the right and left lobes of CDH and control lungs were meaned.

Statistical Analysis

The two-tailed Student's t test was used to compare the body weights, lung volumes, capillary load, and total alveolar and capillary surface area of the control and CDH lambs.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Morphologic Findings

On gross inspection the CDH lungs were visibly smaller than the control lungs and frequently the right and left lungs within the same specimen appeared equal in size. On cut section there was a paucity of bronchial and arterial ramifications within the CDH lungs. To the blinded observer, there did not appear to be differences in tissue gestational maturity or alveolar structure between controls or CDH lungs, by either light or electron microscopy.

Morphometric Findings

Five CDH lambs and five controls were studied. There was no statistically significant difference in body weight between the two groups; the mean body weight for the control and CDH lambs were 3.57 kg ± 0.96 and 3.23 kg ± 0.92, respectively. The lung volume measurements were significantly greater for the control lambs. The volumes of the left lungs of the control and CDH lambs measured 81.10 cc ± 28.26 and 28.66 cc ± 9.27, respectively (p < 0.01). The volumes of the right lungs of the control and CDH lambs were 132.30 cc ± 41.34 and 50.34 cc ± 25.66, respectively (p < 0.01). Statistical significance remained when the total lung volumes for the control and CDH lungs were calculated. When the mean right and left lung volumes within a group were compared, there was no significant difference. The ratio of the mean total lung volume to the mean body weight was significantly greater in the control group; 60.17 cc/kg ± 12.42 for the control and 24.77 cc/kg ± 6.65 for the CDH group, respectively (p < 0.001).

Determination of the capillary load was also performed on five CDH and control lungs. The capillary loads of the control and CDH lungs were 0.025/mm¹ ± 0.003 and 0.024/mm¹ ± 0.001, respectively, which did not reach statistical significance. There were also no differences when the right and left lungs were compared within a group or between groups.

The electron microscopic data for the total alveolar and capillary surface area are shown in Table 1. These morphometric determinations were performed on three specimens in each group as described in the methods section. The data show that the total alveolar surface area and the total capillary surface area of the left and right lungs in CDH are significantly reduced compared with the controls. There was no difference between the right and left lungs within a group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This surgically created model of CDH had been established previously (5, 6). Similar to the human newborn with CDH, this model demonstrated lung hypoplasia, a reduction in pulmonary arterial density (number of arteries per unit area of lung), and increased muscularization of the pulmonary arterial tree (5). As seen from the present data, this model of CDH produces bilateral lung hypoplasia characterized by a reduction in lung volume and total alveolar surface area. The study has successfully demonstrated a diminution in the total capillary surface area which is the gas exchanging surface of the lung. The capillary load (the number of air blood barriers per unit of alveolar surface density) in CDH lungs is not decreased compared with controls, each CDH acinus having the same number of air blood barriers as the control. Therefore, the diminution in total capillary surface area in CDH lungs is not due to any difference in composition at the level of the acinus. Rather, it is explained by global deficits in CDH lung volume and alveolar surface area; reduced alveolar surface area is accompanied by a concomitant reduction in alveolar vascularization as expressed by capillary endothelial surface area.

Most investigators agree that a combination of pulmonary hypoplasia and pulmonary hypertension are limiting factors to survival in newborns with CDH. deLorimer and coworkers (11) produced diaphragmatic hernias in fetal lambs during the early canalicular period of gestation. The morphometric findings revealed a marked decrease in the total internal surface area of the CDH lungs compared with the controls at term. These workers concluded that survival at birth depended on the development of sufficient pulmonary surface area for gas exchange in order to meet the metabolic needs of the newborn. Pringle and coworkers (12) performed morphologic studies on fetal lambs with CDH created during the glandular period of lung development. Using transmission electron microscopy they observed a paucity of capillaries (air blood barriers) in the walls of alveoli. The authors suggested that a diminution in the capillary cross sectional area was a contributing factor to the severe pulmonary hypertension observed in neonates with CDH.

Efforts to improve the survival of infants with CDH have been directed toward optimizing pulmonary blood flow and gas exchange. Most therapeutic interventions utilize various methods of ventilation, extracorporeal membrane oxygenation (ECMO), or pharmacologic agents to promote pulmonary vasodilation of arteries and arterioles. Morphometric analysis on human infants with CDH surviving for days to months on mechanical ventilation and/or ECMO have been performed (13). Infants who survived greater than 8 d exhibited the greatest postnatal vascular remodeling of the arterial bed characterized by a reduction in muscularization and an increase in external diameter. However, there was no increase in the alveolar radial count of the ipsilateral lung in a single infant. The authors suggested that vascular remodeling would contribute to a decrease in pulmonary hypertension over time if invasive life support could continue. Conversely, it is possible that it is the lack of significant alveolar growth and parallel increase in air blood barriers that would prevent a reduction in pulmonary hypertension during the limited time infants can be supported on ECMO. Unlike the systemic circulation, a substantial portion of the total pulmonary vascular resistance resides in capillaries located in the alveolar walls (14, 15). The major pressure drop of blood flowing through the pulmonary circulation occurs across the pulmonary capillaries (14). Pulmonary vascular resistance is also influenced by the volume of the lung with extremes of inflation and atelectasis causing increased vascular resistance (14). Therefore factors that are involved in maintaining lung expansion, normal surface tension and distribution of ventilation could be important in determining vascular resistance and gas exchange.

The present study demonstrates that the capillary surface area of the lung is reduced in animals born with CDH. This reduction in the gas exchanging surface is due to a diminished lung volume and total alveolar surface area. Because the major contribution to pulmonary vascular resistance resides in alveolar wall capillaries, we suggest that this is the major factor contributing to the severe pulmonary hypertension observed in newborns with CDH. These findings could explain the lack of response to pharmacologic agents and ventilatory interventions that promote pulmonary vasodilation of the pre-acinar bed. It could also explain the positive result some infants with CDH experience when administered exogenous surfactant (16). Alveolar recruitment strategies that prevent atelectasis could optimize distribution of ventilation and perfusion and prevent intrapulmonary shunting. We suggest that an inability to reduce pulmonary hypertension and reverse hypoxemia reflects a severely hypoplastic lung with a markedly diminished alveolar vascular bed. Consequently, the rate of gas exchange is inadequate to meet the metabolic needs of the infant, resulting in death.