Imaging lung aeration and lung liquid clearance ... - The FASEB Journal

The FASEB Journal • Research Communication

Imaging lung aeration and lung liquid clearance at birth Stuart B. Hooper,*,1 Marcus J. Kitchen,† Megan J. Wallace,* Naoto Yagi,‡ Kentaro Uesugi,‡ Michael J. Morgan,† Chris Hall,§ Karen K. W. Siu,†,储 Ivan M. Williams,† Melissa Siew,* Sarah C. Irvine,† Konstantin Pavlov,†,§ and Robert A. Lewis§ *Department of Physiology, †School of Physics, §Monash Centre for Synchrotron Science and 储 Department of Medical Imaging and Radiation Sciences, Monash University, Melbourne, Australia; and ‡SPring-8/JASRI, Sayo, Japan Aeration of the lung and the transition to air-breathing at birth is fundamental to mammalian life and initiates major changes in cardiopulmonary physiology. However, the dynamics of this process and the factors involved are largely unknown, because it has not been possible to observe or measure lung aeration on a breath-by-breath basis. We have used the high contrast and spatial resolution of phase contrast X-ray imaging to study lung aeration at birth in spontaneously breathing neonatal rabbits. As the liquid-filled fetal lungs provide little absorption or phase contrast, they are not visible and only become visible as they aerate, allowing a detailed examination of this process. Pups were imaged live from birth to determine the timing and spatial pattern of lung aeration, and relative levels of lung aeration were measured from the images using a power spectral analysis. We report the first detailed observations and measurements of lung aeration, demonstrating its dependence on inspiratory activity and body position; dependent regions aerated at much slower rates. The air/liquid interface moved toward the distal airways only during inspiration, with little proximal movement during expiration, indicating that transpulmonary pressures play an important role in airway liquid clearance at birth. Using these imaging techniques, the dynamics of lung aeration and the critical role it plays in regulating the physiological changes at birth can be fully explored.—Hooper, S. B., Kitchen, M. J., Wallace, M. J., Yagi, N., Uesugi, K., Morgan, M. J., Hall, C., Siu, K. K. W., Williams, I. M., Siew, M., Irvine, S. C., Pavlov, K., Lewis, R. A. Imaging lung aeration and lung liquid clearance at birth. FASEB J. 21, 3329 –3337 (2007)

ABSTRACT

Key Words: phase contrast X-ray imaging 䡠 fetus 䡠 neonate 䡠 airways

The transition to extrauterine life at birth is critically dependent on the ability of the lungs to aerate and initiate gaseous ventilation. Before birth, the airways are liquid-filled (1), pulmonary vascular resistance is high, and the lungs take no part in the gas exchange, which occurs across the placenta (2, 3). At birth, the airways are rapidly cleared of liquid to allow the entry of 0892-6638/07/0021-3329 © FASEB

air and the onset of air-breathing (1, 4), which initiates a cascade of major physiological changes that enable the lung to adopt to its new role of gas exchange (5, 6). These include a surface tension-mediated increase in lung recoil, an increase in oxygenation, a large increase in pulmonary blood flow and closure of the ductus arteriosus, which shunts blood from the pulmonary artery into the aorta (3). Despite the importance of lung aeration in the adaptation to air-breathing at birth, little is known of this process or the regulatory factors involved, because it has not been possible to observe or measure it. Most studies have focused on liquid clearance because airway liquid retention is a major cause of respiratory morbidity in the neonate (1, 4). Airway liquid reduces the gas volume of the lung and causes non-uniform ventilation, leading to a higher risk of volutrauma and respiratory failure, particularly in preterm infants. The principal mechanism by which liquid is thought to be cleared from the airways at birth is via adrenaline-induced sodium reabsorption, which reverses the osmotic gradient that drives liquid secretion across the pulmonary epithelium during fetal life (1, 7). Although numerous other mechanisms have been proposed, the role of trans-pulmonary pressure associated with inspiratory efforts has not been examined before because it has not been possible to follow the process of liquid clearance at birth on a breath-by-breath basis. However, recent developments in phase contrast X-ray imaging (8 –10) have now made it possible to investigate in detail the dynamics of lung aeration and lung liquid clearance as well as the factors that regulate it. Phase contrast X-ray imaging greatly enhances image contrast by exploiting the phase gradients induced on an X-ray wave field as it propagates through an object with an inhomogeneous refractive index (8, 10, 11). Illumination of an object using a coherent X-ray beam produces interference patterns at a finite propagation distance beyond the object that provide strong contrast of boundaries between structures with differing refrac1

Correspondence: Department of Physiology, Monash University, VIC 3800, Australia. E-mail: stuart.hooper@med. monash.edu.au doi: 10.1096/fj.07-8208com 3329

tive indices (9, 11). The phase change of an X-ray beam (within the diagnostic energy range of 20 –90 keV) propagating through soft tissues is more than three orders of magnitude larger than the change in amplitude. As a result, phase contrast greatly increases soft tissue visibility in X-ray images, particularly at the boundaries between structures where the refractive index variation is high (8). The lung is ideally suited to phase contrast X-ray imaging (8, 10, 11) because it is predominantly comprised of air (⬃80% by volume at end expiration), surrounded by thin tissue structures (predominantly water). The air-tissue interfaces yield significant phase shifts, so the air-filled structures of weakly absorbing lung tissue become highly visible (9, 11) (see Figs. 2, 3, and 4). We have used partially coherent synchrotron radiation (12) as the X-ray source because of its unique properties, particularly its brightness, which make it ideal for this type of imaging. Using phase contrast X-ray imaging, we have determined the rate and spatial pattern of lung aeration from birth in rabbit pups and have examined the dynamics of the air/liquid interface as it retracts toward the distal airways. As the liquid-filled fetal lungs exhibit little phase or absorption contrast, they are not visible at birth, but progressively become visible as they aerate.

MATERIALS AND METHODS Animal procedures All procedures involving the use of animals were approved by the SPring-8 Animal Care and Use Committee as well as the Monash University Animal Ethics Committee. Pregnant New Zealand white rabbits (n⫽20) at 30 and 31 days of gestation were anesthetized by an intravenous injection of propofol (Rapinovet; 15 mg/kg). Pups were delivered by cesarean section and randomly allocated into one of three groups. Pups in Group 1 were killed either before their first breath (as a fetus) or 2 h after birth, when the lungs were fully aerated. After death, an endotracheal tube was inserted via a tracheotomy into the midcervical trachea and connected to an air-filled syringe. Movement of the syringe plunger was controlled remotely using a precalibrated syringe driver. Airway pressure was measured using a pressure transducer and recorded digitally using a data acquisition system (Chart, ADI, Sydney, Australia). The lungs were then inflated using known volumes of air in a step-like fashion (0.21 ml/step), with images acquired after each step change in lung volume. Pups in Group 2 were killed (pentobarbitone overdose; 188 mg IP) either before the first breath (n⫽7) or at 30 s (n⫽6), 1 min (n⫽4), 3 min (n⫽7), 5 min (n⫽9), 15 min (n⫽6), 30 min (n⫽7), or 1 h (n⫽6), or 2 h (n⫽7) after birth. During these periods, the position in which the pups lay (left or right sides or prone) was noted. Additional pups (n⫽4) were born by normal vaginal delivery and were killed, then imaged at 48 h after birth. Pups in Group 3 (n⫽10) were imaged live from birth. They were delivered and immediately placed in a prewarmed (37°C) water-filled perspex imaging chamber (head out) while the fetal membranes remained intact over the mouth and nose. The pup and chamber were carefully positioned within the experimental hutch, ensuring that the X-ray beam would align with the pup’s chest before removing the mem3330

Vol. 21

October 2007

branes. Imaging began as soon as possible after this time, although some pups took one or two breaths during the intervening period between membrane removal and imaging onset. Imaging procedures All pups were imaged in water-filled plastic or PMMA chambers (head out) using experimental hutch 3 of beam line 20B2 in the Biomedical Imaging Centre at SPring-8 located ⬃210 m downstream of the source (13). The X-ray energy used was 25 keV and the pup was placed 2.0 m upstream of the detector, which was a GdO2S phosphor-coupled CCD camera (Hamamatsu, C4742–95HR) with an effective pixel size of 22.47 ␮m (2⫻2 binning mode); the active area used for image acquisition was 24.45(H) ⫻ 20.85(V) mm2. An exposure time of 83 ms was used for live imaging to minimize motion blur, whereas exposure times of up to 300 ms were used for imaging pups after death. A square wave pulse was used to trigger both the detector exposure and a preobject shutter to reduce irradiation of live pups. Image analysis After image acquisition, custom software was used to correct for dark current and non-uniform beam intensity effects. This was achieved by recording a flat field with identical illumination to the experimental image, but with no object, as well as a dark image with the X-ray shutter closed; identical exposure times were used for experimental and correction images. Both the experimental image and the flat field image were corrected by subtracting the dark image before the dark corrected experimental image was divided by the dark corrected flat field image. The dimensions of terminal air sacs located within the basal regions of the lung adjacent to the diaphragm were obtained from images of pups from Group 2; these regions were chosen because individual air sacs could be clearly identified (Fig. 3). The dimensions of ⬃60 sacs were measured from six pups with fully aerated lungs using the boundary between the dark and bright bands to define the margins of each saccule. Similar numbers of saccules were measured from each pup, and the same pups were used for the histological analysis (see below). Lengths were determined using the known pixel size of the detector and image magnification. An analysis of sequential images acquired during lung aeration demonstrated that the power spectra of the images significantly changed as the lung aerated after birth, developing a peak between the spatial frequencies of 5.5 mm⫺1 and 9 mm⫺1 (Fig. 1A). The increase in power at this frequency range resulted from the development of “speckle” within the image. Speckle is caused by aeration of multiple overlying small airways as each structure, depending on its shape, acts as an aberrant focusing lens for the projected beam (11). An integration of the change in power over the spatial frequency range of 2 mm⫺1 and 16 mm⫺1 was used as a relative measure of lung aeration. The spatial frequency range that encompasses the observed peak corresponds to a length scale comparable to the dimensions of the terminal air sacs. It is possible, therefore, that the “speckle” size and airway dimensions are related. Histological procedures In nine pups, the lungs were removed and inflation was fixed at 20 cm H2O using 4% paraformaldehyde; these were not additional pups but had been imaged previously. The fixed

The FASEB Journal

HOOPER ET AL.

using a Student’s paired t test. A P value of ⬍0.05 was taken to be statistically significant.

RESULTS Rabbit pups were delivered by cesarean section at either 30 (n⫽42) or 31 (n⫽28) days of gestation or were allowed to deliver vaginally at term (32 days; n⫽8). Average pup weight was similar at 30 days (41.7⫾1.4 g), 31 days (41.1⫾1.9 g), and 32 days (40.9⫾3.4 g) gestation. Phase contrast imaging Imaging during lung inflation with fixed volumes of air

Figure 1. A) Power spectral analysis of images collected from a spontaneously breathing newborn rabbit pup at selected intervals after birth. Images were acquired before lung aeration (fetus; black) and at 9 min (red), 12 min (green), 15 min (blue), and 37 min (pink) after cesarean delivery. B) A graph showing the relationship between the percentage increase in air volume of the lung, measured by integrating the power spectrum of each image between the spatial frequencies of 2 and 16 mm⫺1, and the percentage change in the volume of air infused into the lung. Images were acquired after each step-like inflation of the lung (0.21 ml/inflation; n⫽16). lung volume was determined by volume displacement before the lungs were divided into left and right lobes and embedded in paraffin. Sections were cut at 5 ␮m, mounted onto glass slides, and stained using hematoxylin and eosin. Other sections were stained for elastin using the Hart’s resorcinfuchsin stain, followed by counterstaining with tartrazine (0.25%) in saturated picric acid. These sections were then viewed by light microscopy and images were recorded digitally as described (14). Standard morphometric techniques were used to determine the relative air/tissue volume fractions as well as alveoli/saccular dimensions as described previously (15); tissue shrinkage associated with tissue processing was assumed to be 7% (15). Measurements of airway dimensions were also made on these pups from the phase contrast images. Statistical analysis Results are presented as mean ⫾ se. Changes in thoracic area measured before and after lung aeration were compared AERATION OF THE LUNG AT BIRTH

Pups in Group 1 were killed either before their first breath or after their lungs had fully aerated, and their lungs were inflated with known volumes of air (6⫻0.21 ml inflations); images were acquired after each inflation step. The spatial pattern of lung aeration was similar, with similar increases in airway pressure in liquid-filled fetal lungs compared with fully aerated newborn pup lungs (data not shown). In liquid-filled lungs, air moved rapidly into the smallest, most distal airways, leading to ⬎50% lung aeration after only the third inflation (0.63 ml or ⬃15 ml/kg). Lung aeration values obtained from the power spectrum analysis of the phase contrast images acquired during lung inflation were compared with the known inflation volumes. A highly significant correlation (P⬍0.001; Fig. 1B) was found between the percentage increase in lung aeration (determined from the power spectra) and the percentage increase in air volume used to inflate the lungs; the linear relationship is described by the regression line, y ⫽ 1.01 ⫻ ⫺0.28 (r2⫽0.999). Static imaging of pups at fixed intervals after birth Pups in Group 2 were killed either before the first breath (fetus) or at fixed intervals after birth, then imaged. As the airways were liquid-filled in pups killed before their first breath (n⫽7), no lung structures could be identified (Fig. 2A) in the phase contrast images; the average power spectrum resulting from these images was used as the baseline for the relative lung aeration analysis in this study. After lung aeration, the trachea, major bronchi, and smaller airways, including terminal respiratory units (primary saccules and some alveoli; Fig. 3), were clearly visible in the phase contrast X-ray images (Figs. 2, 3, and 4). At the peripheral margins of the lungs, single structures of the smallest (⬃120 –150 ␮m) respiratory units were resolved, particularly in the basal regions adjacent to the diaphragm (Fig. 3). Narrowing of the upper trachea caused by the glottis (Fig. 2B) is visible, as well as air bubbles in the stomach and esophagus of neonates (Fig. 2B, E). Bubbles, presumably formed by 3331

Figure 2. Phase contrast X-ray images of newborn rabbit pups delivered by cesarean section and killed either before their first breath (fetus, A) or at 0.5 (B), 3 (C), 5 (D), 30 (E), and 120 (F) min after birth. In the fetal image, when the lungs are liquid-filled, no airway structures can be seen. However, in each of the other images, the trachea and division into the major bronchi can be seen as well as some aerated distal airways. As the distal airways aerate, the lungs display a complex “speckle” pattern (11), although the distal airways can be resolved at the outer margins of the lung (see Fig. 4). The outer regions of the lung are clearly defined, particularly along the margin of the diaphragm (at 30 and 120 min, E, F), and some lobular structures can also be seen—for example, the apical lobes (120 min, F). In addition, surfactant bubbles can be seen in the trachea of the pup killed 0.5 min after birth, and the narrowing of the trachea caused by the larynx is clearly evident at the top of the image; bubbles are also present in the stomach.

the presence of surfactant, were occasionally observed, mostly in the trachea and major bronchi of poorly aerated lungs (Fig. 4). The degree of lung aeration between pups at each time point was very variable. Some pups had wellaerated lungs at 30 s, other pups had only partially aerated lungs at 2 h (2 of 7) after birth (Fig. 5). Poorly aerated lobes occurred almost exclusively in the dependent regions (Fig. 4) of pups that displayed limited respiratory activity and few body movements after birth. The measure of lung aeration obtained from the power spectral analysis graphically demonstrates the large variability between pups (Fig. 5). Expressed as a percentage of the mean value determined from pups with fully aerated lungs (imaged at 48 h after birth), some pups achieved ⬃70% of total lung aeration within 30 s of birth. Others clearly took much longer and had achieved only 25–50% of total lung aeration at 15–30 min after birth; one pup had only achieved 23% of total lung aeration by 2 h after birth (Fig. 5). 3332

Vol. 21

October 2007

Live imaging of pups from birth Movies demonstrating the rate and spatial pattern of lung aeration for each pup were constructed by compiling the image sequences acquired at 800 ms intervals; these movies can be downloaded (http://www. fasebj.org) and are replayed at ⬃5 ⫻ normal speed. The spatial pattern of lung aeration was not uniform and depended on body position. In an upright position, lung aeration occurred simultaneously in both lungs, beginning in the major bronchi and rapidly extending to adjacent peripheral airways (Supplemental Movie 1). Medial regions adjacent to the direction of entry of the bronchi aerated first, followed by the basal and apical regions, which aerated considerably slower (Supplemental Movie 1). Occasionally, larger airways refilled with liquid during expiration, as indicated by proximal, rather than distal, movement of the air/ liquid interface during lung recoil (Supplemental movies 3 and 4). Surprisingly, the shape of the meniscus at

The FASEB Journal

HOOPER ET AL.

Figure 4. Phase contrast X-ray image of a newborn rabbit pup, delivered by cesarean section and killed 120 min after birth. This pup was placed on its left side after birth, and as a result the dependent left lung is poorly aerated. Air bubbles in the trachea, left bronchus, and some of the larger airways of the left lung are clearly visible.

Figure 3. Enlarged views of the most caudal region of the right basal lobe (in dorso-ventral view) as well as the most ventral region of the basal lobe in lateral view. At the outer margins of the lung, where only one airway is in projection, the smallest distal airways can be resolved; the small saccularlike structures are 100 –150 ␮m in diameter. In a lateral view, the small conducting airways that terminate in the saccularlike respiratory units are clearly evident.

the air/liquid interface did not invert (e.g., from convex to concave) during expiration, despite the recoil of the lung and in some cases the proximal movement of the interface (see Supplemental Movie 4). Furthermore, although the radius of curvature of the air/liquid interface changed during a breath, it was similar at end expiration between subsequent breaths despite the distal movement of the air/liquid interface. The effect of the heart beat on the airways was also evident in some sequences, causing visible displacement of airway walls (Supplemental Movie 1). To determine the effect of body position on the spatial pattern of lung aeration, some pups were imaged on their side (Supplemental Movie 2). In this position, the dependent region of the lung aerated more slowly than nondependent regions, demonstrating a non-uniform pattern of lung aeration; airway refilling was more common in the dependent lung (cf. Supplemental Movies 3 and 4). Vigorous breathing activity, indicated by blurring of the AERATION OF THE LUNG AT BIRTH

images, rapidly enhanced lung aeration in both dependent and nondependent (medial and basal) regions of the lung. However, despite vigorous breathing activity, the right upper lobe did not appear to aerate in this sequence (Supplemental Movie 2), demonstrating that body position has a significant influence on lung aeration in the apical regions of the lung. The presence of a large air bubble traveling down the esophagus, beginning at the 23rd second of the sequence, is also visible (Supplemental Movie 2).

Figure 5. Relative levels of lung aeration calculated from the power spectrum of images acquired from individual rabbit pups delivered by cesarean section and killed at selected intervals after birth (n⫽57; filled circles). Time is plotted on a logarithmic scale so that the early time points can be delineated; the mean value (⫾sem) for each time point (open squares) is also plotted. The relative degree of lung aeration between individual pups was highly variable over the first 2 h after birth compared with the distribution measured at 2 days. 3333

Quantification of lung aeration using the power spectra of selected images within each sequence demonstrates the variability between pups of the timerelated pattern of lung aeration (Fig. 6). For instance, one pup rapidly aerated its lung within the first 2–3 min (Fig. 6B), whereas in a second pup little aeration had occurred within the first 7 min (Fig. 6A), but by 15 min the lung was well aerated. The process of lung aeration was found to significantly increase the dimensions of the chest wall at end-expiration by 5.13 ⫾ 1.02% (P⬍0.007; n⫽6); determined by the change in position of the ribs (see Supplemental Movie 1 Fig. 7). This is graphically demonstrated in whereby an image acquired at the beginning of lung aeration is superimposed on an image acquired after lung aeration in the same rabbit pup. Histological analysis of lungs All pup lungs examined (n⫽9) were in the early alveolar stage of lung development, demonstrating single capillary layers between adjacent distal airways (16). Elastin bundles were evident at the tips of small, developing secondary septal crests, but these crests

were at an early stage of development and were not readily apparent in hemotoxylin and eosin-stained sections (data not shown). Thus, the small air-filled sacs located at the peripheral margins of the lung in the phase contrast X-ray images (see Fig. 2) are likely to be saccules. Measured by volume displacement, the total volume of the pressure fixed lung (at 20 cm H2O) was 19.5 ⫾ 1.5 ml/kg and the relative air space fraction was 77.8 ⫾ 1.0%. Thus, when fixed ex vivo, the calculated luminal volume of the lung was 15.1 ⫾ 1.2 ml/kg. Measured using the mean linear intercept technique (15), the mean saccule diameter was 84.3 ⫾ 3.9 ␮m (n⫽9; range 70.5–103.5 ␮m), assuming tissue shrinkage of 7% (15). In comparison, the mean saccule diameter measured from the phase contrast X-ray images (see Fig. 2) was 139.8 ⫾ 4.7 ␮m (⬃60 sacs from 6 pups; range 125–150 ␮m); when measured using this technique, individual saccule sizes varied by 29.3 ⫾ 4.2% within a single pup.

DISCUSSION At end-expiration, the lung is predominantly (⬃80% by volume) comprised of air surrounded by thin ribbons of tissue, which confers optical properties enabling it to be imaged using phase contrast X-ray imaging (8, 10, 11, 17). Although lung aeration is the primary hallmark of the successful transition to life after birth, little is known about the dynamics of this process, the factors that regulate it, or the role it plays in the physiological transformation of the lung at birth. Indeed, the respiratory and cardiovascular changes that occur at birth are considerable and unparalleled by any other physiological events that may occur during extrauterine life. Previous studies have attempted to use radiographic techniques to document the pattern of lung aeration from birth (18 –20); however, this is the first study to provide high-resolution images and a quantitative analysis of lung aeration. We have also provided highresolution movies demonstrating the spatial pattern of lung aeration, the effect of body position, and the dynamics of the air/liquid interface as it retracts into the distal airways. Dynamics of lung liquid clearance

Figure 6. Relative levels of lung aeration measured in individual rabbit pups delivered by cesarean section and imaged live from birth. In each pup, the time-related pattern of lung aeration followed a sigmoidal shape that was very similar between pups, although the onset and timing of lung aeration varied considerably between pups. 3334

Vol. 21

October 2007

The movies of lung aeration demonstrate that residual liquid clearance from the airways is closely associated with inspiratory activity, whereas between breaths, no significant distal movement of the air/liquid interface could be detected. Thus, liquid uptake by the distal airways, due to mechanisms such as osmosis (1), are unlikely to be major contributing factors to airway liquid clearance during lung aeration. Instead, as the distal movement of the air/liquid interface is almost exclusively associated with inspiratory activity, it seems more likely that the trans-pulmonary hydrostatic pressure generated by inspiration provides the predominant driving force for residual airway liquid clearance. Furthermore, we found that inflating the lungs with air

The FASEB Journal

HOOPER ET AL.

Figure 7. Increase in thoracic area, bounded by ribs 1 to 8, after lung aeration. Measurements were made from images acquired at the beginning of a sequence, before the lung had aerated, and at the end of the sequence when the lung had fully aerated; each image was acquired between breaths. The image displayed is a composite derived from overlying images derived from the same pup before it had aerated its lung (ribs are white in appearance) and after it had aerated its lungs (ribs are dark). Although the pup has moved slightly during the imaging sequence (left to right), the chest wall expansion associated with lung aeration is clearly evident.

in dead pups caused rapid aeration of the distal airways, with a similar spatial pattern irrespective of whether the lungs were liquid-filled or fully aerated before inflation. This indicates that airway liquid clearance is not totally dependent on active mechanisms and that increasing trans-pulmonary pressure alone can facilitate airway liquid clearance. However, it is possible that transepithelial Na⫹ reabsorption plays a critical role in retaining liquid within the interstitial tissue compartment, thereby preventing its re-entry into the alveolar space during expiration. With each inspiratory effort, the air/liquid interface progressed deeper into the distal airways, presumably due to the movement of liquid into the terminal air sacs, thereby drawing air deeper into the respiratory tree. However, during relaxation, although some proximal movement of the air/liquid interface was occasionally detected, in most instances the air/liquid interface remained stationary even though the lung recoiled to its equilibrium position (supplemental movies 3 and 4). As the meniscus at the air/liquid interface mostly remained stationary and did not invert during expiration, the increase in airway liquid pressure associated with lung recoil during expiration was minimal. For this to occur, significant volumes of liquid must have left the air space and entered the interstitial/vascular space during the preceding inspiration without returning to the airways during lung recoil. This is consistent with the suggestion that the pressure gradient generated by inspiration is responsible for much of the residual liquid movement from the airways that is trapped in the interstitial tissue and does not re-enter the airways during expiration. Although this liquid is progressively AERATION OF THE LUNG AT BIRTH

cleared via the lymphatics/blood vessels (21), the time frame for this clearance is much slower (⬃2 h) than the process of lung aeration (22, 23). For this to be correct, thoracic volume must temporarily increase as the lung aerates because the volume of air increases at a faster rate than water is lost from the thoracic compartment. A comparison of rib positions between the first (nonaerated lung) and last frames (aerated lung) of the image sequences for each pup demonstrates that thoracic dimensions do increase as the lung aerates (Supplemental Movie 1; Fig. 7). Thoracic area was measured from the phase contrast images and was found to increase in all pups after lung aeration. A composite image showing the change in rib position caused by aeration is displayed in Fig. 7. The suggestion that liquid is initially trapped in the interstitial tissue, thereby creating an increase in thoracic volume during lung aeration, is also consistent with the finding that pulmonary interstitial tissue pressure increases from ⬃0.5 cm H2O to ⬃6 cm H2O at 2 h after birth, then gradually decreases to become subatmospheric (24). As indicated above, it is possible that Na⫹ reabsorption acts to retain lung liquid within the interstitial compartment during airway liquid clearance by providing an osmotic pressure that counterbalances the increase in hydrostatic pressure. Spatial pattern of lung aeration In an upright position, both lungs aerated at relatively uniform rates beginning within regions immediately distal to the direction of entry of the major bronchi. 3335

Considering the potential effect of gravity on the distribution of airway fluid, it is surprising that the apical lobes were the slowest to fully aerate in this position. In support of this, we saw no evidence for gravity-related movement of liquid between aerated/ non-aerated regions of the lung, indicating that surface tension prevents significant movement of liquid between regions of the aerating lung. In a horizontal position, the medial and basal lobes of the nondependent lung appeared to aerate uniformly and before the apical lobe. Similarly, although aeration was slower in the dependent lung, the medial and basal lobes of the dependent lung aerated more quickly than the apical lobe, which did not fully aerate in the example provided (Supplemental Movie 2). Furthermore, it appeared that normal tidal breathing was unable to significantly aerate the dependent lung in this example (in contrast to the nondependent lung), which aerated only after vigorous deep inspiratory efforts. This caused blurring of the images and clearly demonstrates the importance and contribution that inspiratory activity makes to the rate and pattern of lung aeration. The underlying basis for the differential spatial pattern of lung aeration is likely to be multifactorial. Although vertical position is important, as indicated by the differential rate of aeration in dependent and nondependent regions, it is not the sole determinant since slower aeration was found in the apical lobes independent of body position. It is likely that proximity to the diaphragm and the non-uniform way the chest wall deforms during inspiratory efforts contribute by causing regional differences in lung expansion. The initial deep inspiratory efforts that occur immediately after birth are accompanied by large, regional deformations in the chest wall that are also known to occur prenatally during fetal breathing movements (25). The deformations are caused by contraction of the diaphragm, which deforms the chest wall predominantly at its points of insertion, thereby limiting lung expansion and lung aeration in adjacent regions. The chest wall gradually stiffens after birth (26, 27), which markedly reduces the deformation caused by the contracting diaphragm and increases breathing efficiency. The time-dependent pattern of lung aeration (Fig. 5) appeared to follow a continuous sigmoidal shape. However, insufficient images were analyzed to demonstrate the “step-like” pattern that is evident in live imaging sequences, because only one image was usually analyzed between each breath. The discontinuous pattern is caused by the intermittent nature of inspiratory activity and, although it is similar to the “avalanche”-like pattern that describes airway reopening from total lung collapse (28), the underlying factors are different. Airways are thought to reopen from a collapsed state in a “burst-like” pattern when pressure in more central airways increases above a critical opening pressure (28), leading to the opening of many smaller airways in the subtended airway tree (28). Although the pattern of air movement down the airway tree is “burst-like” during lung aeration, it is not a function of airway opening, as 3336

Vol. 21

October 2007

airway dimensions do not appear to change. Instead, it is due to the intermittent nature of inspiratory activity, which appears to drive lung liquid clearance and the gradual retraction of the air-liquid interface into the smaller peripheral airways. The forces governing this pattern of liquid movement are currently unknown, but the technique of phase contrast X-ray imaging is ideal for studying them. Morphometry of the distal airways The saccule dimensions derived from the phase contrast images were larger than the dimensions obtained from the histological analysis (139.8⫾4.7 ␮m vs. 84.3⫾3.9 ␮m). Although it is possible that the saccularlike structures measured in the phase contrast images (Fig. 2) were different from those measured histologically, we believe that the histologically derived dimensions underestimate the dimensions in situ. As the lungs were fixed ex vivo, the inflation pressure (20 cm H2O) applied to re-expand them was probably insufficient to re-expand the lungs to the in situ level. Summary We have used phase contrast X-ray imaging to provide a new perspective and a greater understanding of the factors driving lung aeration and the clearance of lung liquid at birth. As the lungs are indistinguishable from surrounding tissue when they are liquid-filled and become visible only when they aerate, the progressive movement of air down the conducting airways into the smaller respiratory units is clearly visible with each successive breath. The live imaging sequences demonstrate the importance of inspiratory activity for airway liquid clearance as well as the effect of body position on the spatial pattern of lung aeration; chest wall deformation during inspiration likely also contributes to the spatial pattern of lung aeration. Airway refilling was occasionally observed during expiration, particularly in dependent regions of the lung, but significant distal retraction of the air/liquid interface was associated exclusively with inspiratory activity. This work was partially funded by grants from the National Health and Medical Research Council of Australia, the Australian Research Council, and Monash University. The Access to Major Research Facilities Program (managed by the Australian Nuclear Science and Technology Organization) also provided financial support by funding the overseas visits of the Australian coauthors to conduct the experiments. The authors gratefully acknowledge and are indebted to the SPring-8 synchrotron facility in Japan for providing the infrastructure support that was an essential component of these studies. M.J.K. and S.C.I. are recipients of an Australian Postgraduate Award. M.J.K. received a Monash University Postgraduate Publications Award. I.W. was awarded postgraduate scholarships from the Monash University Faculty of Science and the Monash Centre for Synchrotron Science. We also gratefully acknowledge the intellectual contribution made by Mr. Andreas Fouras (Monash University) in developing some of the methodology used.

The FASEB Journal

HOOPER ET AL.

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8.

9.

10. 11. 12.

13.

14.

Olver, R. E., Walters, D. V., and Wilson, M. (2004) Developmental regulation of lung liquid transport. Annu. Rev. Physiol. 66, 77–101 Harding, R., and Hooper, S. B. (1996) Regulation of lung expansion and lung growth before birth. J. Appl. Physiol. 81, 209 –224 Rudolph, A. M. (1985) Distribution and regulation of blood flow in the fetal and neonatal lamb. Circ. Res. 57, 811– 821 Bland, R. D. (2001) Loss of liquid from the lung lumen in labor: more than a simple “squeeze”. Am. J. Physiol. 280, L602–L605 Hooper, S. B., and Harding, R. (2005) Role of aeration in the physiological adaptation of the lung to air-breathing at birth. Curr. Resp. Med. Rev. 1, 185–195 Dawes, G. S. (1968) Fetal and Neonatal Physiology, Year Book Inc., Chicago, Illinois Barker, P. M., and Olver, R. E. (2002) Invited review: Clearance of lung liquid during the perinatal period. J. Appl. Physiol. 93, 1542–1548 Kitchen, M. J., Lewis, R. A., Yagi, N., Uesugi, K., Paganin, D., Hooper, S. B., Adams, G., Jureczek, S., Singh, J., Christensen, C. R., et al. (2005) Phase contrast X-ray imaging of mice and rabbit lungs: a comparative study. Br. J. Radiol. 78, 1018 –1027 Lewis, R. A., Yagi, N., Kitchen, M. J., Morgan, M. J., Paganin, D., Siu, K. K., Pavlov, K., Williams, I., Uesugi, K., Wallace, M. J., et al. (2005) Dynamic imaging of the lungs using x-ray phase contrast. Phys. Med. Biol. 50, 5031–5040 Yagi, N., Suzuki, Y., Umetani, K., Kohmura, Y., and Yamasaki, K. (1999) Refraction-enhanced x-ray imaging of mouse lung using synchrotron radiation source. Med. Phys. 26, 2190 –2193 Kitchen, M. J., Paganin, D., Lewis, R. A., Yagi, N., Uesugi, K., and Mudie, S. T. (2004) On the origin of speckle in x-ray phase contrast images of lung tissue. Phys. Med. Biol. 49, 4335– 4348 Snigirev, A., Snigireva, I., Kohn, V., Kuznetsov, S., and Schelokov, I. (1995) On the possibilities of x-ray phase contrast microimaging by coherent high-energy synchrotron radiation. Rev. Sci. Instruments 66, 5486 –5492 Goto, S., Takeshita, K., Suzuki, Y., Ohashi, H., Asano, Y., Kimura, H., Matsushita, T., Yagi, N., Isshiki, M., Yamazaki, H., et al. (2001) Construction and commissioning of a 215-m-long beamline at SPring-8. Nucl. Instrum. Methods Phys. Res. A 467, 682– 685 Joyce, B. J., Wallace, M. J., Pierce, R. A., Harding, R., and Hooper, S. B. (2003) Sustained changes in lung expansion alter

AERATION OF THE LUNG AT BIRTH

15. 16. 17. 18. 19.

20. 21. 22. 23. 24. 25. 26. 27. 28.

tropoelastin mRNA levels and elastin content in fetal sheep lungs. Am. J. Physiol. 284, L643–L649 Nardo, L., Maritz, G., Harding, R., and Hooper, S. B. (2000) Changes in lung structure and cellular division induced by tracheal obstruction in fetal sheep. Exp. Lung Res. 26, 105–119 Kovar, J., Sly, P. D., and Willet, K. E. (2002) Postnatal alveolar development of the rabbit. J. Appl. Physiol. 93, 629 – 635 Suzuki, Y., Yagi, N., and Uesugi, K. (2002) X-ray refractionenhanced imaging and a method for phase retrieval for a simple object. J. Synchrotron Radiat. 9, 160 –165 Hirvonen, L., Peltonen, R., and Peltonen, T. (1975) Disappearance of tracheobronchial fluid at birth in the lamb. Ann. Chir. Gynaecol. Fenn. 64, 170 –175 Fawcitt, J., Lind, J., and Wegelius, C. (1960) The first breath: a preliminary communication describing some methods of investigation of the first breath of a baby and the results obtained from them. Acta Paediatr. Suppl. 49 (Suppl. 123), 5–17 Rimmer, S., and Fawcitt, J. (1982) Delayed clearance of pulmonary fluid in the neonate. Arch. Dis. Child. 57, 63– 67 Raj, J. U., and Bland, R. D. (1986) Lung luminal liquid clearance in newborn lambs— effect of pulmonary microvascular pressure elevation. Am. Rev. Respir. Dis. 134, 305–310 Bland, R. D., McMillan, D. D., Bressack, M. A., and Dong, L. (1980) Clearance of liquid from lungs of newborn rabbits. J. Appl. Physiol. 49, 171–177 Bland, R. D., Bressack, M. A., and McMillan, D. D. (1979) Labor decreases the lung water content of newborn rabbits. Am. J. Obstet. Gynecol. 135, 364 –367 Miserocchi, G., Poskurica, B. H., and Del Fabbro, M. (1994) Pulmonary interstitial pressure in anesthetized paralyzed newborn rabbits. J. Appl. Physiol. 77, 2260 –2268 Harding, R., and Liggins, G. C. (1996) Changes in thoracic dimensions induced by breathing movements in fetal sheep. Reprod. Fertil. Dev. 8, 117–124 Davey, M. G., Johns, D. P., and Harding, R. (1998) Postnatal development of respiratory function in lambs studied serially between birth and 8 weeks. Respir. Physiol. 113, 83–93 Stocks, J. (1999) Respiratory physiology during early life. Monaldi Arch. Chest. Dis. 54, 358 –364 Suki, B., Barabasi, A. L., Hantos, Z., Petak, F., and Stanley, H. E. (1994) Avalanches and power-law behaviour in lung inflation. Nature 368, 615– 618 Received for publication February 21, 2007. Accepted for publication April 26, 2007.

3337