- Email: [email protected]
Leakage of protein in the immature rabbit lung; effect of surfactant replacement
265
Respiration Physiology (1985) 61, 265-276 Elsevier
LEAKAGE OF PROTEIN IN THE IMMATURE RABBIT LUNG; EFFECT OF SURFACTANT REPLACEMENT
BENGT ROBERTSON, DAVID BERRY, TORE CURSTEDT, GERTIE GROSSMANN, MACHIKO IKEGAMI, HARRIS JACOBS, ALAN JOBE and SALLY JONES Departments of Pathology and Pediatrics, St. G6ran's Hospital, and Department of Clinical Chemistry, Karolinska Hospital, Stockholm, Sweden, and Department of Pediatrics, Harbor-UCLA Medical Center, Torrance, CA, U.S.A.
Abstract. Immature newborn rabbits, delivered on day 27 of gestation, were ventilated artificially for 60 min, with or without previous treatment with natural surfactant. Insuffiation pressure was adjusted to maintain an average tidal volume of about 10 ml/kg. All animals received, before the onset of ventilation, 125I-labeled albumin via the airways and ~3q-labeled albumin intravenously. At the end of the experiment 3.1 _+ 1.3~o (2 _+ SD) of the lSq-albumin had permeated into the alveolar compartment of control animals; the corresponding figures for surfactant-treated animals were 1.7 _+ 0.8~o (P < 0.002). In control animals only 18.2 + 4.4~o of the 12SI-albumin could be recovered from the airspaces after 60 min, whereas 69.9 _+ 14.6~o of this label was recovered in surfactant-treated animals (P < 0.002). Alveolar wash samples from control animals also contained significantly increased activity of surfactant inhibitor, as evaluated with pulsating bubble. The bidirectional flux of protein, including surfactant inhibitor, was thus significantly decreased in these immature lungs by surfactant replacement.
Alveolus Epithelium Immaturity
Lung Newborn Pulmonary epithelium
Protein Rabbit Surfactant
The epithelial cells of the alveoli and the airways are linked by 'tight junctions', which determine the permeability of the barrier between air and pulmonary interstitial tissue (Schneeberger, 1976). Under normal conditions, the junctions permit bidirectional diffusion of small molecules only: for example the 'pore equivalent radius' of the
Accepted for publication 1 June 1985 Reprint requests: Dr. B. Robertson, Department of Pathology, St. G6ran's Hospital, S-112 81 Stockholm, Sweden. 0034-5687/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
266
B. ROBERTSONet al.
epithelial barrier of the fullterm lamb lung is of the order of 0.5-0.6 nm (Normand et al., 1971). This means that the epithelium is impermeable to protein and that, therefore, the presence of protein in the alveolar spaces must be the result of pinocytosis, secretion, aspiration, or leakage following disruption of the epithelium. At birth, the permeability of the air-blood barrier depends on the maturity of the lung and on the transpulmonary pressure gradient (Egan et aL, 1975). Immature, surfactantdeficient lamb lungs, which require a high insuffiation pressure and tend to collapse with each expiration, are characteristically 'leaky', with a large bidirectional flux of protein across the pulmonary epithelium (Adams et al., 1971 ; Jobe et al., 1983, 1985; Normand et al., 1970). This leak may be largely due to disruption of the epithelium in bronchioles and alveoli during the ventilatory movements. Studies of premature newborn rabbits have revealed that epithelial lesions develop within a few minutes of artificial (Nilsson et al., 1980) or spontaneous (Nilsson and Robertson, 1985) ventilation. One important consequence of increased epithelial permeability may be the appearance of a surfactant inhibitor protein in the alveolar spaces (Ikegami et al., 1984). The purpose of the present study is to quantify the protein leak in the lungs of artificially ventilated immature newborn rabbits. In particular, we were interested in whether the protein leak could be modified by surfactant replacement as previously shown in lambs (Jobe et aL, 1983), and whether the magnitude of the protein flux would correlate with variations in lung mechanics.
Material and methods
The studies were carried out on preterm newborn rabbits obtained on day 27 of gestation (term = 31 days). The pregnant doe was anesthetized with intravenous pentobarbital and ketamine and the fetuses delivered by hysterotomy. Immediately after birth, a cannula was tied into the trachea of each pup, and alternate pups received either vehicle or surfactant by instillation via the tracheal cannula (see below). They were then transfered to a system of multiple body plethysmographs, heated to 37 °C, paralysed with intraperitoneal pancuronium bromide (0.2 mg/ml), and each litter was ventilated in parallel for 60 min with a pressure-limited ventilator (Infant Ventilator, Sechrist Industries, Anaheim, CA, U.S.A.) delivering 100~ 02 at a frequency of 40 breaths/rain and 50~o inspiration time. Tidal volumes were monitored in each animal with a Fleisch-tube connected to the body plethysmograph (Lachmann et al., 1981). Insufflation pressure was adjusted for each litter to provide an average tidal volume of about 10 ml/kg. After the period of artificial ventilation, ECG recordings were obtained. Blood gases were measured at sacrifice using samples aspirated from the beating heart. Since these samples might have been obtained from either the left or the right ventricle, Pco2 and pH only are reported. The blood samples were also used for determination of hemoglobin and radiolabel content (see below).
A n i m a l experiments; general procedure.
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
267
Preparation and in-vitro evaluation of surfactant. Surfactant was prepared from saline extracts of minced bovine lungs. After filtration of the saline phase, the lipids were extracted with chloroform : methanol 2 : 1 (v/v), and separated by liquid-gel chromatography (LipidexR-5000) into phospholipids, cholesterol, triglycerides and cholesteryl esters. The phospholipid fraction - which also contains about 1 ~o protein with a large proportion ofhydrophobic amino acids - w a s lyophilized and stored at - 18 ° C. Shortly before the experiment, the surfactant material was suspended in saline at a final phospholipid concentration of 80 mg/ml. It was carefully mixed with commercially available 125I-labeled human serum albumin (Mallinckrodt, St. Louis, MO, U.S.A.), added at a concentration of 0.84 mg/ml (8.5/~Ci/mg). The phospholipid composition of our surfactant has been reported elsewhere (Nilsson et al., 1985). Its physical properties were tested with pulsating bubble at room temperature as described by Ikegami et al. (1983). When suspended at a concentration of 3 #mol phosphatidylcholine/ml, our surfactant reduced surface tension to 0 mN/m during 56~o surface compression (rate 16/min). This indicated that although our surfactant preparation only represented one fraction of the complex material usually referred to as 'natural surfactant', it was suitable for further evaluation in vivo.
Surfactant replacement and artificial ventilation. Before the onset of artificial ventilation, experimental animals were treated via the tracheal cannula with 70 #1 surfactant containing 0.5/~Ci ~25I-albumin. Control animals received the same volume of normal saline containing 0.5/~Ci ~zSI-albumin. Concurrently with the onset of artificial ventilation, all animals were injected via an external jugular vein with 0.2 ml saline containing 0.5/~Ci 13q-labeled albumin, made by iodinating bovine serum albumin monomer standard (Miles Laboratories, Elkhart, IN, U.S.A.) (Hunter and Greenwood, 1962). We knew from previous experiments (Berggren et al., 1985; Nilsson et al., 1978) that treatment of-the premature newborn rabbit with natural surfactant leads to a striking improvement in lung-thorax compliance. This was confirmed in pilot studies with the present surfactant preparation (data not shown) indicating that control animals required insuffiation pressures of 25-30 cm H 2 0 to stay alive for 60 min, while such high pressures inevitably caused overinflation of the lungs in surfactant-treated animals, leading to pneumothorax and death. Since one of our basic aims was to maintain proper ventilation during the course of the experiment, we could not ventilate the surfactanttreated animals and controls in parallel with the same pressure. We therefore chose to use alternate litters as surfactant recipients and controls. However, in each surfactanttreated litter we included two non-treated control animals, ventilated with the same 'low pressures'. Failure to obtain adequate tidal volumes in these control animals identified the litter as immature. Similarly, in each control litter, we included two surfactant-treated animals, ventilated with the same 'high pressures'. The regular occurrence of lung rupture in these treated animals confirmed a surfactant response. Our final data were thus obtained from 25 surfactant-treated animals from 6 litters,
268
B. ROBERTSONet al.
ventilated with 'low pressures', and 21 control animals from 6 other litters, ventilated with 'high pressures'. Processing of lungs. The lungs were lavaged in situ with normal saline. The chest was opened and the lungs were filled to total lung capacity by visual inspection; this saline was flushed in and out of the airways three times. This procedure was repeated 4 times, and the recovered lavage fluid was pooled for quantitation of 125I, 131Iand total protein. The lungs were then removed, weighed, and homogenized in distilled water. Aliquots of the homogenate were assayed for the same labels as analysed in the lavage fluid. From the homogenates and the alveolar washes, aliquots were also taken to quantify phosphatidylcholine content by thin-layer chromatography followed by phosphorus determination according to Bartlett (1959). In order to define the contribution of intravascular blood to the radioactivity measured in the homogenates, part of the homogenate was centrifuged at 27 000 x g, and the resulting red supernatant was used for analysis of hemoglobin. We also assayed the hemoglobin content and radioactivity of the blood sampled from the heart at the end of the experiment. By combining the information from these analyses, and assuming that all hemoglobin in the lung homogenate was intravascular, we could determine the contribution of radioactivity in blood to that in the total lung homogenate. This contribution was subtracted in our calculation of radioactivity associated with lung tissue (Jobe et al., 1983). Surfactant inhibitors in alveolar wash. Equal volumes of alveolar wash from surfactanttreated animals were combined into 4 pools. Three of these pools represented separate litters; the fourth was obtained by combining lavage fluid from surfactant-treated animals in the remaining 3 litters. Four similar pools were obtained from control litters. The lavage fluid was centrifuged at 8000 x g and the protein content of the supernatant measured according to Lowry et al. (1951), using bovine serum albumin as standard. Minimum surface tension of our surfactant preparation was determined with the pulsating bubble as described above, in the presence of varying amounts of supernatant protein (0, 0.1, 0.3, 0.6 mg/ml). In these analyses, the phosphatidylcholine content of the surfactant preparation was kept constant, 3/~mol/ml. We also determined the minimal surface tension of the pooled samples of crude alveolar wash, and of surfactant isolated from the same samples by centrifugation at 8000 x g over 0.7 M sucrose in saline. Histologic studies. Six randomly selected animals (3 surfactant-treated, 3 controls) were sacrificed after 60 min of artificial ventilation by transsection of the abdominal aorta. The lungs were then ventilated with air for 3 min, the trachea was clamped at end-inspiration, and the lungs were fixed by immersion in formalin. Paraffin sections from the lower lobes were examined by conventional light microscopy, with particular reference to the alveolar expansion pattern and the extent of epithelial lesions and hyaline membranes.
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
269
Statistical evaluation. Values are given as means + SD. The Wilcoxon two-sample, two-tailed test was applied for evaluation of statistical differences, the limit level of significance being defined as P = 0.05.
Results
Survival, blood gases, lung mechanics. All the 25 surfactant-treated animals ventilated with 'low pressure' and 18 of the 21 controls ventilated with 'high pressure' had regular cardiac activity with frequency > 200/min at the end of the experiment. Body weight, data on lung mechanics after 30 and 60 min, cardiac frequency, and blood gases in survivors are shown in table 1. The figures for tidal volume show no difference between surfactant-treated animals and controls, although the volumes were generally higher than intended at the 30 min interval. After 60 min, the average tidal volume of both groups was close to the desired value of 10 ml/kg. The pressure required to maintain this tidal volume was about 40-50~o lower in the treated group, reflecting improved compliance of the lung-thorax system throughout the period of observation. In both groups, mean values for compliance were lower at 60 min than at 30 min, but this difference was statistically significant only for the surfactant-treated animals. Most animals had normal heart rates and blood gases prior to sacrifice, without differences between the groups, indicating that they were ventilated adequately. Protein leakage. The percent recovery of the 125I-albumin administered into the airways was significantly higher in alveolar wash and lung tissue from animals receiving surfactant than in control samples (fig. 1A), and the percent recovery of the intravenously injected 13q-albumin was lower in alveolar wash from surfactant-treated animals than in controls (fig. 1B). However, the percentage of 131I associated with lung tissue was similar in the two groups (fig. 1B), as were the values for total protein content of alveolar wash (table 2). If the percentage of 1311 in alveolar wash is related to variations in tidal volume, the trends are different in surfactant-treated animals and controls. The amount of label in alveolar wash tends to be inversely related to tidal volume in the surfactant-treated group (y = 14.6-2.73x; r = - 0.39), whereas the opposite seems to be the case among controls (y = 1.46x + 6.23; r = 0.39). However, neither of these correlations is statistically significant. Phospholipids and surfactant inhibitors in lung wash. The phosphatidylcholine content of alveolar wash was 17 times higher in animals treated with surfactant than in controls (table 2). Similarly, the supernatant of alveolar wash (used for assay of surfactant inhibitors) contained significantly more phosphatidylcholine in surfactant-treated animals than in controls (table 2). This difference is, however, small in relation to the total concentration of bovine surfactant (3/~mol/ml) suspended in the samples assayed
270
B. ROBERTSON
e t al.
TABLE 1 Body weight and data on lung mechanics, heart rate, and blood gases in surfactant-treated animals and controls. Abbreviations: VT = tidal volume, P = peak insufflation pressure, C = compliance (VT/P). Experimental condition
Surfactant-treated (n = 25) Controls (n = 21) P< * P vs
% .125
Body weight
30 min
(g)
VT
P
C
(ml/kg)
(cm H20 )
(ml/cm H20.kg )
31 + 6
15.9 + 7.2
14.5 _+2.6
1.09 _+0.39
30 + 5 NS
13.8 + 4.8 NS
26.6 + 2.1 0.002
0.51 _+0.15 0.002
30 min <0.02.
[]
Surfactant-treated
[]
Controls
-K- P <
n=19
n = 16
0.002
80
% I TM
60
[]
Surfactant-treated
[]
Controls P <
40
El'l
20
4
n=19
n = 16
0002
L
3 2 1 0
A l v e o l a r wash
A
A l v e o l a r wash
Lung t i s s u e
Lun( t i s s u e
B
Fig. 1. Percentage of '251-albumin (A) and 131I-albumin (B) in alveolar wash and lung tissue of surfactanttreated animals and controls. 125I-albumin was mixed with bovine surfactant and administered via the airways. ~3q-albumin was injected intravenously with the onset of artificial ventilation.
for inhibitory activity. T h e p r o t e i n c o n t e n t o f the s u p e r n a t a n t w a s similar in s u r f a c t a n t t r e a t e d a n i m a l s a n d c o n t r o l s (table 2). T h e inhibitory activity o f s u p e r n a t a n t f r o m the a l v e o l a r w a s h o f the t w o g r o u p s o f a n i m a l s is s h o w n in fig. 2. A t all three c o n c e n t r a t i o n s o f s u p e r n a t a n t protein, the inhibitory activity is m u c h higher ( m i n i m a l s u r f a c e t e n s i o n is higher) in the c o n t r o l g r o u p (P = 0.05); this activity i n c r e a s e s w i t h the c o n c e n t r a t i o n o f s u p e r n a t a n t protein.
271
PROTEIN L E A K A G E IN THE I M M A T U R E RABBIT LUNG
Heart rate (min 1)
60 min
(ml/kg)
P (cm H20 )
C (ml/cm H 2 0 . k g )
10.8 _+ 5.5*
14.6 _+ 2.6
0.75 + 0.35*
11.3+_4.5 NS
24.8+1.1 0.005
0.45+0.18 0.01
VT
Blood gases pH
Pco2 (mm Hg)
337 + 34
7.29 + 0.23
43.1 _+ 16.1
323_+73 NS
7.23_+0.18 NS
37.5_+12.2 NS
Inhibitory Activity ~ 'rnin , mN/m '
20 15
/
10.
/ /~
'
o.1
~ J q[ Controls n :4
~
i Surfactant/J. Treated
'
6.3
'
'
d.6 mg
Protei~/ml
Fig. 2. Inhibitory activity in supernatants of alveolar wash from surfactant-treated animals and controls. Values for minimal surface tension (7 min) were measured with the pulsating bubble at various protein concentrations. Differences are significantly significant for protein concentrations 0.1-0.6 mg/ml (e = 0.05). TABLE 2 Total phosphatidylcholine (PC) and protein in alveolar wash and in alveolar wash supernatant used for assay of inhibitory activity (see fig. 2). Parameter
Alveolar wash PC (/tmol/kg) protein (mg/kg) Supernatant PC (#mol/ml) protein (mg/ml)
Surfactanttreated animals
94+25 183+84
Controls
(n=25) (n=25)
0.18 + 0.05 (n = 4) 0.76 + 0.10 (n = 4)
5.4+2.8 163+51
P
(n=21) (n=21)
0.03 _+ 0.01 (n = 4) 0.91 + 0.33 (n = 4)
<0.005 NS =0.05 NS
Fig. 3. Histological sections illustrating the effect of surfactant replacement on alveolar expansion and airway epithelium. (A)Control animal with nearly collapsed alveoli and prominent desquamation of bronchiolar epithelium (arrow). (B) Surfactant-treated animal with significantly improved, yet not quite uniform alveolar expansion pattern. Note intact epithelium in airway (*). Hematoxylin and eosin, x 150.
PROTEIN LEAKAGEIN THE IMMATURERABBITLUNG
273
Surfactant isolated from the pooled alveolar washes of treated animals at a concentration of 4/~mol phosphatidylcholine/ml reduced surface tension to 0 mN/m during surface compression in all four samples analysed, indicating that the surface activity was unchanged in comparison with that of the material originally instilled into the airways (see Material and methods). Histologicfindings. Among the six animals studied histologically, the three treated with surfactant had improved alveolar air expansion and less prominent epithelial lesions in comparison with controls, confirming earlier observations under similar experimental conditions (Berggren etal., 1985; Nilsson etal., 1978). There was no evidence of pulmonary hemorrhage. Representative sections from the two groups of animals are illustrated in figs. 3A and B.
Discussion
Our data show that immature newbom rabbits can be kept alive with adequate blood gases for at least 1 h, provided that the ventilation setting is adjusted with respect to the compliance of the lungs. However, whereas surfactant-treated animals survived the period of ventilation without serious lung lesions, control animals characteristically developed widespread desquamation of airway epithelium, with formation of hyaline membranes in bronchioles and alveoli. In all animals there was a bidirectional leakage of protein, but the amount of 13q-albumin entering the airspaces was about twice as large in the control group. Similarly, the loss of 125I-albumin from the airspaces was significantly more rapid in non-treated animals than in those receiving surfactant. Most likely, the increased protein leakage in control animals can be attributed to disruption of the epithelial and endothelial barriers of the peripheral airspaces. Movements of liquid, electrolytes and larger molecules within the lung are usually analysed by means of a three-compartment model (Strang, 1976) comprising the intravascular, interstitial, and intra-alveolar spaces. In the intact fetal lung, the walls separating these compartments have different permeabilities. Thus, the 'pore equivalent radius' of the alveolar capillary endothelium is of the order of 15 nm (Normand et al., 1971), whereas - as mentioned above - that of the alveolar epithelium has been estimated to be about 0.6 nm. This means that, under normal conditions, the endothelial wall is permeable to albumin molecules (diffusion radius 3.4 nm), whereas the alveolar epithelium is not (Olver et al., 1981). The significance of our present observations can be understood from the three-compartment model, taking into account that the epithelial wail is 'leaky' in the ventilated immature lung (Adams et al., 1971 ; Egan et al., 1980, 1984; Jobe et al., 1983, 1985; Normand et al., 1970) and that the permeability of this wall is further increased in control animals by necrosis and disruption of cell junctions. In both surfactant-treated animals and controls, ~25I-labeled albumin instilled into the airways moved from the intra-alveolar compartment (represented by the lung lavage
274
B. ROBERTSONet al.
fluid) into the interstitial space (represented by lung tissue minus blood). However, whereas in control animals altogether about 65~o of this protein permeated into the intravascular compartment and further on to extrapulmonary tissues, nearly 100~o remained lung-associated in surfactant-treated rabbits. The difference in ~25I-albumin content of alveolar wash, shown in fig. 1A, is most probably the result of increased permeability of the epithelial membranes in the control group. The larger amount of the same label in the lung tissue of surfactant-treated animals is less readily understood. To some extent, this latter difference may reflect less effective lung lavage in the surfactanttreated group. An alternative, and perhaps more likely, explanation is that control animals develop an increased protein leakage across the endothelial membrane as well. This would lead to accelerated equilibration between the pulmonary interstitial compartment on one hand, and the intravascular compartment on the other. The observed leakage of protein into the alveolar spaces can be interpreted on the basis of similar assumptions. The fraction of intravascularly injected t31I-albumin recovered in the interstitial space (represented by lung tissue) was the same in surfactant-treated animals and controls, about 3 ~o. This lack of difference may seem at variance with the increased transendothelial protein leakage documented in control animals with the other label. We believe, however, that the ~31I-albumin moving into the interstitial space of control animals rapidly permeated further into the intra-alveolar spaces, leading to accelerated equilibrium between these two compartments. This interpretation is in keeping with the increased permeability of epithelial membranes to 125I-albumin (fig. 1A), as well as with the figures showing enhanced accumulation of ~3q-albumin in the alveolar spaces of the control group (fig. 1B). An increased transendothelial leakage of protein is further indicated by our finding that the amount of 13li_albumin in the combined interstitial and intra-alveolar spaces is about 50 ~o higher in control animals than in the surfactant-treated ones (fig. 1B). Our observation that the total amount of protein, harvested by alveolar washing, was similar in surfactant-treated animals and controls does not invalidate the data shown in fig. 1. Some of these proteins were probably present in the fetal lung liquid before the onset of ventilation. Furthermore, the apparent paradox can be solved by taking into account that there is a significant leakage of protein in both groups of animals but that - for reasons outlined above - the permeation of protein across the endothelial and epithelial membranes is more rapid in controls. This would lead to a faster equilibration between the three compartments in control animals. The net result over a limited period of ventilation could then be a similar intra-alveolar accumulation of total protein in both groups, with a delayed intra-alveolar appearance of the intravenously injected label in animals treated with surfactant. The leakage of protein in immature lungs, documented in the present and other (Jobe et al., 1983, 1985) studies by means or radiolabeled albumin, has potential clinical implications (Jefferies et al., 1984). At least one of the proteins entering the alveolar spaces is a potent inhibitor of pulmonary surfactant (Ikegami et al., 1983 ; Jobe et al., 1983). This particular protein, recently isolated by Ikegami et al. (1984), has a molecular size of 110 000 daltons. As shown in fig. 2, the inhibitory activity of protein in alveolar
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
275
w a s h was significantly higher in the control group. I n c r e a s e d inhibitory activity with loss o f surfactant function leads to g r a d u a l aggravation o f the epithelial lesions, with further leakage o f inhibitory proteins into the airspaces, etc. - a vicious circle p r o b a b l y involved in the pathogenesis o f respiratory insufficiency in immature n e w b o r n babies with surfactant deficiency. Several previous studies have shown that surfactant replacement has a beneficial influence on gas exchange and lung mechanics in clinical and experimental respiratory distress s y n d r o m e (for review, see Jobe, 1984; R o b e r t s o n , 1980, 1983, 1984). T h e s e effects can largely be attributed to r e d u c e d surface tension in the a i r - l i q u i d interfaces o f the lung, with i m p r o v e d aeration o f the alveolar c o m p a r t m e n t . D a t a from the present study indicate that this therapeutic a p p r o a c h might improve lung function also by reducing the leakage o f surfactant-inhibiting proteins into the alveolar spaces.
Acknowledgements This study was s u p p o r t e d by The Swedish M e d i c a l R e s e a r c h Council (Project No. 3351), T h e Swedish N a t i o n a l A s s o c i a t i o n against H e a r t and Chest Diseases, The Swedish Society o f M e d i c a l Sciences, The R e s e a r c h F u n d s o f the K a r o l i n s k a Institute, a n d G r a n t H D - 1 1 9 3 2 from Child H e a l t h a n d D e v e l o p m e n t , D e p a r t m e n t o f Health and H u m a n Services, U . S . A .
References Adams, F.H., T. Fujiwara and H. Latta (1971). 'Alveolar' and whole lung phospholipids of premature newborn lambs. Biol. Neonate 17: 198-218. Bartlett, G.R. (1959). Phosphorus assay in column chromatography. J. Biol. Chem. 234: 466-468. Berggren, P., T. Curstedt, G. Grossmann, R. Nilsson and B. Robertson (1985). Physiological activity of pulmonary surfactant with low protein content; effect of enrichment with synthetic phospholipids. Exp. Lung Res. 8: 29-51. Egan, E.A., R.E. Olver and L.B. Strang (1975). Changes in nonelectrolyte permeability of alveoli and the absorption of lung liquid at the start of breathing in the lamb. J. Physiol. (London) 244: 161-179. Egan, E.A., R.M. Nelson and E.F. Beale (1980). Lung solute permeability and lung liquid absorption in premature ventilated fetal goats. Pediatr. Res. 14: 314-318. Egan, E.A., W.P. Dillon and S. Zorn (1984). Fetal lung liquid absorption and alveolar epithelial solute permeability in surfactant deficient, breathing fetal lambs. Pediatr. Res. 18: 566-570. Hunter, W.M. and F.C. Greenwood (1962). Preparation of 13q-labeled human growth hormone of high specific activity. Nature (London) 194: 495-496. lkegami, M., H. Jacobs and A. Jobe (1983). Surfactant function in the respiratory distress syndrome. J. Pediatr. 102: 443-447. Ikegami, M., A. Jobe, H. Jacobs and R. Lam (1984). A protein from the airways of premature lambs that inhibits surfactant function. J. Appl. Physiol. 57:1134-1142. Jefferies, A.L., G. Coates and H. O'Brodovich (1984). Pulmonary epithelial permeability in hyalinemembrane disease. New Engl. J. Med. 311: 1075-1080. Jobe, A., M. Ikegami, H. Jacobs, S. Jones and D. Conaway (1983). Permeability of premature lamb lungs to protein and the effect of surfactant on that permeability. J. Appl. Physiol. 55: 169-176.
276
B. ROBERTSON et al.
Jobe, A. (1984). Respiratory distress syndrome - new therapeutic approaches to a complex pathophysiology. Adv. Pediatr. (Chicago) 30: 93-130. Jobe, A., H. Jacobs, M. Ikegami and D. Berry (1985). Lung protein leaks in ventilated lambs: effect of gestational age. J. Appl. Physiol. 58: 1246-1251. Lachmann, B., G. Grossmann, J. Freyse and B. Robertson (1981). Lung-thorax compliance in the artificially ventilated premature rabbit neonate in relation to variations in inspiration : expiration ratio. Pediatr. Res. 15: 833-838. Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall (1951). Protein measurement with the Folin phenol reagent. J. Biochem. 193: 265-275. Nilsson, R., G. Grossmann and B. Robertson (1978). Lung surfactant and the pathogenesis of neonatal bronchiolar lesions induced by artificial ventilation. Pediatr. Res. 12: 249-255. Nilsson, R., G. Grossmann and B. Robertson (1980). Bronchiolar epithelial lesions induced in the premature rabbit neonate by short periods of artificial ventilation. Acta Pathol. Microbiol. lmmunol. Scand. (.4) 88: 359-367. Nilsson, R., P. Berggren, T. Curstedt, G. Grossmann, G. Renheim and B. Robertson (1985). Surfactant treatment and ventilation by high frequency oscillation in premature newborn rabbits; effect on survival, lung aeration and bronchiolar epithelial lesions. Pediatr. Res. 19: 143-147. Nilsson, R. and B. Robertson (1985). Bronchiolar epithelial lesions in spontaneously breathing immature newborn rabbits. Biol. Neonate (in press). Normand, I. C. S., E. O. R. Reynolds and L. B. Strang (1970). Passage of macromolecules between alveolar and interstitial spaces in foetal and newly ventilated lungs of the lamb. J. Physiol. (London) 210: 151-164. Normand, I.C.S., R.E. Olver, E.O.R. Reynolds, L.B. Strang and K. Welch (1971). Permeability of lung capillaries and alveoli to nonelectrolytes in the foetal lamb. J. Physiol. ~London) 219: 303-330. Olver, R. E., E.E. Schneeberger and D.V. Waiters (1981). Epithelial solute permeability, ion transport and tight junction morphology in the developing lung of the fetal lamb. J. Physiol. (London) 315: 395-412. Robertson, B. (1980). Surfactant substitution; experimental models and clinical applications. Lung 158: 57-68. Robertson, B. (1983). Lung surfactant for replacement therapy. Clin. Physiol. 3:97-110. Robertson, B. (1984). Choosing the surfactant for replacement therapy. In: Current Concepts in Surfactant Research, edited by P. yon Wichert. Prog. Respir. Res. Vol. 18. Basel, Karger, pp. 240-250. Schneeberger, E.E. (1976). Ultrastructural basis for alveolar-capillary permeability to protein. In: Lung Liquids, edited by R. Porter and M. O'Connor. Ciba Foundation Symposium 38. Amsterdam, Elsevier, pp. 3-21. Strang, L.B. (1976). The permeability of lung capillary and alveolar walls as determinants of liquid movements in the lung. In: Lung Liquids, edited by R. Porter and M. O'Connor. Ciba Foundation Symposium 38. Amsterdam, Elsevier, pp. 49-58.
Respiration Physiology (1985) 61, 265-276 Elsevier
LEAKAGE OF PROTEIN IN THE IMMATURE RABBIT LUNG; EFFECT OF SURFACTANT REPLACEMENT
BENGT ROBERTSON, DAVID BERRY, TORE CURSTEDT, GERTIE GROSSMANN, MACHIKO IKEGAMI, HARRIS JACOBS, ALAN JOBE and SALLY JONES Departments of Pathology and Pediatrics, St. G6ran's Hospital, and Department of Clinical Chemistry, Karolinska Hospital, Stockholm, Sweden, and Department of Pediatrics, Harbor-UCLA Medical Center, Torrance, CA, U.S.A.
Abstract. Immature newborn rabbits, delivered on day 27 of gestation, were ventilated artificially for 60 min, with or without previous treatment with natural surfactant. Insuffiation pressure was adjusted to maintain an average tidal volume of about 10 ml/kg. All animals received, before the onset of ventilation, 125I-labeled albumin via the airways and ~3q-labeled albumin intravenously. At the end of the experiment 3.1 _+ 1.3~o (2 _+ SD) of the lSq-albumin had permeated into the alveolar compartment of control animals; the corresponding figures for surfactant-treated animals were 1.7 _+ 0.8~o (P < 0.002). In control animals only 18.2 + 4.4~o of the 12SI-albumin could be recovered from the airspaces after 60 min, whereas 69.9 _+ 14.6~o of this label was recovered in surfactant-treated animals (P < 0.002). Alveolar wash samples from control animals also contained significantly increased activity of surfactant inhibitor, as evaluated with pulsating bubble. The bidirectional flux of protein, including surfactant inhibitor, was thus significantly decreased in these immature lungs by surfactant replacement.
Alveolus Epithelium Immaturity
Lung Newborn Pulmonary epithelium
Protein Rabbit Surfactant
The epithelial cells of the alveoli and the airways are linked by 'tight junctions', which determine the permeability of the barrier between air and pulmonary interstitial tissue (Schneeberger, 1976). Under normal conditions, the junctions permit bidirectional diffusion of small molecules only: for example the 'pore equivalent radius' of the
Accepted for publication 1 June 1985 Reprint requests: Dr. B. Robertson, Department of Pathology, St. G6ran's Hospital, S-112 81 Stockholm, Sweden. 0034-5687/85/$03.30 © 1985 Elsevier Science Publishers B.V. (Biomedical Division)
266
B. ROBERTSONet al.
epithelial barrier of the fullterm lamb lung is of the order of 0.5-0.6 nm (Normand et al., 1971). This means that the epithelium is impermeable to protein and that, therefore, the presence of protein in the alveolar spaces must be the result of pinocytosis, secretion, aspiration, or leakage following disruption of the epithelium. At birth, the permeability of the air-blood barrier depends on the maturity of the lung and on the transpulmonary pressure gradient (Egan et aL, 1975). Immature, surfactantdeficient lamb lungs, which require a high insuffiation pressure and tend to collapse with each expiration, are characteristically 'leaky', with a large bidirectional flux of protein across the pulmonary epithelium (Adams et al., 1971 ; Jobe et al., 1983, 1985; Normand et al., 1970). This leak may be largely due to disruption of the epithelium in bronchioles and alveoli during the ventilatory movements. Studies of premature newborn rabbits have revealed that epithelial lesions develop within a few minutes of artificial (Nilsson et al., 1980) or spontaneous (Nilsson and Robertson, 1985) ventilation. One important consequence of increased epithelial permeability may be the appearance of a surfactant inhibitor protein in the alveolar spaces (Ikegami et al., 1984). The purpose of the present study is to quantify the protein leak in the lungs of artificially ventilated immature newborn rabbits. In particular, we were interested in whether the protein leak could be modified by surfactant replacement as previously shown in lambs (Jobe et aL, 1983), and whether the magnitude of the protein flux would correlate with variations in lung mechanics.
Material and methods
The studies were carried out on preterm newborn rabbits obtained on day 27 of gestation (term = 31 days). The pregnant doe was anesthetized with intravenous pentobarbital and ketamine and the fetuses delivered by hysterotomy. Immediately after birth, a cannula was tied into the trachea of each pup, and alternate pups received either vehicle or surfactant by instillation via the tracheal cannula (see below). They were then transfered to a system of multiple body plethysmographs, heated to 37 °C, paralysed with intraperitoneal pancuronium bromide (0.2 mg/ml), and each litter was ventilated in parallel for 60 min with a pressure-limited ventilator (Infant Ventilator, Sechrist Industries, Anaheim, CA, U.S.A.) delivering 100~ 02 at a frequency of 40 breaths/rain and 50~o inspiration time. Tidal volumes were monitored in each animal with a Fleisch-tube connected to the body plethysmograph (Lachmann et al., 1981). Insufflation pressure was adjusted for each litter to provide an average tidal volume of about 10 ml/kg. After the period of artificial ventilation, ECG recordings were obtained. Blood gases were measured at sacrifice using samples aspirated from the beating heart. Since these samples might have been obtained from either the left or the right ventricle, Pco2 and pH only are reported. The blood samples were also used for determination of hemoglobin and radiolabel content (see below).
A n i m a l experiments; general procedure.
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
267
Preparation and in-vitro evaluation of surfactant. Surfactant was prepared from saline extracts of minced bovine lungs. After filtration of the saline phase, the lipids were extracted with chloroform : methanol 2 : 1 (v/v), and separated by liquid-gel chromatography (LipidexR-5000) into phospholipids, cholesterol, triglycerides and cholesteryl esters. The phospholipid fraction - which also contains about 1 ~o protein with a large proportion ofhydrophobic amino acids - w a s lyophilized and stored at - 18 ° C. Shortly before the experiment, the surfactant material was suspended in saline at a final phospholipid concentration of 80 mg/ml. It was carefully mixed with commercially available 125I-labeled human serum albumin (Mallinckrodt, St. Louis, MO, U.S.A.), added at a concentration of 0.84 mg/ml (8.5/~Ci/mg). The phospholipid composition of our surfactant has been reported elsewhere (Nilsson et al., 1985). Its physical properties were tested with pulsating bubble at room temperature as described by Ikegami et al. (1983). When suspended at a concentration of 3 #mol phosphatidylcholine/ml, our surfactant reduced surface tension to 0 mN/m during 56~o surface compression (rate 16/min). This indicated that although our surfactant preparation only represented one fraction of the complex material usually referred to as 'natural surfactant', it was suitable for further evaluation in vivo.
Surfactant replacement and artificial ventilation. Before the onset of artificial ventilation, experimental animals were treated via the tracheal cannula with 70 #1 surfactant containing 0.5/~Ci ~25I-albumin. Control animals received the same volume of normal saline containing 0.5/~Ci ~zSI-albumin. Concurrently with the onset of artificial ventilation, all animals were injected via an external jugular vein with 0.2 ml saline containing 0.5/~Ci 13q-labeled albumin, made by iodinating bovine serum albumin monomer standard (Miles Laboratories, Elkhart, IN, U.S.A.) (Hunter and Greenwood, 1962). We knew from previous experiments (Berggren et al., 1985; Nilsson et al., 1978) that treatment of-the premature newborn rabbit with natural surfactant leads to a striking improvement in lung-thorax compliance. This was confirmed in pilot studies with the present surfactant preparation (data not shown) indicating that control animals required insuffiation pressures of 25-30 cm H 2 0 to stay alive for 60 min, while such high pressures inevitably caused overinflation of the lungs in surfactant-treated animals, leading to pneumothorax and death. Since one of our basic aims was to maintain proper ventilation during the course of the experiment, we could not ventilate the surfactanttreated animals and controls in parallel with the same pressure. We therefore chose to use alternate litters as surfactant recipients and controls. However, in each surfactanttreated litter we included two non-treated control animals, ventilated with the same 'low pressures'. Failure to obtain adequate tidal volumes in these control animals identified the litter as immature. Similarly, in each control litter, we included two surfactant-treated animals, ventilated with the same 'high pressures'. The regular occurrence of lung rupture in these treated animals confirmed a surfactant response. Our final data were thus obtained from 25 surfactant-treated animals from 6 litters,
268
B. ROBERTSONet al.
ventilated with 'low pressures', and 21 control animals from 6 other litters, ventilated with 'high pressures'. Processing of lungs. The lungs were lavaged in situ with normal saline. The chest was opened and the lungs were filled to total lung capacity by visual inspection; this saline was flushed in and out of the airways three times. This procedure was repeated 4 times, and the recovered lavage fluid was pooled for quantitation of 125I, 131Iand total protein. The lungs were then removed, weighed, and homogenized in distilled water. Aliquots of the homogenate were assayed for the same labels as analysed in the lavage fluid. From the homogenates and the alveolar washes, aliquots were also taken to quantify phosphatidylcholine content by thin-layer chromatography followed by phosphorus determination according to Bartlett (1959). In order to define the contribution of intravascular blood to the radioactivity measured in the homogenates, part of the homogenate was centrifuged at 27 000 x g, and the resulting red supernatant was used for analysis of hemoglobin. We also assayed the hemoglobin content and radioactivity of the blood sampled from the heart at the end of the experiment. By combining the information from these analyses, and assuming that all hemoglobin in the lung homogenate was intravascular, we could determine the contribution of radioactivity in blood to that in the total lung homogenate. This contribution was subtracted in our calculation of radioactivity associated with lung tissue (Jobe et al., 1983). Surfactant inhibitors in alveolar wash. Equal volumes of alveolar wash from surfactanttreated animals were combined into 4 pools. Three of these pools represented separate litters; the fourth was obtained by combining lavage fluid from surfactant-treated animals in the remaining 3 litters. Four similar pools were obtained from control litters. The lavage fluid was centrifuged at 8000 x g and the protein content of the supernatant measured according to Lowry et al. (1951), using bovine serum albumin as standard. Minimum surface tension of our surfactant preparation was determined with the pulsating bubble as described above, in the presence of varying amounts of supernatant protein (0, 0.1, 0.3, 0.6 mg/ml). In these analyses, the phosphatidylcholine content of the surfactant preparation was kept constant, 3/~mol/ml. We also determined the minimal surface tension of the pooled samples of crude alveolar wash, and of surfactant isolated from the same samples by centrifugation at 8000 x g over 0.7 M sucrose in saline. Histologic studies. Six randomly selected animals (3 surfactant-treated, 3 controls) were sacrificed after 60 min of artificial ventilation by transsection of the abdominal aorta. The lungs were then ventilated with air for 3 min, the trachea was clamped at end-inspiration, and the lungs were fixed by immersion in formalin. Paraffin sections from the lower lobes were examined by conventional light microscopy, with particular reference to the alveolar expansion pattern and the extent of epithelial lesions and hyaline membranes.
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
269
Statistical evaluation. Values are given as means + SD. The Wilcoxon two-sample, two-tailed test was applied for evaluation of statistical differences, the limit level of significance being defined as P = 0.05.
Results
Survival, blood gases, lung mechanics. All the 25 surfactant-treated animals ventilated with 'low pressure' and 18 of the 21 controls ventilated with 'high pressure' had regular cardiac activity with frequency > 200/min at the end of the experiment. Body weight, data on lung mechanics after 30 and 60 min, cardiac frequency, and blood gases in survivors are shown in table 1. The figures for tidal volume show no difference between surfactant-treated animals and controls, although the volumes were generally higher than intended at the 30 min interval. After 60 min, the average tidal volume of both groups was close to the desired value of 10 ml/kg. The pressure required to maintain this tidal volume was about 40-50~o lower in the treated group, reflecting improved compliance of the lung-thorax system throughout the period of observation. In both groups, mean values for compliance were lower at 60 min than at 30 min, but this difference was statistically significant only for the surfactant-treated animals. Most animals had normal heart rates and blood gases prior to sacrifice, without differences between the groups, indicating that they were ventilated adequately. Protein leakage. The percent recovery of the 125I-albumin administered into the airways was significantly higher in alveolar wash and lung tissue from animals receiving surfactant than in control samples (fig. 1A), and the percent recovery of the intravenously injected 13q-albumin was lower in alveolar wash from surfactant-treated animals than in controls (fig. 1B). However, the percentage of 131I associated with lung tissue was similar in the two groups (fig. 1B), as were the values for total protein content of alveolar wash (table 2). If the percentage of 1311 in alveolar wash is related to variations in tidal volume, the trends are different in surfactant-treated animals and controls. The amount of label in alveolar wash tends to be inversely related to tidal volume in the surfactant-treated group (y = 14.6-2.73x; r = - 0.39), whereas the opposite seems to be the case among controls (y = 1.46x + 6.23; r = 0.39). However, neither of these correlations is statistically significant. Phospholipids and surfactant inhibitors in lung wash. The phosphatidylcholine content of alveolar wash was 17 times higher in animals treated with surfactant than in controls (table 2). Similarly, the supernatant of alveolar wash (used for assay of surfactant inhibitors) contained significantly more phosphatidylcholine in surfactant-treated animals than in controls (table 2). This difference is, however, small in relation to the total concentration of bovine surfactant (3/~mol/ml) suspended in the samples assayed
270
B. ROBERTSON
e t al.
TABLE 1 Body weight and data on lung mechanics, heart rate, and blood gases in surfactant-treated animals and controls. Abbreviations: VT = tidal volume, P = peak insufflation pressure, C = compliance (VT/P). Experimental condition
Surfactant-treated (n = 25) Controls (n = 21) P< * P vs
% .125
Body weight
30 min
(g)
VT
P
C
(ml/kg)
(cm H20 )
(ml/cm H20.kg )
31 + 6
15.9 + 7.2
14.5 _+2.6
1.09 _+0.39
30 + 5 NS
13.8 + 4.8 NS
26.6 + 2.1 0.002
0.51 _+0.15 0.002
30 min <0.02.
[]
Surfactant-treated
[]
Controls
-K- P <
n=19
n = 16
0.002
80
% I TM
60
[]
Surfactant-treated
[]
Controls P <
40
El'l
20
4
n=19
n = 16
0002
L
3 2 1 0
A l v e o l a r wash
A
A l v e o l a r wash
Lung t i s s u e
Lun( t i s s u e
B
Fig. 1. Percentage of '251-albumin (A) and 131I-albumin (B) in alveolar wash and lung tissue of surfactanttreated animals and controls. 125I-albumin was mixed with bovine surfactant and administered via the airways. ~3q-albumin was injected intravenously with the onset of artificial ventilation.
for inhibitory activity. T h e p r o t e i n c o n t e n t o f the s u p e r n a t a n t w a s similar in s u r f a c t a n t t r e a t e d a n i m a l s a n d c o n t r o l s (table 2). T h e inhibitory activity o f s u p e r n a t a n t f r o m the a l v e o l a r w a s h o f the t w o g r o u p s o f a n i m a l s is s h o w n in fig. 2. A t all three c o n c e n t r a t i o n s o f s u p e r n a t a n t protein, the inhibitory activity is m u c h higher ( m i n i m a l s u r f a c e t e n s i o n is higher) in the c o n t r o l g r o u p (P = 0.05); this activity i n c r e a s e s w i t h the c o n c e n t r a t i o n o f s u p e r n a t a n t protein.
271
PROTEIN L E A K A G E IN THE I M M A T U R E RABBIT LUNG
Heart rate (min 1)
60 min
(ml/kg)
P (cm H20 )
C (ml/cm H 2 0 . k g )
10.8 _+ 5.5*
14.6 _+ 2.6
0.75 + 0.35*
11.3+_4.5 NS
24.8+1.1 0.005
0.45+0.18 0.01
VT
Blood gases pH
Pco2 (mm Hg)
337 + 34
7.29 + 0.23
43.1 _+ 16.1
323_+73 NS
7.23_+0.18 NS
37.5_+12.2 NS
Inhibitory Activity ~ 'rnin , mN/m '
20 15
/
10.
/ /~
'
o.1
~ J q[ Controls n :4
~
i Surfactant/J. Treated
'
6.3
'
'
d.6 mg
Protei~/ml
Fig. 2. Inhibitory activity in supernatants of alveolar wash from surfactant-treated animals and controls. Values for minimal surface tension (7 min) were measured with the pulsating bubble at various protein concentrations. Differences are significantly significant for protein concentrations 0.1-0.6 mg/ml (e = 0.05). TABLE 2 Total phosphatidylcholine (PC) and protein in alveolar wash and in alveolar wash supernatant used for assay of inhibitory activity (see fig. 2). Parameter
Alveolar wash PC (/tmol/kg) protein (mg/kg) Supernatant PC (#mol/ml) protein (mg/ml)
Surfactanttreated animals
94+25 183+84
Controls
(n=25) (n=25)
0.18 + 0.05 (n = 4) 0.76 + 0.10 (n = 4)
5.4+2.8 163+51
P
(n=21) (n=21)
0.03 _+ 0.01 (n = 4) 0.91 + 0.33 (n = 4)
<0.005 NS =0.05 NS
Fig. 3. Histological sections illustrating the effect of surfactant replacement on alveolar expansion and airway epithelium. (A)Control animal with nearly collapsed alveoli and prominent desquamation of bronchiolar epithelium (arrow). (B) Surfactant-treated animal with significantly improved, yet not quite uniform alveolar expansion pattern. Note intact epithelium in airway (*). Hematoxylin and eosin, x 150.
PROTEIN LEAKAGEIN THE IMMATURERABBITLUNG
273
Surfactant isolated from the pooled alveolar washes of treated animals at a concentration of 4/~mol phosphatidylcholine/ml reduced surface tension to 0 mN/m during surface compression in all four samples analysed, indicating that the surface activity was unchanged in comparison with that of the material originally instilled into the airways (see Material and methods). Histologicfindings. Among the six animals studied histologically, the three treated with surfactant had improved alveolar air expansion and less prominent epithelial lesions in comparison with controls, confirming earlier observations under similar experimental conditions (Berggren etal., 1985; Nilsson etal., 1978). There was no evidence of pulmonary hemorrhage. Representative sections from the two groups of animals are illustrated in figs. 3A and B.
Discussion
Our data show that immature newbom rabbits can be kept alive with adequate blood gases for at least 1 h, provided that the ventilation setting is adjusted with respect to the compliance of the lungs. However, whereas surfactant-treated animals survived the period of ventilation without serious lung lesions, control animals characteristically developed widespread desquamation of airway epithelium, with formation of hyaline membranes in bronchioles and alveoli. In all animals there was a bidirectional leakage of protein, but the amount of 13q-albumin entering the airspaces was about twice as large in the control group. Similarly, the loss of 125I-albumin from the airspaces was significantly more rapid in non-treated animals than in those receiving surfactant. Most likely, the increased protein leakage in control animals can be attributed to disruption of the epithelial and endothelial barriers of the peripheral airspaces. Movements of liquid, electrolytes and larger molecules within the lung are usually analysed by means of a three-compartment model (Strang, 1976) comprising the intravascular, interstitial, and intra-alveolar spaces. In the intact fetal lung, the walls separating these compartments have different permeabilities. Thus, the 'pore equivalent radius' of the alveolar capillary endothelium is of the order of 15 nm (Normand et al., 1971), whereas - as mentioned above - that of the alveolar epithelium has been estimated to be about 0.6 nm. This means that, under normal conditions, the endothelial wall is permeable to albumin molecules (diffusion radius 3.4 nm), whereas the alveolar epithelium is not (Olver et al., 1981). The significance of our present observations can be understood from the three-compartment model, taking into account that the epithelial wail is 'leaky' in the ventilated immature lung (Adams et al., 1971 ; Egan et al., 1980, 1984; Jobe et al., 1983, 1985; Normand et al., 1970) and that the permeability of this wall is further increased in control animals by necrosis and disruption of cell junctions. In both surfactant-treated animals and controls, ~25I-labeled albumin instilled into the airways moved from the intra-alveolar compartment (represented by the lung lavage
274
B. ROBERTSONet al.
fluid) into the interstitial space (represented by lung tissue minus blood). However, whereas in control animals altogether about 65~o of this protein permeated into the intravascular compartment and further on to extrapulmonary tissues, nearly 100~o remained lung-associated in surfactant-treated rabbits. The difference in ~25I-albumin content of alveolar wash, shown in fig. 1A, is most probably the result of increased permeability of the epithelial membranes in the control group. The larger amount of the same label in the lung tissue of surfactant-treated animals is less readily understood. To some extent, this latter difference may reflect less effective lung lavage in the surfactanttreated group. An alternative, and perhaps more likely, explanation is that control animals develop an increased protein leakage across the endothelial membrane as well. This would lead to accelerated equilibration between the pulmonary interstitial compartment on one hand, and the intravascular compartment on the other. The observed leakage of protein into the alveolar spaces can be interpreted on the basis of similar assumptions. The fraction of intravascularly injected t31I-albumin recovered in the interstitial space (represented by lung tissue) was the same in surfactant-treated animals and controls, about 3 ~o. This lack of difference may seem at variance with the increased transendothelial protein leakage documented in control animals with the other label. We believe, however, that the ~31I-albumin moving into the interstitial space of control animals rapidly permeated further into the intra-alveolar spaces, leading to accelerated equilibrium between these two compartments. This interpretation is in keeping with the increased permeability of epithelial membranes to 125I-albumin (fig. 1A), as well as with the figures showing enhanced accumulation of ~3q-albumin in the alveolar spaces of the control group (fig. 1B). An increased transendothelial leakage of protein is further indicated by our finding that the amount of 13li_albumin in the combined interstitial and intra-alveolar spaces is about 50 ~o higher in control animals than in the surfactant-treated ones (fig. 1B). Our observation that the total amount of protein, harvested by alveolar washing, was similar in surfactant-treated animals and controls does not invalidate the data shown in fig. 1. Some of these proteins were probably present in the fetal lung liquid before the onset of ventilation. Furthermore, the apparent paradox can be solved by taking into account that there is a significant leakage of protein in both groups of animals but that - for reasons outlined above - the permeation of protein across the endothelial and epithelial membranes is more rapid in controls. This would lead to a faster equilibration between the three compartments in control animals. The net result over a limited period of ventilation could then be a similar intra-alveolar accumulation of total protein in both groups, with a delayed intra-alveolar appearance of the intravenously injected label in animals treated with surfactant. The leakage of protein in immature lungs, documented in the present and other (Jobe et al., 1983, 1985) studies by means or radiolabeled albumin, has potential clinical implications (Jefferies et al., 1984). At least one of the proteins entering the alveolar spaces is a potent inhibitor of pulmonary surfactant (Ikegami et al., 1983 ; Jobe et al., 1983). This particular protein, recently isolated by Ikegami et al. (1984), has a molecular size of 110 000 daltons. As shown in fig. 2, the inhibitory activity of protein in alveolar
PROTEIN LEAKAGE IN THE IMMATURE RABBIT LUNG
275
w a s h was significantly higher in the control group. I n c r e a s e d inhibitory activity with loss o f surfactant function leads to g r a d u a l aggravation o f the epithelial lesions, with further leakage o f inhibitory proteins into the airspaces, etc. - a vicious circle p r o b a b l y involved in the pathogenesis o f respiratory insufficiency in immature n e w b o r n babies with surfactant deficiency. Several previous studies have shown that surfactant replacement has a beneficial influence on gas exchange and lung mechanics in clinical and experimental respiratory distress s y n d r o m e (for review, see Jobe, 1984; R o b e r t s o n , 1980, 1983, 1984). T h e s e effects can largely be attributed to r e d u c e d surface tension in the a i r - l i q u i d interfaces o f the lung, with i m p r o v e d aeration o f the alveolar c o m p a r t m e n t . D a t a from the present study indicate that this therapeutic a p p r o a c h might improve lung function also by reducing the leakage o f surfactant-inhibiting proteins into the alveolar spaces.
Acknowledgements This study was s u p p o r t e d by The Swedish M e d i c a l R e s e a r c h Council (Project No. 3351), T h e Swedish N a t i o n a l A s s o c i a t i o n against H e a r t and Chest Diseases, The Swedish Society o f M e d i c a l Sciences, The R e s e a r c h F u n d s o f the K a r o l i n s k a Institute, a n d G r a n t H D - 1 1 9 3 2 from Child H e a l t h a n d D e v e l o p m e n t , D e p a r t m e n t o f Health and H u m a n Services, U . S . A .
References Adams, F.H., T. Fujiwara and H. Latta (1971). 'Alveolar' and whole lung phospholipids of premature newborn lambs. Biol. Neonate 17: 198-218. Bartlett, G.R. (1959). Phosphorus assay in column chromatography. J. Biol. Chem. 234: 466-468. Berggren, P., T. Curstedt, G. Grossmann, R. Nilsson and B. Robertson (1985). Physiological activity of pulmonary surfactant with low protein content; effect of enrichment with synthetic phospholipids. Exp. Lung Res. 8: 29-51. Egan, E.A., R.E. Olver and L.B. Strang (1975). Changes in nonelectrolyte permeability of alveoli and the absorption of lung liquid at the start of breathing in the lamb. J. Physiol. (London) 244: 161-179. Egan, E.A., R.M. Nelson and E.F. Beale (1980). Lung solute permeability and lung liquid absorption in premature ventilated fetal goats. Pediatr. Res. 14: 314-318. Egan, E.A., W.P. Dillon and S. Zorn (1984). Fetal lung liquid absorption and alveolar epithelial solute permeability in surfactant deficient, breathing fetal lambs. Pediatr. Res. 18: 566-570. Hunter, W.M. and F.C. Greenwood (1962). Preparation of 13q-labeled human growth hormone of high specific activity. Nature (London) 194: 495-496. lkegami, M., H. Jacobs and A. Jobe (1983). Surfactant function in the respiratory distress syndrome. J. Pediatr. 102: 443-447. Ikegami, M., A. Jobe, H. Jacobs and R. Lam (1984). A protein from the airways of premature lambs that inhibits surfactant function. J. Appl. Physiol. 57:1134-1142. Jefferies, A.L., G. Coates and H. O'Brodovich (1984). Pulmonary epithelial permeability in hyalinemembrane disease. New Engl. J. Med. 311: 1075-1080. Jobe, A., M. Ikegami, H. Jacobs, S. Jones and D. Conaway (1983). Permeability of premature lamb lungs to protein and the effect of surfactant on that permeability. J. Appl. Physiol. 55: 169-176.
276
B. ROBERTSON et al.
Jobe, A. (1984). Respiratory distress syndrome - new therapeutic approaches to a complex pathophysiology. Adv. Pediatr. (Chicago) 30: 93-130. Jobe, A., H. Jacobs, M. Ikegami and D. Berry (1985). Lung protein leaks in ventilated lambs: effect of gestational age. J. Appl. Physiol. 58: 1246-1251. Lachmann, B., G. Grossmann, J. Freyse and B. Robertson (1981). Lung-thorax compliance in the artificially ventilated premature rabbit neonate in relation to variations in inspiration : expiration ratio. Pediatr. Res. 15: 833-838. Lowry, O.H., N.J. Rosebrough, A.L. Farr and R.J. Randall (1951). Protein measurement with the Folin phenol reagent. J. Biochem. 193: 265-275. Nilsson, R., G. Grossmann and B. Robertson (1978). Lung surfactant and the pathogenesis of neonatal bronchiolar lesions induced by artificial ventilation. Pediatr. Res. 12: 249-255. Nilsson, R., G. Grossmann and B. Robertson (1980). Bronchiolar epithelial lesions induced in the premature rabbit neonate by short periods of artificial ventilation. Acta Pathol. Microbiol. lmmunol. Scand. (.4) 88: 359-367. Nilsson, R., P. Berggren, T. Curstedt, G. Grossmann, G. Renheim and B. Robertson (1985). Surfactant treatment and ventilation by high frequency oscillation in premature newborn rabbits; effect on survival, lung aeration and bronchiolar epithelial lesions. Pediatr. Res. 19: 143-147. Nilsson, R. and B. Robertson (1985). Bronchiolar epithelial lesions in spontaneously breathing immature newborn rabbits. Biol. Neonate (in press). Normand, I. C. S., E. O. R. Reynolds and L. B. Strang (1970). Passage of macromolecules between alveolar and interstitial spaces in foetal and newly ventilated lungs of the lamb. J. Physiol. (London) 210: 151-164. Normand, I.C.S., R.E. Olver, E.O.R. Reynolds, L.B. Strang and K. Welch (1971). Permeability of lung capillaries and alveoli to nonelectrolytes in the foetal lamb. J. Physiol. ~London) 219: 303-330. Olver, R. E., E.E. Schneeberger and D.V. Waiters (1981). Epithelial solute permeability, ion transport and tight junction morphology in the developing lung of the fetal lamb. J. Physiol. (London) 315: 395-412. Robertson, B. (1980). Surfactant substitution; experimental models and clinical applications. Lung 158: 57-68. Robertson, B. (1983). Lung surfactant for replacement therapy. Clin. Physiol. 3:97-110. Robertson, B. (1984). Choosing the surfactant for replacement therapy. In: Current Concepts in Surfactant Research, edited by P. yon Wichert. Prog. Respir. Res. Vol. 18. Basel, Karger, pp. 240-250. Schneeberger, E.E. (1976). Ultrastructural basis for alveolar-capillary permeability to protein. In: Lung Liquids, edited by R. Porter and M. O'Connor. Ciba Foundation Symposium 38. Amsterdam, Elsevier, pp. 3-21. Strang, L.B. (1976). The permeability of lung capillary and alveolar walls as determinants of liquid movements in the lung. In: Lung Liquids, edited by R. Porter and M. O'Connor. Ciba Foundation Symposium 38. Amsterdam, Elsevier, pp. 49-58.