INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

LUNGS ARE POTENTIALLY very susceptible to fluid disturbances because they have a large surface area, high blood flow, and relatively leaky capillary endothelial cells. However, the amount of fluid in the lung is very tightly regulated to enable adequate exchange of gases. It is generally agreed that in the mammalian lung there is a small but constant leakage of fluid from the pulmonary capillaries into the lung interstitium. This transudate is continuously removed by the lymphatic circulation (19). It is believed that fluid accumulation on the inside of the lung, i.e., within the alveolar air space, is prevented by the presence of pulmonary surfactant (PS) (6, 11, 15). The monolayer of surfactant lipids at the air-liquid interface lowers and also varies the interfacial surface tension with changing radii of curvature (both of different alveoli and within different regions of an alveolus), such that the negative fluid pressure in the liquid lining of the alveoli is balanced and minimized, thereby preventing the net flux of interstitial fluid into the alveolar air space (11). When the PS system is impaired, the alveolar surface tension increases and pulmonary edema results (4, 14). Furthermore, surfactant replacement can reverse the pulmonary edema that results from lung injury or surfactant dysfunction (12).

However, to our knowledge, the role of PS in fluid balance in nonmammalian vertebrates has not been examined. Fluid balance studies of reptilian and amphibian lungs suggest that these lungs could be prone to fluid disturbances, because they possess a comparatively high degree of pulmonary capillary filtration (at least an order of magnitude higher than mammals), presumably resulting from both higher capillary pressures and leakier capillary membranes (5, 18). Furthermore, lungs of reptiles possess little interstitial tissue, and the capillaries bulge into the air spaces (13, 16), possibly increasing the potential for alveolar flooding. However, when extreme pulmonary vascular fluid filtration was induced in the marine toad by bilateral denervation of the recurrent laryngeal nerves, there was no evidence of either interstitial or alveolar pulmonary edema (18).

Reptiles also possess 7-70 times more PS per square centimeter of respiratory surface area than mammals (9). We have shown in numerous studies that reptilian PS functions as an anti-glue to prevent adjacent epithelial surfaces of lungs from sticking together at low lung volumes (7-10). The removal of surfactant significantly elevates lung opening pressures but has only a minor effect on the compliance of these unicameral lungs (7, 9, 10, 20, 21). At this time, the anti-glue feature is the only validated function of surfactant in reptilian lungs. However, it is possible that PS also plays a major role in the prevention of pulmonary edema. An alveolar fluid lining with exceptionally low surface tension (provided by surfactant) could reduce the negative pressure in the hypophase, potentially affecting both fluid extravasation from the capillary and the elimination of excess fluid in the hypophase.

The present study directly tests the role of reptilian PS first in the entry of vascularly derived fluid into the alveolar air space and second in the removal of accumulated alveolar fluid by the pulmonary circulation and the pulmonary lymphatic system. We first introduced fluorescently labeled inulin [fluorescein isothiocyanate (FITC)-inulin] as a marker of bulk fluid flow into the perfusate of an isolated lung from the bearded dragon lizard, Pogona vitticeps. We initiated pulmonary fluid filtration by raising venous pressure both in the presence and absence of PS to investigate the flux of vascularly derived fluid into the alveolar space of the lung. In the second part of our study, we instilled FITC-inulin into the lung air space of the red-eared turtle, Trachemys scripta, both in the presence and in the absence of PS to investigate the pathways responsible for removing accumulated alveolar fluid from the lung.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals and Animal Care

Part 2. Red-eared turtles (Trachemys scripta) (mass range: 1,055-1,736 g, n = 11) were obtained from a commercial supplier (Lemberger, Oshkosh, WI) and housed in the animal care facility in the Department of Biology at the University of Texas at Arlington. Turtles were kept in 1.5-m-diameter water tanks at a water temperature of 21°C and fed pieces of beef heart weekly. Animals were kept in captivity for a period of 2-8 wk before experimentation.

Experimental Procedure

The lung vasculature was perfused with a mixture of equal portions of autologous blood and Hemaccel (Behring Institut, Marburg, Germany) propelled by a peristaltic pump (model RL175-110, Extracorporeal Medical specialists). Atropine sulfate (Sigma Chemical) was added to the perfusate to approximate a 10⁴ M concentration to cause maximal dilation of pulmonary arterial vasculature. The venous outflow was directed into a reservoir suspended from a Grass FT-03 force transducer. This venous reservoir, constantly mixed by magnetic stirring, was the source for reperfusion by the peristaltic pump, and its mass measured on a continuous basis was the direct indicator of lung capillary filtration (mass loss) or absorption (mass gain). Arterial and venous perfusion pressures were measured with Statham-Gould (model P23 DJc) pressure transducers modified for flow-through design and placed within 5 cm of the common pulmonary artery and vein. Venous reservoir mass and pulmonary arterial and venous pressure were continuously recorded on a Grass polygraph (model 7C). FITC-inulin (molecular weight 2,000-5,000; Sigma Chemical) was added to the perfusate (1 ml of a 5 mg/ml solution) as a sensitive indicator of bulk fluid flow into the alveolar space. After the addition of FITC-inulin, the preparation was maintained isogravimetric by adjustment of arterial flow (pump velocity) and venous outflow pressure (P_v; height of animal with respect to height of venous reservoir) for 10 min before the first measurement of capillary filtration coefficient (CFC). During isogravimetric conditions, venous pressure was held at 0 to 1 cmH₂O. The established perfusate flow and resulting arterial pressure to maintain isogravimetric conditions averaged 1-2 ml/min and 25-30 cmH₂O, respectively. Body temperatures of all experimental animals approximated that of room temperature (22-24°C) throughout all procedures.

To test whether PS might influence net transcapillary fluid movement in the lizard lung, measurement of pulmonary CFC was performed three times on each lizard, once before lung lavage and two more times after each of two subsequent lung lavages (see below). After a 10-min isogravimetric state, a small blood sample (0.5 ml) was taken from the venous reservoir, and the P_v was sharply elevated by 12-14 cmH₂O by lowering the lizard with respect to the reservoir by the appropriate amount for 10 min. This increase in P_v initiated pulmonary capillary filtration (determined by a rapid, then gradual mass decrease of the reservoir). The time-dependent change in reservoir mass, taken as an average over the final 5 min, was normalized as a function of the change in P_v and the wet mass of the lungs, to express CFC in milliliters per minute per centimeter H₂O per 100 gram. Control experiments on three animals on which lung lavages were not performed confirmed that CFC did not change as a function of perfusion or experimental duration. Lungs were returned to an isogravimetric state for 5 min before the perfusion was halted for lung lavage. Lavage usually took 5 min.

Part 2. Turtles were anesthetized with an intraperitoneal injection of Nembutal (50 mg/kg body mass; Abbott Laboratories, Chicago, IL). A 5- to 6-cm-diameter hole was drilled into the plastron to expose the heart. The trachea was cannulated and the lungs were ventilated (6 breaths/min, tidal volume = 50 ml) with a CWE ventilator (SAR-830, Ardmore, PA). After heart puncture, ~30 ml of blood was removed for later use in lung perfusion. Atropine sulfate (10⁴ M, Sigma Chemical) and 10⁴ M isoproterenol (Sigma Chemical) were injected intracardially in 5 ml of 3.7% bovine serum albumin (BSA) in Ringer to cause maximal dilation of pulmonary vasculature. A 1.5-cm-diameter hole was then drilled into the left side of the carapace (Dremel Moto-Tool, Racine, WI), enabling access to the anterior and dorsal region of the left lung. To vascularly isolate the left lung for perfusion, the left pulmonary artery and vein were cannulated (PE-160, 1.14 mm ID) as close as possible to the point at which these vessels enter and exit the lung. The left subclavian artery was then cannulated (PE-60, 0.76 mm ID) for later sampling of systemic blood.

The left lung was perfused with a mixture of whole blood, extra red blood cells from a donor animal, and 3.7% BSA-Ringer. Between 20 and 30 ml of fresh autologous blood were obtained by heart puncture from the experimental animal. Washed and spun red blood cells from ~20 ml of blood from a donor animal were then added to the fresh blood. The perfusate was then made up to a total volume of 250 ml and an average hematocrit of 5-6% with 3.7% BSA-Ringer, containing heparin (100 units/ml), 10⁴ M atropine, and 10⁴ M isoproterenol. The perfusate was filtered through several layers of tissues to remove any blood clots and placed into the perfusate reservoir (50-ml plastic syringe) at a height of 2-3 cm above the animal preparation. The perfusate was constantly mixed by magnetic stirring and topped up with fresh perfusate to maintain a constant volume (and therefore constant arterial pressure) by a peristaltic pump. By subsequently calibrating the pump, we were able to determine the perfusion flow rate. With the perfusate at a height of 12-13 cm above the animal preparation, our gravity perfusion setup resulted in an average flow rate of 2-3 ml/min.

As a marker of fluid removal from the inner surface of the lung, we sprayed FITC-inulin into the lung. Using the hole drilled into the anterior carapace of the animal, we cauterized a 0.3-0.4-cm-diameter hole in the left lung, using a Birtcher hyfrecator (model 733). Immediately before inulin instillation, we took "time zero" samples of pulmonary venous effluent and subclavian arterial (systemic) blood. We sprayed an average of 0.1-0.2 ml of FITC-inulin (3.2 mg/ml) into the anterior region of the left lung, using a modified artist's airbrush. The inulin solution was instilled into the lung from a 1-ml syringe, the needle of which pierced the shaft of a 10-cm hard plastic tube (OD = 3.0 mm, ID = 0.5 mm) at a 45° angle halfway along the tube. The proximal end of the tube was connected to the artist's airbrush, which in turn was connected to the laboratory air propellant (20 lb/in.²). When the airbrush was activated, a constant stream of air passed through the small-diameter tube and into the lung. By slowly and simultaneously injecting the inulin solution from the syringe into the tube, a fine, aerosolized spray was delivered into the lung. We maneuvered the apparatus to vary the angle of spraying to cover as large an area of the inner surface of the lung as possible. As soon as the inulin was sprayed into the lung, the hole in the lung was tied off with cotton. Pulmonary venous effluent (1.5 ml) was usually sampled every 2 min for the first 15 min and thereafter every 5 min. Systemic blood (0.3-0.4 ml) was usually sampled from the subclavian artery every 5 min. The fluorescence in the systemic circulation was an indicator of the amount of inulin removed by the lymphatic circulation. Pulmonary perfusion and sampling continued for 60-70 min, and the lung was ventilated continuously. Samples were later analyzed for FITC-inulin fluorescence.

Lung Lavage and Analyses

FITC-inulin fluorescence in the lavages was quantified to determine the degree of fluid flux from the blood space to the alveolar space. An aliquot (3 ml) of fluid from each rinse of each lavage was placed in 5-ml fluorimetry tubes, and the relative fluorescence was measured on a Sequoia-Turner fluorimeter (model 450). We determined the concentration of FITC-inulin (µg/ml) in the three lavages using a standard curve of known amounts of FITC-inulin dissolved in Ringer. Prelavage fluid (0.15 M NaCl or 3.7% BSA) was used as a blank. Perfusate concentrations of FITC-inulin were also measured to confirm the constancy of perfusate FITC-inulin throughout the experiment. The 3-ml sample was returned to the entire lavage volume for subsequent analyses.

Estimation of PS removal from the lungs in the sequential lavages was accomplished by lipid extraction and phospholipid quantitation. Lavage samples were lyophilized to dryness, and the total lipids were extracted according to the method of Bligh and Dyer (3). Total phospholipids were quantified by phosphorus assay (1) and expressed as total PS phospholipids recovered per lavage, in micrograms per gram wet lung mass.

An aliquot (100 µl) from each lavage was analyzed for LDH content with the use of an LDH analysis kit for automated clinical chemistry analyzers (Boehringer Mannheim, Indianapolis, IN) and calculated as units of LDH per milliliter of lavage fluid. Perfusate (LDH) was also measured to provide an estimate of the LDH gradient between the blood and alveolar spaces. The protocol sequence of all measurements is illustrated in Fig. 1.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   Sequence of isogravimetric perfusion protocol for the bearded dragon (part 1). One group of lizards was lavaged 3 times with 0.9% saline, and a second group was lavaged 3 times with 3.7% bovine serum albumin (BSA)-Ringer. After each lavage, there was a 10-min period during which capillary filtration was elicited by raising pulmonary venous pressure. The capillary filtration coefficient (CFC) was measured each time, as was the amount of fluorescein isothiocyanate (FITC)-inulin fluorescence, of phospholipid, and of lactate dehydrogenase (LDH).

Part 2. We tested the effect of PS on the removal of fluid from the inner surface of the left lung of the turtle by removing the surfactant of six animals before inulin instillation. We tied off the right bronchus and then lavaged the left lung via the tracheal cannula. We used six separate 50-ml volumes of 0.9% saline, each instilled and withdrawn three times to remove as much of the PS as possible. After pulmonary perfusion, the left lung of all 11 animals was lavaged in an identical fashion to gain a measure of the amount of FITC-inulin remaining within the lung. The six rinses of the postperfusion lavage were centrifuged at 150 g for 5 min to remove cells and cellular debris. The volume of the supernatant was recorded, and the fluid was kept for later analysis of FITC-inulin fluorescence.

FITC-inulin fluorescence in pulmonary venous effluent, systemic blood, and lavage samples was measured on a Sequoia-Turner fluorimeter (model 450). All blood and perfusate samples were centrifuged to pellet red blood cells. The plasma supernatants of time zero samples of venous effluent and systemic blood were used as blanks for the samples taken after inulin instillation. Isotonic saline was used as the blank for lavage samples. In the case of the systemic blood samples, 200 µl of plasma was diluted with Ringer to a volume of 1 ml before fluorescence was measured. Two standard curves, one using Ringer and the other using 3.7% BSA-Ringer as the diluting solvent, were constructed to determine the concentration of FITC-inulin (µg/ml) in lavage and the diluted blood samples on the one hand, and pulmonary venous effluent on the other hand. The concentration of inulin in the six rinses of each lavage was extrapolated to the amount of inulin (in µg) present in the total lavage volume.

Data Analysis

Part 2. The amount of lavage inulin in the two experimental groups was compared with a one-way ANOVA. The inulin concentration (µg/ml) in the venous effluent and the systemic circulation was plotted against the time of perfusion. Because it was difficult to sample always at the same time in different animals, we constructed a regression curve of inulin concentration in venous effluent and systemic blood against time of perfusion for each animal. We then obtained the concentration of inulin at the predetermined time points for both venous effluent and systemic blood for each animal within the two experimental groups. We first compared the two curves using an analysis of covariance (ANCOVA). If this test revealed a significant difference between the two slopes, we further analyzed the curves with a one-way ANOVA with repeated measures, followed by Tukey's post hoc t-tests to reveal which time points were significantly different. The significance level was set at 0.05.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Part 1

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Measurements (means ± SE) of pulmonary CFC in ml · min1 · cmH2O1 · 100 g1 of wet lung tissue before each of 3 sequential lavages. Because there was no significant effect of lavage fluid used on the CFC [2-way analysis of variance (ANOVA), P = 0.83], data for the 2 groups of lizards were combined (n = 8). There was no significant effect of the sequential lavages on the measurements of pulmonary CFC (2-way ANOVA, P = 0.99).

There was no difference in the amount of surfactant phospholipid that could be removed from the lung, whether the lungs were lavaged with saline or with 3.7% BSA (2-way ANOVA, P = 0.67). After the data for the eight animals (4 saline lavage, 4 BSA lavage) were combined, there was a significant decrease in the amount of phospholipids between the first and second (t = 2.67, P <0.5) and the first and third lavages (t = 3.89, P < 0.01) (Fig. 3).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Amount of surfactant phospholipid (means ± SE) in µg phospholipid/g wet lung mass (WL) present in 3 sequential lavages. Because there was no significant effect of lavage fluid used (2-way ANOVA, P = 0.67), data from 2 groups of lizards were combined (n = 8). Pairs of * or # indicate significant decrease in amount of phospholipid (t = 2.67, P < 0.05, and t = 3.89, P < 0.01, respectively).

There was a measurable amount of FITC-inulin in all three lavages. The two-way ANOVA revealed no significant effect of the lavage fluid used (P = 0.92). When the data were combined, there was no significant difference in the amount of inulin between the first and second lavage but a significant increase between the second and third (t = 2.19, P < 0.05) and between the first and third lavage (t = 3.99, P < 0.01) (Fig. 4).

View larger version (9K): [in this window] [in a new window]   Fig. 4.   Amount of FITC-inulin (µg; means ± SE) in 3 sequential lavages. Because there was no significant effect of lavage fluid used on the amount of inulin that could be recovered from the alveolar space, data were combined (2-way ANOVA, P = 0.92). Pairs of * or # indicate significant increase in inulin (t = 2.19, P < 0.05, and t = 3.99, P < 0.01, respectively).

LDH concentration in lavage fluid was ~25 times lower than in the blood perfusate (lavage: 4.08 ± 0.53 U; plasma: 110.9 ± 32.4 U). There was no effect either of the lavage fluid used (2-way ANOVA, P = 0.39) or of the three sequential lavages (2-way ANOVA, P = 0.26) on LDH concentration (saline lavage 1: 4.0 ± 1.0 U, lavage 2: 3.3 ± 0.3 U, lavage 3: 6.3 ± 2.1 U, BSA lavage 1: 5.0 ± 0.7 U, lavage 2: 3.0 ± 0.0 U, lavage 3: 3.0 ± 0.0 U).

Part 2

Some of the inulin that was instilled into the lung appeared in the venous circulation after as little as 5 min in both experimental groups. The inulin concentration of the venous effluent increased with time in both experimental groups. However, the rate of increase (i.e., the slope) was smaller in the group of turtles that had been lavaged, as revealed by the ANCOVA (P = 0.0194) (Fig. 5). The Tukey's t-tests revealed no significant differences (t always <1.91) between the two experimental groups at any time point over the 60 min of perfusion (Fig. 5).

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Change in concentration of inulin ([inulin], µg/ml; means ± SE) recovered in pulmonary venous effluent of the 2 experimental groups of turtles (with surfactant vs. without surfactant) during the perfusion period. Analysis of covariance (ANCOVA) revealed a significant difference between the slopes of the 2 curves (P = 0.0194). ANOVA with repeated measures followed by Tukey's post hoc t-tests demonstrated no significant differences (t always <1.91) between the 2 experimental groups at any time point over the 60 min of perfusion.

The amount of inulin that appeared in the systemic circulation (sampled from the subclavian artery) was used as an indirect indicator of the amount of inulin that was removed from the alveolar compartment via the pulmonary lymphatic circulation. The inulin concentration ([inulin]) in the systemic circulation was an order of magnitude greater than that in the pulmonary venous circulation, and measurable amounts of inulin appeared after as little as 10 min in both experimental groups (Fig. 6). The [inulin] in the group without surfactant remained at very low levels throughout the 70-min perfusion period, whereas the [inulin] in the group of animals with the surfactant system intact increased significantly over time. ANCOVA revealed a significant difference in the slopes of the two curves (P < 0.0001). The Tukey's post hoc t-tests revealed that the systemic [inulin] in the group with the surfactant system intact was significantly greater than in the group lacking surfactant at all time points after 40 min of perfusion (40 min: t = 2.40, P < 0.05; 50 min: t = 3.01, P < 0.01; 60 min: t = 2.88, P < 0.01; 70 min, t = 3.68, P < 0.01) (Fig. 6).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Change in [inulin] (µg/ml; means ± SE) recovered in systemic circulation (used as indicator of inulin removal by pulmonary lymphatics) in the 2 experimental groups of turtles (with surfactant vs. without surfactant) during perfusion period. [Inulin] of the group of animals with the surfactant system intact increased significantly over time (ANCOVA; P < 0.0001) and from 40 min of perfusion was significantly greater than in group lacking surfactant (40 min: t = 2.40, P < 0.05; 50 min: t = 3.01, P < 0.01; 60 min: t = 2.88, P < 0.01; 70 min, t = 3.68, P < 0.01). * P < 0.05; ** P < 0.01.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Reptiles and amphibians are remarkable in their capacity for pulmonary plasma filtration (5) without the formation of either alveolar or interstitial pulmonary edema (18). Furthermore, reptilian and amphibian lungs contain between one and two orders of magnitude more PS per square centimeter of respiratory surface area than do those of mammals (9). Because the low surface tensions produced by PS in mammals are required to prevent pulmonary edema, it is possible that the large amounts of PS present in reptilian lungs fulfill a similar function. We therefore investigated the role of reptilian PS in the control of fluid movement between the alveolar, interstitial, and vascular compartments.

Validation of Methods

For the isolated perfused lung preparation, it was imperative that the pulmonary perfusion was completely isolated from the systemic perfusion. In both cases we checked whether a thebesian vessel system supplying the trachea and bronchi existed adjacent to the catheter site by perfusing the lungs with a vegetable dye. In the case of the bearded dragon, the system was not adjacent to the heart, thereby enabling us to cannulate the main pulmonary arterial and venous trunks. In the case of the turtle, there were smaller vessels that branched off the main pulmonary arterial trunk near the heart. Consequently, we bypassed these vessels by cannulating only the left pulmonary artery and left pulmonary vein, as close as possible to their points of entry to and exit from the left lung.

Because the lavage process only removes the surfactant that is present in the air space at that time and there are significant tissue stores that can be released subsequent to the lavage, we performed repeated lavages separated by 10-15 min (time needed to initiate fluid filtration and then restore isogravimetric perfusion). We wished to determine the effect of surfactant removal on fluid flux across the alveolar epithelium and were concerned that the act of the saline lavage may reduce the colloid osmotic pressure, thereby decreasing the normal flux of fluid into the alveoli. Alternatively, the consecutive lavages may damage the alveolar epithelium, thereby increasing fluid flux. In a second group of animals, we therefore lavaged the lungs with 3.7% BSA in 0.15 M NaCl (plasma protein concentration of the bearded dragon) to determine whether alveolar colloid osmotic pressure influenced fluid filtration. There were no differences in amount of surfactant PL or concentration of lavage inulin between the two experimental groups of bearded dragons. We also determined the concentration of LDH in the lavage fluid of all animals to establish whether there was substantial damage to the alveolar epithelium. The LDH concentration in the three lavages was similar.

The method used to instill inulin into the turtle lung was validated in pilot experiments by instilling a vegetable dye. This established that the inside of the anterior approximately one-third of the left lung received an even coating of inulin. Although it would have been preferable to coat the entire inner surface of the lung, we needed to balance the extent of the area sprayed with inulin and the pressure with which the spray was delivered. However, because the anterior chamber of the turtle lung is the largest and the anterior portion of reptilian lungs appears to contain the majority of the respiratory tissue (13, 16), it is the most likely portion of the lung to be involved in fluid uptake.

Part 1: Fluid Entry Into the Alveolus

Although the removal of surfactant resulted in a significant increase in the amount of inulin that could be removed from the alveolar compartment, the perturbation did not alter the amount of fluid leaving the capillaries, i.e., CFC did not change. It therefore appears that a perturbation of alveolar liquid pressure does not affect the permeability of the endothelial barrier, presumably because most of it is protected by the interstitium. However, reptiles contain very little interstitial tissue, and the capillaries are usually seen bulging very significantly into the alveolus (13, 16). Moreover, it is generally believed that interstitial liquid pressure is in equilibrium with alveolar liquid pressure (6, 11), and any perturbation of the alveolus should be transmitted to the interstitium. We expected such a transmission of the perturbation to be especially relevant in reptilian lungs. The accumulation of inulin in the alveolus following lavage, without an increase in net vascular fluid filtration, indicates that there is increased fluid movement from the interstitium into the alveolus when surfactant is absent (Fig. 7). Because the lavage medium (saline or BSA) did not affect the accumulation of alveolar inulin, the increased accumulation cannot be explained on the basis of a change in colloid oncotic pressure. It is likely that the virtually complete removal of surfactant increases the surface tension of the hypophase fluid, resulting in a more negative liquid pressure in the alveolus, causing more fluid to be drawn into this compartment from the interstitium. For example, in the dog lung, an increase in surface tension causes a decrease in alveolar liquid pressure (2). If alveolar liquid pressure and interstitial liquid pressure are in equilibrium, then the increase in surface tension should also lead to a more negative interstitial fluid pressure and hence an increase in fluid filtration. That an increase in filtration did not occur may reflect the very acute nature of our study. It is possible that we did not allow enough time to observe the equilibration between alveolar liquid pressure and interstitial liquid pressure. Therefore, a longer time period of filtration in the absence of PS may be required to elicit a change in CFC. Nevertheless, the removal of surfactant by lavage did increase the amount of fluid retained in the air spaces.

View larger version (33K): [in this window] [in a new window]   Fig. 7.   Schematic diagram indicating fluid movement between alveolar air space, capillaries, interstitium, and lymphatics and summarizing results of part 1 of the present study. Diagram includes part of an alveolar septum, a lymph vessel, a longitudinal capillary with arteriolar (a) and venous (v) ends, and a transverse capillary located within the interstitium. Hypophase fluid covers the alveolar surface, and surfactant is either present (A; before lavage) or absent (B; after lavage). Fluid movement that is known to occur is indicated by solid arrows, whereas fluid movement that is unknown or presumed is indicated by open arrows. Fluid is shown as leaving the capillary over its entire length because in this experimental protocol we raised venous pressure to induce capillary filtration. Experimental part 1 demonstrated that after lavage the layer of hypophase fluid is thickened because of increased fluid movement from the interstitium into the alveolar space (thick solid arrows in B). Extravasation of fluid from capillaries, however, did not change in absence of surfactant.

We have therefore established that vascularly derived fluid is able to enter the alveolar space during increased fluid filtration in the presence of PS. Furthermore, the almost complete removal of surfactant exacerbates this fluid flux and results in an accumulation of fluid in the air space (Fig. 7). Although we did not determine the presence of pulmonary edema in these bearded dragons, an increase in fluid filtration in the toad did not cause any alveolar or interstitial edema over 45 min (18). If vascularly derived fluid is therefore free to enter the alveolus, the question arises as to how the excess accumulated alveolar fluid is cleared and how PS influences the pathways.

Part 2: Fluid Exit From the Alveolus

Inulin from the alveolar compartment appeared in the pulmonary venous circulation 5 min after inulin instillation into the lung. Over the 60 min of perfusion, the pulmonary venous [inulin] increased at a slower rate in the group of turtles lacking surfactant compared with the group that had an intact surfactant system. This implies that the surfactant system is necessary for the effective removal of fluid from the alveolar compartment by the pulmonary venous circulation. Furthermore, over the 60-min time frame, a perturbation of alveolar liquid pressure is capable of affecting the flux of alveolar fluid into the vascular compartment. In contrast, in the bearded dragon (part 1), we found that a 10-min period of fluid filtration was not sufficient for a perturbation of the alveolar liquid pressure to influence the flux of vascularly derived fluid into the alveolar compartment, presumably because alveolar and interstitial liquid pressures had not equilibrated. It is possible that it takes at least 60 min for alveolar and interstitial liquid pressures to come into equilibrium. That there was no difference between the two experimental curves at any of the time points, even though the slopes were different, suggests that the time period over which we tested fluid flux was still not sufficient to indicate when the difference in venous inulin becomes significantly different between the two treatments.

It appears that the pulmonary lymphatic circulation is a more important route for alveolar fluid clearance than the pulmonary venous circulation, because the [inulin] was an order of magnitude greater in the systemic circulation than in the pulmonary venous circulation. However, because the systemic circulation is a closed system and the venous circulation is a nonrecirculating pool, we cannot be certain how the absolute amounts of inulin compare between these two pathways. Nevertheless, systemic [inulin] was greater in that group of lungs in which the surfactant system remained intact. Therefore, despite the very indirect manner in which we measured lymphatic inulin, we were able to determine that surfactant enhances the clearance of alveolar fluid via this pathway.

The two studies presented here have therefore clarified how surfactant controls the exit of fluid from the reptilian alveolus. The removal of surfactant in the bearded dragon caused an increase in fluid movement from the interstitium into the alveolus, presumably in response to the two pressures (alveolar and interstitial liquid) approaching equilibrium. Hence, there would be a "drying" effect of the interstitium. Although this was not tested in the turtle, there is no reason to expect that the turtle lung would behave any differently. On the contrary, the ultrastructure of turtle and lizard lungs is very similar (16), and also the pulmonary filtration rates of a range of nonmammalian vertebrates, including toads, lizards, and turtles, are all similar (17). The drying effect of the interstitium would have the consequence that fluid flux from the interstitium both to the pulmonary venous capillaries and to the lymphatic system would be greatly reduced (Fig. 8). This would explain the reduction in the clearance of inulin from the alveolar compartment by both pathways after the removal of the surfactant system.

View larger version (31K): [in this window] [in a new window]   Fig. 8.   Schematic diagram indicating fluid movement between alveolar air space, capillaries, interstitium, and lymphatics and summarizing results of part 2 of the present study. Structures and labels are the same as for Fig. 7. A, before lavage. B, after lavage. Experimental part 2 demonstrated first that fluid from the alveolar space was removed via the pulmonary venous circulation and that this pathway was reduced by the removal of surfactant. Second, the lymphatic circulation removed fluid from the alveolar space, and the amount of fluid was reduced in the absence of surfactant (indicated by dotted arrow in B). This was possibly due to the progressive drying of the interstitium, which would result in a reduction in lymphatic fluid flow (indicated by reduced lymphatic vessel in B).

Because there were significant levels of inulin in the pulmonary venous circulation after only 5 min of perfusion, even in the presence of surfactant, it is likely that there is a small but normal flux of fluid directly between the alveolar compartment and the pulmonary venous circulation, i.e., a pathway that does not necessarily involve the interstitial space. This possible pathway is further supported by the fact that in the lungs of nonmammalian vertebrates the capillaries bulge into the air space, with very little evidence for any interstitial tissue along the convex surface of the capillary (13, 16). Hence, it is also likely that the clearance of alveolar inulin directly by the pulmonary venous circulation is impaired after surfactant removal, because the increasingly negative alveolar liquid pressure would oppose the normal tendency for fluid to move from the alveolus into the pulmonary venous circulation. This would also contribute to the reduction in alveolar inulin clearance by the pulmonary venous circulation after surfactant removal. However, the relative contributions of the direct and indirect routes of alveolar fluid clearance by the pulmonary venous circulation cannot be determined.

In summary, we determined that reptilian PS functions to prevent excess fluid accumulation in the air spaces of the lung, thereby acting as an anti-edema agent. We found that increased pulmonary vascular fluid filtration caused by raising pulmonary venous pressure resulted in movement of fluid into the alveolar space both in the presence and absence of PS. The removal of surfactant did not affect the extent of fluid filtration over the 10-min time frame, but increased the flux of vascularly derived fluid into the alveolar space. This increased fluid flux must be occurring indirectly between the interstitium and the alveolus and could cause a drying of the interstitium (Fig. 7). We further demonstrated that the clearance of alveolar inulin by both the pulmonary venous circulation and the pulmonary lymphatic system was reduced after the removal of surfactant (Fig. 8). It appeared that the pulmonary lymphatic system represented the major pathway of alveolar inulin clearance, and it was more severely affected by the removal of surfactant. The mechanism by which the removal of surfactant reduced the amount of alveolar fluid cleared by both the lymphatic circulation and the pulmonary venous circulation is presumably related to the drying effect of the interstitium.

Perspectives