Simultaneous measurement of fluid and protein permeability in isolated rabbit lungs during edema PETER A. VINCENT, PAUL B. KREIENBERG, THOMAS M. SABA, AND DONALD R. BELL
FRED
Department of Physiology and Cell Biology, The Neil Hellman Albany Medical College, Albany, Nezu York 12208
VINCENT,PETER
A., PAUL B. KREIENBERG,FRED L. MINM. SABA, AND DONALD R. BELL. Simultaneous measurement of fluid and protein permeability in isolated rubbit lungsduring edema.J. Appl. Physiol. 73(6): 2440-2447,1992.Fluid conductance and protein permeability have been studied in isolated perfused lung modelsof pulmonary edema. However, previous studies have not investigated changesof both fluid conductance and protein permeability in the sameisolated lung preparation after injury. Arachidonic acid (AA) metabolites are involved in the inflammatory processesthat lead to the development of pulmonary edema. The hemodynamic effects of AA have beenwell established;however, controversy exists concerning the ability of AA to alter the permeability of the pulmonary microvasculature to fluid and protein. The purpose of this study was to simultaneously determine whether transvascular fluid conductance and protein permeability are increased in isolated perfused rabbit lungs with pulmonary edemainducedby AA. Indomethacin (80 PM) was addedto the perfusate to inhibit the hemodynamic effects of AA and producea pressure-independentmodelof pulmonary edema.Fluid conductancewas assessed by determination of the capillary filtration coefficient (KJ, and protein permeability wasevaluated by measurementof 1251-albumin clearance.The injection of AA (3 mg/200 ml of perfusate) into the pulmonary arterial catheter resulted in an increasein lung weight over the remaining 30min experimental period. Kf (~1 s-l cmH,O-’ g dry lung-l) was increased(P < 0.05) in AA-treated lungs at 10 and 30 min post-AA injection when comparedwith control lungsand baseline values (determined 10 min before AA injection). Albumin clearancewasalsogreater (P < 0.05) in lungs that received AA. 1251-albumin clearancewas measuredat different rates of fluid flux produced by elevation of venous pressure.Linear regression of albumin clearancevs. fluid flux was used to calculate the permeability-surface area product (y-intercept) and the reflection coefficient (1 - slope) for each group. The calculated permeability-surface area product was greater and the calculated reflection coefficient was lower in the AA-treated lungs than in control lungs. These results show that AA increases both protein permeability and fluid conductance but that the increasein protein permeability (117%) is larger than the increasein fluid conductance (42%). NEAR,
THOMAS
l
l
l
arachidonic acid; capillary filtration coefficient; albumin; permeability-surface area product; reflection coefficient
RESPIRATORY DISTRESS SYNDROME (ARDS) is characterized by the development of pulmonary edema that can lead to ventilation-perfusion mismatch and hypoxia (20). Pulmonary edema in ARDS is believed to result in part from an increased permeability of the lung
ADULT
2440
L. MINNEAR, Medical Research Building,
microvasculature (1, 20). Studies have investigated the permeability of the pulmonary microvasculature in isolated lung preparations utilizing techniques that assess either the capillary filtration coefficient (Kf) (2,16,28), a measure of fluid conductance, or albumin clearance (8), a measure of protein permeability. These techniques have also been used to determine changes in lung vascular permeability after lung injury models (9, 11, 12, 24, 27). However, they have not been used simultaneously to determine whether both protein permeability and fluid conductance are altered in the same lung in models of pulmonary edema. Kern and Malik (9) described a technique that uses the extravascular uptake of albumin to calculate the permeability-surface area product (PS) and the reflection coefficient (a) in isolated lungs. This method, however, requires that the lungs be isogravimetric, a condition that often does not exist in lungs that have an altered microvascular permeability. Indeed, when arachidonic acid is used to produce a model of lung injury, the measurement of permeability must be made in an nonisogravimetric condition (11, 22). Oxygenation products of arachidonic acid have a major role in the inflammatory processes involved in the development of pulmonary edema with ARDS. Changes in hemodynamic parameters and an increase in microvascular permeability have been implicated in arachidonic acidinduced lung injury (11, 22, 24, 27). The hemodynamic effects of arachidonic acid are well characterized. The generation of cyclooxygenase products, particularly thromboxane A,, after the addition of arachidonic acid to the perfusate of isolated lungs results in vasoconstriction and an elevation of pulmonary capillary pressure (l&12, 23,24,27) that contributes to the development of pulmonary edema (11, 12, 27). Addition of arachidonic acid or its metabolite, leukotriene D,, to the perfusate of isolated rabbit lungs also results in an increased transvascular fluid flux that is independent of hemodynamic changes (3,11,21,22). Controversy exists, however, concerning the ability of arachidonic acid to alter fluid conductance and protein permeability. Seeger et al. (22) attributed the increase in fluid accumulation in the lung to a change in Kf after administration of arachidonic acid. Other studies, however, have not found a change in Kf after the addition of arachidonic acid to isolated lungs (24, 27). Controversy also exists concerning changes in solute flux after the generation of arachidonic acid metabolites. Shasby et al. (25) found no change in albumin
0161-7567/92 $2.00 Copyright 0 1992 the American Physiological Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLIJID AND PROTEIN
PERMEABILITY
flux in isolated rabbit lungs, whereas Littner et al. (11) did report a change in solute permeability in isolated dog lungs. The purpose of this investigation was to study the changes in both fluid conductance and protein permeability in lungs with pulmonary edema induced by arachidonic acid. This model of pulmonary edema was selected because controversy exists as to its ability to directly alter either fluid conductance or protein permeability. Both fluid conductance and protein permeability were assessed by measuring the Kf and the extravascular uptake of 1251-albumin, respectively, in the same isolated perfused rabbit lung. We also present a different approach to examine the changes in PS and c in nonisogravimetric edematous lungs.
AND PULMONARY
2441
EDEMA
AA 3 mg 18018tlon
I I
Kfl I I
Kf2 I I
I I
Kf3 I I
10
20
90
40
SO
kl--ss~Y
0
JsICp
I
Kf4 I 1
I 1
60
70
80
1
Kf5 90
Time (minutes) Experimental protocol consisted of taking lary filtration coefficient (KJ measurements, injecting (AA) into pulmonary arterial catheter, and determining min post-AA injection. 1251-albumin clearance (JJC,) mined 30 min post-AA injection. FIG.
1.
3 baseline capilarachidonic acid Kf at 10 and 30 was also deter-
achidonic acid was added to the perfusate of isolated rabbit lungs to produce pulmonary edema as evaluated by an increase in lung weight. Indomethacin was added to the perfusate to inhibit changes in hemodynamic parameters MATERIALS AND METHOIX and thus produce a pressure-independent model of pulIsolated perfused rabbit lung. Lungs were isolated as monary edema. Indomethacin was added to the perfusate previously described (7). New Zealand White rabbits, to give a final concentration of 80 ,uM. Lungs were isoweighing 1.5-2.5 kg, were anesthetized with xylazine (0.3 lated, and the flow rate and venous pressure were admg/kg im), acepromazine (0.75 mg/kg im), and ketamine justed in the first 10 min. Lungs were then allowed to (30 mg/kg im) and then heparinized (1,000 U/kg) via an stabilize for 10 min. & was determined at 20,30, and 50 ear vein. A tracheostomy was performed, the abdominal min. At 60 min, 3 mg of arachidonic acid (Sigma Chemicavity was opened, and the rabbit was exsanguinated via cal) in 30 ~1 of ethanol were injected into the arterial transection of the abdominal aorta. The chest was catheter, giving a final concentration of 50 PM arachiquickly opened, and the heart and lungs were carefully donic acid in the perfusate. Arachidonic acid was adminexcised to avoid injury to the lungs and were suspended istered in this manner because addition of arachidonic from the arm of a beam balance. The pulmonary artery acid to the reservoir resulted in no change in lung weight. was cannulated and perfusion started with a Ringer solu- Kf was then determined IO and 30 min after the injection tion of the following composition (in mM): 137 NaCl, 1.8 of arachidonic acid (70 and 90 min, respectively). AlbuCaCl,, 1.05 MgCl,, 2.68 KCl, 0.06 NaHCO,, 0.13 min clearance (J&J was determined at the end of the NaH,PO,, 0.869 Na,HPO,, and 5.55 dextrose, as well as experiment. This experimental protocol was designed to 0.5 g/100 ml of bovine serum albumin fraction V (Sigma test the effects of various agents given between g and G Chemical). A large-bore catheter was then placed in the on arachidonic acid-induced lung injury. left atrium. The lungs were inflated with a maximum Determination of J$. & was determined as described inflation pressure of 25 cmH,O and allowed to deflate previously (2). Capillary pressure was estimated by simulpassively. This procedure was repeated several times to taneously occluding the arterial and venous lines. The aid in clearing blood from the lungs. Airway pressure was venous outflow was then raised 4-6 cm. The resulting maintained at 2 cmH,O with humidified compressed air. increase in pulmonary capillary pressure (3-4 cmH,O) Recirculation of perfusate was started when the pulmoinduced an increase in weight that was recorded on the nary effluent was clear of blood. The lungs were inflated polygraph and simultaneously sampled 5 times/s by an every 10 min with an inflation pressure 125 cmH,O. on-line microcomputer. At the end of the pressure tranPulmonary arterial and left atria1 pressures were con- sient, the capillary pressure was measured again by doutinuously recorded on a multichannel strip chart re- ble occlusion and the venous outflow tubing was lowered corder (model 7D polygraph, Grass), Pulmonary arterial to its original position. and left atria1 cannulas had in-line solenoids allowing for Because Kf is the change in fluid filtration due to a rapid simultaneous occlusion to determine capillary change in driving force, the filtration rate must be deterpressure (10). The perfusate was pumped by a variable mined immediately after the increase in venous pressure speed pump (model 1215, Harvard Apparatus) at a flow before the other driving forces start to oppose filtration. rate of 90-100 ml/min. The temperature of the perfusate However, an increase in venous pressure results not only was maintained at 39”C, pH was maintained between in fluid filtration but also in an increase in vascular vol7.35 and 7.40, and venous pressure was set at 4 cmH,O to ume. Therefore, the change in weight that is due to an keep the lungs in zone 3 of West. Changes in lung weight increase in vascular volume must be separated from that were detected by a sensitive displacement transducer at- which is due to fluid filtration. The change in vascular volume is believed to occur rapidly and is completed tached to the beam balance. Weight change was continuously recorded on the polygraph with a deflection of 5 cm within seconds after the change in capillary pressure. equal to 5 g. The slower rate of weight gain takes longer to equilibrate Experimental protocol. The experiments in this study and requires minutes to become isogravimetric. We have were conducted according to the protocol shown in Fig. 1. chosen the graphic method of Drake et al. (2) to calculate Time in Fig. 1 is referenced to the start of exsanguination the Kf from the weight gain curve. Linear regression of accomplished by transection of the abdominal aorta. Ar- weight gain vs. time was performed using a 20-s interval Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2442
FLUID
AND PROTEIN
PERMEABILITY
of the weight gain curve. The rate of weight gain was determined in this manner at every 5 s along the curve. The log of the rate of weight gain was plotted as a function of time. This plot showed that the rapid vascular filling phase was completed within the 1st min after the increase in venous pressure. Thus, the time period from 1 to 3 min was used to determine the initial rate of weight gain by extrapolating the slope of the slow component back to time zero. Because the lungs that received arachidonic acid were not isogravimetric, the rate of weight gain just before t;he increase in venous pressure was subtracted from the rate of weight gain determined by extrapolation to time zero. This value was then divided by the change in capillary pressure. Kf was expressed in units of microliters per second per centimeter of water per gram dry lung. Normalization to dry lung weight allowed for comparison of control lungs to the edematous lungs of the arachidonic acid group. Determination of ‘251-albumin clearance. Kern et al. (8) described a single sample technique for measuring JJC, in the isolated lung preparation. We used a similar method for determining JJC, in these experiments. At 30 min after arachidonic acid injection, 1251-albumin was infused into the arterial catheter at a rate to achieve w lo5 counts/min (cpm) 1251-albumin/ml of perfusate. Aliquots of perfusate were obtained at 1, 2, and 3 min during tracer infusion. After 3 min, the infusion was stopped, and the lung vasculature was washed for 3 min with perfusate not containing tracer. Aliquots obtained at l-min intervals during this wash showed that >99% of the counts were removed from the perfusate. Perfusion was then stopped, and the vessels and airways were tied to prevent loss of fluid from the lung. The lungs were cut down and weighed. The right lung was homogenized in a total fluid volume of 20 ml. One milliliter of homogenate was pipetted into three tubes to be counted, and three l-ml aliquots were dried to obtain the dry weight of the aliquot. J,Ic, was then calculated as J,/C p = AI(C,t) (0 where A is the amount of 12*1-albumin in the tissue (cpm/ g dry lung), t is the exposure time of the tracer in minutes, and C, is the perfusate concentration of ‘?-albumin. J,lC, is expressed as microliters per minute per gram dry lung. J,/C, was used to calculate the PS and c of albumin. The relationship between these parameters is shown in the linear flux equation of Kedem and Katchalsky (6)
J, = PS(C, - Ci) + J&l - o)C,
(2)
where J, is the transvascular 1251-albumin flux, C, is the 1251-albumin concentration in the plasma, Ci is the interstitial concentration of 1251-albumin, and Jv is transvascular fluid flux (pl/min). In this model the initial tracer concentration in the interstit,ium (Ci) is considered negligible during the 3-min infusion of tracer; thus Ci can be deleted from the equation. Ct, is then divided into each term of the equation leaving J,/C p = PS + J&l - a)
(3)
If the lung is isogravimetric (J, = 0), then JJC, is equivalent to PS. If the lung is gaining weight (J, # 0), then the
AND PULMONARY
EDEMA
J, (weight gain) during the albumin infusion is measured. A regression line of J,/C, vs. J, can then be determined using the J,/C, and Jv from all experiments. The slope of this regression line equals I - g, and t,he y-intercept equals PS. Calculation of PS and 0 can also be performed in control lungs by producing a weight gain (J, > 0) with increased venous pressure. The Jv after the increase in venous pressure is calculated from the slow component of the weight gain curve determined in the calculation of Kf. ?-albumin was prepared using Na1251 and bovine serum albumin with the use of the chloramine-T procedure as previously described (8). Bovine serum albumin was passed over a Blue Sepharose affinity column to remove immunoglobin G contamination. To ensure that monomeric albumin was used in these experiments, the albumin was passed over a Sephacryl200 column and the albumin purity was checked with use of sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After iodination, the 125I-albumin was separated from free 125I by dialysis (Spectrapore: 25,000 mol wt cutoff) against normal saline (0.09%) and was maintained under dialysis until the day of experiment. The 1251-albumin was only used in experiments if the percent free was ~0.1% as determined by comparing isotope stock to filtrate from ultrafiltration cones (Amicon, CF30; 30,000 mol wt cutoff). Statistics. We studied 12 rabbits in each group. Hemodynamic and Kf data were analyzed using a two-way analysis of variance (ANOVA) with repeated measures over time. A Dunnett’s test was then used to evaluate differences from baseline and a Scheffh’s test was used to compare control and arachidonic acid groups at 10 and 30 min postarachidonic acid. J,/C, was assessed using a one-way ANOVA followed by Scheffb’s test to determine specific differences between groups. A Student’s t‘ test was performed to determine significant differences between the slopes and y-intercept of the two regression lines of J,lC, vs. J,. RESULTS
Figure 2 shows the changes in lung weight in a representative experiment during the first 8 min after the bolus injection of arachidonic acid. There was a slight loss in lung weight during the first 2 min after arachidonic acid injection. This was followed by a sustained increase in lung weight that continued for the remaining 30-min experimental period in all lungs receiving arachidonic acid. Because the lungs were not isogravimetric after the addition of arachidonic acid, the rate of weight gain (J,) was measured before the determination of Z?Lfat 10 and 30 min after arachidonic acid injection. The mean Jv at 10 and 30 min after arachidonic acid was 147 t 14 and 88 t 11 pl min-’ g dry lung-l, respectively. Control lungs were isogravimetric fur the entire experimental period. Figure 2 also shows the changes in arterial and venous pressures during the first 8 min after the injection of arachidonic acid. There was an initial slight increase in arterial pressure, corresponding temporally to the decrease in lung weight, which returned toward baseline. There were no changes in arterial, capillary, and venous l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2444
FLUID
AND PROTEIN
PERMEABILITY
+
, a
AA J,=
0
control
49*4
244f39
FIG. 4. JJC, of lungs that were isogravimetric [transvascular fluid flux (J,) = 0, n = 41 or gaining weight because of an increase in venous pressure (J, = 244 t 49 &‘min, n = 7) and of lungs receiving AA (Jv = 49 t 4 pl/min, n = 6). * P < 0.05 from control lungs.
acid was added, however, the lungs were no longer isogravimetric during the determination of JJC, (J, = 49 t 4 ~1 min-l g dry lung-l; Fig. 4). Because an elevated J, may have increased the convective transport of albumin in lungs that received arachidonic acid, we also measured JJC, in control lungs during a period of elevated J, induced by increasing venous pressure. J,/C, determined in control lungs during increased venous pressure was 85 ? 14 ~1 min-’ . g dry lung-l (Fig. 4). This clearance value was lower (P < 0.05) than that of arachidonic acid lungs, even though the mean J, in control lungs (244 t 49 ~1 min-l 4g dry lung-l) was greater than J, in the arachidonic acid group (49 t 4 ~1 min-l g dry lung-‘; Fig. 4). Figure 5 is a plot of JJC, vs. corresponding J, for control and arachidonic acid-treated lungs. Using the equation JJC, = PS + (1 - a)& we calculated the PS (y-intercept) and of (1 - slope) of albumin with use of the regression line described by these points (calculation outlined in METHODS). There was a 117% increase (P -c 0.05) in PS from a control value of 64 t 32 to 139 t 40 ~1 min-’ g dry lung-l in lungs receiving arachidonic acid. There was a 26% decrease in g from a control value of 0.92 t 0.07 to 0.68 t 0.08 in lungs receiving arachidonic acid. l
l
l
l
l
l
l
l
AND PULMONARY
EDEMA
leukotriene D,, results in an increased J, that is independent of hemodynamic changes (3, 11, 22) . In this study, 3 mg arachidonic acid were injected to yield a final concentration of 50 PM. The time course of the changes in lung weight after the addition of arachidonic acid in this study is similar to that reported by Seeger et al. (22) when they added arachidonic acid (100 PM) during a hydrostatic challenge. Shasby et al. (25) found that arachidonic acid given via the reservoir to yield a final perfusate concentration of 100 PM did not result in the accumulation of fluid in isolated rabbit lungs. Our observations support the findings of Shasby et al. (25), because we did not observe a response to 50 PM arachidonic acid if it was administered via the reservoir. Because the arachidonic acid was injected into the arterial catheter in the present study, the concentration of arachidonic acid on first pass through the lung would be higher than the final diluted perfusate concentration. The ability of similar doses of arachidonic acid to produce an increase in lung weight when added via the arterial catheter, but not the reservoir, supports the pre VfOUS findings that the degree of arachidonic acid-induced lung injury is dose dependent (11, 22). Previous investigators have proposed that low concentrations of arachidonic acid will produce cyclooxygenase-mediated hemodynamic changes and that high doses can result in leukotriene- ,mediated permeability changes (1 2, 22). We used the a bove dose of arach .idonic acid to look at changes in fluid conductance 1 and protei n perm .eability because it produced a consistent model of nonhydrostatic pulmonary edema. We found a 53 and 42% increase in Kf at 10 and 30 min, respectively, after the injection of arachidonic acid. This increase is similar to that of Seeger et al. (22), who found small changes in Kf when arachidonic acid was added in the absence of a hydrostatic pressure challenge. Other investigators did not find a ch .ange in K,; however, these investigations did not have a pressure-independent weight-gain on the addition of araohidonic acid (24, 27). The findings of this investigation support the concept that a change in Kf is associated with a pressu re-independent weight gain after the addition of arachidonic acid. Measuring Kf in a lung that is gaining weight makes the determination of Kf more complicated. When the
DISCUSSION
In this investigation we used arachidonic acid to produce a pressure-independent model of pulmonary edema as monitored by an increase in lung weight. This model was chosen to study the changes in both fluid conductance and protein permeability when lungs are edematous and continuously gaining weight. The injection of arachidonic acid into isolated lungs resulted in edema as demonstrated by a continuous weight gain during the experimental period. The constant rate of weight gain that started between 4 and 5 min post-arachidonic acid was not the result of an increase in capillary pressure because the hemodynamic effects of arachidonic acid were inhibited by adding indomethacin to the perfusate. These data are consistent with previous findings that arachidonic acid in the presence of indomethacin, or the metabolite
200 3v (phnin-l
300 *g dry lung-
’)
FE. 5. Plot of J,/C, vs. Jv at time of perfusion Dashed line, best fit line for AA group; solid line, best lungs. Permeability-surface area product (PS) was andreflection coefficient (u) was lower (P < 0.05) in
with 1251-albumin. fit line for control greater (P < 0.05) AA-treated lungs.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLUID
AND PROTEIN
PERMEABILITY
AND PULMONARY
EDEMA
2445
lungs are isogravimetric, as in baseline conditions, the movement of fluid must be taken into consideration rate of weight gain estimated from the time zero.extrapuwhen determining protein permeability. Solute transport lation is equal to the change in fluid filtration (J.) be- in isogravimetric lungs is believed to be due solely to difcause the initial J, is zero. This is not true during a period fusive transport, whereas the transport of solutes in nonof constant weight gain, because the initial J, is not zero isogravimetric lungs, i.e., those receiving arachidonic and, therefore, contributes to the estimated time zero acid, would be the result of both diffusive and convective rate of weight gain determined by extrapolation. For this transport (18). Thus, the difference in clearance between reason, we subtracted the constant rate of weight gain isogravimetric control lungs and nonisogravimetric arafound after the addition of arachidonic acid from the rate chidonic acid-treated lungs as seen in Fig. 4 may be due, of weight gain found by extrapolation to time zero. in part, to convective transport. To investigate whether The measurement of Kf in this experimental model of increased clearance was due solely to an increase in conpulmonary edema may be difficult to interpret because vective transport, J,/C, was determined in lungs in which part of the weight gain observed was due to alveolar J, was elevated by increasing venous pressure. As seen in flooding. Thus, when Kf was being determined, fluid was Fig. 4, the J,IC, in control lungs during a period of elenot only moving into the interstitium, but it was also vated J, was considerably less than that in the arachimoving through the interstitium into the alveolus. When donic acid group even though Jv was greater in control the log of the rate of weight gain was plotted vs. time, the lungs after an increase in venous pressure. curve was linear between 1 and 3 min, indicating that The results shown in Fig. 4 demonstrate that the inchanges in vascular volume were completed by 1 min. crease in protein flux in the arachidonic acid-treated During this time, there was an exponential decrease in lungs was not only due to an increase in convective transthe rate of weight gain as the semilog plot decreased linport but also to a change in permeability. This can be early with time. The exponential decrease in the rate of demonstrated by comparing the PS and G of control weight gain suggests that the forces that oppose filtraand arachidonic acid-treated lungs. J,IC, in control isotion still had a role in the injured lung and that we were gravimetric lungs is equivalent to PS. This is not true for primarily measuring fluid movement from the vascular arachidonic acid-treated lungs because these lungs were space into the interstitium. Also, by extrapolating to time gaining weight during the determination of J,/C,. In the zero, we are estimating the initial flux of water across the method used by Kern and Malik (9) to determine PS and endothelium immediately after the increase in venous g’t the lungs must be isogravimetric. Presented in Fig. 5 is pressure. the PS and g for isogravimetric control lungs and noniTo assess protein permeability, we measured the extrasogravimetric arachidonic acid-treated lungs as detervascular uptake of 1251-labeled albumin by use of a tech- mined using the linear flux equation. This method is siminique similar to that previously described (8,12). In this lar to that used by Renkin et al. (19) to estimate 0 in skin study, the 3-min uptake of 1251-albumin was divided by and skeletal muscle. In this method, linear regression of the perfusate concentration so that the albumin uptake J,/C, vs. J, was performed using clearances determined was expressed as a clearance. The JJC, shown in Fig. 4 at various J, induced by increasing venous pressure (Fig. for isogravimetric control lungs is similar to that re- 5). The intercept of this line is equivalent to PS. After the ported previously (8,9,12). After the addition of arachiaddition of arachidonic acid, the extrapolated PS was donic acid, J,/C, was increased twofold over control twofold higher than that of control lungs, indicating that lungs. In the injured lung, the increase in protein permethe permeability of the vascular barrier for albumin had ability resulted in a greater flux of tracer over the 3-min been altered. As seen in Fig. 5 the slope of the regression tracer infusion. This increase in the interstitial concenline for the arachidonic acid-treated group was greater tration of 1251-albumin, along with an increase in permethan that of the control group. The increase in the slope ability of the end&helium, may allow for a greater indicates that the convective flux of albumin is greater in amount of diffusion back into the vasculature than seen the arachidonic acid-treated lungs than in control lungs in control lungs. This would mean that we have underesat a similar J,. This agrees with the findings of Littner et timated the increase in J,/C, in the arachidonic acid- al. (ll), who found increases in convective flux when they treated lung. However, the presence of some alveolar plotted wet-to-dry weight ratios vs. albumin uptake in flooding may provide a larger interstitial pool for the di- arachidonic acid-treated dog lungs. The slope of the relution of the tracer, resulting in a decreased gradient for gression line in Fig. 5 is equivalent to 1 - G, After the back flux during the wash. In control lungs, the interstiaddition of arachidonic acid, G decreased from a value of tial concentration of radiolabeled albumin has been 0.92 to 0.68. The linear flux equation described by Kedem shown nut to be a factor, This is because the increase in and Katchalasky (6) was used in this analysis as PS and c the interstitial concentration of tracer is so small during had been calculated in a similar system using this equathe 3-min tracer infusion that the concentration gradient tion. The use of the linear flux equation permits easy for the labeled albumin during the wash period is miniseparation of the diffusive and convective terms. In this mal and a negligible amount of the interstitial tracer is equation, however, the diffusive term may be overestilost by back-diffusion (8). mated because an increase in fluid movement may alter The ability of exogenous arachidonic acid to increase the concentration gradient through the microvascular albumin flux has also been reported for dog lungs (11). wall. Fluid movement also makes it difficult to define the solute concentration within the pore. The nonlinear flux However, as in this study, the arachidonic acid-treated lungs were gaining weight during the measurement of equation (Eq 4) does not have these constraints (18). albumin uptake. Similar to the determination of K,, the When the nonlinear flux equation (Eq. 4) was fit to the Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2446
FLUID
AND PROTEIN
PERMEABILITY
data in Fig. 5 by nonlinear least-square regression and assuming Ci = 0, the control PS was estimated to be 65 t 15 ~1 min-l g dry lung-’ and the control CJwas O&6 t 0.12 l
l
Js = J,( 1 - a)[Cp - Cie-“/( 1 - eWx)]
(4)
After the addition of arachidonic acid, the PS as calculated using Eq. 4 was 142 -t 19 ,~l. min. g dry lung-’ and CTwas 0.48 t 0.12. Our value of 0 for control lungs is similar to that found with use of tracer techniques in isolated perfused lungs (0.96) (9). This value of G is higher than that calculated by lymph protein measurements (0.61-0.70) (4, 15, 17) and the integral mass balance technique (0.77) (14). Our estimated value for c is also higher than the 0.81 reported by Ishibashi et al. (5), who used tracer kinetics to determine 0 in dog lungs. Our studies differ in that Ishibashi et al. (5) delivered 1251-albumin as a bolus and albumin flux was assessed 2 h later, whereas we infused the tracer at a const,ant rate for 3 min and assessed albumin uptake after a 3-min wash. As discussed above, the short exposure time prevents a significant increase in the interstitial tracer concentration, thereby allowing the concentration gradient for the 1251-albumin to remain constant during tracer infusion and for back flux to be negligible. An increase in protein flux, as found after addition of arachidonic acid, would increase interstitial tracer concentration. A large increase in concentration of the interstitial tracer during its infusion may result in an increase in back flux of tracer during the wash period. This would cause a greater overestimation of 0 in the arachidonic acid group compared with the control group; thus the results of a decrease in G after arachidonic acid injection are valid. This study is the first to measure both fluid conductance and protein permeability in the same lung during acute injury leading to lung edema. By measuring both of these parameters in the same lung, we found that protein permeability is changed to a greater degree than is fluid conductance atier arachidonic acid. This finding is not difficult to explain if one compares the different transvascular exchange pathways used by fluid and protein. The pathway available for transvascular exchange of water is much larger than that of protein as demonstrated by a high g in normal lungs. Although we have estimated PS and g as if the vascular barrier was a homoporous membrane, the movement of fluid and protein is believed to be through large and small pores. The vascular barrier of the lung is thought to consist of two sized pores of 80 and 200 A, with the ratio of small to large pores being 195:l (26). The small pores have been estimated to be responsible for 80% of the hydraulic conductance, whereas 16% of fluid conductance is through the large pores (26). The majority of protein transport is believed to be through the large pores. Therefore, a change in protein permeability due to the formation of a few large leaks would result in a small change in the total water pathway and the calculated fluid conductance. This suggests that a change in protein permeability could be measured in the absence ,f? 1--- -,,,,,,,Ll, -1, ---I\ :, A?l,,:A #%~-A~F&~yIcI* TL.,o
AND PULMONARY
EDEMA
changes in protein permeability, as opposed to fluid conductance, may be a more sensitive marker of lung vascular injury in models of pulmonary edema. Moreover, as shown by the present study, both of these parameters can be simultaneously measured in the same isolated lung preparation. The authors thank Dr. David Kern for assistance in the development of the isolated lung preparation in our laboratory. We also thank Ed Lewis for excellent technical support and Deborah Moran and Wendy Ward for editorial assistance. This work was supported by National Institute of General Medical Sciences Grant GM-21447, and National Heart, Lung, and Blood Institute Grants HL-38894 and HL-26807. P. A. Vincent was a National Institutes of Health (NIH) Postdoctoral Fellow supported by National Research Service Award (NRSA) F32-GM-14026. P. B. Kreienberg was a NIH Postdoctoral Fellow supported by NRSA F32-CM-14025 F. L. Minnear was a recipient of Research Career Development Award K04HL-01830. Address for reprint requests: P. A. Vincent, Dept. of Physiology and Cell Biology (A-134), Albany Medical College, 47 New Scotland Ave., Albany, NY 12208. Received 31 October 1991; accepted in final form 14 July 1992. REFERENCES 1. ANDERSON,
R. R., R. L. HOLLIDAY, A. A. DRIEDGER, M. LEFCOE, 3. W. J. SIBBALD. Documentation of pulmonary capillary permeability in adult respiratory distress syndrome accompanying human sepsis. Am. Rev. Respir. Dis. 119: 869-877, 1979. DRAKE, R., K. A. GAAR, AND A. E. TAYLOR. Estimation of the filtration coefficient of pulmonary exchange vessels. Am. J. Physiol. 234 (Heart Circ. Physiol. 3): H266-H274, 1978. FARRUKH, I. S., A. M. SCIUTO, E. W. SPANNHAKE, G. H. GURTNER, AND J. R. MICHAEL. Leukotriene D, increases pulmonary vascular permeability and pressure by different mechanisms in the rabbit. Am. Rev. Respir. Dis. 134: 229-232, 1986. GABEL, J. C., B. D. BUTLER, R. L. SCOTT, AND R. E. DRAKE. Pulmonary microvascular permeability after coronary artery ligation in dogs. J. Appl. Physiol. 55: 372-375, 1983. ISHIBASHI, M., R. K. REED, M. I. TOWNSLEY, J. C. PARKER, AND A. E. TAYLOR. Albumin transport across pulmonary capillary-interstitial barrier in anesthetized dogs. J. Appl. Physiol. 70: 21042110,1991. KEDEM, O., AND A. KATCHALSKY. Thermodynamic analysis of the permeability of biological membranes to non-electrolytes. Biochim. REID,
2. 3.
4. 5.
6.
Biophys.
AND
Acta
27: 229-246,
1958.
7. KERN, 8.
9. 10. 11.
12. 13. 14.
D. F. Blood Volume Changes and Z’ranscapillary Fluid Flow in Isolated Rabbit Lung (MS thesis). Minneapolis: Univ. of Minnesota, 1978. KERN, D. F., D. LEVITT, AND D. WANGENSTEEN. Endothelial albumin permeability measured with a new technique in perfused rabbit lung. Am. J. Physiol. 245 (Heart Circ. Physiol. 14): H229-H236, 1983. KERN, D. F., AND A. B. MALIK. Microvascular albumin permeability in isolated perfused lung: effects of EDTA. J. Appl. Physiol. 58: 372-375, 1985. LINEHAN, J. H., C. A. DAWSON, AND D. A. RICKABY. Distribution of vascular resistance and compliance in a dog lung lobe. J. Appl. Physiol, 53: 158-168, 1982. LITTNER, M. R., G. M. KAZMI, AND F. D. LOTT. Inhibition of cyclooxygenase production does not prevent arachidonate from increasing extravascular lung water and albumin in an isolated dog lung. Prostaglandins Leukotrienes Med. 15: 53-68, 1984. LITTNER, M. R., AND F. D. LOTT. Edema from cyclooxygenase products of endogenous arachidonic acid in isolated lung. J. Appl. Physiol. 67: 846-855, 1989. MAGNO, M., B. ATKINSON, A. KATZ, AND P, FISHMAN. Estimation of pulmonary interstitial fluid space compliance in isolated perfused rabbit lung. J. Appl. Physiol. 48: 677-683, 1980. MARON, M. B., C. F. PILATI, AND K. C. MEANDER. Effect of hemolysis on reflection coefficient determined by endogenous blood iniI*.,,+,,... T A ,,I DL, ,*:*I Cc).c)c)KF) r)OK7 lOQ7
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLUID
AND PROTEIN
PERMEABILITY
15. PARKER, J. C., R. E. PARKER, N. GRANGER, AND A. E. TAYLOR. Vascular permeability and transvascular fluid and protein transport in the dog lung. Circ. &s. 48: 549-561, 1981, 16. PARKER, J. C., M. I. TOWNSLEY, AND J. T. CARTLEDGE. Lung edema increases transvascular filtration rate but not filtration eoefficient. J. Appl. PhysioZ. 66: 1553-1560, 1989. 17. PARKER, R.E.,R. J. ROSELLI, T.R. HARRIS,AND K. L. BRIGHAM. Effects of graded increases in pulmonary vascular pressures on lung fluid balance in unanesthetized sheep. Circ. Res. 49: 11641172,198l. 18. RENKIN, E. M. Capillary transport of macromolecules: pores and other endothelial pathways. J. Appl. Physiol. 58: 315-325, 1985. 19. RENKIN, E.M.,M. GUSTAFSON-SGRO,AND L. SIBLEY. Couplingof albumin flux to volume flow in skin and muscles of anesthetized rats. Am. J. Physiol. 255(Heart Cire. Physiol. 24): H458-H466,1988. 20. RINALDO, J. E., AND R. Me ROGERS. Adult respiratory-distress syndrome: changing concepts of lung injury and repair. N. Engl. J. Med. 306: 900-909, 1982. 21. SEEGER, W.,M. MENGER,D.WALMRATH,G.BECKER, F. GRIMMINGER, AND H. NEUHOF. Arachidonic acid lipoxygenase pathways and increased vascular permeability in isolated rabbit lungs. Am. Rev. Respir. Dis. 136: 964-972, 1987.
AND PULMONARY
EDEMA
2447
22. SEEGER, W.,D. WALMRATH, M. MENGER, AND H. NEUHOF. Increased lung vascular permeability after arachidonic acid and hydrostatic challenge. J. Appl. Physiol. 61: 1781-1789, 1986. 2% SEEGER, W.,H. WOLF,G. STAHLER, H. NEUHOF,AND L. ROKA. Increased pulmonary vascular resistance and permeability due to arachidonate metabolism in isolated rabbit lungs. Prostaglandins 23: 157-173, 1982. 24* SELIG, W. M.,T.C. NOONAN, D.F. KERN,AND A.B. MALIK. Pulmonary microvascular responses to arachidonic acid in isolated perfused guinea pig lung. J. Appl. Physiol. 60: 1972-1979, 1985. 25* SHASBY, D. M., S. S. SHASBY, AND M. J. PEACH. Polymorphonuclear leukocyte: arachidonate edema. J. Appt. Physiol. 59: 47-55, 26 1985. TAYLOR, A. E., AND D. N. GRANGER. Exchange of macromolecules across the microcirculation. In: Handbook of Physiology. The Cardiovascular System. Microcirculation. Bethesda, MD: Am. Physiol. Sot., 1984, sect. 2, vol. IV, pt. 1, chapt. 11, p. 467-520. 27 TOWNSLEY, M. I., R.J. KORTHUIS,AND A.E. TAYLOR. Effects of arachidonate on permeability and resistance distribution in canine lungs. J. Appl. Physiol. 58: 206-210, 1985. 28. WANGENSTEEN,~. II., E. LYSAKER,AND P. SAVARYN. Pulmonary capillary filtration and reflection coefficients in the adult rabbit. Microvasc. Res. 14: 81-97. 1977. l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FRED
Department of Physiology and Cell Biology, The Neil Hellman Albany Medical College, Albany, Nezu York 12208
VINCENT,PETER
A., PAUL B. KREIENBERG,FRED L. MINM. SABA, AND DONALD R. BELL. Simultaneous measurement of fluid and protein permeability in isolated rubbit lungsduring edema.J. Appl. Physiol. 73(6): 2440-2447,1992.Fluid conductance and protein permeability have been studied in isolated perfused lung modelsof pulmonary edema. However, previous studies have not investigated changesof both fluid conductance and protein permeability in the sameisolated lung preparation after injury. Arachidonic acid (AA) metabolites are involved in the inflammatory processesthat lead to the development of pulmonary edema. The hemodynamic effects of AA have beenwell established;however, controversy exists concerning the ability of AA to alter the permeability of the pulmonary microvasculature to fluid and protein. The purpose of this study was to simultaneously determine whether transvascular fluid conductance and protein permeability are increased in isolated perfused rabbit lungs with pulmonary edemainducedby AA. Indomethacin (80 PM) was addedto the perfusate to inhibit the hemodynamic effects of AA and producea pressure-independentmodelof pulmonary edema.Fluid conductancewas assessed by determination of the capillary filtration coefficient (KJ, and protein permeability wasevaluated by measurementof 1251-albumin clearance.The injection of AA (3 mg/200 ml of perfusate) into the pulmonary arterial catheter resulted in an increasein lung weight over the remaining 30min experimental period. Kf (~1 s-l cmH,O-’ g dry lung-l) was increased(P < 0.05) in AA-treated lungs at 10 and 30 min post-AA injection when comparedwith control lungsand baseline values (determined 10 min before AA injection). Albumin clearancewasalsogreater (P < 0.05) in lungs that received AA. 1251-albumin clearancewas measuredat different rates of fluid flux produced by elevation of venous pressure.Linear regression of albumin clearancevs. fluid flux was used to calculate the permeability-surface area product (y-intercept) and the reflection coefficient (1 - slope) for each group. The calculated permeability-surface area product was greater and the calculated reflection coefficient was lower in the AA-treated lungs than in control lungs. These results show that AA increases both protein permeability and fluid conductance but that the increasein protein permeability (117%) is larger than the increasein fluid conductance (42%). NEAR,
THOMAS
l
l
l
arachidonic acid; capillary filtration coefficient; albumin; permeability-surface area product; reflection coefficient
RESPIRATORY DISTRESS SYNDROME (ARDS) is characterized by the development of pulmonary edema that can lead to ventilation-perfusion mismatch and hypoxia (20). Pulmonary edema in ARDS is believed to result in part from an increased permeability of the lung
ADULT
2440
L. MINNEAR, Medical Research Building,
microvasculature (1, 20). Studies have investigated the permeability of the pulmonary microvasculature in isolated lung preparations utilizing techniques that assess either the capillary filtration coefficient (Kf) (2,16,28), a measure of fluid conductance, or albumin clearance (8), a measure of protein permeability. These techniques have also been used to determine changes in lung vascular permeability after lung injury models (9, 11, 12, 24, 27). However, they have not been used simultaneously to determine whether both protein permeability and fluid conductance are altered in the same lung in models of pulmonary edema. Kern and Malik (9) described a technique that uses the extravascular uptake of albumin to calculate the permeability-surface area product (PS) and the reflection coefficient (a) in isolated lungs. This method, however, requires that the lungs be isogravimetric, a condition that often does not exist in lungs that have an altered microvascular permeability. Indeed, when arachidonic acid is used to produce a model of lung injury, the measurement of permeability must be made in an nonisogravimetric condition (11, 22). Oxygenation products of arachidonic acid have a major role in the inflammatory processes involved in the development of pulmonary edema with ARDS. Changes in hemodynamic parameters and an increase in microvascular permeability have been implicated in arachidonic acidinduced lung injury (11, 22, 24, 27). The hemodynamic effects of arachidonic acid are well characterized. The generation of cyclooxygenase products, particularly thromboxane A,, after the addition of arachidonic acid to the perfusate of isolated lungs results in vasoconstriction and an elevation of pulmonary capillary pressure (l&12, 23,24,27) that contributes to the development of pulmonary edema (11, 12, 27). Addition of arachidonic acid or its metabolite, leukotriene D,, to the perfusate of isolated rabbit lungs also results in an increased transvascular fluid flux that is independent of hemodynamic changes (3,11,21,22). Controversy exists, however, concerning the ability of arachidonic acid to alter fluid conductance and protein permeability. Seeger et al. (22) attributed the increase in fluid accumulation in the lung to a change in Kf after administration of arachidonic acid. Other studies, however, have not found a change in Kf after the addition of arachidonic acid to isolated lungs (24, 27). Controversy also exists concerning changes in solute flux after the generation of arachidonic acid metabolites. Shasby et al. (25) found no change in albumin
0161-7567/92 $2.00 Copyright 0 1992 the American Physiological Society
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLIJID AND PROTEIN
PERMEABILITY
flux in isolated rabbit lungs, whereas Littner et al. (11) did report a change in solute permeability in isolated dog lungs. The purpose of this investigation was to study the changes in both fluid conductance and protein permeability in lungs with pulmonary edema induced by arachidonic acid. This model of pulmonary edema was selected because controversy exists as to its ability to directly alter either fluid conductance or protein permeability. Both fluid conductance and protein permeability were assessed by measuring the Kf and the extravascular uptake of 1251-albumin, respectively, in the same isolated perfused rabbit lung. We also present a different approach to examine the changes in PS and c in nonisogravimetric edematous lungs.
AND PULMONARY
2441
EDEMA
AA 3 mg 18018tlon
I I
Kfl I I
Kf2 I I
I I
Kf3 I I
10
20
90
40
SO
kl--ss~Y
0
JsICp
I
Kf4 I 1
I 1
60
70
80
1
Kf5 90
Time (minutes) Experimental protocol consisted of taking lary filtration coefficient (KJ measurements, injecting (AA) into pulmonary arterial catheter, and determining min post-AA injection. 1251-albumin clearance (JJC,) mined 30 min post-AA injection. FIG.
1.
3 baseline capilarachidonic acid Kf at 10 and 30 was also deter-
achidonic acid was added to the perfusate of isolated rabbit lungs to produce pulmonary edema as evaluated by an increase in lung weight. Indomethacin was added to the perfusate to inhibit changes in hemodynamic parameters MATERIALS AND METHOIX and thus produce a pressure-independent model of pulIsolated perfused rabbit lung. Lungs were isolated as monary edema. Indomethacin was added to the perfusate previously described (7). New Zealand White rabbits, to give a final concentration of 80 ,uM. Lungs were isoweighing 1.5-2.5 kg, were anesthetized with xylazine (0.3 lated, and the flow rate and venous pressure were admg/kg im), acepromazine (0.75 mg/kg im), and ketamine justed in the first 10 min. Lungs were then allowed to (30 mg/kg im) and then heparinized (1,000 U/kg) via an stabilize for 10 min. & was determined at 20,30, and 50 ear vein. A tracheostomy was performed, the abdominal min. At 60 min, 3 mg of arachidonic acid (Sigma Chemicavity was opened, and the rabbit was exsanguinated via cal) in 30 ~1 of ethanol were injected into the arterial transection of the abdominal aorta. The chest was catheter, giving a final concentration of 50 PM arachiquickly opened, and the heart and lungs were carefully donic acid in the perfusate. Arachidonic acid was adminexcised to avoid injury to the lungs and were suspended istered in this manner because addition of arachidonic from the arm of a beam balance. The pulmonary artery acid to the reservoir resulted in no change in lung weight. was cannulated and perfusion started with a Ringer solu- Kf was then determined IO and 30 min after the injection tion of the following composition (in mM): 137 NaCl, 1.8 of arachidonic acid (70 and 90 min, respectively). AlbuCaCl,, 1.05 MgCl,, 2.68 KCl, 0.06 NaHCO,, 0.13 min clearance (J&J was determined at the end of the NaH,PO,, 0.869 Na,HPO,, and 5.55 dextrose, as well as experiment. This experimental protocol was designed to 0.5 g/100 ml of bovine serum albumin fraction V (Sigma test the effects of various agents given between g and G Chemical). A large-bore catheter was then placed in the on arachidonic acid-induced lung injury. left atrium. The lungs were inflated with a maximum Determination of J$. & was determined as described inflation pressure of 25 cmH,O and allowed to deflate previously (2). Capillary pressure was estimated by simulpassively. This procedure was repeated several times to taneously occluding the arterial and venous lines. The aid in clearing blood from the lungs. Airway pressure was venous outflow was then raised 4-6 cm. The resulting maintained at 2 cmH,O with humidified compressed air. increase in pulmonary capillary pressure (3-4 cmH,O) Recirculation of perfusate was started when the pulmoinduced an increase in weight that was recorded on the nary effluent was clear of blood. The lungs were inflated polygraph and simultaneously sampled 5 times/s by an every 10 min with an inflation pressure 125 cmH,O. on-line microcomputer. At the end of the pressure tranPulmonary arterial and left atria1 pressures were con- sient, the capillary pressure was measured again by doutinuously recorded on a multichannel strip chart re- ble occlusion and the venous outflow tubing was lowered corder (model 7D polygraph, Grass), Pulmonary arterial to its original position. and left atria1 cannulas had in-line solenoids allowing for Because Kf is the change in fluid filtration due to a rapid simultaneous occlusion to determine capillary change in driving force, the filtration rate must be deterpressure (10). The perfusate was pumped by a variable mined immediately after the increase in venous pressure speed pump (model 1215, Harvard Apparatus) at a flow before the other driving forces start to oppose filtration. rate of 90-100 ml/min. The temperature of the perfusate However, an increase in venous pressure results not only was maintained at 39”C, pH was maintained between in fluid filtration but also in an increase in vascular vol7.35 and 7.40, and venous pressure was set at 4 cmH,O to ume. Therefore, the change in weight that is due to an keep the lungs in zone 3 of West. Changes in lung weight increase in vascular volume must be separated from that were detected by a sensitive displacement transducer at- which is due to fluid filtration. The change in vascular volume is believed to occur rapidly and is completed tached to the beam balance. Weight change was continuously recorded on the polygraph with a deflection of 5 cm within seconds after the change in capillary pressure. equal to 5 g. The slower rate of weight gain takes longer to equilibrate Experimental protocol. The experiments in this study and requires minutes to become isogravimetric. We have were conducted according to the protocol shown in Fig. 1. chosen the graphic method of Drake et al. (2) to calculate Time in Fig. 1 is referenced to the start of exsanguination the Kf from the weight gain curve. Linear regression of accomplished by transection of the abdominal aorta. Ar- weight gain vs. time was performed using a 20-s interval Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2442
FLUID
AND PROTEIN
PERMEABILITY
of the weight gain curve. The rate of weight gain was determined in this manner at every 5 s along the curve. The log of the rate of weight gain was plotted as a function of time. This plot showed that the rapid vascular filling phase was completed within the 1st min after the increase in venous pressure. Thus, the time period from 1 to 3 min was used to determine the initial rate of weight gain by extrapolating the slope of the slow component back to time zero. Because the lungs that received arachidonic acid were not isogravimetric, the rate of weight gain just before t;he increase in venous pressure was subtracted from the rate of weight gain determined by extrapolation to time zero. This value was then divided by the change in capillary pressure. Kf was expressed in units of microliters per second per centimeter of water per gram dry lung. Normalization to dry lung weight allowed for comparison of control lungs to the edematous lungs of the arachidonic acid group. Determination of ‘251-albumin clearance. Kern et al. (8) described a single sample technique for measuring JJC, in the isolated lung preparation. We used a similar method for determining JJC, in these experiments. At 30 min after arachidonic acid injection, 1251-albumin was infused into the arterial catheter at a rate to achieve w lo5 counts/min (cpm) 1251-albumin/ml of perfusate. Aliquots of perfusate were obtained at 1, 2, and 3 min during tracer infusion. After 3 min, the infusion was stopped, and the lung vasculature was washed for 3 min with perfusate not containing tracer. Aliquots obtained at l-min intervals during this wash showed that >99% of the counts were removed from the perfusate. Perfusion was then stopped, and the vessels and airways were tied to prevent loss of fluid from the lung. The lungs were cut down and weighed. The right lung was homogenized in a total fluid volume of 20 ml. One milliliter of homogenate was pipetted into three tubes to be counted, and three l-ml aliquots were dried to obtain the dry weight of the aliquot. J,Ic, was then calculated as J,/C p = AI(C,t) (0 where A is the amount of 12*1-albumin in the tissue (cpm/ g dry lung), t is the exposure time of the tracer in minutes, and C, is the perfusate concentration of ‘?-albumin. J,lC, is expressed as microliters per minute per gram dry lung. J,/C, was used to calculate the PS and c of albumin. The relationship between these parameters is shown in the linear flux equation of Kedem and Katchalsky (6)
J, = PS(C, - Ci) + J&l - o)C,
(2)
where J, is the transvascular 1251-albumin flux, C, is the 1251-albumin concentration in the plasma, Ci is the interstitial concentration of 1251-albumin, and Jv is transvascular fluid flux (pl/min). In this model the initial tracer concentration in the interstit,ium (Ci) is considered negligible during the 3-min infusion of tracer; thus Ci can be deleted from the equation. Ct, is then divided into each term of the equation leaving J,/C p = PS + J&l - a)
(3)
If the lung is isogravimetric (J, = 0), then JJC, is equivalent to PS. If the lung is gaining weight (J, # 0), then the
AND PULMONARY
EDEMA
J, (weight gain) during the albumin infusion is measured. A regression line of J,/C, vs. J, can then be determined using the J,/C, and Jv from all experiments. The slope of this regression line equals I - g, and t,he y-intercept equals PS. Calculation of PS and 0 can also be performed in control lungs by producing a weight gain (J, > 0) with increased venous pressure. The Jv after the increase in venous pressure is calculated from the slow component of the weight gain curve determined in the calculation of Kf. ?-albumin was prepared using Na1251 and bovine serum albumin with the use of the chloramine-T procedure as previously described (8). Bovine serum albumin was passed over a Blue Sepharose affinity column to remove immunoglobin G contamination. To ensure that monomeric albumin was used in these experiments, the albumin was passed over a Sephacryl200 column and the albumin purity was checked with use of sodium dodecyl sulfate-polyacrylamide gel electrophoresis. After iodination, the 125I-albumin was separated from free 125I by dialysis (Spectrapore: 25,000 mol wt cutoff) against normal saline (0.09%) and was maintained under dialysis until the day of experiment. The 1251-albumin was only used in experiments if the percent free was ~0.1% as determined by comparing isotope stock to filtrate from ultrafiltration cones (Amicon, CF30; 30,000 mol wt cutoff). Statistics. We studied 12 rabbits in each group. Hemodynamic and Kf data were analyzed using a two-way analysis of variance (ANOVA) with repeated measures over time. A Dunnett’s test was then used to evaluate differences from baseline and a Scheffh’s test was used to compare control and arachidonic acid groups at 10 and 30 min postarachidonic acid. J,/C, was assessed using a one-way ANOVA followed by Scheffb’s test to determine specific differences between groups. A Student’s t‘ test was performed to determine significant differences between the slopes and y-intercept of the two regression lines of J,lC, vs. J,. RESULTS
Figure 2 shows the changes in lung weight in a representative experiment during the first 8 min after the bolus injection of arachidonic acid. There was a slight loss in lung weight during the first 2 min after arachidonic acid injection. This was followed by a sustained increase in lung weight that continued for the remaining 30-min experimental period in all lungs receiving arachidonic acid. Because the lungs were not isogravimetric after the addition of arachidonic acid, the rate of weight gain (J,) was measured before the determination of Z?Lfat 10 and 30 min after arachidonic acid injection. The mean Jv at 10 and 30 min after arachidonic acid was 147 t 14 and 88 t 11 pl min-’ g dry lung-l, respectively. Control lungs were isogravimetric fur the entire experimental period. Figure 2 also shows the changes in arterial and venous pressures during the first 8 min after the injection of arachidonic acid. There was an initial slight increase in arterial pressure, corresponding temporally to the decrease in lung weight, which returned toward baseline. There were no changes in arterial, capillary, and venous l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2444
FLUID
AND PROTEIN
PERMEABILITY
+
, a
AA J,=
0
control
49*4
244f39
FIG. 4. JJC, of lungs that were isogravimetric [transvascular fluid flux (J,) = 0, n = 41 or gaining weight because of an increase in venous pressure (J, = 244 t 49 &‘min, n = 7) and of lungs receiving AA (Jv = 49 t 4 pl/min, n = 6). * P < 0.05 from control lungs.
acid was added, however, the lungs were no longer isogravimetric during the determination of JJC, (J, = 49 t 4 ~1 min-l g dry lung-l; Fig. 4). Because an elevated J, may have increased the convective transport of albumin in lungs that received arachidonic acid, we also measured JJC, in control lungs during a period of elevated J, induced by increasing venous pressure. J,/C, determined in control lungs during increased venous pressure was 85 ? 14 ~1 min-’ . g dry lung-l (Fig. 4). This clearance value was lower (P < 0.05) than that of arachidonic acid lungs, even though the mean J, in control lungs (244 t 49 ~1 min-l 4g dry lung-l) was greater than J, in the arachidonic acid group (49 t 4 ~1 min-l g dry lung-‘; Fig. 4). Figure 5 is a plot of JJC, vs. corresponding J, for control and arachidonic acid-treated lungs. Using the equation JJC, = PS + (1 - a)& we calculated the PS (y-intercept) and of (1 - slope) of albumin with use of the regression line described by these points (calculation outlined in METHODS). There was a 117% increase (P -c 0.05) in PS from a control value of 64 t 32 to 139 t 40 ~1 min-’ g dry lung-l in lungs receiving arachidonic acid. There was a 26% decrease in g from a control value of 0.92 t 0.07 to 0.68 t 0.08 in lungs receiving arachidonic acid. l
l
l
l
l
l
l
l
AND PULMONARY
EDEMA
leukotriene D,, results in an increased J, that is independent of hemodynamic changes (3, 11, 22) . In this study, 3 mg arachidonic acid were injected to yield a final concentration of 50 PM. The time course of the changes in lung weight after the addition of arachidonic acid in this study is similar to that reported by Seeger et al. (22) when they added arachidonic acid (100 PM) during a hydrostatic challenge. Shasby et al. (25) found that arachidonic acid given via the reservoir to yield a final perfusate concentration of 100 PM did not result in the accumulation of fluid in isolated rabbit lungs. Our observations support the findings of Shasby et al. (25), because we did not observe a response to 50 PM arachidonic acid if it was administered via the reservoir. Because the arachidonic acid was injected into the arterial catheter in the present study, the concentration of arachidonic acid on first pass through the lung would be higher than the final diluted perfusate concentration. The ability of similar doses of arachidonic acid to produce an increase in lung weight when added via the arterial catheter, but not the reservoir, supports the pre VfOUS findings that the degree of arachidonic acid-induced lung injury is dose dependent (11, 22). Previous investigators have proposed that low concentrations of arachidonic acid will produce cyclooxygenase-mediated hemodynamic changes and that high doses can result in leukotriene- ,mediated permeability changes (1 2, 22). We used the a bove dose of arach .idonic acid to look at changes in fluid conductance 1 and protei n perm .eability because it produced a consistent model of nonhydrostatic pulmonary edema. We found a 53 and 42% increase in Kf at 10 and 30 min, respectively, after the injection of arachidonic acid. This increase is similar to that of Seeger et al. (22), who found small changes in Kf when arachidonic acid was added in the absence of a hydrostatic pressure challenge. Other investigators did not find a ch .ange in K,; however, these investigations did not have a pressure-independent weight-gain on the addition of araohidonic acid (24, 27). The findings of this investigation support the concept that a change in Kf is associated with a pressu re-independent weight gain after the addition of arachidonic acid. Measuring Kf in a lung that is gaining weight makes the determination of Kf more complicated. When the
DISCUSSION
In this investigation we used arachidonic acid to produce a pressure-independent model of pulmonary edema as monitored by an increase in lung weight. This model was chosen to study the changes in both fluid conductance and protein permeability when lungs are edematous and continuously gaining weight. The injection of arachidonic acid into isolated lungs resulted in edema as demonstrated by a continuous weight gain during the experimental period. The constant rate of weight gain that started between 4 and 5 min post-arachidonic acid was not the result of an increase in capillary pressure because the hemodynamic effects of arachidonic acid were inhibited by adding indomethacin to the perfusate. These data are consistent with previous findings that arachidonic acid in the presence of indomethacin, or the metabolite
200 3v (phnin-l
300 *g dry lung-
’)
FE. 5. Plot of J,/C, vs. Jv at time of perfusion Dashed line, best fit line for AA group; solid line, best lungs. Permeability-surface area product (PS) was andreflection coefficient (u) was lower (P < 0.05) in
with 1251-albumin. fit line for control greater (P < 0.05) AA-treated lungs.
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLUID
AND PROTEIN
PERMEABILITY
AND PULMONARY
EDEMA
2445
lungs are isogravimetric, as in baseline conditions, the movement of fluid must be taken into consideration rate of weight gain estimated from the time zero.extrapuwhen determining protein permeability. Solute transport lation is equal to the change in fluid filtration (J.) be- in isogravimetric lungs is believed to be due solely to difcause the initial J, is zero. This is not true during a period fusive transport, whereas the transport of solutes in nonof constant weight gain, because the initial J, is not zero isogravimetric lungs, i.e., those receiving arachidonic and, therefore, contributes to the estimated time zero acid, would be the result of both diffusive and convective rate of weight gain determined by extrapolation. For this transport (18). Thus, the difference in clearance between reason, we subtracted the constant rate of weight gain isogravimetric control lungs and nonisogravimetric arafound after the addition of arachidonic acid from the rate chidonic acid-treated lungs as seen in Fig. 4 may be due, of weight gain found by extrapolation to time zero. in part, to convective transport. To investigate whether The measurement of Kf in this experimental model of increased clearance was due solely to an increase in conpulmonary edema may be difficult to interpret because vective transport, J,/C, was determined in lungs in which part of the weight gain observed was due to alveolar J, was elevated by increasing venous pressure. As seen in flooding. Thus, when Kf was being determined, fluid was Fig. 4, the J,IC, in control lungs during a period of elenot only moving into the interstitium, but it was also vated J, was considerably less than that in the arachimoving through the interstitium into the alveolus. When donic acid group even though Jv was greater in control the log of the rate of weight gain was plotted vs. time, the lungs after an increase in venous pressure. curve was linear between 1 and 3 min, indicating that The results shown in Fig. 4 demonstrate that the inchanges in vascular volume were completed by 1 min. crease in protein flux in the arachidonic acid-treated During this time, there was an exponential decrease in lungs was not only due to an increase in convective transthe rate of weight gain as the semilog plot decreased linport but also to a change in permeability. This can be early with time. The exponential decrease in the rate of demonstrated by comparing the PS and G of control weight gain suggests that the forces that oppose filtraand arachidonic acid-treated lungs. J,IC, in control isotion still had a role in the injured lung and that we were gravimetric lungs is equivalent to PS. This is not true for primarily measuring fluid movement from the vascular arachidonic acid-treated lungs because these lungs were space into the interstitium. Also, by extrapolating to time gaining weight during the determination of J,/C,. In the zero, we are estimating the initial flux of water across the method used by Kern and Malik (9) to determine PS and endothelium immediately after the increase in venous g’t the lungs must be isogravimetric. Presented in Fig. 5 is pressure. the PS and g for isogravimetric control lungs and noniTo assess protein permeability, we measured the extrasogravimetric arachidonic acid-treated lungs as detervascular uptake of 1251-labeled albumin by use of a tech- mined using the linear flux equation. This method is siminique similar to that previously described (8,12). In this lar to that used by Renkin et al. (19) to estimate 0 in skin study, the 3-min uptake of 1251-albumin was divided by and skeletal muscle. In this method, linear regression of the perfusate concentration so that the albumin uptake J,/C, vs. J, was performed using clearances determined was expressed as a clearance. The JJC, shown in Fig. 4 at various J, induced by increasing venous pressure (Fig. for isogravimetric control lungs is similar to that re- 5). The intercept of this line is equivalent to PS. After the ported previously (8,9,12). After the addition of arachiaddition of arachidonic acid, the extrapolated PS was donic acid, J,/C, was increased twofold over control twofold higher than that of control lungs, indicating that lungs. In the injured lung, the increase in protein permethe permeability of the vascular barrier for albumin had ability resulted in a greater flux of tracer over the 3-min been altered. As seen in Fig. 5 the slope of the regression tracer infusion. This increase in the interstitial concenline for the arachidonic acid-treated group was greater tration of 1251-albumin, along with an increase in permethan that of the control group. The increase in the slope ability of the end&helium, may allow for a greater indicates that the convective flux of albumin is greater in amount of diffusion back into the vasculature than seen the arachidonic acid-treated lungs than in control lungs in control lungs. This would mean that we have underesat a similar J,. This agrees with the findings of Littner et timated the increase in J,/C, in the arachidonic acid- al. (ll), who found increases in convective flux when they treated lung. However, the presence of some alveolar plotted wet-to-dry weight ratios vs. albumin uptake in flooding may provide a larger interstitial pool for the di- arachidonic acid-treated dog lungs. The slope of the relution of the tracer, resulting in a decreased gradient for gression line in Fig. 5 is equivalent to 1 - G, After the back flux during the wash. In control lungs, the interstiaddition of arachidonic acid, G decreased from a value of tial concentration of radiolabeled albumin has been 0.92 to 0.68. The linear flux equation described by Kedem shown nut to be a factor, This is because the increase in and Katchalasky (6) was used in this analysis as PS and c the interstitial concentration of tracer is so small during had been calculated in a similar system using this equathe 3-min tracer infusion that the concentration gradient tion. The use of the linear flux equation permits easy for the labeled albumin during the wash period is miniseparation of the diffusive and convective terms. In this mal and a negligible amount of the interstitial tracer is equation, however, the diffusive term may be overestilost by back-diffusion (8). mated because an increase in fluid movement may alter The ability of exogenous arachidonic acid to increase the concentration gradient through the microvascular albumin flux has also been reported for dog lungs (11). wall. Fluid movement also makes it difficult to define the solute concentration within the pore. The nonlinear flux However, as in this study, the arachidonic acid-treated lungs were gaining weight during the measurement of equation (Eq 4) does not have these constraints (18). albumin uptake. Similar to the determination of K,, the When the nonlinear flux equation (Eq. 4) was fit to the Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
2446
FLUID
AND PROTEIN
PERMEABILITY
data in Fig. 5 by nonlinear least-square regression and assuming Ci = 0, the control PS was estimated to be 65 t 15 ~1 min-l g dry lung-’ and the control CJwas O&6 t 0.12 l
l
Js = J,( 1 - a)[Cp - Cie-“/( 1 - eWx)]
(4)
After the addition of arachidonic acid, the PS as calculated using Eq. 4 was 142 -t 19 ,~l. min. g dry lung-’ and CTwas 0.48 t 0.12. Our value of 0 for control lungs is similar to that found with use of tracer techniques in isolated perfused lungs (0.96) (9). This value of G is higher than that calculated by lymph protein measurements (0.61-0.70) (4, 15, 17) and the integral mass balance technique (0.77) (14). Our estimated value for c is also higher than the 0.81 reported by Ishibashi et al. (5), who used tracer kinetics to determine 0 in dog lungs. Our studies differ in that Ishibashi et al. (5) delivered 1251-albumin as a bolus and albumin flux was assessed 2 h later, whereas we infused the tracer at a const,ant rate for 3 min and assessed albumin uptake after a 3-min wash. As discussed above, the short exposure time prevents a significant increase in the interstitial tracer concentration, thereby allowing the concentration gradient for the 1251-albumin to remain constant during tracer infusion and for back flux to be negligible. An increase in protein flux, as found after addition of arachidonic acid, would increase interstitial tracer concentration. A large increase in concentration of the interstitial tracer during its infusion may result in an increase in back flux of tracer during the wash period. This would cause a greater overestimation of 0 in the arachidonic acid group compared with the control group; thus the results of a decrease in G after arachidonic acid injection are valid. This study is the first to measure both fluid conductance and protein permeability in the same lung during acute injury leading to lung edema. By measuring both of these parameters in the same lung, we found that protein permeability is changed to a greater degree than is fluid conductance atier arachidonic acid. This finding is not difficult to explain if one compares the different transvascular exchange pathways used by fluid and protein. The pathway available for transvascular exchange of water is much larger than that of protein as demonstrated by a high g in normal lungs. Although we have estimated PS and g as if the vascular barrier was a homoporous membrane, the movement of fluid and protein is believed to be through large and small pores. The vascular barrier of the lung is thought to consist of two sized pores of 80 and 200 A, with the ratio of small to large pores being 195:l (26). The small pores have been estimated to be responsible for 80% of the hydraulic conductance, whereas 16% of fluid conductance is through the large pores (26). The majority of protein transport is believed to be through the large pores. Therefore, a change in protein permeability due to the formation of a few large leaks would result in a small change in the total water pathway and the calculated fluid conductance. This suggests that a change in protein permeability could be measured in the absence ,f? 1--- -,,,,,,,Ll, -1, ---I\ :, A?l,,:A #%~-A~F&~yIcI* TL.,o
AND PULMONARY
EDEMA
changes in protein permeability, as opposed to fluid conductance, may be a more sensitive marker of lung vascular injury in models of pulmonary edema. Moreover, as shown by the present study, both of these parameters can be simultaneously measured in the same isolated lung preparation. The authors thank Dr. David Kern for assistance in the development of the isolated lung preparation in our laboratory. We also thank Ed Lewis for excellent technical support and Deborah Moran and Wendy Ward for editorial assistance. This work was supported by National Institute of General Medical Sciences Grant GM-21447, and National Heart, Lung, and Blood Institute Grants HL-38894 and HL-26807. P. A. Vincent was a National Institutes of Health (NIH) Postdoctoral Fellow supported by National Research Service Award (NRSA) F32-GM-14026. P. B. Kreienberg was a NIH Postdoctoral Fellow supported by NRSA F32-CM-14025 F. L. Minnear was a recipient of Research Career Development Award K04HL-01830. Address for reprint requests: P. A. Vincent, Dept. of Physiology and Cell Biology (A-134), Albany Medical College, 47 New Scotland Ave., Albany, NY 12208. Received 31 October 1991; accepted in final form 14 July 1992. REFERENCES 1. ANDERSON,
R. R., R. L. HOLLIDAY, A. A. DRIEDGER, M. LEFCOE, 3. W. J. SIBBALD. Documentation of pulmonary capillary permeability in adult respiratory distress syndrome accompanying human sepsis. Am. Rev. Respir. Dis. 119: 869-877, 1979. DRAKE, R., K. A. GAAR, AND A. E. TAYLOR. Estimation of the filtration coefficient of pulmonary exchange vessels. Am. J. Physiol. 234 (Heart Circ. Physiol. 3): H266-H274, 1978. FARRUKH, I. S., A. M. SCIUTO, E. W. SPANNHAKE, G. H. GURTNER, AND J. R. MICHAEL. Leukotriene D, increases pulmonary vascular permeability and pressure by different mechanisms in the rabbit. Am. Rev. Respir. Dis. 134: 229-232, 1986. GABEL, J. C., B. D. BUTLER, R. L. SCOTT, AND R. E. DRAKE. Pulmonary microvascular permeability after coronary artery ligation in dogs. J. Appl. Physiol. 55: 372-375, 1983. ISHIBASHI, M., R. K. REED, M. I. TOWNSLEY, J. C. PARKER, AND A. E. TAYLOR. Albumin transport across pulmonary capillary-interstitial barrier in anesthetized dogs. J. Appl. Physiol. 70: 21042110,1991. KEDEM, O., AND A. KATCHALSKY. Thermodynamic analysis of the permeability of biological membranes to non-electrolytes. Biochim. REID,
2. 3.
4. 5.
6.
Biophys.
AND
Acta
27: 229-246,
1958.
7. KERN, 8.
9. 10. 11.
12. 13. 14.
D. F. Blood Volume Changes and Z’ranscapillary Fluid Flow in Isolated Rabbit Lung (MS thesis). Minneapolis: Univ. of Minnesota, 1978. KERN, D. F., D. LEVITT, AND D. WANGENSTEEN. Endothelial albumin permeability measured with a new technique in perfused rabbit lung. Am. J. Physiol. 245 (Heart Circ. Physiol. 14): H229-H236, 1983. KERN, D. F., AND A. B. MALIK. Microvascular albumin permeability in isolated perfused lung: effects of EDTA. J. Appl. Physiol. 58: 372-375, 1985. LINEHAN, J. H., C. A. DAWSON, AND D. A. RICKABY. Distribution of vascular resistance and compliance in a dog lung lobe. J. Appl. Physiol, 53: 158-168, 1982. LITTNER, M. R., G. M. KAZMI, AND F. D. LOTT. Inhibition of cyclooxygenase production does not prevent arachidonate from increasing extravascular lung water and albumin in an isolated dog lung. Prostaglandins Leukotrienes Med. 15: 53-68, 1984. LITTNER, M. R., AND F. D. LOTT. Edema from cyclooxygenase products of endogenous arachidonic acid in isolated lung. J. Appl. Physiol. 67: 846-855, 1989. MAGNO, M., B. ATKINSON, A. KATZ, AND P, FISHMAN. Estimation of pulmonary interstitial fluid space compliance in isolated perfused rabbit lung. J. Appl. Physiol. 48: 677-683, 1980. MARON, M. B., C. F. PILATI, AND K. C. MEANDER. Effect of hemolysis on reflection coefficient determined by endogenous blood iniI*.,,+,,... T A ,,I DL, ,*:*I Cc).c)c)KF) r)OK7 lOQ7
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
FLUID
AND PROTEIN
PERMEABILITY
15. PARKER, J. C., R. E. PARKER, N. GRANGER, AND A. E. TAYLOR. Vascular permeability and transvascular fluid and protein transport in the dog lung. Circ. &s. 48: 549-561, 1981, 16. PARKER, J. C., M. I. TOWNSLEY, AND J. T. CARTLEDGE. Lung edema increases transvascular filtration rate but not filtration eoefficient. J. Appl. PhysioZ. 66: 1553-1560, 1989. 17. PARKER, R.E.,R. J. ROSELLI, T.R. HARRIS,AND K. L. BRIGHAM. Effects of graded increases in pulmonary vascular pressures on lung fluid balance in unanesthetized sheep. Circ. Res. 49: 11641172,198l. 18. RENKIN, E. M. Capillary transport of macromolecules: pores and other endothelial pathways. J. Appl. Physiol. 58: 315-325, 1985. 19. RENKIN, E.M.,M. GUSTAFSON-SGRO,AND L. SIBLEY. Couplingof albumin flux to volume flow in skin and muscles of anesthetized rats. Am. J. Physiol. 255(Heart Cire. Physiol. 24): H458-H466,1988. 20. RINALDO, J. E., AND R. Me ROGERS. Adult respiratory-distress syndrome: changing concepts of lung injury and repair. N. Engl. J. Med. 306: 900-909, 1982. 21. SEEGER, W.,M. MENGER,D.WALMRATH,G.BECKER, F. GRIMMINGER, AND H. NEUHOF. Arachidonic acid lipoxygenase pathways and increased vascular permeability in isolated rabbit lungs. Am. Rev. Respir. Dis. 136: 964-972, 1987.
AND PULMONARY
EDEMA
2447
22. SEEGER, W.,D. WALMRATH, M. MENGER, AND H. NEUHOF. Increased lung vascular permeability after arachidonic acid and hydrostatic challenge. J. Appl. Physiol. 61: 1781-1789, 1986. 2% SEEGER, W.,H. WOLF,G. STAHLER, H. NEUHOF,AND L. ROKA. Increased pulmonary vascular resistance and permeability due to arachidonate metabolism in isolated rabbit lungs. Prostaglandins 23: 157-173, 1982. 24* SELIG, W. M.,T.C. NOONAN, D.F. KERN,AND A.B. MALIK. Pulmonary microvascular responses to arachidonic acid in isolated perfused guinea pig lung. J. Appl. Physiol. 60: 1972-1979, 1985. 25* SHASBY, D. M., S. S. SHASBY, AND M. J. PEACH. Polymorphonuclear leukocyte: arachidonate edema. J. Appt. Physiol. 59: 47-55, 26 1985. TAYLOR, A. E., AND D. N. GRANGER. Exchange of macromolecules across the microcirculation. In: Handbook of Physiology. The Cardiovascular System. Microcirculation. Bethesda, MD: Am. Physiol. Sot., 1984, sect. 2, vol. IV, pt. 1, chapt. 11, p. 467-520. 27 TOWNSLEY, M. I., R.J. KORTHUIS,AND A.E. TAYLOR. Effects of arachidonate on permeability and resistance distribution in canine lungs. J. Appl. Physiol. 58: 206-210, 1985. 28. WANGENSTEEN,~. II., E. LYSAKER,AND P. SAVARYN. Pulmonary capillary filtration and reflection coefficients in the adult rabbit. Microvasc. Res. 14: 81-97. 1977. l
l
Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 27, 2018.
