INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

INCREASES IN LEFT ATRIAL PRESSURES cause edema accumulation in perivascular and peribronchial tissues of the lungs (13, 10). Much less fluid appears in the tissues surrounding the capillaries, and it is not known whether leakage occurs from alveolar or extra-alveolar vessels. Weibel and Bachofen (14) suggested that fluid is lost from the capillary surfaces and then makes its way along interstitial fibers to lymphatics, which terminate at the respiratory bronchioles. However, leakage through other vessels is also possible. In the early studies of Iliff (8), increased alveolar pressures were associated with leakage from extra-alveolar vessels. Subsequent studies in which air space pressures were increased or the arteries or veins were embolized with polystyrene beads suggested that extra-alveolar filtration is about the same as alveolar filtration (1). Interpretation of these experiments is complicated by the possibility that high airway pressures or embolism might alter the permeability of the extra-alveolar vessels.

Filtration in individual venules on the surface of isolated lungs is significantly greater than that from arterioles, and the filtration coefficients of the extra-alveolar vessels greatly exceed that of the capillaries (2). Freeze-fracture studies of Schneeberger and Karnovsky (12) indicate that there are discontinuities in the sealing strands that join adjacent cells at the venous end of the capillaries, suggesting that increases in hydrostatic pressure might cause greater filtration through these regions than through the capillaries. It is also possible that leakage of fluid may occur through bronchial vessels.

To obtain information concerning the sites of edema formation in the pulmonary vasculature, a series of stop-flow studies was conducted. These experiments provided information regarding the relative distribution of filtration and diffusion along the pulmonary vasculature.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Twenty-six Sprague-Dawley rats [300 ± 25 (SD) g] were anesthetized with an intraperitoneal injection of 0.7 ml of a 64.8-mg/ml pentobarbital sodium solution. The chest was opened, and catheters were placed in the trachea, pulmonary artery, and left atrium. Blood was flushed from the vasculature at 37°C with 5 ml of perfusion fluid that contained (in mM) 25 NaHCO₃, 125 NaCl, 4 KCl, 2.5 CaCl₂, and 0.8 MgSO₄, as well as 150 mg/dl glucose, 10 mg/dl urea, and 5 g/dl bovine serum albumin (Cohn Fraction V, 98-99% albumin, Sigma Chemical, St. Louis, MO) adjusted to pH 7.4 with 1 N NaOH when exposed to 5% CO₂. The perfusate also contained 2,000 mg/l fluorescein isothiocyanate (FITC)-labeled dextran, with an average molecular weight (mol wt) of 2,000,000 (Sigma). In some experiments the albumin was labeled with Evans blue (100 mg/l, Aldrich, Milwaukee, WI). The lungs were kept inflated with a 5% CO₂-95% O₂ mixture at a constant pressure of 5 cmH₂O. Pulmonary vascular pressures were measured just proximal to the pulmonary artery catheter and distal to the left atrial catheter.

The apparatus used in these experiments is shown in Fig. 1. The valves were used to 1) begin perfusion of the lungs, 2) stop flow and expose the left atrium to a pressure of 20 cmH₂O, and 3) flush the pulmonary vasculature after the stop-flow period in an anterograde or retrograde direction. The dead space volume of the tubing between the left atrium and the collection site was 1.5 ml. Temperatures were measured with a thermocouple inserted into the arterial line just proximal to the point at which the catheter entered the pulmonary artery.

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Stop-flow apparatus. Lungs were initially perfused for 10 min with recirculation (not shown). Thereafter, lungs were perfused for 1 min without recirculation, and samples were collected from outflow (initial perfusion). Flow was then stopped, and lungs were exposed to 20-cmH2O pressure for 10 min (stop-flow). Flow was resumed from pulmonary artery (pul art) to left atrium (atr; anterograde flush) or from left atrium to pulmonary artery (retrograde flush), and samples were again collected. Perfusate and lungs were kept warm with fluid recirculated from a water bath around lower reservoir, tubing, and lungs. Fluid was placed in upper reservoir just before stop-flow period to ensure temperature was also kept close to 37°C.

Perfusate was initially recirculated through the lungs at 6 ml/min for 10 min from a reservoir containing 100 ml of perfusate at 37°C. After this initial period of recirculation, the lungs were perfused without recirculation at 6 ml/min for 1 min (see Fig. 1) from the pulmonary artery to left atrium, and baseline samples were collected. Left atrial pressures were kept at zero during these intervals. Perfusion was then stopped for 10 min, during which time the left atrial outflow was connected to an upper reservoir in which the top of the fluid column was kept 20 cm above the lungs. At the end of the 10-min period, 0.5 ml of perfusate containing 9 µCi of ³HOH (DuPont, Boston, MA) was instilled into the air spaces. Then, the perfusate in the lungs was flushed with either anterograde (pulmonary artery to left atrium) or retrograde (left atrium to pulmonary artery) perfusion at 3 ml/min for 4 min. Outflow samples were collected by hand at 10-s intervals, and the volume of fluid collected in these samples was determined from the weight of the collection cups. The volume of fluid remaining in the upper reservoir was continuously monitored from the weight of a hollow 10-cm-long, 0.5-cm-diameter glass cylinder that was suspended in the reservoir from a strain gauge.

Four sets of experiments were conducted (see Table 1). In group 1, the fluid remaining within the lungs was flushed from the vasculature after the stop-flow period from the pulmonary artery to the left atrium (anterograde flush). In group 2, the fluid was flushed from the left atrium to the pulmonary artery (retrograde flush). In group 3, 1.0 ml of 0.1 N HCl in 154 mM NaCl was instilled into the air spaces before the lung was perfused, and the lung was flushed in an anterograde direction. Three control experiments (not shown in Table 1) were also conducted in which 1.0 ml of 154 mM NaCl was instilled into the lungs without HCl. In group 4, Evans blue was not incorporated in the perfusate and the lungs were perfused in an anterograde direction.

Samples were diluted 1:9 and centrifuged at 13,000 g for 10 s in a Fisher Scientific Marathon Centrifuge-13K (Pittsburgh, PA), and optical densities were measured at 495 nm (FITC) and 620 nm (Evans blue) in a Milton Roy 1001 Plus spectrophotometer (St. Louis, MO). In addition, 0.1-ml aliquots from the collection vials were added to 3 ml of Biosafe II (Research Products International, Mount Prospect, IL). Radioactivity of the sample was determined with an automated beta scintillation counter and corrected for background counts. Radioactivity of the samples was divided by that in the solution instilled into the lungs and plotted separately against the volume of fluid recovered from the left atrium.

Concentrations of protein (only albumin was present in the perfusate) were measured by the Bradford technique (3) in five experiments that did not include Evans blue.

Concentrations of dye and unlabeled albumin in samples flushed from the lungs after the stop-flow period were divided by the concentrations in the control outflow, and these were plotted against the volume of fluid recovered from the left atrium. The volume (V) of fluid not containing FITC-dextran that was lost from the ith sample of perfusate was calculated with use of the equation

(1)

where [dex]_in designates the concentration of FITC-dextran in the perfusate fluid entering the lungs, [dex]_out refers to the concentration of FITC-dextran in the ith sample of the perfusate flushed out of the lung after the stop-flow period, V_out is the volume of fluid in this collection sample, and V_in is the volume of fluid that would have been in the sample had none been lost by filtration from the perfusate.

The mean outflow volumes () of the increases in FITC-dextran and ³HOH above baseline were calculated from the equation

(2)

where c_i is the concentration of FITC-dextran or ³HOH in the ith sample, c₀ is the concentration of the indicator in the baseline period just before the stop-flow interval (this is zero for ³HOH), V_i is the volume of the ith sample, and V_D is the dead space of the tubing.

Mean values of _dex/_³HOH for each of the sets of experiments were compared by a one-way analysis of variance and a Neuman-Keuls test of differences between individual mean values (SigmaStat I, Jandel, San Rafael, CA).

To test binding of Evans blue to albumin, samples of the original perfusate containing Evans blue were filtered by centrifugation at 1,000 g at room temperature through a 30,000-mol-wt-cutoff polysulfonated filter (Microcon, Beverly, MA). No blue dye was detected in the filtrate. As a further test of binding, 1 ml of perfusate without dye was placed in a 10,000-mol-wt-cutoff dialysis chamber (Slide-A-Lyzer dialysis cassette, Pierce, Rockford, IL) and immersed in a flask containing 200 ml of the perfusate with Evans blue that was slowly stirred overnight at room temperature. None of the Evans blue dye was found in the dialysis chamber. Samples of the perfusate containing FITC-dextran (mol wt 2,000,000) were also filtered by centrifugation at 1,000 g at room temperature through a 100,000-mol-wt-cutoff polysulfonated filter (Microcon). None of the FITC-dextran was found in the filtrate.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Elevation of left atrial pressure to 20 cmH₂O during the stop-flow period resulted in increases in the concentrations of FITC-dextran, (mol wt 2,000,000) in the perfusate flushed from the pulmonary vasculature, regardless of whether an anterograde or retrograde flush was used (Figs. 2 and 3). These increases in the concentrations of FITC-dextran were attributed to transudation of fluid that contained relatively low concentrations of FITC-dextran from the vasculature during the stop-flow period. If it is assumed that none of the FITC-dextran is lost from the capillaries when edema formation is occurring, then the amount of fluid that entered the interstitium of the lungs during the stop-flow period averaged 0.22 ± 0.02 (SE) ml in the anterograde experiments and 0.19 ± 0.04 ml in the retrograde experiments.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Group 1 experiments. After stop-flow interval, lung was flushed from artery to vein (anterograde flush). Note that increases in outflow concentrations (Conc) of both fluorescein isothiocyanate (FITC)-dextran and Evans blue-albumin precede those in 3HOH. This is consistent with filtration from vessels that are more venous than those that were responsible for 3HOH exchange. Increases in Evans blue-albumin concentrations are less than one-half of those in FITC-dextran, indicating losses of Evans-blue albumin from pulmonary vasculature during stop-flow period. Dotted line here and in Figs. 3 and 4 indicates the position of 3HOH curve shown below. Maximal values of this curve were set to those of Dextran to facilitate comparison.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Group 2 experiments. After stop-flow, pulmonary vasculature was flushed from left atrium to pulmonary artery (retrograde flush). Increases in concentrations of FITC-dextran and Evans-blue albumin occur at about same time as those in 3HOH. This suggests that relatively little filtration occurred from vessels that are more arterial than from sites of 3HOH exchange from air spaces.

Instillation of 0.1 M HCl into the lungs before the stop-flow period significantly increased the rise in FITC-dextran concentrations (compare Figs. 2 and 4, noting difference in ordinate scaling). This corresponds to a loss of 1.01 ± 0.31 ml (P < 0.05) from the vasculature to the interstitium.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Group 3 experiments. Before perfusion was commenced, 1.0 ml of 0.1 N HCl was instilled into lungs (anterograde flush). Increases in concentrations of FITC-dextran were ~5 times greater than in control experiments (e.g., compare with Fig. 2, noting difference in scale of ordinate). Increases in Evans-blue albumin were <10% those of FITC-dextran, indicating increased losses of Evans-blue from vasculature.

Increases in concentrations of Evans blue-albumin were considerably less than those of FITC-dextran (Figs. 2-4), suggesting more of the labeled albumin was lost from the vasculature than was FITC-dextran. Increases in Evans blue-albumin averaged only 38 ± 8% of those of FITC-dextran in the anterograde and retrograde experiments (no differences were noted in this regard between the anterograde and retrograde experiments). After instillation of HCl into the lungs, increases in Evans blue-albumin fell to 5.4 ± 3.0% of those in FITC-dextran, indicating a rise in the fraction of Evans blue-albumin lost from the pulmonary vasculature. Instillation of 1 ml of isotonic saline without HCl (in 3 experiments) did not cause any increase in FITC-dextran concentrations and did not influence increases in Evans blue-albumin or ³HOH concentrations.

No evidence was found for any detectable dissociation of Evans blue from albumin in the infusion solution by either ultrafiltration or dialysis (see METHODS). It is possible that dissociation of Evans blue from albumin was due to metabolism or uptake in the pulmonary vasculature during the stop-flow period. However, increases in unlabeled albumin approximated those in FITC-dextran during the stop-flow period (see Fig. 5), suggesting that there was no metabolism or uptake of unlabeled albumin during this interval.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Group 4 experiments. Evans blue was not incorporated in perfusate of these experiments (anterograde flush). Increases in unlabeled albumin were comparable to those in FITC-dextran, suggesting little loss of unlabeled albumin.

When the fluid remaining in the pulmonary vasculature was flushed out of the lungs in an anterograde fashion, increases in FITC-dextran concentrations occurred before those in ³HOH (Fig. 2), which had been instilled into the air spaces [_dex/_³HOH = 0.56 ± 0.04 (SE), n = 6]. This observation is consistent with the hypothesis that filtration had occurred at more-venous sites than those through which the ³HOH had exchanged. When the lungs were flushed in a retrograde direction, increases in FITC-dextran and ³HOH were observed in the outflow at about the same time (_dex/_³HOH = 0.97 ± 0.05, n = 6) (Fig. 3), suggesting that there had been relatively little ultrafiltration from vessels that were more arterial than those through which exchange of ³HOH had occurred. Differences in _dex/_³HOH between the retrograde and each of the anterograde groups were significant (P < 0.05).

In the lungs that were flushed in an anterograde fashion after exposure to 0.1 N HCl, increases in FITC-dextran concentrations occurred before those in ³HOH (_dex/_³HOH = 0.80 ± 0.06, n = 5). This ratio was significantly more than that observed in the control anterograde study, but it was still less than that observed when retrograde flow was used to flush the lungs (P < 0.05). It can therefore be concluded that overall filtration was increased after exposure to HCl, but the increase in prevenous filtration may have exceeded that lost from the veins. Evidence for venous filtration was also noted when an anterior flush was used in the lungs that were perfused with a solution that did not contain Evans blue (_dex/_³HOH = 0.76 ± 0.02, n = 6).

More fluid was lost from the upper reservoir during the stop-flow period than could be accounted for by increases in concentrations of FITC-dextran (see Fig. 6). An average of 4.2 ± 0.1 ml was lost from the reservoir in the seven anterograde and retrograde experiments in which this parameter was measured, or ~20 times as much as that calculated from the FITC-dextran data. Even more fluid (12.0 ± 2.3 ml) was lost from the pulmonary vasculature during the stop-flow period after exposure of the air spaces to 0.1 N HCl in isotonic saline.

View larger version (32K): [in this window] [in a new window]   Fig. 6.   Volume of fluid lost from upper reservoir during stop-flow interval. Instillation of 0.1 N HCl into the air spaces before start of experiments (group 3 experiments) accelerated losses of perfusate from upper reservoir compared with experiments without HCl exposure (groups 1 and 2 experiments).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Stop-flow experiments have been used for many years to determine the sites of solute and water exchange in the nephron (9). In the latter experiments (9), urine flow was arrested for a period of time and then resumed, allowing serial samples to be analyzed to determine whether solute and fluid movement occurred more proximally or distally. Application of this approach to the lungs has been much more limited, despite the fact that there are potentially some intrinsic advantages to using stop-flow studies in the lungs. Because access to both the upstream and downstream compartments is available, it is possible to use either anterograde or retrograde flows to flush the contents of the vasculature from the pulmonary circulation.

The observation that, when the lungs are flushed in an anterograde direction after the stop-flow period, increases in FITC-dextran and albumin precede those in ³HOH is consistent with the conclusion that some filtration occurs at sites which are more venous than the site at which ³HOH enters the pulmonary vasculature after instillation into the lungs. On the other hand, the failure to observe a similar pattern when the lungs are flushed from the venous to arterial vessels suggests that relatively little filtration occurred through vessels more proximal than the capillaries. Even in lungs injured with instillation of 0.1 N HCl in isotonic saline, filtration, which was greatly augmented, appeared to occur in more-venous vessels, although _dex/ _³HOH was not as low in the group 3 or group 4 experiments as it was in the initial series of anterograde flush studies.

Evans blue has been used for many years as a convenient tracer for albumin in physiological experiments. In single-pass studies with this dye, intravenous injections of hypertonic solutions resulted in comparable dilution of Evans blue-labeled protein and hemoglobin, suggesting that little had leaked out of the vasculature (5). However, recent studies of Evans blue-labeled albumin and ¹²⁵I-labeled albumin conducted in perfused rat lungs showed considerably greater pulmonary uptake of Evans blue than ¹²⁵I-albumin over a 3-min period (4). No significant dissociation of dye from albumin was found in the present studies when the perfusate was ultrafiltered through a membrane with a cutoff of 30,000 mol wt or when it was dialyzed against a protein-free solution. The unlabeled albumin was placed in the dialysis bag in the latter experiments so that any free Evans blue in the large volume of fluid in the external solution would have the opportunity to equilibrate with the small volume of fluid in the dialysis bag and then bind with albumin in this compartment. This would appear to argue against significant dissociation of Evans blue from albumin before it reached the lungs.

It is possible that exposure of the lungs to an albumin solution containing Evans blue increased the permeability of the lungs to albumin. However, increases in the concentrations of FITC-dextran (mol wt 2,000,000) in experiments in which Evans blue was included were similar to those in which unlabeled albumin was used. This suggests that the Evans blue did not increase the filtration coefficient of the pulmonary vasculature.

The similarity of outflow emergence of unlabeled protein (albumin, mol wt 68,000) and FITC-dextran (mol wt 2,000,000) from the lungs suggests that these molecules provide a reasonable index of filtration in undamaged lungs. Ideally, a comparison with a cellular element such as red blood cells would permit assessment of how many of these macromolecules may have been lost through the capillary walls (11, 15). Unfortunately, settling of red blood cells during the stop-flow period made it difficult to utilize this approach. Nevertheless, the close correspondence of these curves suggests that neither macromolecule crosses normal capillary walls during a 10-min stop-flow period. More albumin than FITC-dextran would probably be lost if the lungs were sufficiently injured: it is likely that the failure of Evans blue concentrations to increase as much as FITC-dextran after instillation of HCl into the air spaces reflects increased loss of albumin molecules.

Excision of the lungs from the chest inevitably resulted in tearing of bronchial vessels, and much of the fluid from the upper reservoir was presumably lost through disrupted vessels. Instillation of HCl into the air spaces before perfusion increased losses of perfusate from the upper reservoir.

Most studies of filtration in the lungs have been made by collecting lymphatic fluid or measuring changes in organ weight. Interpretation of lymphatic experiments has been complicated by uncertainty concerning the tissues drained by the lymphatics and the effect of the nodes themselves on lymph constituents. In gravimetric studies, it is difficult to distinguish alterations in the vascular, interstitial, and cellular compartments from increases in lung weight when venous pressures are increased. Measurements of filtration in continuously perfused preparations require a delay before changes in indicator concentration can be detected because of the large amount of perfusate needed to fill the reservoirs and tubing used in these experiments. The stop-flow approach may prove useful because it permits measurement of filtration out of the circulation over a 10-min interval from a relatively small volume of perfusate, i.e., that which is within the vasculature during the stop-flow period. Furthermore, it can potentially provide information regarding the relative arterial and venous sites of filtration, diffusion, and metabolism in the intact lung, which would be difficult to obtain by other experimental approaches.

There are a number of modifications that might be made in the stop-flow approach that could provide additional information regarding the distribution of diffusion and filtration in the pulmonary vasculature. Rather than the use of instillation of ³HOH into the air spaces to mark the microvascular environment, potassium analogs such as ²⁰¹Tl⁺ or ⁸⁶Rb could be included in the perfusate (6, 7). Because these indicators become concentrated in lung cells, concentrations within the vasculature fall to very low levels, and the site of solute exchange would be indicated by a fall in the concentrations of these indicators. This would avoid any possible alterations in local flow induced by instilling small volumes of fluid into the air spaces. The fact that increases in outflow concentrations of ³HOH and FITC-dextran occurred in the same samples in the retrograde experiments suggests that instillation of small volumes of fluid does not have a significant effect on regional flow in these experiments. Although the sites of diffusion of ³HOH on the one hand and ²⁰¹Tl⁺ on the other cannot be precisely defined, the fact that the capillary surface area greatly exceeds that of the arterial and venous vessels makes it likely that most of these indicators diffuse across capillary membranes. The use of a particulate indicator that remains in suspension during the stop-flow period represents another potentially valuable modification in the stop-flow approach because it would make it possible to detect losses of high-mol-wt molecules such as FITC-dextran and albumin.