INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

THE PULMONARY ENDOTHELIUM is capable of reducing certain blood-borne electron acceptors as they pass through the pulmonary capillary bed. The mechanisms involved include transplasma membrane electron transport systems that utilize intracellular electron donors to reduce extracellular acceptors (1, 10, 12, 34, 50). These redox systems control the pulmonary disposition of certain redox active drugs (10), and, by analogy with the functions of transplasma membrane transport systems in other cell types, they may be mechanisms by which the endothelial cells influence and/or sense the redox status of the blood (15, 19, 22, 28, 32, 35-37, 46, 47, 51). Studying metabolic processes such as these electron transport systems within an intact functioning organ is complicated by the many factors that can influence the disposition of probes for such processes within an organ (4, 7, 8, 23, 29), and probe disposition can be affected by factors other than the targeted process. The bolus injection multiple-indicator dilution (MID) method has been used for studies of other pulmonary endothelial surface reactions (9, 16-18, 20, 21, 24, 30, 49). A principal feature of this MID method is that it provides the temporal information needed to identify separately the contributions of various factors that may influence the overall disposition of a probe for a particular metabolic process. Identification of useful probes is one step in the development of an MID method for a specific cellular function. The purpose of the present study was to evaluate two thiazine compounds, methylene blue (MB) and toluidine blue O (TBO), as electron acceptor probes for the pulmonary endothelium in intact lungs. Both were previously shown to be electron acceptors for transplasma membrane electron transport in bovine pulmonary arterial endothelial cells grown in cell culture (12, 34, 40). A useful MID probe must fulfill several criteria; some are specific to the particular study conditions, but some are general. An important general criterion is that the probe must be a good enough substrate that the rate of the reaction under study is fast enough to have a significant impact on probe disposition within the 0.5- to 2-s pulmonary capillary transit time but slow enough that the reaction kinetics of interest, rather than only the rate of probe delivery, dominates the probe disposition (4, 23). Therefore, the determination of the kinetics of MB and TBO disposition on passage through the pulmonary circulation was a central focus of the present study. We addressed the problem of quantifying the kinetics of probe disposition on passage through the lungs by representing the hypothesized processes involved in the form of a mathematical model.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

Isolated Lung Preparation

To measure the venous effluent indicator concentration versus time curves, the venous effluent could be diverted into the sample tubes of a Gilson-Escargot fraction collector modified to sample at 6.67 samples/s. For experiments in which TBO was studied in PSS-albumin-perfused lungs, the venous effluent tubing passed through a previously described photodetector (40), measuring optical absorbance at wavelengths of 490 and 590 nm, before flowing into the fraction collector tubes. Each fraction collector tube contained 20 µl of 0.8 mM potassium ferricyanide solution so that complete oxidation of any reduced MB (MBH) or TBO (TBOH) in the effluent would occur before the absorbance was measured. Thus total MB, MBH + oxidized MB (MB⁺), total TBO, or TBOH + oxidized TBO (TBO⁺) concentration in the venous effluent was measured in the fraction collector tubes. The photodetector measured TBO⁺ rapidly enough after leaving the lung that autoxidation of any TBOH present was small. FITC-dextran, which was also measured by the photodetector, was included in the bolus for the purpose of determining the transport function between the two sampling sites as indicated in Bolus Injections.

Bolus Injections

To prepare the TBOH injectates, 3 ml of the respective perfusate containing 160 µM TBO, 1.5 mg FITC-dextran and 200 µM NADH were added to a test tube having two ports. The mixture was bubbled through one of the ports for 10 min with 5.8% CO₂ in N₂ to deoxygenate while maintaining the pH at 7.4. Diaphorase (0.2 U in 0.02 ml) was then added to the solution under anoxic conditions. After the disappearance of blue color from the solution, the ports were sealed and low-dead space syringes were filled with the injectate directly from the sealed tube. The injection loop was prefilled with 3 ml of deoxygenated perfusate just before the addition of the 0.9-ml TBOH injectate.

At the end of each experiment, the lungs were removed and the pulmonary arterial and left atrial cannulas were connected directly together. Boluses containing the ¹²⁵I-labeled albumin and FITC-dextran were injected and fraction collector samples and photodetector output were collected at each flow rate used in the experiment. At least one such bolus also included the respective thiazine compound. The data obtained from these samples were used to measure the part of the bolus transit time and dispersion that was caused by the injection, tubing, and sampling system and to affirm that any separation between the thiazine test indicator concentration curves and the reference indicator concentration curves was caused only by passage through the lungs and not some interaction with the perfusion system connected to the lungs.

To measure the indicator concentration versus time curves over a range of flow rates, the first step was to raise the flow rate to the highest level to be used in the experiment, with the left atrial pressure set equal to atmospheric pressure at the level of the left atrium. The resulting arterial pressure divided by 2 was the value at which the average of the arterial left atrial pressure was set for each of the lower flow rates by adjusting the height of the reservoir. This was done so that the vascular volume was approximately the same at each flow rate. The flow rate was then set at ~200, 400, 600, or 800 ml/min in TBO experiments or 50, 100, 200, or 300 ml/min in MB experiments, and a bolus was injected and samples collected as indicated above. Injections were then made at two or three of the other designated flow rates in similar fashion.

To determine whether dextran in the perfusate might have any influence on TBO or MB disposition in the lungs, after the above-described protocol had been completed in one of the lungs from the TBO and MB groups perfused with PSS-dextran, the perfusate was changed to PSS with neither dextran nor albumin. Then a TBO⁺ or MB⁺ bolus (also dextran free) was injected at a flow rate of 800 or 300 ml/min, respectively. Also, in one lung perfused with PSS-dextran the photodetector was included to confirm that effluent MBH and TBOH concentrations were below detectable levels when there was no plasma albumin in the perfusate.

Additional Measurements

The apparent octanol-to-water partition coefficient for TBOH and TBO⁺ at room temperature was determined as previously described (10) by using a concentration of 10.4 µM TBO⁺ or TBOH in 0.1 M phosphate buffer (pH 7.4). TBOH was prepared as indicated in Bolus Injections for the ultrafiltration studies. The values of the partition coefficients were 0.95 and 10.3 for TBO⁺ and TBOH, respectively. The previously reported values for MB⁺ and MBH were 0.22 and 16, respectively (10).

To determine the TBOH autoxidation rate, TBOH was prepared in a manner similar to the TBOH injectate described above. TBO and NADH (final concentration of 160 µM for both) were added to tubes containing 3 ml Hanks' balanced salt solution and 10 mM HEPES at pH 7.4 with or without 5% BSA or 5% dextran. Each solution was bubbled with N₂ for 10 min, and diaphorase (0.2 U) was added. The resulting TBOH solution (0.1 ml) was added to a spectrophotometric cuvette containing 2 ml Hanks'-HEPES buffer with or without 5% BSA or 5% dextran at pH 7.4 and equilibrated with atmospheric PO₂. Absorbance at 590 nm was recorded at 5-s intervals for 3 min in a Beckman DU 7400 spectrophotometer at 37°C.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

Experimental Results

Examples of graphs of indicator concentrations normalized to their respective injected amounts versus time, for the various combinations of PSS-dextran, PSS-albumin, MB⁺, TBO⁺, and TBOH and at different flow rates, are shown in Figs. 1-5. With PSS-dextran perfusion, only the oxidized forms were detected in the venous effluent. The presence of TBOH in the effluent after TBO⁺ injections when the lungs were perfused with PSS-albumin is reflected by the fact that the total TBO (TBOH + TBO⁺) concentrations were higher than the concentrations of TBO+ alone (Fig. 3). The fractions of MB⁺ and TBO⁺ that disappeared from the PSS-dextran and PSS-albumin perfusates on passage through the lungs are summarized as extractions in Fig. 6. The TBO⁺ extractions were higher than the MB⁺ extractions even though the flow rate range was also higher for TBO⁺. The extraction of TBO⁺ from PSS-albumin was lower than from PSS-dextran, reflecting the effects of the TBO⁺-BSA binding. When the lungs were perfused with PSS with neither albumin nor dextran, the effluent TBO⁺ and MB⁺ were indistinguishable from those when the lungs were perfused with PSS-dextran, indicating that the dextran had no detectable effect on TBO⁺ or MB⁺ extraction. When TBOH rather than TBO⁺ was injected with PSS-albumin as the perfusate, the effluent TBOH concentration curves tended to be displaced to the right in comparison to the reference indicator or to the TBO⁺ curves after TBO⁺ injection (Fig. 4). This difference reflects the tissue distribution of the TBOH (see DISCUSSION). When TBOH was injected with PSS-dextran as the perfusate, the TBOH almost completely disappeared from the perfusate (Fig. 5).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Concentration (normalized to amount of injected indicator) vs. time curves for oxidized methylene blue (MB+) and reference indicator (CR) obtained at 4 flow rates in 1 lung perfused with albumin-free perfusate [physiological salt solution (PSS)-dextran]. Solid lines are model fits to data.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Concentration (normalized to amount of injected indicator) vs. time curves for oxidized toluidine blue O (TBO+) and reference indicator (CR) obtained at 3 flow rates in 1 lung perfused with albumin-free perfusate (PSS-dextran). Solid lines are model fits to data.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Concentration (normalized to amount of injected indicator) vs. time curves for TBO+, total TBO [reduced TBO (TBOH) + TBO+], and reference indicator (CR) obtained at 3 flow rates in 1 lung perfused with albumin-containing perfusate (PSS-albumin). Solid lines are model fits to data.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Concentration (normalized to amount of injected indicator) vs. time curves for TBOH and reference indicator (CR) obtained at 3 flow rates in 1 lung perfused with albumin-containing perfusate (PSS-albumin). Solid lines are model fits to data.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Concentration (normalized to amount of injected indicator) vs. time curves for TBOH and reference indicator (CR) obtained at 3 flow rates in 1 lung perfused with albumin-free perfusate (PSS-dextran). Solid lines are model fits to data.

View larger version (29K): [in this window] [in a new window]   Fig. 6.   Average extraction (±SE) of MB+ from PSS-dextran (n = 6-8) and of TBO+ from PSS-dextran (n = 8) or PSS-albumin (n = 4) vs. flow rate. Extraction = 1  ([MB+] or [TBO+])/ CR(t), where concentrations were obtained at time of maximum of CR(t). Dashed lines are isopleths for constant values of F × ln(1  extraction) indicated.

Further evaluation of these results was carried out by expressing the hypothesized interactions between these thiazine compounds and the lung tissue in the following model.

Model

Stoichiometric equations.

1

2

3

Species balance equations. A single capillary element of the kinetic model is composed of a capillary volume (Q_c) and a surrounding tissue volume (Q_t). The spatial and temporal variations in the concentrations of the vascular reference indicator and B⁺ and BH are described by the following species balance equations based on the above stoichiometric equations and the following assumptions. 1) The reference indicator, having perfusate concentration [R], was convected through the pulmonary vascular bed without interacting with the tissue. 2) When the plasma protein (P) was present in the perfusate, the equilibration between P and B⁺ or BH was rapid relative to the capillary transit time (_c). 3) The endothelial cell reduction of B⁺ mediated by E was irreversible. 4) The association of B⁺ with Z was rapid relative to _c. 5) The equilibration of BH between Q_c and Q_t with tissue to perfusate partition coefficient  was rapid relative to _c. 6) The PO₂ and [DH] were constant under the study conditions.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

1

2

(8)

3

(9)

(10)

1

Whole organ. To construct an organ model from the single capillary element model, the distribution of pulmonary capillary transit times [h_c(t)] needs to be taken into account (4, 29). Previously, we (5) estimated that for normal rabbit lungs in this perfusion system, the pulmonary capillary mean transit time (_c) was ~44% of the total vascular mean transit time ( ), the relative dispersion of h_c(t) (RD_c = _c/_c) was ~0.9, and the skewness coefficient of h_c(t) (m³_c/³_c) was ~2, where m³_c and ³_c are the third central moment and standard deviation of h_c(t), respectively. For the present analysis, we used these values to approximate h_c(t) using a shifted random walk function, which is a probability density function whose functional values are determined by these three moments as previously described (2).

Autoxidation of BH in arteries, veins, and connecting tubing. Because BH autoxidation can occur outside the capillary region within the conducting arteries and veins and the tubing connecting the lungs to the injection and sampling sites, this autoxidation was also addressed in the model. In this part of the system, Eqs. 9 and 10 reduce to

(11)

(12)

which can be analytically solved for and [⁺](,t), where x =  is the inlet to either the capillaries or to the photodetector

(13a)

(13b)

2

2

1

View larger version (17K): [in this window] [in a new window]   Fig. 7.   TBO+ vs. time in 5% dextran and 5% albumin solutions when respective deoxygenated TBOH solutions were rapidly equilibrated with atmospheric oxygen.

Estimation of Model Parameters for Specific Experiments

TBO⁺ injections in lungs perfused with PSS-albumin. In the lungs perfused with PSS-albumin, the outflow concentrations of TBO⁺ and total TBO (TBOH + TBO⁺) were measured at different sampling sites as described in METHODS. The first step in the model interpretation of these data was to reconstruct the TBO⁺ curve measured by the photodetector as it would have appeared had it been measured at the fraction collector sampling site where the total TBO was measured. This was carried out as follows. The photodetector concentration curves acquired at 25 Hz for TBO⁺ and FITC were converted to discretely sampled curves with the same sampling time interval as the sample collector time interval using a infinite impulse response filter and decimator. The transfer function from the photodetector sampling site to the fraction collector was obtained by numerically deconvolving (13) the photodetector FITC tubing curve and the ¹²⁵I fraction collector tubing curve. The resulting transfer function was then convolved with the discretized photodetector TBO⁺ curve to estimate the TBO⁺ curve at the sample collector site.

TBO⁺ and MB⁺ injections in lungs perfused with PSS-dextran. For TBO⁺ and MB⁺ injections in lungs perfused with PSS-dextran, because [P] = 0, ₁ = 1, Q_F = Q_f, and K_red = Ek_r1 and Eqs. 9 and 10 simplify to

(14)

The model fit to the TBO⁺ or MB⁺ data from each injection was obtained by solving Eq. 14 with the initial condition [B⁺](x,0) = 0 and boundary condition [B⁺](0,t) = C_in(t)q/F. For these data, the identifiable model parameters are Q_F and K_red.

TBOH injections in lungs perfused with PSS-albumin or PSS-dextran. For the TBOH injection data, the model was fit to the total TBO data from each injection by solving Eqs. 9 and 10 with the initial conditions [⁺](x,0) = (x,0) = 0 and boundary conditions
and
The above boundary conditions account for the autoxidation of the TBOH during transit time _o upstream from the capillary inlet. For the PSS-albumin data, the identifiable model parameters are Q_T and K_seq, with Q_F and K_red fixed to the mean values estimated from the TBO⁺ injections given in Table 3. For the PSS-dextran data, the identifiable model parameter is Q_T, with Q_F and K_red fixed to the mean values estimated from the TBO⁺ injections given in Table 4 and K_seq fixed to the value given in Table 5.

Model Results

View larger version (29K): [in this window] [in a new window]   Fig. 8.   Sensitivity functions [S(t)] for 4 parameters, Kred, QF, QT, and Kseq, from model fit to TBO+ and total TBO (TBOH + TBO+) concentration data obtained after injection of TBO+ into a lung perfused with PSS-albumin. Reference indicator concentration curves [CR(t)] are also provided for timing perspective. See text for definition of model parameters.

View larger version (14K): [in this window] [in a new window]   Fig. 9.   Sensitivity functions [S(t)] for the 2 parameters QT and Kseq from model fit to TBOH concentration data obtained after injection of TBOH into a lung perfused with PSS-albumin.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES APPENDIX

Both TBO⁺ and MB⁺ were extracted from the PSS-dextran perfusate on passage through the lungs. However, the TBO⁺ extractions were considerably greater than those of MB⁺. The TBO⁺ extractions were >70% even at flow rates above those reported for the normal resting pulmonary flow rate for rabbits in this size range (276-540 ml/min; Ref. 38). This is in contrast to MB⁺, for which the extraction was only ~15% at only 300 ml/min. The role of reduction in the extraction of MB⁺ during its passage through the lungs was demonstrated previously (10). The fact that TBO⁺ was also reduced during passage through the lungs was revealed only when albumin was present in the perfusate. This is because the highly lipophilic TBOH has such a high solubility in the lung tissue that, with no protein in the perfusate, the effluent contained little trace of any TBOH that might have been formed. The presence of TBOH in the effluent when the lungs were perfused with the albumin-containing perfusate resulted from its affinity for the albumin, which offset its affinity for the tissue. The present study does not identify the location of the TBO⁺ reduction. However, a previous study, with a TBO-containing polymer that could not enter the cells, established the ability of endothelial cells to reduce TBO⁺ on the cell surface (12). Therefore, it appears likely that this surface reduction accounts for at least part of the disappearance of the relatively hydrophilic TBO⁺ within the lungs.

Another key observation was that the extractions of both MB⁺ and TBO⁺ were inversely proportional to flow rate to the extent that variations in the reduction rate parameter, K_red, with flow rate were small over the fairly wide range of flow rates studied. The implication of this observation is that the reduction rate, rather than the rate of convective supply of the indicators, dominates the extraction. One might visualize the reduction reaction as a barrier that must be traversed for TBO or MB to enter the tissue. In the terminology commonly used in the indicator dilution field, their extractions tend toward "barrier-limited" rather than "flow-limited" (23) behavior. The isopleths in Fig. 6 are for the most parsimonious nested version of the model developed herein, commonly referred to as the Crone-Renkin model (7), for which the solution for K_red is K_red = F × ln(1  extraction). For this model, barrier-limited extraction as a function of flow rate produces curves parallel to these isopleths. Thus the extent to which the TBO⁺ and MB⁺ extraction data parallel these isopleths may be thought of as a preliminary indication of the extent to which the extractions correspond to barrier- or reaction-limited behavior. The alternative of flow-limited behavior would produce horizontal lines in Fig. 6. As discussed below, the inclusion of additional detail in the model can explain most of the deviation from the Crone-Renkin model prediction observable in Fig. 6.

The model developed herein to interpret the test indicator probe disposition represents a simplistic view of the overall processing of the indicators by the lung. Steps have been left out of the stoichiometric equations that could be added if warranted by experimental manipulations of the relevant variables in future studies. The attempt was to accommodate aspects of the tissue disposition of TBO and MB predicted by previous studies under the present range of experimental conditions. Even so, the model is fairly complex, having several parameters that were either input, i.e., obtained from sources separate from the data from an individual bolus, or estimated by fitting the model to the bolus data. The fitted parameters include terms that are not separately identifiable without a more complex set of experimental conditions. Furthermore, all of the model parameters do not contribute equally to the fits to the data from each type of experiment performed or from the same type of experiment but at different flow rates. In fact, the effects of some phenomena represented in the model are not detectable in all experimental conditions. To help put this in perspective, the sensitivity functions for the model fits to the TBO⁺ and total TBO data from the PSS-albumin experiments are shown in Fig. 8. The sensitivity functions reveal the extent, and the time epoch, to which the optimized model parameters make their contributions to the fit of the model to the data (2, 6). Thus the dominant role of K_red in the fits to the TBO⁺ data, especially near the peak of the reference indicator curve can be seen in Fig. 8. Q_F has most of its influence at early and late times with respect to the reference indicator passage and little influence near the peak of the reference indicator. The very different shapes of the sensitivity functions for these two parameters also imply their relatively independent contributions to the model fit. This is expressed quantitatively for the examples in Fig. 8 by the fact that the correlations (3, 26) between Q_F and K_red were <0.34 at all flow rates studied. The physical counterpart of Q_F is not certain. It may include a volume into which at least some TBO⁺ diffuses before its reduction, possibly inside the cell. However, a compound that binds to plasma protein may likewise have an affinity for molecules on the cell surface, presumably including plasma albumin itself associated with the endothelial surface or glycocalyx (2, 43, 44).

The contribution of Q_T to the fit to the TBO⁺ data is small and barely detectable on the same scale as S(t) for K_red and Q_F at the high end of the range of flow rates studied. On the other hand, when the model is fit to the total TBO data, the contribution of Q_T is substantial and the impact of K_seq, which does not contribute to the TBO⁺ fit, is evident.

K_red, Q_F, and Q_T are the dominant parameters in the PSS-albumin experiment. The sensitivity functions for the MB⁺ and TBO⁺ experiments in the PSS-dextran-perfused lungs are not shown because they are qualitatively similar to those for TBO⁺ in PSS-albumin sensitivity functions, but without Q_T. Thus, for the PSS-dextran experiments, wherein the returning flux of TBOH is insignificant, K_red and Q_F are the controlling parameters.

The sensitivity functions for the TBOH injected into PSS-albumin reveal the importance of Q_T and K_seq (Fig. 9). TBOH is a lipophilic amine compound, and except for the small effects due to autoxidation, the model for its tissue disposition is the same as the general model developed for lipophilic amines in a previous study (3). The reasons for variations in the model parameters Q_F, Q_T, and K_seq for the MB⁺, TBO⁺, and TBOH injections with flow rate have been discussed previously (2, 3). In essence, the phenomenon occurs because two or more reversible processes having different dissociation rate constants, which are not sufficiently different for the processes to be distinguishable in the data obtained at a given flow rate, become lumped into one process in an MID model. When the flow rate is changed the relative contribution of each process to the data also changes, and the lumped parameter(s) representing these processes will be flow rate dependent. One conclusion with regard to the present study is that the reversible processes represented by Q_F and Q_T include distributed dissociation rate constants not explicitly represented in the model or evident in the data from a single flow rate (3). Likewise, the term sequestration applied to K_seq means that the dissociation rates are long relative to the capillary transit time, not that the processes are necessarily irreversible.

The overall tissue disposition of TBO and MB apparently involves the reduction, and nonspecific binding of the oxidized forms, and the partitioning of the reduced forms between tissue and perfusate volumes as represented in the model. However, it has also been demonstrated in endothelial cells in culture that TBOH can be reoxidized and sequestered within the cells (33). The sequestration apparently occurs because the oxidized forms of the thiazines are strongly polar, cationic (14), relatively hydrophilic compounds and thus relatively membrane impermeant. Thus the oxidized forms cannot so readily escape their intracellular site(s) of oxidation, and partitioning in accordance with membrane potentials of the sequestering intracellular organelle(s) may play a role as well (27). This sequestration contributes relatively little to the disposition of the TBO within the time frame of the bolus transient in normal lungs, as evidenced by the small value of K_seq (<3 ml/s) in comparison to K_red (>17 ml/s). However, preliminary results suggest that the sequestration rate may be increased in the lungs of rats adapted to hyperoxia, to the extent that it can have a substantial influence even during the bolus transient (39). Thus K_seq may take on greater significance in conditions in which intracellular reactions that reoxidize the reduced thiazine compounds may be affected.

General aspects of the sensitivity to the input parameters controlling bolus dispersion and capillary transit times have been discussed at length previously (4, 29). The main reason for the larger values of K_red obtained from the present model, reported in Tables 3, 4, and 6, in comparison to the Crone-Renkin model predictions, which can be read in Fig. 6, is the effect of capillary perfusion heterogeneity not included in the Crone-Renkin model. One of the important reasons for parameterizing the data using the particular modeling approach described herein is the potential for preventing a change in perfusion heterogeneity from being aliased by a change in tissue parameters (4). In the present study, the capillary transport function was assumed to be typical of the normal rabbit lung (5). However, in future studies of abnormal conditions, in which that assumption would not necessarily hold, inclusion of a flow-limited indicator such as labeled water in the bolus would provide the necessary information to account for a change in perfusion heterogeneity as discussed previously (4, 42).

Another input parameter was the TBOH autoxidation rate (k_o), which was included to evaluate the assumption that the TBO⁺ data measured on-line can represent the TBO⁺ exiting the lung capillaries without contamination by autoxidation of effluent TBOH. The autoxidation turned out to be a minimal contributor to the on-line signal as can been seen in Fig. 10, in which a model simulation of the effluent TBO⁺, the effluent TBOH, and the TBO⁺ resulting from autoxidation of effluent TBOH is shown for the lowest flow rate in the range studied (i.e., the flow rate providing the longest time for autoxidation to occur).

View larger version (16K): [in this window] [in a new window]   Fig. 10.   Simulated effluent concentration vs. time curves after injection of TBO+ in a lung perfused with PSS-albumin. Curves are for TBO+ that passed through lungs without being reduced, TBOH generated by reduction of TBO+, and TBO+ resulting from autoxidation within perfusate of TBOH formed in lungs.

As indicated above, it was the presence of the albumin in the perfusate that revealed the reduction of TBO⁺. Because even without albumin binding the MB⁺ extraction was low, we did not pursue the consequences of including albumin in the perfusate for MB uptake. On the other hand, TBO uptake was high enough to provide an adequate window for examining the influence of albumin binding. As expected from the fact that TBO⁺ associates with plasma albumin, the presence of albumin decreased the rate of TBO⁺ disappearance from the perfusate. However, the decrease was no