INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ALVEOLI IN THE UPPER LUNG are larger than in the lower lung due to the stretching effect of gravity (3). We hypothesize that the expanded alveoli in the upper lung stretch the capillaries embedded within their walls into oval cross sections. In turn, the oval-shaped capillaries cause the disk-shaped red blood cells to flow with their broad sides facing alveolar gas (Fig. 1). This orientation favors gas exchange, because the large surface area of the cell is placed in close proximity to the alveolar gas surrounding both sides of the cell. In contrast, lower lung alveoli are less expanded, and their walls are less stretched (Fig. 1). This permits the capillaries in those alveolar walls to assume circular cross sections, a configuration that allows red blood cells to transit the capillaries in many different orientations (Fig. 1). Some orientations, e.g., edge facing alveolar gas, are less favorable for gas exchange because of the reduction in surface area in close proximity to alveolar gas and the consequent increased diffusion distance necessary to reach all of the hemoglobin in the red blood cell. We tested this hypothesis by rapidly freezing rat lungs and histologically determining red blood cell orientations in the alveolar capillaries. The small size of rat lungs precluded an appreciable effect of gravity on regional alveolar expansion but did permit rapid freezing. We mimicked the effect of gravity on lung expansion by using either high- or low-inflation pressures in two groups of rats immediately before rapidly freezing the lungs. We estimated the effect of red blood cell orientation on gas diffusion using finite-element analysis. To simplify the analysis, we modeled carbon monoxide, rather than oxygen, uptake.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Alveoli during high- and low-inflation pressure. We hypothesize that expanded alveolar walls will stretch the capillaries into oval cross sections, causing the disk-shaped red blood cells to flow with their broad sides facing alveolar gas. This orientation favors gas exchange, because the large surface area of the cell is placed in close proximity to the alveolar gas surrounding both sides of the cell. In contrast, the less expanded alveolus permits the capillaries to assume a circular cross section, a shape that allows red blood cells to transit the capillaries in many different orientations. Some orientations (e.g., edge-on to alveolar gas) are less favorable for gas exchange, because of the reduction in surface area immediately adjacent to alveolar gas and the increased diffusion distance to reach the hemoglobin in the red blood cell.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Red blood cell orientation. Eight adult female Sprague-Dawley rats, 272 ± 18 (SD) g, were anesthetized by intraperitoneal injection of pentobarbital sodium (90 mg/kg) dissolved in 0.9% saline. After tracheal cannulation and a sternotomy, the lungs were held for 10 s at an inflation pressure of either 6 or 20 cmH₂O. These airway pressures were based on work done by Godbey et al. (4) on isolated dog lung lobes. Liquid propane, cooled to 189.9°C via liquid nitrogen, was steadily poured into the open thorax, rapidly freezing the lungs. Chiseled pieces of the frozen lungs were removed and immediately placed in liquid nitrogen. Only small lung pieces that had a pleural surface were quickly transferred to a freeze-substitution, solvent-fixative solution (1% HgCl₂ in 100% ethanol) that was cooled to 70°C. The lung pieces were kept at 70°C for 7 days with four changes of fixative solution. After 7 days, the lung pieces were warmed to room temperature, rinsed with 100% ethanol, and infiltrated and embedded in Polysciences JB-4 medium. The lung pieces were oriented in the embedding medium so that each pleural surface would be cut at a right angle. Sections (1-2 µm thick, Sorvall MT-1 microtome) were stained with Biebrich's scarlet-acid fuchsin.

View larger version (56K): [in this window] [in a new window]   Fig. 2.   Orientation of red blood cells in alveolar walls was determined by measuring the length of the visual aspect of cell that was parallel to the alveolar wall (major axis) and the aspect perpendicular to the alveolar wall (minor axis). The major axis was divided by the minor axis to obtain the aspect ratio. High-aspect ratios indicated that the diameter of the red blood cell was nearly parallel to the alveolar wall. Low-aspect ratios indicated red blood cell diameter was at a high angle with respect to the alveolar wall with its broad side tipped away from alveolar gas.

Finite-element analysis. Hsia et al. (8-10) have developed an application of finite-element analysis applied to models of red blood cells in pulmonary capillaries. Our finite-element modeling was based on their work. We compared two cases of red blood cell orientation in capillaries. The first orientation occurred in an oval-shaped capillary, causing the red blood cell to pass along the capillary with its broad side facing alveolar gas (Fig. 3, top). In the second case, the disk-shaped red blood cell had its edge toward alveolar gas (Fig. 3, bottom), an orientation permitted by a circular capillary cross section. Although a circular cross-section capillary would not constrain the orientation of red blood cells flowing within it, we chose to model a red blood cell with its disk edge toward alveolar gas to determine the diffusing capacity for this least favorable orientation.

View larger version (37K): [in this window] [in a new window]   Fig. 3.   Two orientations for finite-element analysis were used. Top: the first orientation occurred in an oval-shaped capillary, causing a red blood cell to pass along the capillary with its broad side facing alveolar gas. Bottom: in the second case, a disk-shaped red blood cell had its edges toward alveolar gas, an orientation permitted by a circular capillary cross section. The model was simplified by making the surface of the red blood cells flat, instead of biconcave. Computing time was reduced by using only one-fourth of the red blood cell for the finite-element analysis. FEM area, synonymous with finite-element analyses.

(1)

(2)

(3)

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Region of model was divided into a minimum of 5,000 adjacent triangles. Using Eq. 3, software calculated PCO at each of the nodal points formed by the mesh of triangles. After the spatial distribution of the PCO gradient was determined, the flux of carbon monoxide (arrows) along the gradient was calculated. Diffusing capacity for carbon monoxide is shown for a red blood cell with its broad side toward alveolar gas (top) and for a red blood cell with its disk edge toward alveolar gas (bottom).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 5 shows photomicrographs of red blood cells in alveolar walls in both high and low lung inflation groups. Aspect ratios were determined for 200 red blood cells in four rats (50 cells in each animal) that received 6-cmH₂O pressure lung inflation and for 200 red blood cells in another four rats (also 50 cells in each animal) with 20-cmH₂O pressure lung inflation. The average red blood cell aspect ratio was 1.44 ± 0.40 (SE) for the low-inflation pressure group and was 2.55 ± 0.06 (SE) for the high-inflation pressure group (P < 0.00001). The distribution of aspect ratios for both groups of rats was skewed toward greater aspect ratios (Fig. 6). By finite-element analysis, the diffusing capacity for carbon monoxide (DL_CO) for a red blood cell with its broad side toward alveolar gas (Fig. 4, top) was 0.911 µm³ · s¹ · Torr¹. For a red blood cell with its disk edge toward alveolar gas (Fig. 4, bottom), the DL_CO was 0.543 µm³ · s¹ · Torr¹, a 40% decrease.

View larger version (48K): [in this window] [in a new window]   Fig. 5.   Photomicrographs of red blood cells in alveolar septa during both high (top) and low lung inflation (bottom; same magnification).

View larger version (42K): [in this window] [in a new window]   Fig. 6.   Distribution of aspect ratios for both groups of rats was of similar shape. The high lung inflation pressure group was shifted toward higher aspect ratios (P < 0.00001).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The way red blood cells are oriented with respect to alveolar gas depends on the level of alveolar inflation. We conclude this from rat lungs inflated with high airway pressure before they were rapidly frozen. The enlarged alveoli stretched capillaries into oval cross sections. The oval-shaped capillaries constrained the red blood cells to move through them with the broad sides of the cells oriented toward alveolar gas, thereby minimizing intravascular diffusion distance and reducing the total diffusion distance from gas into hemoglobin. In lungs rapidly frozen during low-pressure inflation, alveolar walls were not as stretched, and the capillaries assumed a nearly circular cross section. The circular shape permitted the red blood cells to flow through the capillaries in various orientations, including edge-on toward alveolar gas, an orientation that added to the intravascular diffusion distance for alveolar gas to reach hemoglobin. We determined the effect on gas uptake of these orientations by finite-element analysis to estimate the DL_CO of single red blood cells in broad-side vs. edge-on orientations. Red blood cells in the edge-on orientation had a 40% lower diffusing capacity.

There are several issues that need to be considered in interpreting these data. First is whether our rapid freezing technique arrested the moving red blood cells with sufficient rapidity to maintain their orientation. We optimized the rapid freezing in the following ways. 1) Rats were chosen for their small lung size. 2) The chest of each animal was widely opened to expose the lungs as completely as possible, and the chest was rapidly filled with liquid propane. 3) Liquid propane was used because of its wetting property. The propane was cooled 145.8°C below its boiling point by placing its container in a bath of liquid nitrogen (195.8°C), which ensured that the propane would remain a liquid and continue to freeze the tissue as it warmed. Typically we cooled the liquid propane until propane ice crystals formed at 189.9°C. 4) Only red blood cells that were within three alveoli of the pleural surface were measured. All red blood cells in these alveoli had smooth surfaces when viewed by light microscopy, indicating that the cells were frozen in a vitrified state. Some red blood cells in the walls of alveoli more distant from the pleural surface had pointed edges, suggesting that freezing was slow enough for ice crystals to form (15). The further inward from the pleural surface, the greater was the proportion of irregular-shaped red blood cells.

The estimated freezing time for blood in these near-surface alveoli was 100 ms (15). Presson et al. (12) showed that the average capillary transit time for red blood cells in pentobarbital-anesthetized dogs was 4.1 s when cardiac output was basal. During this time, the cells traveled across approximately three alveoli, each ~100 µm in diameter. Therefore, their velocity was ~73 µm/s (300 µm  4.1 s). If we assume a similar velocity for red blood cells in an anesthetized rat, it would mean that a freezing time of ~100 ms would stop the cells completely in ~7 µm and would begin the slowing process in an even shorter distance. Because there are not large twisting forces affecting the red blood cells as they pass through the pulmonary capillaries, as shown by the lack of spinning and flipping of the cells viewed with in vivo microscopy, we think that the cells being frozen in place within their own body length is a short enough distance to maintain the in vivo orientation.

When the pulmonary microcirculation is viewed with in vivo microscopy, the red blood cells traversing the pulmonary capillaries on the upper surface of the lung are oriented with their broad sides toward alveolar gas. Although the orientation is visually obvious, the cells move too rapidly for quantitative measurements to be made. We have obtained a few high-speed photomicrographs, which demonstrate the orientation in the living lung (16). A typical example is shown in Fig. 7, which supports the idea that these orientations exist in intact animals.

View larger version (79K): [in this window] [in a new window]   Fig. 7.   Typical in vivo, high-speed photomicrograph of red blood cells (RBCs) traversing a well-expanded alveolus in a canine lung. This subpleural alveolar wall is flat with alveolar gas just below the optical plane. The disk-shaped objects are RBCs flowing through a single capillary. The orientation of the RBCs with respect to alveolar gas is uniform and shows that capillaries in a living lung under zone 2 conditions constrain the cells to be oriented with their broad sides toward alveolar gas.

To determine the importance of red blood cell orientation with respect to gas diffusion in pulmonary capillaries, we used a computer model to analyze how gas diffusion was influenced by different red blood cell orientations. Modeling of gas diffusion on a microlevel became possible with high-speed computers and the elegant application of finite-element analysis developed by Hsia et al. (8-10), which they applied to models of red blood cells in pulmonary capillaries. Our analysis was based almost entirely on their method. Once our computational routines were running, we checked our model by using the same variables they did (8) and obtained the same results. Hsia et al. (8) studied the effect of hematocrit on diffusing capacity using these methods and showed that, when red blood cells were close to each other, the cells competed for gas, which resulted in a decreased rate of uptake for each cell. In another study (9), they found that, when cells were distorted into paraboloids, a condition that exists in the systemic microcirculation (6, 14), they had a lower diffusing capacity for oxygen compared with disk-shaped cells. Later, they showed that cellular diffusing capacity was reduced when red blood cells were clumped together compared with being evenly spaced (10). Their findings showed the importance of red cell distribution and shape. In our study, we focused our finite-element analysis on red blood cell orientation measured in the capillaries of rapidly frozen lungs. In our diffusion analysis, for purposes of simplification, we ignored the effect of hematocrit by modeling only a single red blood cell in a capillary.

When we applied the analysis to our data, we found that the edge-on orientation produced a 40% decrease in diffusing capacity. Because the diffusing distance for alveolar gas appears to be significantly greater when red blood cells are in the edge-on orientation toward alveolar gas, we think the actual decrease is likely to be significantly greater. Our underestimate of the effect of orientation may be due, in some measure, to three simplifying assumptions. First, we calculated the DL_CO, rather than the diffusing capacity for oxygen, which made the modeling simpler. In our calculations, we assumed, as did Hsia et al. (8), that the PCO everywhere within the red blood cell remained zero and that the diffusion path for carbon monoxide ended just inside the membrane. This assumption is justified by the high-binding affinity of hemoglobin for carbon monoxide and by the low PCO used in diffusing capacity measurements. Oxygen, however, would have to travel not only through the red blood cell membrane but also to all hemoglobin binding sites within the cell to saturate the hemoglobin. In the broad-side orientation, when gas enters the cell through the broad side, the middle of the cell is much closer to the alveolar membrane than when the gas enters the cell through the edge in the edge-on orientation. This would make the total diffusion path for gas even longer in the edge-on orientation than for the broad-side orientation, if we had applied the finite-element method to oxygen, instead of carbon monoxide diffusion. The analysis for oxygen, however, is a considerably more complicated problem (2), so we elected to focus on the simpler carbon monoxide model in this initial study. Second, the red blood cell orientation, and thus diffusion distance, would be more important in the case of oxygen than for carbon monoxide, because there is oxygen backpressure, but not carbon monoxide backpressure, within the red blood cell. Backpressure slows the net rate of diffusion and thus places greater importance on diffusion distance. Third, our finite-element analysis was based on a two-dimensional geometric model, which could not take into account the much larger surface area in close proximity to alveolar gas for the red blood cell in the broad-side orientation. For these reasons, we think the 40% decrease in diffusing capacity for the edge-on orientation is a lower bound estimate.

This is the first study to show by direct microvascular measurements in rapidly frozen lungs that the geometric orientation of red blood cells with respect to alveolar gas depends on alveolar size. If alveoli are highly expanded, red blood cells will transit the resulting oval alveolar capillaries with their broad sides facing alveolar gas. We assume that the same capillary shape and red blood cell orientation occur whether small-animal lungs are expanded by high-airway pressure, as in this study, or whether alveoli in the upper lung of larger animals are expanded by the weight of the lung below zone 2 (3). Figure 7 shows this red blood cell orientation by in vivo microscopy in the subpleural capillaries of alveoli in zone 2 of a dog. If alveoli are less expanded, either because of lower lung inflation or by being in zone 3 of a large animal (3), red blood cells may transit the more circular cross sections of alveolar capillaries with their edges facing alveolar gas. We have also observed this latter red blood cell orientation with in vivo microscopy in the subpleural capillaries of canine alveoli in zone 3, but we have not taken high-resolution photographs of these fast-moving red blood cells. Our finite-element modeling of these two extremes in red blood cell orientation showed that the edge-on orientation reduces the diffusive gas uptake by the red blood cells. This adds to the list of pulmonary intravascular factors shown by Hsia et al. (8-10) to reduce diffusive gas uptake.

During resting conditions or during moderate exercise, the orientation of red blood cells is of little consequence with regard to saturation with oxygen, because of the long capillary transit times (1, 12, 17). During heavy exercise, however, we think red blood cell orientation is likely to be of greater importance. One reason is the well-established observation that left atrial pressures rise to >30 mmHg during maximal exercise (13). That raises capillary transmural pressures, which will distend the capillaries into circular cross sections. Red blood cells are then free to traverse the capillaries in unfavorable orientations, including flowing in the middle of the stream surrounded by a relatively thick layer of plasma. This suggests that it may be a combination of too rapid a transit time and the unfavorable orientation of red blood cells that causes oxygen desaturation during maximal exercise and during some diseases in which capillary pressure is elevated. To our knowledge, this is a new and potentially important concept, i.e., that capillary cross-sectional shape determines the red blood cell orientation, which in turn can be a significant contributor to oxygen desaturation.