INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CLINICIANS HAVE LONG SUSPECTED that high airway pressures used to maintain adequate oxygenation of patients with adult respiratory distress syndrome (ARDS) may actually contribute to the lung pathophysiology, and subsequent experimental studies have confirmed this suspicion (21). Webb and Tierney (28) first demonstrated pulmonary edema formation in rats ventilated with high peak inflation pressures (PIPs), and both Parker et al. (23) and Dreyfuss et al. (3) demonstrated that an increased microvascular permeability contributed significantly to the edema formation in lungs ventilated at high PIP. Later studies indicated that overdistension with high tidal volumes was more injurious to the lung than was high airway pressure per se (4, 12), and subsequent clinical application of low-tidal-volume ventilation in ARDS patients has resulted in reduced mortality and morbidity in some studies (13).

Increases in vascular permeability induced by the mechanical stress of high vascular and airway pressures have generally been attributed to a passive structural failure of the vascular wall. Shirley et al. (25) first proposed that the increased transcapillary fluid and protein leak from peripheral capillaries subjected to high venous pressures was produced by a "stretched-pore phenomenon" caused by forcing apart of the endothelial intracellular junctions. The elegant electron-microscopic studies of West and associates (7, 17, 27, 31) later revealed the presence of epithelial and endothelial "breaks" or openings in lungs subjected to high venous pressures. These openings penetrated the cell bodies rather than through open intercellular junctions, and the ruptures often extended through the basement membranes. These breaks generally occurred parallel to the intercellular junctions between pulmonary microvascular endothelial cells, and their numbers were proportional to the capillary transmural pressure at both high airway and vascular pressures (27). West et al. (31) proposed that "stress failure" of the alveolar capillary barrier was determined primarily by the tensile strength of the basement membrane.

In contrast, recent evidence from our laboratory suggests that mechanical stress-induced permeability increases may have a significant active cellular component (20, 22). Infusion of isoproterenol significantly attenuated the capillary filtration coefficient (K_fc) increases in isolated rat lungs induced by high venous pressures, possibly by decreasing endothelial cytoskeletal tone (22). In addition, inhibition of phosphotyrosine phosphatase to increase intracellular tyrosine phosphorylation of focal adhesion proteins significantly increased the susceptibility of isolated rat lungs to high PIP-induced injury (20). In receptor-induced permeability increases, an increase in intracellular calcium initiates an active increase in tension of actin-myosin fibrils and causes retraction of endothelial cell margins (2, 11). Mechanical stretch also initiates calcium entry into endothelial monolayers through stretch-activated, nonspecific cation channels (18), and this increase could induce a similar cell retraction. Therefore, prevention of calcium entry might block the signal-transduction process and attenuate the increased permeability secondary to mechanical injury (9).

The purpose of the present study was to determine whether gadolinium, a trivalent lanthanide element that effectively blocks nonselective, stretch-activated cation channels (9), would attenuate the increase in K_fc induced by high PIPs in the isolated perfused rat lung preparation. Gadolinium infusion prevented any increase in microvascular permeability due to alveolar overdistension at the airway pressures investigated in this preparation.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Isolated Rat Lung Preparation

K_fc and Total Vascular Resistance (RT)

(1)

(2)

1

1

(3)

Experimental Protocols

High-PIP control group (control group; n = 5). The rat lungs were isolated and prepared as described in Isolated Rat Lung Preparation. The general time course of airway and venous pressure increases are shown in Fig. 1. After a baseline period of 30 min, a baseline K_fc was performed with a nominal PIP of ~7 cmH₂O, and then the lungs were ventilated for 30-min periods each at 20 cmH₂O PIP with 3 cmH₂O PEEP, 30 cmH₂O PIP with 3 cmH₂O PEEP, and 35 cmH₂O PIP with 8 cmH₂O PEEP. The tidal volumes used to achieve these peak pressures were ~5.1 ml (14 ml/kg), 5.8 ml (16 ml/kg), and 5.1 ml (14 ml/kg), respectively. Each ventilation period was followed by a K_fc measurement for 20 min. This combination of airway pressures and ventilation times produced a consistent three- to fourfold increase in K_fc in this preparation. After perfusion was stopped, the lungs were weighed, sectioned, and frozen.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Diagram of time course of peak inflation and venous pressures as function of time used for both high peak inflation pressure control and gadolinium groups.

High-PIP gadolinium group (gadolinium group; n = 5). The rat lungs were isolated and prepared as described in Isolated Rat Lung Preparation). After a baseline period of 30 min, a baseline K_fc was measured at PIP of ~7 cmH₂O. Gadolinium chloride (20 µM) was added to the venous reservoir and the same protocol shown in Fig. 1 and described in High-PIP control group (control group; n = 5) was performed. After perfusion was stopped, the lungs were weighed, sectioned, and frozen.

Histology

Statistics

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Hemodynamics

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Total pulmonary vascular resistance in high peak inflation pressure control and gadolinium groups as function of peak inflation pressure. Values are means ± SE. * P < 0.05 vs. baseline in the same group.

K_fc Values

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Capillary filtration coefficients in control and gadolinium groups as a function of peak inflation pressure. Values are means ± SE. * P < 0.05 vs. baseline in the same group. # P < 0.05 vs. all previous states in the same group. § P < 0.05 vs. gadolinium group at the same peak inflation pressure.

Histology

View larger version (145K): [in this window] [in a new window]   View larger version (150K): [in this window] [in a new window]   Fig. 4.   Histology sections of rat lungs stained with hematoxylin and eosin. A: section of high-peak inflation pressure control lung showing a bronchus and adjacent artery. Large perivascular edema cuff with large lucent areas surrounds the artery. Original magnification ×80. B: section of high-peak inflation pressure gadolinium lung at same magnification showing a perivascular edema cuff around a comparable-sized vessel. Edema cuffs were generally smaller with more stroma present. Original magnification ×80.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The major new findings of the present study were the threshold for high PIP increases in K_fc in isolated rat lungs and the total block of high PIP increases in K_fc by infusion of gadolinium chloride. The susceptibility of the untreated rat lungs to high PIP injury appears to be comparable to that of rabbits and significantly less than that of dogs, because K_fc increased by 77% after 30 min of 30 cmH₂O PIP ventilation in the present study. In two previous studies, K_fc in the lungs of young rabbits (1 kg) ventilated in situ with closed chests for 1 h at a PIP of 30 cmH₂O increased significantly to 160 (12) and 177% of baseline (1). In contrast, Parker et al. (23) reported an increase in K_fc in isolated dog lungs only with PIP exceeding 42 cmH₂O. A species difference was also observed by Mathieu-Costello et al. (17), who reported an increased number of epithelial breaks in rabbit lungs compared with dog lungs subjected to comparable increases in microvascular transmural pressures.

The dramatic effect of 20 µM gadolinium chloride in preventing any increase in K_fc in response to a PIP which consistently increased K_fc from baseline by three- to fourfold, supports the hypothesis that the increased microvascular permeability associated with mechanical injury is an active rather than a passive process (22). Neither the stretched-pore separation of intercellular junctions (25) nor stress failure of basement membranes by mechanical stress (31) can account for the effect of infusion of gadolinium in preventing even the slightest increase in K_fc after ventilation with 30 and 35 cmH₂O PIP for 30 min. Gadolinium is the most commonly used blocker of mechanogated cation channels and produces an effective blockade of mechanogated cation channels in atrial and ventricular myocytes, renal and pituitary secreting cells, nerve cells, and plant cells at concentrations of 1-100 µM (9). Complete block of mechanogated cation channels in most mammalian cells occurs at doses of 10-20 µM. Gadolinium is a member of the lanthanide series of elements with an atomic number of 64 and an ionic radius of 0.938 Å (9). All lanthanides are trivalent cations and have ionic radii similar to that of calcium (0.99 Å) as well as similar bonding and donor atom preferences (6). Despite the similarity of size and charge between gadolinium and calcium, the action of gadolinium in blocking mechanogated cation channels is not a simple physical blockade because it apparently acts on multiple ion-channel sites (9). Although most studies demonstrate selective blockade of mechanogated nonselective cation channels, other studies show blockade of L- and T-type calcium channels at micromolar concentrations and blockade of mechanogated potassium channels at much higher doses (9).

The essential role of increased endothelial intracellular calcium in active, receptor-mediated increases in paracellular endothelial permeability has been well established. Curry (2) has demonstrated the necessary coupling of increased intracellular calcium to increased permeability in single frog capillaries, where an influx of extracellular calcium is required for an increase in hydraulic conductivity (11). Regional endothelial calcium spikes were always associated with macromolecular leak sites in their single capillary preparation (19), but endothelial calcium transients could be uncoupled from the hydraulic conductivity increase by increasing intracellular adenosine 3',5'-cyclic monophosphate, indicating an active actin-myosin interaction in the permeability increase (10). In the case of mechanical distension of capillaries, the most likely nonreceptor sources of increased endothelial intracellular calcium are the stretch-activated cation channels. Naruse and Sokabe (18) stretched endothelial monolayers on a silicone membrane and measured intracellular calcium after fura-2 loading. A calcium peak occurred which was essentially abolished either by removal of calcium from the medium or by adding 10 µM gadolinium to block nonselective stretch-activated cation channels. A small, persistent component of the calcium transient was blocked by ryanodine, indicating mobilization of calcium from cellular stores by the calcium influx. Manganese also entered during endothelial stretch to quench the fura 2 fluorescence, indicating a nonselective cation channel, but lithium had no effect, suggesting that the sodium-calcium exchanger was not involved. The recent observation of Ying and Bhattacharya (32) of an increase in intracellular calcium in capillary venular endothelium in isolated blood perfused rat lungs during increased venous pressure may have resulted from calcium entry through such stretch- activated cation channels. Because the K_fc increases were completely prevented by gadolinium infusion even at the highest PIP of 35 cmH₂O in the present study, an active endothelial response to high PIP mediated by calcium entry through stretch-activated cation channels is strongly implied.

Although tensile strength of the basement membrane and its stress failure at high vascular transmural pressures undoubtedly determine the amount of hemorrhage at very high pulmonary vascular pressures in racehorses and elite athletes (29, 30), the endothelial cellular layer is generally considered to be the major barrier to microvascular fluid filtration (26). The implication of the present study is that there is a considerable active contribution by endothelium to the increase in microvascular permeability induced by the mechanical stress of alveolar distension. Two previous studies from our laboratory also indicate an active participation of endothelium (20, 22). In isolated rat lungs, K_fc increased 6.2 times baseline at a sustained venous pressure of 31 cmH₂O (22). Infusion of isoproterenol attenuated this increase by 64%. This response suggests an active cytoskeletal retractile response to mechanical injury because the increased intracellular adenosine 3',5'-cyclic monophosphate induced by isoproterenol may inhibit myosin light chain kinase activity and reduce intracellular calcium to reduce increases in microvascular permeability (14-16). With use of a high-PIP protocol similar to that in the present study, K_fc was significant increased by protein tyrosine phosphatase inhibition and was significantly attenuated by tyrosine kinase inhibition compared with untreated lungs (20). Modulation of tyrosine phosphorylation may have altered focal adhesion integrity or other activities mediated by tyrosine kinases to induce these changes.

An active endothelial component of injury is also suggested by the reversibility of lung injury induced by high vascular and airway transmural pressures. In morphological studies of rabbit lungs, Elliott et al. (5) observed a rapid reversal in the number of transepithelial and transendothelial breaks when high vascular transmural pressures were decreased to baseline pressures. Also, Rippe et al. (24) noted that high vascular pressure-induced increases in K_fc returned toward control values over a 20-min period when Ppv was decreased in isolated dog lungs. Similarly, Parker et al. (23) observed time-dependent decreases in K_fc after high PIP injury in isolated dog lungs. Taken together these studies suggest an active "stretch-recoil" effect of endothelium in microvessels subjected to mechanical stretch. Signal transduction through stretch-activated cation channels and active actin-myosin tension development may then increase transcapillary fluid flux. Such a mechanism might serve to release transient increases in pressure to preserve the basement membrane and collagen scaffolding supporting cellular layers so that the barrier function of capillaries could be rapidly reestablished. A true stress-failure injury of the collagen and matrix support structures would presumably require much longer to repair.