Reopening the injured lung with deep inflation (DI) and positive end-expiratory pressure (PEEP) likely depends on the duration and severity of acute lung injury (ALI), key features of which include increased alveolar permeability and fibrin accumulation. We hypothesized that the response to DI and PEEP would worsen as ALI evolves and that this would correspond with increasing accumulation of alveolar fibrin. C57BL/6 mice were anesthetized and aspirated 75 μl of HCl (pH 1.8) or buffered normal saline. Subgroups were reanesthetized 4, 14, 24, and 48 h later. Following DI, tissue damping (G) and elastance (H) were measured periodically at PEEP of 1, 3, and 6 cmH2O, and air within the lung (thoracic gas volume) was quantified by microcomputed tomography. Following DI, G and H increased with time during progressive lung derecruitment, the latter confirmed by microcomputed tomography. The rise in H was greater in acid-injured mice than in controls (P < 0.05) and also increased from 4 to 48 h after acid aspiration, reflecting progressively worsening injury. The rise in H was reduced by PEEP, but this effect was significantly blunted by 48 h (P < 0.05), also confirmed by thoracic gas volume. Lung permeability and alveolar fibrin also increased over the 48-h study period, accompanied by increasing levels and transcription of the fibrinolysis inhibitor plasminogen activator inhibitor-1. Lung injury worsens progressively in mice during the 48 h following acid aspiration. This injury manifests as progressively increasing alveolar instability, likely due to surfactant dysfunction caused by increasing levels of alveolar protein and fibrin.
- acute lung injury
- lung mechanics
- respiratory impedance
- plasminogen activator inhibitor-1
acute lung injury (ali) is characterized by hypoxemic respiratory failure secondary to noncardiogenic pulmonary edema. Although not a requisite component of the clinical definition (7), lung compliance is invariably reduced in ALI (45, 47). This reduction in compliance is thought to reflect both a reduction in resting lung volume from alveolar flooding or collapse, and an increase in alveolar surface forces in the remaining aerated lung (46, 47). In addition to the contribution from edema fluid, these increased surface forces are thought to result from inhibitory binding of surfactant by plasma proteins, particularly fibrin, a potent inhibitor of surfactant function in vivo (51, 52) and a chief component of the hyaline membranes that line injured alveoli (4).
The hypoxemia and low lung compliance in ALI can, in some patients, however, be reversed by a recruitment maneuver, delivered as either a brief or sustained deep inflation (DI) with the intention of reopening collapsed regions of the lung. Once open, the lung may be kept open by the application of sufficient positive end-expiratory pressure (PEEP) (23, 38). Regrettably, these maneuvers have met with variable success in clinical trials (10, 11), perhaps owing to a disparity in response to recruitment among patients with ALI of differing origin (17, 39, 44) or at varying stages of injury (21). Although not yet determined, this impaired response to recruitment may, in fact, be a marker of injury severity. We thus hypothesized that the lungs’ mechanical response to recruitment and PEEP would deteriorate as the severity of injury progresses over time. We tested this hypothesis by examining the response of the lung to DI and PEEP at various points during the pathogenesis of ALI in a clinically relevant mouse model of acid aspiration, while simultaneously measuring indexes of lung permeability, fibrin accumulation, and markers of depressed fibrinolysis (2).
Female C57BL/6 mice, 8–10 wk old (18.2 ± 1.0 g), were anesthetized with 400 mg/kg intraperitoneal (IP) avertin (tribromo-ethyl alcohol, Aldrich, Milwaukee, WI) and positioned vertically upright, their tongues were retracted, and their deep oropharynx was instilled with 75 μl of either hydrochloric acid (HCl) (pH 1.8) or sterile phosphate-buffered normal saline (PBS) (pH 7.4) using a syringe and bulb-tipped steel feeding catheter. The tongue was held retracted until the mice fully aspirated the fluid. Then the mice were briefly rotated to the left and right lateral decubitus position, allowed to recover from anesthesia, and monitored periodically thereafter. Mice predetermined to have had a poor aspiration at the time of injury were automatically excluded from the study. Lung mechanics were later measured among subgroups at 4, 14, 24, or 48 h following acid (n = 8 per group) or saline (n = 5 per group) aspiration. At this point, the mice were anesthetized with IP pentobarbital (90 mg/kg), then underwent surgical tracheal intubation with an 18-gauge metal cannula, and were ventilated in a quasi-sinusoidal fashion at 180 breaths/min on a flexiVent (SCIREQ, Montreal, Canada) small-animal ventilator. The cylinder piston displacement was set at 0.25 ml, which resulted in tidal volumes of 0.20 ml (∼10 ml/kg) when accounting for gas compression. PEEP was controlled by submerging the expiratory limb from the ventilator to the desired depth in a water trap. The mice were paralyzed with an IP injection of pancuronium bromide (0.5 ml/kg) and allowed 5 min to adjust to the ventilator at a PEEP of 3 cmH2O. To ensure adequate anesthesia, heart rate was monitored by continuous electrocardiogram (Silogic International), measured via transcutaneous needle electrodes. The entire protocol was reviewed and approved by the Institutional Animal Care and Use Committee at the University of Vermont.
Following a 5-min stabilization period, the level of PEEP was set at either 1 or 6 cmH2O, and two 1.0-ml DIs were sequentially delivered over 4 s (each lasting 2 s) at constant flow, with a pressure limit of 25 cmH2O. This pressure limit represents the typical pressure achieved during a stepwise pressure-volume (PV) curve and was chosen to limit volutrauma and risk of pneumothorax. The mice were then returned to quasi-sinusoidal ventilation at 180 breaths/min. Respiratory system input impedance (Zrs) was measured via a forced-oscillation technique (described later) immediately following the two DIs, then subsequently every 15 s for 5 min, and then every 30 s for an additional 2 min. Each post-DI measurement period was accompanied by PEEP, delivered in the order of 1, 3, and 6 cmH2O to one-half of the mice randomly in each group and in the reverse order in the other one-half. At the end of each post-DI measurement period, a quasi-static PV curve was obtained from functional residual capacity by dropping PEEP to 0 cmH2O and immediately delivering seven steps of inspiratory volume to a total volume of 0.8 ml, followed by seven equal expiratory steps, pausing at each step for 1 s. Plateau cylinder pressure was measured during each pause and plotted against piston displacement (corrected for gas compression and slow leak).
Mice that died prematurely on the ventilator before the end of the protocol were excluded from the analysis. This number came to five, four, and six mice at the 14-, 24-, and 48-h time points, respectively. No mice died prematurely at the 4-h time point, and no mice died prematurely before mechanical ventilation at any time point. At the end of the timed ventilator protocol, the abdomen was opened, and the left ventricle was visualized and punctured through the diaphragm to obtain blood, which was spun down, and the serum was stored at −80°C. Immediately following thoracotomy, bronchoalveolar lavage fluid (BALF) was obtained by instilling 1 ml of PBS into the lungs via the tracheal cannula and slowly suctioned back for a return of ∼0.8 ml. The left atrium was then cut, and 10 ml of PBS were slowly perfused through the right ventricular outflow tract to blanch the lungs of intravascular blood. Blanched lungs were then surgically removed. The left lung was tied off with suture, dissected away, flash frozen in liquid nitrogen, and stored at −80°C. The right lung was instilled with 10% buffered formalin to a pressure of 30 cmH2O, fixed with 70% ethanol, and later embedded in paraffin, cut, and mounted for staining.
Immediately following collection, BALF was centrifuged, and the supernatant was stored at −80°C. The cell pellet was resuspended, and the total cell count was determined by an Advia 120 hematology analyzer (Bayer, Tarrytown, NY). Cytospun slides were stained with hematoxylin and eosin for differential count determination. Protein content was calculated using a colorimetric assay (Bio-Rad Laboratories, Hercules, CA), standardized to graded concentrations of BSA. ELISA kits for D-dimer (Diagnostica Stago), murine total plasminogen activator inhibitor-1 (PAI-1), and active murine PAI-1 (Molecular Innovations, Southfield, MI) were used according to the manufacturer's protocol for BALF measurements.
Immediately before anesthesia, mice received a tail vein injection with a 2.5 mg/ml solution of FITC-labeled dextran (4,000 Da) (Sigma Chemical, St. Louis, MO) at a dose of 25 mg/kg, according to a protocol previously described (24). Following BALF and serum collection, 50 μl of BALF and diluted serum were loaded on a microwell plate, excited at 485 nm, and read at 528 nm, and the relative fluorescence of BALF to serum (expressed as a percentage) was used as an index of combined endothelial and epithelial permeability.
Tissue fibrin immunohistochemical staining.
Formalin-fixed, cut, and mounted slides were deparaffinized with xylene and graded ethanol series and rinsed with water. Antigen was unmasked with sodium citrate, and slides were treated with a M.O.M. immunodetection kit (Vector Laboratories, Burlingame, CA) to block unwanted background staining. Fibrin and fibrinogen were stained using a mouse anti-fibrin(ogen) β-chain antibody (Accurate Chemical & Scientific, Westbury, NY), followed by a biotinylated anti-mouse antibody, a Vectastain avidin-based alkaline phosphatase solution, and a vector red alkaline phosphatase substrate. Following fibrin(ogen) staining, hematoxylin was used for background staining.
Snap-frozen lungs were pulverized using liquid nitrogen-chilled mortars and pestles. RNA was extracted using TRIzol, then DNase was treated using RNeasy columns (Qiagen), and 1.0 μg of total RNA was used as a template to synthesize the first-strand cDNA using random primers and Superscript II reverse transcriptase mix, according to instructions by the manufacturer (GIBCO-BRL). Real-time semiquantitative RT-PCR was performed using the Taqman Universal PCR Master Mix and the ABI PRISM 7700 Sequence Detection System. The Assay-On-Demand primers and probes used were mouse hypoxanthine guanine phosphoribosyl transferase 1 (Hprt1) (assay Mm00446948_m1) and mouse serpine-1 (assay Mm00435860_m1) genes, both purchased from Applied Biosystems (Foster City, CA). cDNA levels were measured using the ΔΔthreshold cycle method and normalized to Hprt1, with the data presented as mean expression relative to the housekeeping gene, Hprt1, and then calculated as a quotient relative to a randomly chosen naive control.
Fibrin Western blot.
Frozen lung tissue specimens were individually crushed, weighed, and suspended in an extraction buffer composed of sodium phosphate (10 mM), EDTA (5 mM), ε-aminocaproic acid (100 mM), aprotinin (10 U/ml), heparin (10 U/ml), and PMSF (2 mM). Then each sample was homogenized at 2,600 rpm (POLYTRON, Kinematica) for 1 min and spun at 10,000 g for 10 min at 4°C, after which the supernatant was removed, and the pellet was resuspended in sodium phosphate (10 mM) and EDTA (5 mM). Samples were spun again at 10,000 g for 10 min, the supernatant was removed, the pellet was resuspended in urea and spun at 14,000 g, and the supernatant was again discarded. The pellet was then resuspended in an SDS buffer and incubated at 65°C for 1.5 h before Beta-mercaptoethanol (β-ME) was added and loaded into the gel. A fibrin standard was generated by mixing 5 mg of murine fibrinogen (Sigma Chemical, St. Louis, MO) with 5 units of bovine thrombin (Sigma Chemical), incubated at 37°C for 10 min, then solubilized in SDS buffer and β-ME, and diluted to a concentration such that the greatest total amount of fibrin loaded into the well was 2,000 ng. Following gel electrophoresis, each gel was first blocked with 1% BSA on Tris-buffered saline/Tween (TBST) and incubated overnight at 4°C with a 1:500 dilution of monoclonal mouse anti-β-chain fibrin antibody (MAb350, American Diagnostica, Stamford, CT), previously conjugated to biotin, according to manufacturer protocol (EZ-Link Sulfo-NHS-LC-Biotin, Pierce, Rockford, IL). The MAb350 antibody specifically binds to a heptapeptide sequence on the β-chain of fibrin after cleavage of fibrinopeptide B by thrombin, thus binding specifically to previously polymerized fibrin monomers, and not to fibrinogen (54). The gel was then washed in TBST and incubated for 30 min with a 1:1,000 dilution of avidin-horseradish peroxidase (R&D), washed in TBST, and exposed with Amersham ECL reagent (GE, Healthcare) for 10 min before chemiluminescence measurements.
Impedance data analysis.
Zrs was determined by measuring piston volume displacement and cylinder pressure while delivering 2-s oscillatory volume perturbations to the airway opening in a manner described previously (1). These perturbations were composed of 13 superimposed sine waves with frequencies ranging from 1.0 to 20.5 Hz, all mutually prime to reduce harmonic distortion that can occur in nonlinear systems (25). Initial dynamic calibration signals were obtained to correct for the physical characteristics of the ventilator and tubing in subsequent measurements of Zrs (26, 50). Zrs itself was determined via Fourier transform from the signals of ventilator piston volume and cylinder pressure, as described previously (20, 26). Zrs was interpreted by being fit with the model (1) where (2) and i represents the square root of −1 and f represents frequency. The parameters RN and Iaw largely characterize the resistive and inertive properties, respectively, of the airways, while G and H characterize the dissipative and elastic properties of the lung tissues, respectively (25). In particular, the parameter H is equal to respiratory elastance at an oscillation frequency of 1/2π Hz. Hysteresivity (η) is the quotient G/H. Increases in η are believed to reflect changes in intrinsic tissue properties and/or increased regional heterogeneity in lung function (34, 40). We invoked the normalization scheme of Ito et al. (32) to express G and H in the same units as RN (cmH2O·s·ml−1) without changing their numerical values.
Computed tomography imaging and lung volume.
Using the identical protocol described previously, a separate group of mice (n = 5 per time point) were exposed to 75 μl of HCl (pH 1.8) and placed on mechanical ventilation at either 4 or 48 h after injury. The mice were ventilated with a PEEP of 1 cmH2O for 7 min and then given a lethal dose of pentobarbital before being switched over to 100% inhaled nitrogen (N2) gas for 3 min (to prevent further resorption atelectasis). Once cardiac death was confirmed by electrocardiogram, the trachea was tied off at end-exhalation (PEEP 1 cmH2O). High-resolution computed tomography (CT) images (47 μm/voxel edge, 80 kVp, 450 mA, over 80 min) were obtained using a GE Medical Systems eXplore Locus laboratory volumetric cone-beam micro-CT scanner. Lung volumes [thoracic gas volume (VTG)] were calculated using Microview visualization software, version 2.0.29 (GE HealthCare, London, ON, Canada). Two-dimensional (2D) regions of interest (ROIs) were created on ∼10 cross-sectional images selected from a range of slices between the proximal trachea and lung bases. The 2D ROIs were defined by contours drawn freehand closely around the lungs and trachea to exclude all extrathoracic gas (bowel gas and air outside the body was not included). 2D ROIs were then automatically created for all cross sections by linear interpolation, and three-dimensional (3D) ROIs were then generated by compiling all 2D regions. Frequency histograms of Hounsfield units (HU) were calculated for the voxels contained within each 3D ROI. The frequencies of the HUs between −1,000 and 0 (corresponding to N2 and water, respectively) were then converted to fractions of gas by multiplying each HU value by its corresponding number of voxels and then dividing by −1,000. These fractions were then summed and multiplied by the voxel volume of 1.038 × 10−7 ml to yield an estimate for VTG. Isosurface renderings were created using an algorithm in the Microview software package, wherein a 3D surface is drawn over all contiguous voxels (within the previously defined thoracic ROI), having a gray-scale value at or above a given threshold. Image data were inverted so that the high end of gray-scale values corresponded to voxels representing regions of low X-ray attenuation (low density). A threshold value corresponding to −500 HUs in noninverted data was selected, so that the spaces enclosed by the surface were considered to be occupied by at least 50% air. The resulting isosurface rendering constitutes a “virtual casting” of the airways and parenchyma.
All graphing and statistical analyses were performed using Origin software (version 7.5, Northampton, MA). ANOVA was used to compare values among all groups, followed by post hoc Bonferroni tests for means comparison between groups. A graphic representation of the protocol design is outlined in Fig. 1.
Lung function and derecruitment.
Mean values for G and H rose immediately following DI in all acid-injured mice, but G and H rose more by 24 and 48 h from injury compared with 4 and 14 h (Fig. 2). Immediate post-DI values for RN (data not shown) were never significantly elevated compared with saline controls at any level of PEEP at any time point of injury, but did rise over time significantly following DI at all time points during PEEP of 1 cmH2O (P < 0.05). Immediate post-DI measures of H were significantly greater than those of saline controls at PEEP of 1 cmH2O for all time points, and at PEEP of 3 and 6 cmH2O by 14, 24, and 48 h (P < 0.05). Immediate post-DI measures of G were not significantly elevated by 4 h, but were significantly elevated at PEEP of 1 cmH2O by 14 h, and at PEEP 1, 3, and 6 cmH2O by 24 and 48 h (P < 0.05). Of particular interest to our hypothesis, the total rise in H following DI (Fig. 2A) was greater than that of saline controls in all groups except at PEEP of 6 cmH2O, 4 h from injury (P = 0.08). The total rise in H following DI tended to increase as the time following acid instillation progressed, while the response of H to added PEEP decreased as time from injury progressed (Fig. 2A). The total rise in H at PEEP 3 cmH2O was significantly greater at 48 h (32.2 ± 4.5 cmH2O·s·ml−1), compared with that at 4 h (15.4 ± 3.4 cmH2O·s·ml−1). The total rise in H at PEEP 6 cmH2O became significantly greater at 48 h (15.7 ± 3.3 cmH2O·s·ml−1), compared with both 4 h (5.9 ± 1.0 cmH2O·s·ml−1) and 14 h (6.4 ± 0.8 cmH2O·s·ml−1). A similar pattern in the response to PEEP by 48 h was observed for the total rise in G following DI (Fig. 2B). By 4 and 14 h, the total rise in G following DI was only significantly higher than that of saline controls during PEEP of 1 cmH2O. By 24 and 48 h, the total rise in G following DI became significantly higher than control during PEEP of 1, 3, and 6 cmH2O. At 48 h, the total rise in G during PEEP 3 cmH2O (1.77 ± 0.34 cmH2O·s·ml−1) was greater than at 4 and 14 h (0.71 ± 0.30 and 0.66 ± 0.12 cmH2O·s·ml−1, respectively, ANOVA, P < 0.05).
Increasing the level of PEEP always had a significant effect on the post-DI rises in G and H at every time point studied (ANOVA, P < 0.0001). However, the magnitude of difference in the mean rise in H between increasing levels of PEEP was less at 48 h than at all other time points, with the exception of the difference between PEEP of 3 and 6 cmH2O at 4 h. This latter exception was likely due to the greater effect of PEEP 3 cmH2O at 4 h (see Fig. 2).
The PV curves demonstrated a significant decline in hysteresis following ventilation at a PEEP of 6 cmH2O compared with either 1 or 3 cmH2O at all time points (Fig. 3, A and B). However, PEEP was less effective at reducing hysteresis at 24 and 48 h compared with the earlier time points (Fig. 3, C and D). The mean area enclosed by the PV loops obtained following ventilation at PEEP 3 cmH2O was significantly different from that following ventilation at 1 cmH2O at 4 h, but this comparison was not significantly different at the later time points.
Mean values for BALF total protein concentrations increased progressively as time advanced from injury (Fig. 4, left axis) and were significantly elevated by 24 and 48 h compared with values from 4 h (ANOVA, P < 0.05). Relative ratios of BALF to serum FITC-dextran concentrations were significantly elevated above saline controls at every time point, and there was a statistical trend for these values to increase over time (ANOVA, P = 0.092) (Fig. 4, right axis). BALF demonstrated a significant rise in total cells by 4 h and a significant rise in neutrophils by 24 h, relative to saline controls (P < 0.05) (Fig. 5), and both remained elevated through to 48 h.
The clearance of fibrin is chiefly governed by the relative quantity and activity of fibrinolysis promoters and fibrinolysis inhibitors; the latter of such is represented by PAI-1 (28). As an indirect measure of fibrinolysis inhibition, total PAI-1 antigen and active PAI-1 levels in the BALF were measured and were noted to be significantly elevated by 48 h (Fig. 6) compared with saline controls and all other time points following acid injury (ANOVA, P < 0.05). Mean relative lung tissue mRNA levels for the PAI-1 gene, serpine-1 (Fig. 6), demonstrated a significant increase in transcription of mRNA for PAI-1 by 24 h (relative to saline controls) and particularly by 48 h (relative to all other groups) (ANOVA, P < 0.05). Immunohistochemical staining signified greater levels of fibrinogen and fibrin at 24 and 48 h compared with 4 and 14 h (representative specimens from 4 and 48 h shown in Fig. 7). BALF levels of the fibrin split product, D-dimer, were significantly elevated at 14, 24, and 48 h (Fig. 8A) compared with both saline controls and specimens from the 4-h time point (ANOVA, P < 0.05). Total lung fibrin levels also increased progressively over time and were significantly elevated at 48 h compared with 4 h (Fig. 8B). A representative image from the immunoblot gels is provided in Fig. 8C.
CT lung volumes.
Mean VTG from nitrogen-fixed lungs (Fig. 9A) following 7-min ventilation at a PEEP of 1 cmH2O was significantly lower at 48 h following acid aspiration (0.085 ± 0.010 ml) compared with 4 h following aspiration (0.206 ± 0.043 ml, P = 0.017). An inverse linear correlation was found between the post-DI rise in H (over 7-min ventilation at PEEP 1 cmH2O) and the corresponding postventilation estimates of VTG for the separate groups of mice imaged at 4 and 48 (Fig. 9B, R = 0.90, P < 0.001). Representative 3D isosurface renderings at 4 and 48 h are shown in Fig. 9C and illustrate the reduction in the amount of open lung at 48 h compared with 4 h following aspiration.
The results of this study demonstrate that acid aspiration in mice, as expected, leads to a substantial deterioration in lung function. However, contrary to previous studies in acid-injured rats demonstrating that indexes of lung permeability recover by 14 h (35), we found that injury worsened progressively over the 48 h following aspiration in our model. The primary mechanical derangement we observed was an increased tendency for derecruitment of the lung to occur during mechanical ventilation. This was evidenced by progressive and proportionate increases in the tissue parameters G and H following a DI (Fig. 2) and visualized loss of aerated lung units on micro-CT images (Fig. 9). Furthermore, the degree of post-DI lung derecruitment increased progressively with time after aspiration, while PEEP became less effective at both ameliorating this derecruitment (Fig. 2) and reducing PV hysteresis (Fig. 3), particularly by 48 h. Together, these findings demonstrate that acid aspiration in mice causes an alveolar instability, and that this instability continues to progress for at least 48 h. This suggests that alveolar surfactant function, which is crucial for alveolar stability, was being progressively disrupted in the mice. Plasma proteins, and fibrin in particular, are potent inhibitors of surfactant function (51). Hence it is noteworthy that we established that both permeability and alveolar fibrin increased in concert with the observed mechanical derangements over the 48 h following acid aspiration. While no single parameter of injury increased significantly at each successive time point, taken together the parameters demonstrate a consistent trend of injury progression to account for the worsening response to recruitment and PEEP over time.
Increased permeability from both epithelial and endothelial injury is a well-recognized feature of ALI (60), and one would expect this to lead to elevated levels of plasma proteins, including fibrin, in the alveolar space. However, alveolar fibrin may have been further elevated in our mice through the inhibition of fibrinolysis, as evidenced by increased lung transcription and alveolar elaboration of PAI-1. The increased levels of the fibrin split product D-dimer in BALF could be due to either increased levels of total fibrin or an increased breakdown of fibrin, and thus the lack of difference between 14, 24, and 48 h (Fig. 8A) does not lend as much insight as the levels of PAI-1 and total fibrin. Both active and total PAI-1 levels in BALF, and mRNA levels from tissue, became increasingly elevated with time, with the highest levels occurring at 48 h (Fig. 6), when total fibrin levels were also greatest (Fig. 8). These results are consistent with the dysregulated alveolar fibrin accumulation that has long been recognized to occur in ALI (4), manifesting as impaired fibrinolysis and increased procoagulant activity within the BALF of patients with ALI (8, 15, 22, 29). Fibrin formation and clearance in the lung is in part governed by the relative quantities and activities of fibrinolysis inhibitors, such as the serine protease PAI-1 (28), which can prevent fibrinolysis via direct binding and inhibition of tissue and urokinase plasminogen activator (58). Inhibition of urokinase plasminogen activator activity by PAI-1 in the BALF of ALI patients was first noted by Bertozzi and colleagues in 1990 (8), and it has subsequently been shown that mice genetically deficient in PAI-1 fail to accumulate alveolar fibrin and are more resistant to hyperoxia-induced ALI (5). Elevated plasma and edema fluid levels of PAI-1 in human subjects with ALI have also recently been linked to higher mortality (48, 59), imparting value to PAI-1 as a prognostic marker in ALI.
Our findings of a temporal variation in the response of the lung to DI and PEEP are in agreement with clinical findings that lasting responses to recruitment maneuvers in ALI patients are typically confined to the early course of their disease (9, 21). The lung mechanical derangement of ALI manifests mainly as an increase in pulmonary elastance, particularly in the more direct injury forms of ALI such as pneumonia (17). At the bedside, this increase in elastance is typically observed as an increase in peak and plateau airway pressures, but it can also be observed as an expansion in the hysteresis of PV curves obtained during graded inflation of the lung (47). Increased elastance is thought to reflect both an increase in surface tension within the alveolar lining fluid and a reduction in lung volume from alveolar flooding and collapse, the latter process being colloquially termed “baby lung” (16). Affected regions of the lungs may often be so damaged that they remain fluid-filled or collapsed throughout the entire course of inflation (27), and CT imaging has demonstrated elastance to correspond to the amount of remaining aerated tissue in the lung (18). This has lead to the notion that the increased lung elastance seen in ALI is more a reflection of the degree of lung derecruitment than of changes in intrinsic tissue stiffness (16). We have made the same assumption in our laboratory's previous work (3) and continue to do so in the present study by using G and H as surrogate markers of open lung. This assumption is supported by previous findings that derecruitment causes G and H to increase in the same proportion (3), such that their ratio (G/H), also termed “hysteresivity”(14), remains relatively unchanged over time. However, it should also be noted that such rises in H, and particularly G, could also represent an element of increasing heterogeneity within the lung, which has been shown to occur in models of ALI (33) and may, in turn, contribute to variations in mechanical stress and bioinjury (53). In turn, if the process of derecruitment occurs during the measurement of Zrs, we have shown that hysteresivity can potentially even decrease (6). Nevertheless, the strong correlation between the rise in H and CT estimates of VTG shown in Fig. 9B suggest that the majority of these changes represents derecruitment of lung volume. This can be further appreciated by both CT-based calculations of VTG (Fig. 9A) and the air surface reconstructions from acid-injured lungs (Fig. 9C), which demonstrate that the effects of DI and PEEP become less effective at recruiting lung volume between 4 and 48 h following acid aspiration.
Our mouse model of acid aspiration was designed to simulate a widely recognized and common cause of ALI in humans, namely aspiration of acidic gastric contents (60). Acid aspiration involves a direct injury to the airways and parenchyma, and thus represents a subset of direct or “pulmonary” ALI in which recruitment has been historically shown to be less effective (44), particularly in its later phases (21). One potential explanation for the differential effect of recruitment on early and late injury in our model is a divergence in the character of alveolar injury over time. For instance, the early phase of acid injury in our model is characterized by an immediate disruption of epithelial barrier function, with an early increase in permeability and alveolar protein, but negligible accumulation of neutrophil exudates or fibrin. This likely leads to an injury pattern dominated by edema and atelectasis, a type of injury demonstrated in previous studies to be more responsive to recruitment and PEEP (57). Conversely, the later phase of injury in our model was characterized by an accumulation of fibrin-laden exudates (Fig. 7). This may be a histological pattern more in keeping with the late phase of ALI (21) or direct forms of ALI (44) that are typically less responsive to recruitment and PEEP.
We assert that the progressively increased transcription and elaboration of PAI-1 and increasing alveolar fibrin accumulation in our mice further support their validity as a clinically relevant animal model of ALI. Indeed, this model may have a number of advantages over the commonly used approach of delivering endotoxin to the air spaces. Although endotoxin reliably generates a robust accumulation of alveolar neutrophils (1, 55), the injury it causes often subsides within 24–48 h (42, 43). Furthermore, repeated endotoxin exposures are often required to generate significant derangement in lung mechanical function (31) and sometimes fail to generate any significant derangement at all (1). Even intravenous endotoxin, which generates a profound inflammatory response with increased permeability and derangement in lung function (12), subsides within 48 h (49). In contrast, acid aspiration appears to instigate a crescendo of injury in the form of increasing air space protein, permeability, and fibrin with a corresponding derangement in lung function that continues to evolve through to 48 h. Interestingly, the number of air space neutrophils remains stable between 14 and 48 h, possibly due to an attenuation of neutrophil apoptosis and clearance (41).
Nevertheless, there are aspects of our mouse model of ALI that were unexpected. In particular, rat models of acid-induced lung injury (35) have been shown to recover barrier function by 14 h and to exhibit a peak in neutrophil accumulation by 4 h. Both features of our mouse model extended much further out from the time of aspiration. The reasons for these differences are unclear. One possibility is that a delayed neutrophil response in our mice may have helped drive a correspondingly delayed derangement in epithelial/endothelial barrier function (13, 19, 37). Some studies have demonstrated a negligible contribution of neutrophil elastases to the disruption of barrier function during the earlier phase of acid-induced lung injury (56), suggesting that early injury is due to a direct caustic effect of acid on the epithelium. Other studies examining acid injury at a later time point (24 h) have demonstrated a more important role for neutrophils by demonstrating an attenuation in injury through neutrophil inhibition and depletion (13). The concept that neutrophils might drive a second phase of lung injury is also supported by the finding that neutrophil-elastase inhibitors only protect the lung in regions remote from the original site of aspiration injury (19). It is possible, therefore, that neutrophil dynamics following acid aspiration differ between mice and rats. However, little is known about this. Although mouse models of acid aspiration are becoming more widely utilized (30, 61), few investigators (36) have followed this model out as far as in the present study. Furthermore, none to our knowledge has investigated the effect of this injury on lung mechanical function, an important clinical marker of injury.
In conclusion, we have developed an acid-aspiration mouse model of ALI that exhibits many of the key physiological and biological features of direct ALI in patients. The injury in this model continues to worsen out to 48 h after aspiration, as reflected in progressive derangements in lung mechanical function with commensurate reductions in the effectiveness of DI recruitment maneuvers and PEEP. Furthermore, the mice exhibited increasing indexes of permeability and accumulation of alveolar fibrin, the latter known to be pathognomonic for human ALI and likely involved in the inhibition of surfactant function (51). Increasing levels of pulmonary PAI-1 and PAI-1 mRNA indicate that alveolar fibrin accumulation may have been further augmented through inhibition of fibrinolysis. Our study thus helps to elucidate the complex interplay between host defense mechanisms in the lung and their possible pathological consequences following a sufficiently egregious insult, and in particular supports the notion that alveolar protein and fibrin accumulation chiefly determine the mechanical response of the lung to recruitment maneuvers and PEEP.
This study was supported by National Institutes of Health Grants K08HL074107, R01 HL75593, and Centers of Biomedical Research Excellence (COBRE) P20RR15557.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society