Linking Ventilator Injury-Induced Leak across the Blood-Gas Barrier to Derangements in Murine Lung Function

Abstract

Mechanical ventilation is vital to the management of acute respiratory distress syndrome, but it frequently leads to ventilator-induced lung injury (VILI). Understanding the pathophysiological processes involved in the development of VILI is an essential prerequisite for improving lung-protective ventilation strategies. The goal of this study was to relate the amount and nature of material accumulated in the airspaces to biomarkers of injury and the derecruitment behavior of the lung in VILI. Forty-nine BALB/c mice were mechanically ventilated with combinations of tidal volume and end-expiratory pressures to produce varying degrees of overdistension and atelectasis while lung function was periodically assessed. Total protein, serum protein, and E-Cadherin levels were measured in bronchoalveolar lavage fluid (BALF). Tissue injury was assessed by histological scoring. We found that both high tidal volume and zero positive end-expiratory pressure were necessary to produce significant VILI. Increased BALF protein content was correlated with increased lung derecruitability, elevated peak pressures, and histological evidence of tissue injury. Blood derived molecules were present in the BALF in proportion to histological injury scores and epithelial injury, reflected by E-Cadherin levels in BALF. We conclude that repetitive recruitment is an important factor in the pathogenesis of VILI that exacerbates injury associated with tidal overdistension. Furthermore, the dynamic mechanical behavior of the injured lung provides a means to assess both the degree of tissue injury and the nature and amount of blood-derived fluid and proteins that accumulate in the airspaces.

Keywords: ARDS; alveolar leak; lung injury; mechanical ventilation; surfactant dysfunction.

Figure 1

Dynamic pressure-volume loops showing the location of the upper corner point determined by fitting to the Venegas equation (filled circles) and the overdistension volume before (solid lines) and after (dotted lines) 100 min of ventilation for representative mice in the Low-Vt/PEEP0 (A), Low-Vt/PEEP3 (B), Mid-Vt/PEEP0 (C), High-Vt/PEEP0 (C), and Mid-Vt/PEEP3 (D) groups.

Figure 2

(A) Change in dynamic pressure-volume loop peak pressure (ΔP_Max) during ventilation. ^*Indicates that the rate of increase of ΔP_Max (P_S) was significantly greater than all other groups; significant increases from Mid-Vt/PEEP3 (#), Low-Vt/PEEP3 (~), and Low-Vt/PEEP0 (%). Error bars show standard deviation. (B) Changes in derecruitment dynamics at PEEP = 0, 3, and 6 cm H₂O caused by mechanical ventilation. The statistical significance of intra-group elastance changes over the course of ventilation are shown in Table 2, intra-group differences in elastance at different derecruitment test PEEPs are shown in Table 3, and post-ventilation inter-group elastance changes are shown in Supplementary Tables 2–4. (C) P_S increases as a function of BALF total protein [y = 0.18(1 – e^{−0.89(x-0.10)}), R² = 0.81, p < 0.027, black line]. (D) Increase in initial elastance during the PEEP = 0 derecruitability test ($H_1^0$) at the conclusion of the experiment is correlated with increased BALF total protein (y = 18.0 + 14.1x, R² = 0.78, p < 10⁻¹⁸). Data points for $H_1^0$ and P_S show the mean with error bars showing standard deviations; data points for BALF protein show the median with error bars showing the interquartile range.

Figure 3

(A) Ratio of bronchoalveolar lavage fluid (BALF):Serum fluorescense demonstrates significantly increased leak (with respect to the control group) of retro-orbitally injected 4 kDa dextran molecules into airspace. The fluorescence ratio was linearly related to BALF protein (y = 0.112 + 0.269x, R² = 0.92, p < 0.005). Panels (B,C) show representitave Coomassie G-250 blue staining used for quantitative densitometry. (D) Distribution of small (10–15 kDa, black), medium (65–70 kDa, light gray), large (150–200 kDa, dark gray), and all other protein species in the 10–250 kDa range (white) in the BALF as determined by densitometric analysis for the control group. (E) Percent changes in the protein content for each ventilation group compared to control showing significant increases in medium and large proteins caused by Mid-Vt/PEEP0 and High-Vt/PEEP0 ventilation. Significant difference from all other groups (^*), from control (†), Low-Vt/PEEP3 (~), Low-Vt/PEEP0 (%), Mid-Vt/PEEP3 (#), Mid-Vt/PEEP0 (@), and High-Vt/PEEP0 (^∧). Error bars show the standard deviation.

Figure 4

Blood-borne molecules in the bronchoalveolar lavage fluid (BALF) normalized to the control group: (A) serum albumin, (B) immunoglobulin G heavy chain (IgG-HC), (C) and immunoglobulin G light chain (IgG-LC). Error bars show the standard deviation. Significant difference from all other groups (^*), from control (†), Mid-Vt/PEEP3 (#), Mid-Vt/PEEP0 (@), and High-Vt/PEEP0 (^∧). Error bars show standard deviation.

Figure 5

Epithelial damage marker soluble E-Cadherin content in bronchoalveolar lavage fluid (BALF). (A) The 85 kDa E-Cadherin level is normalized to the control group; Mid-Vt/PEEP0, Mid-Vt/PEEP3, and High-Vt/PEEP0 are significantly different from all other groups (^*). (B) BALF E-Cadherin is linearly related to BALF total protein (R² = 0.996). Data points show the median with error bars indicating the interquartile range.

Figure 6

Representative microphotographs of H&E stained mouse lung sections at 10x and 40x magnifications. Black arrows point to macrophages, red arrows point to weakened, damaged, and thinner alveolar walls, and blue arrows point to dilated or congested capillaries. The already broken walls with retracted tissue are circled.

Figure 7

Correlation between injury scores obtained from histological evaluation and total protein (A), normalized E-Cadherin (B), normalized IgG-HC (C), and normalized serum albumin (D). Normalized total protein and E-Cadherin show the median and interquartile range; all other data are mean and standard deviation.

Figure 8

Bronchoalveolar lavage fluid (BALF) total protein plotted against the total cumulative lung overdistension (Sε_T) for the six experimental groups (A). The sigmoidal best fit (black line) y = 0.32 + 2.80/(1 + e^{−(x-1802.2)/180.8}), R² = 0.92, p < 10⁻⁶. Sε_T is quantified as the total volume delivered over the upper corner point [determined by fitting the Venegas equation (Venegas et al., 1998) as described in Section Pressure-Volume Loop Analysis] shown with a filled circle in inset (B) and in Figure 1.