This research investigated whether stretching of lung tissue due to increased positive alveolar pressure swings during mechanical ventilation (MV) at various tidal volumes (Vt) might affect the composition and/or structure of the glycosaminoglycan (GAG) components of pulmonary extracellular proteoglycans. Experiments were performed in 30 healthy rats: 1) anesthetized and immediately killed (controls, C-0); 2) anesthetized and spontaneously breathing for 4 h (C-4h); and 3) anesthetized, paralyzed, and mechanically ventilated for 4 h with air at 0-cmH2O end-expiratory pressure and Vt of 8 ml/kg (MV-1), 16 ml/kg (MV-2), 24 ml/kg (MV-3), or 32 ml/kg (MV-4), adjusting respiratory rates at a minute ventilation of 270 ml/min. Compared with C-0 and C-4h, a significant reduction of dynamic and static compliance of the respiratory system and of the lung was observed only in MV-4, while extravascular lung water significantly increased in MV-3 and MV-4, but not in MV-1 and MV-2. However, even in MV-1, MV induced a significant fragmentation of pulmonary GAGs. Extraction of covalently bound GAGs and wash out of loosely bound or fragmented GAGs progressively increased with increasing Vt and was associated with increased expression of local (matrix metalloproteinase-2) and systemic (matrix metalloproteinase-9) activated metalloproteases. We conclude that 1) MV, even at “physiological” low Vt, severely affects the pulmonary extracellular architecture, exposing the lung parenchyma to development of ventilator-induced lung injury; and 2) respiratory mechanics is not a reliable clinical tool for early detection of lung injury.
- ventilator-induced lung injury
- pulmonary extracellular matrix
- respiratory mechanics
- lung proteoglycans
mechanical ventilation is essential to sustain respiratory function, both during general anesthesia and in patients with respiratory failure. Although a satisfactory arterial blood oxygenation is achieved through a variable combination of modified gas mixture, tidal volume (Vt), respiratory rate, and positive end-expiratory pressure (PEEP), these same factors may be hazardous to lung parenchyma, favoring the development of pulmonary edema. In fact, it has been shown that large Vt values and high inspiratory airway pressures (Paw) with increased transpulmonary pressures (Pl) are associated with increased mechanical stress that may damage the endothelial (7) and epithelial cells (37) with the development of inflammatory response (3, 35) and/or with the inactivation of surfactant (16). Other studies showed that ventilation not only with high, but also with low, Vt values can induce endothelial and vascular alterations (13), interstitial cellular engorgement (10), and bronchiolar damage (8, 9).
The majority of these studies focused on the effects of mechanical ventilation on the alveolar-capillary layer, while studies investigating the role of the macromolecular components of the extracellular tissue matrix (ECM) are, at present, scant. The ECM represents the three-dimensional scaffold of the alveolar wall, which is composed of a layer of epithelial and endothelial cells, their basement membrane, and a thin layer of interstitial space lying between the capillary endothelium and the alveolar epithelium (38). In the lung, the ECM plays several roles, providing 1) mechanical tensile and compressive strength and elasticity; 2) low mechanical tissue compliance, contributing to the maintenance of normal interstitial fluid dynamics (21); 3) low resistive pathway for effective gas exchange (31); and 4) control of cell behavior by binding of growth factors, chemokines, cytokines, and interaction with cell-surface receptors (17).
The pulmonary ECM is composed of a three-dimensional mesh filled with different macromolecules, mainly hyaluronic acid (HA), proteoglycan (PG) families, and other macromolecules, such as collagen and elastin. PGs include families of multidomain core proteins covalently linked to glycosaminoglycan (GAG) chains, like chondroitin sulfate (CS), dermatan sulfate, heparan sulfate (HS), and keratan sulfate (29). In the lung matrix, the most abundant GAGs are CS-GAGs and HS-GAGs, which play different roles: the former being mainly involved in structural support, the latter in the regulation of membrane permeability and cell-to-cell and cell-to-matrix interaction. Unlike HA, CS-GAGs and HS-GAGs are covalently linked to a core protein. Therefore, the proteolytic activity of tissue metalloproteinases (MMPs) may release the GAG chains. It has been shown that interstitial pulmonary edema of different etiologies results from the loss of the protective role played by PGs in organizing the ECM structure (23, 25–27). In the present study, we investigated whether stretching of the lung tissue and/or positive alveolar pressure swings during mechanical ventilation at low- or high-Vt values might have a direct effect on the composition and/or structure of the main pulmonary PG families.
Experiments were performed on 30 adult male Wistar rats (body weight = 345 ± 59 g), anesthetized with an intraperitoneal injection of 2.5 ml/kg of saline containing 0.25 g/ml of urethane, 10 mg/ml of pentobarbital sodium, and 0.03 mg/ml of fentanyl. All experimental procedures and protocols were approved by the Institutional Animal Care Committee at the University of Insubria, Varese, Italy, in accordance with the Health Research Extension Act. The adequacy of anesthesia level was assessed on the basis of the disappearance of the corneal reflexes. Additional boluses of the anesthetic cocktail were added every 60 min. Rats were tracheostomized and instrumented for measurements of esophageal pressure (Pes) and Paw at the tracheal cannula inlet. The correct position of the esophageal catheter was assessed by the Baydur test (2). Saline-filled plastic catheters (PE-50) were inserted into the carotid artery and the jugular vein to continuously monitor systemic arterial and central venous pressure.
Airflow was measured with a heated Fleish pneumotachograph (model 8420, Hans Rudolph) connected to the tracheal cannula and to a differential pressure transducer (Validyne MP45, Northridge, CA). Tracheal pressure was measured with a physiological pressure transducer (model P23XL; Gould Electronics) connected to the side arm of the tracheal cannula; there was no appreciable shift in the signal or alteration in amplitude up to 20 Hz. All recorded signals were amplified, sampled at 100 Hz by a 14-bit analog-to-digital converter, and stored on a desktop computer that also integrated airflow, allowing continuous recording of Vt.
Six experimental groups were used: 1) anesthetized and immediately killed untreated controls (C-0, n = 5); 2) anesthetized rats suppressed after 4 h of spontaneous breathing (C-4h, n = 5); and 3) four groups of rats were left to breathe spontaneously for 15 min after induction of anesthesia and catheters instrumentation. Subsequently, after being paralyzed with 0.2 ml/kg of pancuronium bromide, they were mechanically ventilated (ventilator model 7025; Ugo Basile, Comerio, Italy) with room air at 0 cmH2O end-expiratory pressure using four different Vt values: 8 ml/kg [mechanical ventilation (MV)-1, n = 5], 16 ml/kg (MV-2, n = 5), 24 ml/kg (MV-3, n = 5), or 32 ml/kg (MV-4, n = 5). After paralysis, the adequacy of anesthesia level was assessed judging from the stability of the heart rate and systemic arterial pressure. Respiratory rate was set to keep minute ventilation constant among groups and comparable with C-4h: 90, 45, 30, and 23 breaths/min, respectively.
At the end of the experiment, after sampling arterial blood for gas and serum analysis, dynamic and static pressure-volume (PV) curves were obtained in all groups except C-0. In MV-1, MV-2, MV-3, and MV-4, the dynamic PV curves of the respiratory system were computed off-line at the beginning and at the end of the experiment. For each rat, the dynamic PV plots of 10 subsequent respiratory cycles were averaged. The dynamic compliance of the respiratory system at the beginning (Crs,dyn-b) and at the end (Crs,dyn-f ) of the period of mechanical ventilation was calculated as the ratio between the imposed Vt and the peak Paw (Pawpeak) at end-inspiration, i.e., Crs,dyn-b = Vt/Pawpeak-b and Crs,dyn-f = Vt/Pawpeak-f. The dynamic end-inspiratory Pl (ΔPl,dyn) at baseline and at the end of the experiment was calculated as: ΔPl,dyn = Pawpeak − (Pespeak − Pesexp), where Pespeak and Pesexp are the peak inspiratory and end-expiratory Pes, respectively. At the end of the experiment, the static PV curves of the respiratory system, the lung, and the chest wall were performed in paralyzed animals. After connecting the tracheal cannula to a 20-ml syringe, the respiratory system was passively inflated and deflated three times while simultaneously measuring Pes and Paw. The static PV curve of the lung was obtained by plotting the attained volume as a function of Pl. The final static lung compliance (Clstat-f,) was calculated as the slope of the relaxation PV curve at a volume corresponding to 8 ml above functional residual capacity (≈70% of maximum volume above functional residual capacity).
Then the animals were suppressed with an anesthesia overdose. The right lung was tied off at the hilum and immediately frozen in liquid nitrogen for subsequent biochemical analysis, while the left lung was fixed by intratracheal instillation of fixative to be subsequently processed for morphological analysis (15).
The biochemical analysis consisted of the isolation and purification of GAGs and in the detection of MMP-2, MMP-9, and interleuklin-6 (IL-6) from lung extracts (27, 28). Right lungs were cut in two pieces. One piece was used for measurements of MMP-2 and MMP-9 through zymography and of IL-6 by Western blotting. The presence of IL-6, an early marker of inflammation, was tested to verify whether possible modification of GAG composition and/or MMP activity might be related to inflammatory processes associated with the ventilatory strategy. The other specimen was immediately weighted (wet weight) by means of a precision balance and dried in a speed vacuum apparatus for 24 h and weighted again (dry weight) to obtain the wet-to-dry weight ratio (W/D) as an index of extracellular extravascular lung water. The dry tissue was then used to proceed with GAG analysis.
The lyophilized lung samples were extracted with guanidinium hydrochloride (GuHCl), 0.4 M in 50 mM sodium acetate buffer, pH 5.6, containing protease inhibitors (5 mM benzamidine, 0.1 M ε-aminocaproic acid, 10 mM EDTA, 5 μl/ml PMSF) (22) at 4°C for 48 h and centrifuged, and the pellet was reextracted with 4 M GuHCl in 50 mM sodium acetate buffer, pH 5.6, containing the already described protease inhibitors at 4°C for 48 h (32).
Sequential treatment with increasing concentration of GuHCl is a standardized procedure to break the intermolecular noncovalent bonds: 0.4 M GuHCl still permits the formation of PG aggregates and thus allows extraction of GAG fragments loosely bound to the other ECM components, while exposure to 4 M GuHCl determines the cleavage of most of the intermolecular noncovalent bonds and thus allows the extraction of whole GAGs. Therefore, the amount of PG recovery after tissue extraction with 0.4 M and 4 M GuHCl allows evaluation of 1) the strength of the noncovalent interactions linking PGs to the other macromolecular ECM components, and 2) the ECM molecular integrity.
The two supernatants were collected after centrifuging and used to determine the GAG content with the Farndale et al. method (14). To characterize the GAG content, two aliquots of 100 μl of each extraction were precipitated with four volumes of ethanol at −20°C for 16–18 h and centrifuged at 11,000 g for 30 min, and the pellets were dried and digested at 60°C for 2 h in 300 μl of 100 mM ammonium acetate buffer, pH 7.0, containing 20 U/ml of protease K. The enzymatic treatment was terminated by boiling for 5 min. Four volumes of 96% ethanol per sample volume were added, and the GAGs in the mixture were precipitated at −20°C overnight. Ethanol-precipitated GAGs were centrifuged at 11,000 g at 4°C for 15 min.
Distinction between CS-GAGs, extracted from large PGs typical of the solid matrix structures, and HS-GAGs, extracted from PGs preferentially localized in the basal membrane of the endothelial and epithelial membranes, was performed by selective enzymatic digestion followed by HPLC analysis. HA is a minor share of total GAGs and, in the present analysis, is included in the CS-GAG fraction.
After proteolytic digestion of the tissue, the extracted free GAGs were precipitated with cold ethanol, and the pellets were subsequently dried, dissolved in 100 μl of 100 mM ammonium acetate, pH 7.0, containing 100 mU/ml of chondroitinase avidin-biotin complex (EC 22.214.171.124) and hyaluronidase Streptococcus dysgalactiae (EC 126.96.36.199), or a mix of heparinases I, II, and III, (EC: I, 188.8.131.52; II, no number; III, 184.108.40.206), and digested at 37°C for 16–18 h. The samples were then frozen at −80°C and lyophilized. Unsaturated Δ-disaccharide obtained by enzymatic digestions was then derivatized with 2-aminoacridone (AMAC) and analyzed by HPLC (18, 36).
Derivatization of GAG Δ-disaccharides was done as described by Calabro et al. (5), using the dried sample derived from 100 μl of the enzymatic digestion. A 40-μl volume of 12.5 mM AMAC solution in glacial acetic acid/DMSO (3:17 vol/vol) was added, and samples were incubated for 10–15 min at room temperature. A 40-μl volume of a freshly prepared solution of 1.25 M NaBH3CN in water was added to each sample followed by an overnight incubation at 37°C.
Separation and analysis of AMAC-derivatives of Δ-disaccharides were done as described by Karousou et al. (18) with a Jasco-Borwin chromatograph system with a fluorophore detector (Jasco FP-920, excitation wavelength = 442 nm and emission wavelength = 520 nm). Chromatography was carried out using a reversed-phase column (C-18, 4.6 × 150 mm, Bischoff) at room temperature, equilibrated with 0.1 M ammonium acetate buffer, pH 7.0, filtered through a 0.22-μm membrane filter. A gradient elution was done using a binary solvent system composed of 0.1 M ammonium acetate buffer, pH 7.0 (eluent A), and acetonitrile (eluent B). The flow rate was 1 ml/min, and the following program was used: prerun of column with 100% eluent A for 20 min, isocratic elution with 100% eluent A for 5 min, and gradient elution to 30% eluent B for 30 min and from 30 to 50% for 5 min. Sample peaks were identified and quantified comparing the fluorescence spectra with standard Δ-disaccharides, using Jasco-Borwin software.
For the detection of MMP-2 and MMP-9 activities (20), lung specimens were homogenized in 10 mM Tris·HCl, 150 mM NaCl, 20 mM EDTA, pH 7.5, and added with 100 μl of gelatin-sepharose resin for 2 h at 4°C. After centrifuging, the gelatin sepharose, recovered as a pellet, was washed four times with a buffer of 10 mM Tris·HCl, 200 mM NaCl, 20 mM EDTA, pH 7.5 (20). At the end of washings, the resin was than added with 2% SDS, 10% glycerol, and 0.01% bromophenol blue, and an amount corresponding to 10 mg of initial sample was loaded on a SDS polyacrylamide gels containing 1 mg/ml gelatin. The samples were loaded in the gels without heat denaturization and reducing agents and run at 150 V for 1 h in a minigel apparatus. After the run, the gels were washed at room temperature for 2 h in 2.5% Triton X-100 and incubated overnight at 37°C in 10 mM CaCl2, 150 mM NaCl, and 50 mM Tris·HCl, pH 7.5 buffer. The gel was stained in 2% (vol/vol) Coomassie blue G-250 in fixing solution and photographed on a light box using an imaging densitometer (model GS700, BioRad Laboratories, Hercules, CA). Proteolysis was detected as white bands in a dark blue field.
IL-6 content was assessed by immunoblot (Western blot), performed by the method of Towbin et al. (34). The whole tissue extracts, homogenized in 10 mM Tris·HCl, 150 mM NaCl, 20 mM EDTA, pH 7.5, were separated in 15% SDS-PAGE gels under reducing conditions and transferred onto nitrocellulose sheets. Nitrocellulose membrane, blocked with 5% BSA in Tris-buffered saline/0.1% Tween 20, was probed with polyclonal anti-IL-6 (R-19, sc 1266; Santa Cruz Biotechnology; 1:200 dilution) and anti-actin to confirm equal loading (I-19, sc 1616; Santa Cruz Biotechnology; 1:200 dilution) antibodies. Signals were revealed using secondary peroxidase-conjugated antibodies (1:20,000 diluted horseradish peroxidase-conjugated donkey anti-goat IgG antibody; sc 2020; Santa Cruz Biotechnology), and the band visualization was carried out by the chemiluminescence method (ECL, Amersham Pharmacia Biotech). Pixel intensity of each sample band was measured by densitometric analysis using imaging densitometer (model GS700, BioRad Laboratories, Hercules, CA).
One lung was inflated with 4% Karnovsky's solution, with the fixing solution introduced via the endotracheal tube at a Pl of 5 cmH2O. Finally, the lung was excised and processed for light microscopy analysis. Specimens were fixed in 10% formalin, dehydrated in graded concentrations of alcohol, embedded in paraffin, and cut in 7-μm-thick sections. Serial sections were stained with Azan-Mallory and observed with a light microscope.
Total wall thickness consisted of medial plus adventitial thickness, since the thickness of the intima was minimal in all cases. Measurements of the arteriolar wall in micrometers were done on Azan-Mallory-stained sections. Eight randomly selected images at a final magnification of ×25 were chosen from each lung. In all images, total wall thickness was measured in three different points along the shortest arteriolar diameter. The image analysis system includes an optical microscope (Ortholux 2 Leitz, Wetzlar, Germany) and a high-resolution digital camera (Nikon DS-L1, Nikon, Japan) transmitting image data to a personal computer equipped with appropriate software for image acquisition and analysis (Qwin, Leica Imaging Solution, Cambridge, UK).
Data are reported as means ± 1 SE. Comparison between values from different groups were performed by one-way ANOVA using pairwise multiple-comparison procedures (Bonferroni t-test). Comparison between data obtained in the same animals at different times were performed by paired t-test. Differences between mean values were considered significant at P < 0.05.
At baseline, the mean systemic arterial and venous pressures were 112 ± 8 mmHg and 2.9 ± 0.5 cmH2O, respectively, and remained essentially unaltered throughout the whole experiment. Animals belonging to C-4h, MV-1, MV-2, and MV-3 survived the expected 4 h of spontaneous or mechanical ventilation and were euthanized at the end of the experiment. Vice versa, animals ventilated with the highest Vt (MV-4) spontaneously died at 105 ± 20 min.
The arterial blood-gas analysis performed at the end of the experiment is reported in Table 1. Arterial blood sample was withdrawn before animal suppression with the anesthesia overdose in groups C-4h, MV-1, MV-2, and MV-3 and before dying in group MV-4. With respect to C-4h, a significantly (P < 0.05) decreased arterial pH and arterial Po2 and an increased arterial Pco2 associated with increased lactate was observed in MV-1, but not in MV-2. Increased Vt in MV-2, MV-3, and MV-4 resulted in an increased pH and decreased arterial Pco2 with respect to MV-1.
To evaluate possible changes induced by the ventilatory strategies on respiratory mechanics, the dynamic PV curves of the respiratory system and the static PV curves of the lung were determined. In MV-1, MV-2, and MV-3, the dynamic PV curves of the respiratory system (Fig. 1) were attained at the beginning of the experiment (solid lines) and after 4 h of mechanical ventilation (dotted lines); in animals of group MV-4 that died unexpectedly, the second dynamic PV curve was recorded immediately before death. The dynamic PV curves of the respiratory system shifted to the right only in groups MV-1 and MV-4. In fact, as shown in Table 2, in these two groups the Crs,dyn-f was significantly decreased (P < 0.05) with respect to corresponding initial values (Crs,dyn-b).
The static PV curves of the lung obtained at the end of the experiment are presented in Fig. 2 for MV-1 (A), MV-2 (B), MV-3 (C), and MV-4 (D). In all panels, the static PV curve of the lung obtained at the end of the experiments in group C-4h (open circles) is shown as a reference control curve. In MV-1, the first inflation curve (open triangles) was significantly (P < 0.05) shifted to the right compared with C-4h; however, the two subsequent PV curves of the lung (open inverted triangles) returned toward baseline values. No significant change in the static PV curves of the lung was found in MV-2 (open squares) and MV-3 (solid squares) compared with C-4h. An evident shift to the right of both the inflation (right arm) and the deflation (left arm) PV curves of the lung, accompanied by a significant reduction of the average Clstat obtained at the end of experiment (Clstat-f, Table 2), was instead observed in MV-4 (solid diamonds) with respect to C-4h. This evident modification suggests that the highest Vt induced considerable lung tissue injury, as also indicated by the fact that, in face of their satisfactorily arterial blood-gas analysis (Table 1), MV-4 animals did not survive 4 h of mechanical ventilation.
The amount of GAGs extracted with 0.4 M and 4 M GuHCl is shown in Fig. 3. Material extracted with 0.4 M GuHCl quantifies free GAG chains, which represent fragments from tissue PGs loosely bound to other ECM macromolecules, whereas treatment with 4 M GuHCl allows recovery of all GAGs, including the chains covalently linked to PGs. The total amount of GAGs extracted from the tissue, given by the sum of the material extracted with 0.4 M and 4 M GuHCl, significantly decreased in all mechanically ventilated lungs (MV-1, MV-2, MV-3, and MV-4) compared with C-0 and C-4h. In fact, in MV-1 and MV-4, total GAGs were, respectively, 57.1 and 7% of the corresponding value in C-0. The amount of GAGs extracted with 0.4 M GuHCl increased (by 150%) in MV-1 compared with controls. With further increasing Vt, both the 0.4 M GuHCl and the 4 M GAGs extracted amount progressively decreased; this suggests that the higher the Vt, the stronger the damaging effect on PGs, whose progressively smaller GAG fragments were more promptly washed out of the lung tissue.
As indicated at the bottom of Fig. 3, the W/D significantly increased above control values (C-0 and C-4h) only in MV-3 (6.5 ± 0.5, P < 0.005) and MV-4 (7.6 ± 0.2, P < 0.001).
HPLC characterization of the extracted GAGs is shown in Fig. 4. As the most abundant GAG molecules in the lung tissue are CS-GAGs and HS-GAGs, their percent composition was assessed. The CS-GAG and HS-GAG quantities were obtained by multiplying the CS-GAG and HS-GAG percentage by the total extracted GAG amount expressed per unit dry tissue weight (Fig. 4A). Figure 4B shows the percent change of extracted CS-GAGs and HS-GAGs, with respect to C-0. In MV-1 and MV-2, the decreased amount of GAGs extracted with 4 M GuHCl and the simultaneous increase of the 0.4 M fraction indicate the progressive fragmentation of the strongly bound GAGs into smaller fragments, extractable with 0.4 M GuHCl. The decrease in both CS-GAGs and HS-GAGs with further increasing Vt (MV-3 and MV-4) indicates the occurrence of additional GAG fragmentation and, possibly, fragment wash out.
The evolution toward lung injury at progressively increasing Vt was also evident from lung histology (Fig. 5). In fact, with respect to C-0, the interalveolar lung septa (arrowheads) appeared thickened and slightly congested, even in C-4h and MV-1. The lung appearance progressively worsened with increasing Vt, showing formation of periarteriolar fluid cuff and capillary congestion. In fact, as reported at the bottom of Fig. 5, the average thickness of the perivascular interstitial space surrounding the arterioles of diameter between ∼50 and 300 μm was significantly lower in C-0 compared with C-4h and further increased with increasing Vt. The perivascular interstitial space was slightly stained in MV-1, MV-2, MV-3, and MV-4 with respect to C-0 and C-4h, indicating that protein and mucopolysaccharide components were thinning.
To assess whether the observed PG degradation reflected a purely mechanical tissue stress or an enzymatic activation, we measured the activity of metalloproteases MMP-2 and MMP-9 and their zymogens (pro-MMP-2 and pro-MMP-9) in the lung tissue. As indicated by the zymograph examples of Fig. 6A and by the corresponding histogram (Fig. 6B) referring to the densitometric analysis of each band, the enzymatic activity of pro-MMP-2 and pro-MMP-9 and of MMP-2 and MMP-9 significantly increased with mechanical ventilation, following a Vt-dependent fashion. A slight MMP activation, with respect to C-0, was observed even in C-4h, likely as a result of the surgical insult suffered during tracheostomy, intubation, and vessel catheterization.
To test whether the observed PG fragmentation and MMP increase might depend on an inflammatory phenomenon, IL-6 activity was measured. The Western blot analysis performed on lung tissue specimen (Fig. 7, top) shows that IL-6 expression was increased already in C-4h compared with C-0, suggesting that the induction of anesthesia and preparative surgery caused “per se” an inflammatory response. However, as indicated by the histogram at the bottom of Fig. 7, IL-6 activity did not significantly change with respect to MV-1, when the lungs were ventilated at higher Vt values.
Mechanical Ventilation and GAGs
The main result of the present study is that, in healthy lungs, shifting from spontaneous breathing to mechanical ventilation leads to a marked fragmentation of pulmonary GAGs. The net effect of mechanical ventilation at the same Vt as in spontaneous breathing was reflected by the differences between MV-1 and C-0 and/or C-4h. Although a mild activation of MMPs was observed in C-4h with respect to C-0 (Fig. 6), GAG extraction in C-4h was unaltered with respect to C-0 (Figs. 3 and 4). Conversely, significant GAG fragmentation and activation of MMP-9 and MMP-2 was observed in MV-1. This finding suggests that, in mechanically ventilated MV-1, tissue stress enhanced the proteolytic effect of released MMPs, leading to GAG degradation. Therefore, the increased GAG degradation in MV-1 compared with C-4h might mainly depend on the higher inspiratory Paw and Pl (Table 2), being Vt, end-expiratory pressure, and MMP activity similar in both groups. In addition, the modified pressure distribution related to the shift from spontaneous breathing with negative pressure-induced inspiration around the lung to mechanical ventilation with positive Paw might have contributed to the development of GAG fragmentation.
It is worth noting that, although GAGs extracted with 4 M GuHCl (intact GAGs covalently linked to PG core protein) and the total (0.4 M + 4 M) GAGs decreased in MV-1 with respect to controls, the 0.4 M GuHCl fraction was increased. Since the 0.4 M GuHCl GAGs include noncovalently linked PGs and fragments resulting from cleavage of the covalently linked GAGs, this finding suggests that the ongoing fragmentation of PGs is associated with formation of smaller, more extractable molecules.
The lesional effect of mechanical ventilation on pulmonary GAGs worsened with increasing Vt and Pl (Fig. 3) and was accompanied by a marked activation of MMPs for Vt ≥ 16 ml/kg (MV-2, Fig. 6). The present data do not allow us to distinguish between the net effect of mechanical load and of increased proteolytic activity on GAG cleavage. However, MMP and pro-MMP activities were similar in MV-2, MV-3, and MV-4; therefore, the Vt-dependent tissue damage might depend on an increasing tissue stretch, caused by higher peak inspiratory Paw and Pl (Table 2), applied to a matrix scaffold already weakened by the proteolytic action of MMPs.
In line with previous observations (4, 9), GAG degradation during mechanical ventilation of previously healthy lungs did not seem to be mediated by an inflammatory process. In fact, IL-6 expression increased after anesthesia and mild surgery (C-4h), but remained substantially unaltered in mechanical ventilation at increasing Vt values. However, since we limited our observation to IL-6, we cannot exclude the possibility that other inflammatory mediators are activated and/or that GAG breakdown fragments might provide by themselves an undetected proinflammatory activity.
Although the damaging effect of low Vt values on pulmonary GAGs in healthy lungs has, at least to our knowledge, never been described before, other effects of mechanical ventilation not mediated by proinflammatory processes (9) have been reported, such as 1) epithelial cell injuries with leucocyte infiltration in the alveolar septa and increased abnormal alveolar-bronchiolar attachments (10); 2) epithelial cell damage depending on distending pressure (19); 3) endothelial cell damage promoting right ventricular dysfunction with increased microvascular leakage (13); and 4) peripheral airway injury (8). On the other hand, mechanical ventilation at high Vt has been reported to determine a significant increase in the protein component of CS-PG and HS-PGs and a strengthening of the GAG bonds (1). A possible explanation of this discrepancy might be a mild proteolytic activity induced by mechanical ventilation after a shorter time (2 h) compared with ours.
Mechanical Ventilation and Lung Fluid Balance
Thickening of interalveolar septa and appearance of periarteriolar cuffs (Fig. 5) were observed even at low Vt values in face of a normal lung W/D (Fig. 3), which increased significantly only in MV-3 and MV-4. This indicates that fluid accumulation in the lung tissue became evident only at high Vt values, being instead undetectable in terms of W/D changes at low Vt values.
This observation may be explained by considering the specific roles of HS-PGs and CS-PGs in controlling tissue fluid balance and the consequences of their fragmentation. HS-PG in the normal capillary basement membranes provides a selective sieve with respect to fluid and solutes, minimizing fluid filtration from capillary into the interstitial space. Structural CS-PGs are instead responsible for the considerable stiffness, or low compliance, of the pulmonary matrix (21, 25). A low compliant tissue is well protected against edema development: in fact, even small increases in interstitial fluid volume raise interstitial fluid pressure, preventing further fluid accumulation in the tissue. Therefore, cleavage of HS-PGs and CS-PGs (Fig. 4) favored fluid filtration during mechanical ventilation by increasing both endothelial permeability and tissue compliance. In MV-1 and MV-2, a normal W/D suggests that, despite a partial fragmentation of CS- and HS-GAGs, the matrix structure was likely able to oppose fluid filtration through the capillary endothelium, limiting fluid accumulation to the peri-microvascular space (Fig. 5). However, further degradation and washout of CS- and HS-GAGs, as in MV-3 and MV-4, led to a deeper disorganization of the whole tissue scaffold (22–26), likely augmenting interstitial tissue compliance favoring development of interstitial edema.
Mechanical Ventilation and Lung Mechanics
Changes in pulmonary mechanics have been proposed as a noninvasive tool to monitor progression of lung injury in patients. A recent study (30) performed on healthy lungs showed that lung injury correlated with the compliance of the respiratory system. However, the experimental protocol was quite different from the present one by using 1) higher Vt values and inspired O2 fraction; 2) continuous intravenous fluid load; and 3) repeated recruitment maneuvers at high alveolar pressures.
Our study suggests that, in the face of a Vt-dependent GAG degradation, a significant reduction of the dynamic and static compliance of the respiratory system and of the lung was observed only at the highest Vt (MV-4), when both the 0.4 M GuHCl and 4 M GuHCl GAG fractions were severely degraded and washed out (Fig. 3). A mild shift to the right of the dynamic and static PV curves was actually observed at low Vt (MV-1, Fig. 1A and 2A); however, unlike what was found in MV-4, in MV-1 the static PV curves returned to baseline just after a single recruitment maneuver, suggesting the occurrence, in MV-1, of just a transitory alveolar de-recruitment. In MV-1, MV-2, and MV-3, where GAGs were not as fragmented as in MV-4, no modification of the static PV curves was encountered. In line with conclusions drawn by Cavalcante et al. (6) on lung tissue strips, our data indicate that even a limited fraction of functional GAGs is able to stabilize the collagen-elastic tissue network, preserving the mechanical behavior of the lung parenchyma. On the other hand, an almost complete loss of the GAG/PG component, as in MV-4, severely compromises the mechanical properties and the macromolecular stability of the matrix, leading to tissue injury.
Different from previous studies, we did not apply any external PEEP. It is possible that the increased GAG degradation observed in MV-1 compared with C-4h was a consequence of the lack of PEEP. In fact, it is assumed that, during anesthesia and mechanical ventilation, the end-expiratory lung volume decreases, causing an increased closure of the peripheral airways and/or atelectasis, mainly in the dependent lung (33). Therefore, mechanical ventilation “per se” might induce lung damage and trigger inflammatory processes, due to the continuous and higher than normal tissue stretching, even at “physiological” low Vt values. PEEP is supposed to avoid such negative consequences by restoring the end-expiratory lung volume, at least in injured lungs (11). However, although taking into account this possible limitation, we decided to perform the experiments in the absence of PEEP for the following reasons. 1) It has been shown that lung volume reduction and atelectasis formation were similar during spontaneous breathing and mechanical ventilation at the same level of anesthesia (33). In addition, there is no evidence that muscle paralysis “per se” further affects lung volumes when added to general anesthesia. Therefore, no volume reduction ought to be expected between the spontaneously breathing or mechanically ventilated groups. 2) Most of the studies showing the protective effect of PEEP were performed in experimental models of lung injury (11). On the contrary, we studied rats with healthy lungs. Only a few experimental studies in mechanically ventilated, anesthetized animals with healthy lungs showed protective effects of PEEP, when ventilated with “physiological” low Vt values. However, the experimental conditions were very different from the present ones: either the chest wall was open (8, 10), or negative pressures were applied at end expiration (9), or intermittent recruitment maneuvers were performed throughout the experiments (12). 3) Conflicting data are available on the protective role of PEEP toward systemic and local inflammatory response during general anesthesia in patients with healthy lungs (39, 40). Therefore, we decided not to apply an external PEEP, because 1) there are no conclusive indications to apply an external PEEP during general anesthesia in healthy lungs, and 2) the aim of the present study was not to propose a new ventilatory strategy during mechanical ventilation for general anesthesia, but to investigate the response of tissue GAGs to increased tissue stress.
In conclusion, the present study suggests that mechanical ventilation “per se” exposes the healthy lungs to tissue injury and GAG fragmentation, even at low Vt values, with further progressive worsening in a Vt-dependent fashion. Monitoring of the respiratory mechanics and gas-exchange does not seem to be reliable tools for an early detection of ventilatory-induced lung injury in previously healthy lungs.
This research was funded by the Italian Ministry of the University and of Scientific and Technological Research (COFIN-2003, grant no. 2003055193; FAR 2004–2005) and by a contribution for Research from the University of Insubria.
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