Hypervolemia induces and potentiates lung damage after recruitment maneuver in a model of sepsis-induced acute lung injury

This study was approved by the Ethics Committee of the Health Sciences Center, Federal University of Rio de Janeiro. All animals received humane care in compliance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the National Academy of Sciences, USA.

Sixty-six adult male Wistar rats (270 to 300 g) were kept under specific pathogen-free conditions in the animal care facility at the Laboratory of Pulmonary Investigation, Federal University of Rio de Janeiro. In 36 rats, Est,L, histology, and molecular biology were analyzed. The remaining 30 rats were used to evaluate lung W/D ratio. Animals were fasted for 16 hours before the surgical procedure. Following that, sepsis was induced by cecal ligation and puncture (CLP) as described in previous studies [19]. Briefly, animals were anesthetized with sevoflurane and a midline laparotomy (2 cm incision) was performed. The cecum was carefully isolated to avoid damage to blood vessels, and a 3.0 cotton ligature was placed below the ileocecal valve to prevent bowel obstruction. Finally, the cecum was punctured twice with an 18 gauge needle [20] and animals recovered from anesthesia. Soon after surgery, each rat received a subcutaneous injection of 1 ml of warm (37°C) normal saline with tramadol hydrochloride (20 μg/g body weight).

Figure 1 depicts the time-course of interventions. Forty-eight hours after surgery, rats were sedated (diazepam 5 mg intraperitoneally), anesthetized (thiopental sodium 20 mg/kg intraperitoneally), tracheotomized, and a polyethylene catheter (PE-10; SCIREQ, Montreal, Canada) was introduced into the carotid artery for blood sampling and monitoring of mean arterial pressure (MAP). The animals were then paralyzed (vecuronium bromide 2 mg/kg, intravenously) and mechanically ventilated (Servo i, MAQUET, Switzerland) with the following parameters: V_T = 6 ml/kg, respiratory rate (RR) = 80 breaths/min, inspiratory to expiratory ratio = 1:2, fraction of inspired oxygen (FiO₂) = 1.0, and PEEP equal to 0 cmH₂O (zero end-expiratory pressure (ZEEP)). Blood (300 μl) was drawn into a heparinized syringe for measurement of arterial oxygen partial pressure (PaO₂), arterial carbon dioxide partial pressure (PaCO₂) and arterial pH (pHa) (i-STAT, Abbott Laboratories, North Chicago, IL, USA) (BASELINE-ZEEP). Afterwards, mechanical ventilation was set according to the following parameters: V_T = 6 ml/kg, RR = 80 bpm, PEEP = 5 cmH₂O, and FiO₂ = 0.3 (Figure 1). Est,L was then measured (BASELINE) and the animals were randomly assigned to one of the following groups: 1) hypovolemia (HYPO); 2) normovolemia (NORMO), and 3) hypervolemia (HYPER). Hypovolemia was induced by blood drainage in order to achieve a MAP of about 70 mmHg. Normovolemia was maintained at a MAP of about 100 mmHg. Hypervolemia was obtained with colloid administration (Gelafundin^®; B. Braun, Melsungen, Germany) at an infusion rate of 2 ml/kg/min to achieve a MAP of about 130 mmHg. Following that, the colloid infusion rate was reduced to 1 ml/kg/min in order to maintain a constant MAP. Depth of anesthesia was similar in all animals and a comparable amount of sedative and anesthetic drugs were given in all groups. After achieving volemic status, animals were further randomized to be recruited, with a single RM consisting of continuous positive airway pressure (CPAP) of 40 cmH₂O for 40 seconds (RM-CPAP), or not (NR) (n = 6 per group; Figure 1). After one hour of mechanical ventilation (END), Est,L was measured. FiO₂ was then increased to 1.0, and after five minutes arterial blood gases were analyzed (END). Finally, the animals were euthanized and lungs, kidney, liver and small intestine were prepared for histology. IL-6, IL-1β, caspase-3, and PCIII mRNA expressions were measured in lung tissue. The experiments took no longer than 80 minutes.

Airflow, airway and esophageal pressures were measured [9, 21]. Changes in esophageal pressure, which reflect chest wall pressure, were measured with a water-filled catheter (PE205) with side holes at the tip connected to a SCIREQ differential pressure transducer (SC-24, Montreal, Canada). Before animals were paralyzed, the catheter was passed into the stomach, slowly returned into the esophagus, and its proper positioning was assessed using the 'occlusion test' [22, 23]. Transpulmonary pressure was calculated by the difference between airway and esophageal pressures [9]. All signals were filtered (100 Hz), amplified in a four-channel conditioner (SC-24, SCIREQ, Montreal, Quebec, Canada), sampled at 200 Hz with a 12-bit analogue-to-digital converter (DT2801A, Data Translation, Marlboro, MA, USA) and continuously recorded throughout the experiment by a personal computer. To calculate Est,L, airways were occluded at end-inspiration until a transpulmonary plateau pressure was reached (at the end of five seconds), after which this value was divided by V_T[9, 21]. All data were analyzed using ANADAT data analysis software (RHT-InfoData, Inc., Montreal, Quebec, Canada).

Volemic status and cardiac function were assessed by an echocardiograph equipped with a 10 MHz mechanical transducer (Esaote model, CarisPlus, Firenze, Italy). Images were obtained from the subcostal and parasternal views. Short-axis B-dimensional views of the left ventricle were acquired at the level of the papillary muscles to obtain the M-mode image. The inferior vena cava (IVC) and right atrium (RA) diameters were measured from the subcostal approach. Cardiac output, stroke volume, and ejection fraction were obtained from the B-mode according to Simpson's method [24].

A laparotomy was performed immediately after determination of lung mechanics and heparin (1,000 IU) was intravenously injected in the vena cava. The trachea was clamped at end-expiration (PEEP = 5 cmH₂0), and the abdominal aorta and vena cava were sectioned, yielding a massive hemorrhage that quickly killed the animals. Right lung, kidney, liver, and small intestine were then removed, fixed in 3% buffered formaldehyde and paraffin-embedded. Four-μm-thick slices were cut and stained with H&E.

Lung morphometric analysis was performed using an integrating eyepiece with a coherent system consisting of a grid with 100 points and 50 lines (known length) coupled to a conventional light microscope (Olympus BX51, Olympus Latin America-Inc., São Paulo, Brazil). The volume fraction of the lung occupied by collapsed alveoli or normal pulmonary areas or hyperinflated structures (alveolar ducts, alveolar sacs, or alveoli, all wider than 120 μm) was determined by the point-counting technique [25] at a magnification of × 200 across 10 random, non-coincident microscopic fields [26].

Three slices measuring 2 × 2 × 2 mm were cut from three different segments of the left lung and fixed (2.5% glutaraldehyde and phosphate buffer 0.1 M (pH = 7.4)) for electron microscopy (JEOL 1010 Transmission Electron Microscope, Tokyo, Japan) analysis. For each electron microscopy image (15 per animal), the following structural damages were analyzed: a) alveolar capillary membrane, b) type II epithelial cells, and c) endothelial cells. Pathologic findings were graded according to a five-point semi-quantitative severity-based scoring system as: 0 = normal lung parenchyma, 1 = changes in 1 to 25%, 2 = changes in 26 to 50%, 3 = changes in 51 to 75%, and 4 = changes in 76 to 100% of examined tissue [9, 21].

Terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) staining was used in a blinded fashion by two pathologists to assay cellular apoptosis. Apoptotic cells were detected using In Situ Cell Death Detection Kit, Fluorescin (Boehringer, Mannheim, Frankfurt, Germany). The nuclei without DNA fragmentation stained blue as a result of counterstaining with hematoxylin [20]. Ten fields per section from the regions with apoptotic cells were examined at a magnification of × 400. A five-point semi-quantitative severity-based scoring system was used to assess the degree of apoptosis, graded as: 0 = normal lung parenchyma; 1 = 1-25%; 2 = 26 to 50%; 3 = 51 to 75%; and 4 = 76 to 100% of examined tissue.

Quantitative real-time RT-PCR was performed to measure the expression of IL-6, IL-1β, caspase-3, PCIII, VCAM, and ICAM genes. Central slices of left lung were cut, collected in cryotubes, quick-frozen by immersion in liquid nitrogen and stored at -80°C. Total RNA was extracted from the frozen tissues using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to manufacturer's recommendations. RNA concentration was measured by spectrophotometry in Nanodrop^® ND-1000 (Thermo Fisher Scientific, Wilmington, DE, USA). First-strand cDNA was synthesized from total RNA using M-MLV Reverse Transcriptase Kit (Invitrogen, Carlsbad, CA, USA). PCR primers for target gene were purchased (Invitrogen, Carlsbad, CA, USA). The following primers were used: IL-1β (sense 5'-CTA TGT CTT GCC CGT GGA G-3', and antisense 5'-CAT CAT CCC ACG AGT CAC A-3'); IL- 6 (sense 5'-CTC CGC AAG AGA CTT CCA G-3' and antisense 5'-CTC CTC TCC GGA CTT GTG A-3'); PCIII (sense 5'-ACC TGG ACC ACA AGG ACA C-3' and antisense 5'-TGG ACC CAT TTC ACC TTT C-3'); caspase-3 (sense 5'-GGC CGA CTT CCT GTA TGC-3' and antisense 5'-GCG CAA AGT GAC TGG ATG-3'); VCAM-1 (sense 5'-TGC ACG GTC CCT AAT GTG TA-3' and antisense 5'-TGC CAA TTT CCT CCC TTA AA-3'); ICAM-1 (sense 5'-CTT CCG ACT AGG GTC CTG AA-3' and antisense 5'-CTT CAG AGG CAG GAA ACA GG-3'); and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; sense 5'-GGT GAA GGT CGG TGTG AAC- 3' and antisense 5'-CGT TGA TGG CAA CAA TGT C-3'). Relative mRNA levels were measured with a SYBR green detection system using ABI 7500 Real-Time PCR (Applied Biosystems, Foster City, CA, USA). All samples were measured in triplicate. The relative expression of each gene was calculated as a ratio compared with the reference gene, GAPDH and expressed as fold change relative to NORMO-NR.

W/D ratio was determined in the right lung as previously described [27]. Briefly, the right lung was separated, weighed (wet weight) and then dried in a microwave at low power (200 W) for five minutes. The drying process was repeated until the difference between the two consecutive lung weight measurements was less than 0.002 g. The last weight measurement represented the dry weight.

Normality of data was tested using the Kolmogorov-Smirnov test with Lilliefors' correction, while the Levene median test was used to evaluate the homogeneity of variances. If both conditions were satisfied, one-way analysis of variance (ANOVA) for repeated measures was used to compare the time course of MAP, IVC and RA dimensions. To compare arterial blood gases, Est,L, and echocardiographic data at BASELINE and after one hour of mechanical ventilation (END), the paired t-test was used. Lung mechanics (END) and morphometry, echocardiographic data (END), arterial blood gases (END), W/D ratio, and inflammatory and fibrogenic mediators were analyzed using two-way ANOVA followed by Tukey's test. To compare non-parametric data, two-way ANOVA on ranks followed by Dunn's post-hoc test was selected. The relations between functional and morphological data were investigated with the Spearman correlation test. Parametric data were expressed as mean ± standard error of the mean, while non-parametric data were expressed as median (interquartile range). All tests were performed using the SigmaStat 3.1 statistical software package (Jandel Corporation, San Raphael, CA, USA), and statistical significance was established as P < 0.05.

To our knowledge, this is the first study investigating the combined effects of RMs and volemic status in sepsis-induced ALI. We used a CLP model of sepsis because it is reproducible and leads to organ injury that is comparable with that observed in human surgical sepsis [28, 29].

Volemic status was assessed by echocardiography. It has been shown that echocardiography provides valuable information on preload and cardiac output [30, 31]. An inspired oxygen fraction of 0.3 was used throughout the study to minimize possible iatrogenic effects of high inspiratory oxygen concentration on the lung parenchyma [32]. To avoid possible confounding effects of ventilation/perfusion mismatch on the interpretation of the gas-exchange data, inspiratory oxygen fraction was increased to 1.0 just before arterial blood sampling [33]. All animals underwent protective mechanical ventilation to minimize possible interactions between conventional mechanical ventilation, volemic status, and RMs.

The mRNA expressions of IL-6 and IL-1β in lung tissue were determined due to the role of these markers in the pathogenesis of sepsis and ventilator-induced lung injury (VILI) [34]. Although IL-6 has been implicated in the triggering process of sepsis and correlates with its severity [35], IL-1β has been associated with the degree of VILI [32]. On the other hand, mRNA expression of PCIII was determined because it is the first collagen to be remodeled in the development/course of lung fibrogenesis [36], as well as being an early marker of lung parenchyma remodeling [32, 37]. We also measured the levels of mRNA expression of caspase-3, because it represents a surrogate parameter for the final step of apoptosis [38]. Finally, the effects of volemic status and RM on mRNA expressions of ICAM-1 and VCAM-1 were determined because these adhesion molecules are involved in the accumulation of neutrophils in the lung tissue, playing a crucial role in the pathogenesis of VILI [39].

In severe sepsis aggressive fluid resuscitation is recommended [40]. However, in ALI/ARDS the optimal fluid management protocol is yet to be established. Conservative management of ALI/ARDS prescribes that fluid intake be restricted in an attempt to decrease pulmonary edema, shorten the duration of mechanical ventilation, and improve survival. A possible risk of this approach is a decrease in cardiac output and worsening of distal organ function, both of which are reversed with the liberal approach.

Our data show that a hypervolemic status led to increased lung, but not distal organ injury. In fact, hypervolemia was associated with a more pronounced detachment of the alveolar-capillary membrane as well as injury of endothelial cells. On the other hand, fluid restriction did not increase distal organ injury. Different mechanisms could explain the adverse effects of hypervolemia on lung injury: 1) increased hydrostatic pressures; and 2) augmented capillary blood flow and volume.

During hypervolemia, increased pulmonary edema was induced by altered permeability of the alveolar capillary membrane, which is a common finding in sepsis [41], combined with higher hydrostatic pressure. In the presence of pulmonary edema, the increase in hydrostatic pressures along the ventral-dorsal gradient promoted a reduction in normally aerated tissue, contributing to increased stress/strain and cyclic collapse/reopening [42].

Hypervolemic groups were characterized by impaired oxygenation and higher Est,L. The reduction in oxygenation can be attributed to increased edema and atelectasis. The increase in Est,L suggested higher lung stress in aerated lung areas during inflation. In addition, as the same V_T was applied in all groups and hypervolemia decreased the normally aerated tissue, the strain in the hypervolemic group may be increased. However, even if stress/strain were higher, we did not observe hyperinflation probably because low V_T and moderate PEEP levels were applied. In this line, cyclic collapse/reopening has also been recognized as a determinant of VILI [43].

Cardiac output, stroke volume, and ejection fraction were increased during hypervolemia. Increased pulmonary perfusion may also directly damage the lungs. In a model of VILI, Lopez-Aguilar and colleagues [44] have shown that the intensity of pulmonary perfusion contributes to the formation of pulmonary edema, adverse distribution of ventilation, and histological damage.

In hypervolemia, we observed an increase in IL-6 mRNA expression in lung tissue, but PCIII mRNA expression did not change, which may be explained by the absence of hyperinflation [12]. Additionally, VCAM-1 and ICAM-1 mRNA expressions were elevated in HYPER group suggesting endothelial activation due to vascular mechanical stretch.

Despite increased lung injury and activation of the inflammatory process, hypervolemia was not associated with increased distal organ injury. Furthermore, hypovolemia and normovolemia did not contribute to distal organ injury, but rather protected the lungs from further damage. Our observation supports the claim that the lungs are particularly sensitive to fluid overload [45]. Lung-borne inflammatory mediators can spill over into the circulation and promote distal organ injury. However, when protective mechanical ventilation is used, decompartmentalization of the inflammatory process is limited [46].

The low V_T and airway pressure concept has been shown to decrease the mortality in ALI/ARDS patients [1]. Given the uncertain benefit of RMs on clinical outcomes, the routine use of RMs in ALI/ARDS patients cannot be recommended at this time. However, RMs have been shown to improve oxygenation without serious adverse events [11]. Furthermore, other papers suggested that RMs may be useful before PEEP setting, after inadvertent disconnection of the patient from the mechanical ventilator or airways aspiration [47]. Finally, RMs have been proposed to further improve respiratory function in ALI/ARDS patients in prone position [48]. Thus, in our opinion, their judicious use in the clinical setting may be justified.

In our animals, RMs reduced alveolar collapse and increased normal aerated tissue independent of the degree of volemia. Along this line, experimental and clinical studies have shown that improvement in lung aeration is associated with better lung mechanics [49‐51]. RMs improved oxygenation during hypervolemia, probably because of the higher amount of collapsed lung tissue, which may increase the effectiveness of RMs reversing atelectasis and decreasing intrapulmonary shunt. Gattinoni and colleagues [51] have shown that the beneficial effects of RMs are more pronounced in patients with higher lung weight and atelectasis. The lack of correlation between reduction in atelectasis and oxygenation after RMs in the HYPO and NORMO groups could also be explained by the redistribution of perfusion [52, 53]. After RM, Est,L was reduced in HYPO but not in NORMO or HYPER groups. The improvement in Est,L in HYPO group could be explained by alveolar recruitment, whereas the lack of improvement in the other groups may be related to the combination of alveolar recruitment and the increase in interstitial and/or alveolar edema, with consequent increase in specific Est,L.

RMs increase alveolar fluid clearance [8] and aerated tissue, which may lead to reduced lung stretch and inflammatory mediator release [54]. Our data suggest that RMs in the HYPO and NORMO groups did not result in further damage of epithelial and endothelial cells or increased expression of inflammatory and fibrogenic mediators. In addition, RMs induced higher mRNA expression of VCAM-1 in NORMO and HYPER groups, but not of ICAM-1, which was presented higher in HYPER group regardless of RM. Conversely, in HYPO group after RM, the mRNA expression of VCAM-1 and ICAM-1 decreased, probably reflecting reduced shear stress.

RMs transiently increase lung stress [50], probably damaging the alveolar capillary membrane triggering inflammatory and fibrogenic responses [9, 12] and impairing net alveolar fluid clearance [8]. However, the potential of RMs to damage the lung is still a matter of debate [11]. In hypervolemia, our results suggest that despite an improvement in functional parameters, RMs are associated with increased detachment of the alveolar capillary membrane, injury of epithelial type II and endothelial cells, as well as an activation of inflammatory and fibrogenetic response. As previously discussed, hypervolemia per se may worsen lung injury, especially at the level of the alveolar capillary membrane. Our results suggest that the negative effects of hypervolemia on lung damage are potentiated by increased stress/strain induced by RMs.

The increase in different inflammatory mediators after RMs in hypervolemia cannot be explained by increased atelectasis and/or cyclic opening and closing of collapsed units. In fact, atelectasis was reduced after RMs in hypervolemia. Thus, the increase in gene expression of inflammatory mediators in the lung may have resulted from a single sustained inflation RM.

There are conflicting data on the potential of RMs to decompartmentalize lung inflammation [55, 56]. Our results suggest that the combination of RMs with hypervolemia does not result in distal organ injury. Nevertheless, we cannot extrapolate these results to longer periods of ventilation and/or the application of other strategies to recruit the lungs. Theoretically, the inflammatory process could spread to distal organs in the long term. On the other hand, more frequent RMs could accelerate and exacerbate our findings. Also, RMs with pressure profiles different from the sustained inflation, for example gradual increase of airway pressure, could lead to reduced stress and reduce the biological impact of the maneuver. Certainly, this issue deserves further investigation.

We observed greater injury of type II epithelial cells and gene expression of PCIII when lungs were recruited in hypervolemia. Not only are type II cells involved in surfactant production, they are also associated with repairing mechanisms of injured lungs [57]. Re-expansion of collapsed lung units may expose the alveoli to tensile and shear stresses stimulating fibroblasts and macrophages to synthesize collagen fibers [58]. Our results are in accordance with previous reports demonstrating increased procollagen mRNA expression in lungs submitted to high airway pressures [37].

This study has several limitations. Firstly, we used a CLP model of sepsis. Thus, our results cannot be extended to other experimental models of sepsis or directly extrapolated to the clinical scenario. Secondly, the mortality of our sepsis model was relatively high (40%). Thus, we cannot completely exclude that there was a kind of natural bias and a 'sepsis-tolerating' population has been unintentionally selected. However, if hypervolemia was able to produce and potentiate lung damage after RMs in this subgroup, effects would have been even more pronounced in a less 'sepsis-tolerating' population. Thirdly, the observation time was relatively short (one hour), precluding extrapolation of our findings to longer periods of ventilation. The one-hour period was chosen based on our experience with this model and taking the time needed to detect alterations in the proinflammatory and fibrogenetic response of the lungs due to mechanical ventilation in rats [21, 59]. As we identified that the proinflammatory response was activated and the alveolocapillary membrane was damaged in the short period, we speculate that the protein levels of the inflammatory cytokines would be higher in the lungs with hypervolemia (specially after RMs) and achieve distal organs due to decompartmentalization if the observation period would have been extended. Fourthly, hypervolemia was achieved by infusion of gelatin. Different results may be observed with other types of colloids or even crystalloids. Finally, the RM was performed as sustained inflation. Recent studies have reported reduced lung injury and fewer adverse hemodynamic effects with other types of RM [12]. However, sustained inflation is the most commonly used RM in clinical practice [11].