Injurious mechanical ventilation affects neuronal activation in ventilated rats

This study was approved by the Animal Ethics Committee at the Corporació Sanitaria Parc Taulí, Barcelona, Spain. We studied 24 adult male Sprague-Dawley rats weighing 350 to 370 grams housed in standard conditions with food and water ad libitum. Animals were anesthetized with intraperitonial ketamine (75 mg/kg) and xylazine (10 ml/kg), tracheotomized, and paralyzed with intravenous succinylcholine (2 ml/kg). An endotracheal tube (2 mm inner diameter) was inserted and tightly tied to avoid air leaks, and rats were ventilated using a Servo 300 ventilator (Siemens, Solna, Sweden). Vt was set and measured using the ventilator's pneumotachograph. Airway pressure was monitored via a side port in the tracheal tube using a pressure transducer (Valydine MP45, Valydine Engineering, Northridge, CA, USA). The left carotid artery was cannulated and connected to a pressure transducer (Transpac Monitoring Kit; Abbot, Sligo, Ireland) to monitor mean arterial pressure (MAP). The right jugular vein was cannulated for fluid infusion. Blood and airway pressures were routed to an amplifier (Presograph, Gould Godart, Netherlands), converted to digital (Urelab, Barcelona, Spain), and recorded in a personal computer (Anadat-Labdat Software, RTH InfoDat, Montreal, QC, Canada). Then, animals were randomly assigned to one of three experimental groups (n = 8 in each group): (i) Basal group (BAS), unventilated animals, were immediately exanguinated after induction of the anesthesia (ii) Low Vt group (LVt), ventilated with 8 ml/kg and a positive end-expiratory pressure (PEEP) of 0 cmH₂O (ZEEP) for three hours, and (iii) High Vt group (HVt), ventilated with 30 ml/kg and ZEEP for three hours. To maintain normocapnia without decreasing respiratory rate, instrumental dead space was increased in the HVt group. An additional group of spontaneously breathing rats (Spont) was added to discriminate the effect of the surgical procedure and anesthesia on brain activation. The spontaneously breathing animals were anaesthesized and the same surgical procedure as ventilated groups (cannulation and tracheostomy) was performed. No mechanical ventilation was applied in this group. Identical patterns of fluid infusion and anaesthesia were applied in the three groups maintained in protocol during three hours (spont, LVt, HVt).

At baseline, animals in the MV groups underwent volume-controlled ventilation with 8 ml/kg Vt and 2 cmH₂O PEEP. Inspired oxygen fraction (FiO₂) was kept at 0.4 throughout the experiment, and the respiratory rate was adjusted for normocapnia. We measured values of MAP, arterial blood gases, and respiratory system parameters 15 minutes after initiating MV (baseline) and hourly thereafter after randomization. Inspiratory and expiratory pauses were applied to calculate static lung compliance (Crs). Fluid management was identical in all groups (Ringer-lactate, 10 mL/kg/h) to prevent differences that might favour edema formation and vasoactive drugs were not used in any group. At the end of the three-hour period, rats were euthanatized by exsanguination. We centrifuged 7 ml of blood from each animal and stored the plasma at -80°C for protein determinations. Hearts and lungs were removed en bloc, and the right lung was frozen for additional tissue analyses of proteins. Rats' brains were removed from the cranium by careful dissection and immediately frozen and stored at -80°C.

Left lungs were fixed by instillation of 4% buffered formaldehyde into the airway at a pressure of 5 cmH₂O and immersed in the same fixative. Two investigators blinded to experimental groups calculated histological scores after hematoxylin-eosin (HE) staining as described elsewhere [11] and assessed intra-alveolar neutrophil infiltration by counting the number of neutrophils in fifty fields per animal at a magnification of X400 using ImageJ v1.40g (Wayne Rasband, NIH, USA).

Lung damage was determined using a Lung Injury Score (LIS), based on the evaluation of alveolar edema, hemorrahage, neutrophil infiltration and alveolar septal thickening in each animal. Each parameter was scored from 0 to 4. Subsequently, the total LIS was calculated by adding the indicidual score for each parameter, up to a maximum score of 16 [12].

Commercially available enzyme-linked immunosorbent assay (ELISA) kits (Biosource, Camarillo, CA, USA) were used to determine the following plasma/lung protein levels: tumor necrosis factor-alpha (TNF)-α, macrophage inflammatory protein (MIP-2), interleukin (IL)-6, IL-1β, monocyte chemoattractant protein (MCP-1), and IL-10. Analyses of all samples, standards, and controls were run in duplicate following the manufacturer's recommendations.

The brain areas of interest were cut into 20-μm coronal sections (CM1900, Leica Microsystems, Wetzlar, Germany) and stored at -80°C. Sections were processed for single immunohistochemistry using a rabbit polyclonal antibody against c-fos (c-fos (4), Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) diluted 1:250 The immunoreaction was visualized with diaminobenzidine and H₂O₂[13]. Additional sections were stained with cresyl violet to identify the regions of interest: thalamus, retrosplenial cortex (RS), central amygdala (CeA), hippocampus, paraventricular hypothalamic nuclei (PVN), and supraoptic nucleus (SON). After immunostaining, specific activated areas were identified by light microscopy (DM250, Leica, Wetzlar, Germany) with the aid of a stereotaxic atlas [14]. Brain sections were digitized and c-fos-positive cells were evaluated according to the intensity of staining and then semiquantified using Image J software (ImageJ 1.40g, W. Rasband, NIH, USA). An optimal threshold was set for all sections to minimize background signals.

MAP remained stable within the normal range throughout the three-hour period in all groups (Figure 1). Respiratory system compliance (Crs) and plateau pressure (Pplateau) increased with HVt MV, but both remained unchanged throughout the experimental period (Figure 1). Respiratory rates were not significantly different between LVt and HVt animals (mean 47.3 vs 47, respectively; P = 0.7). Significant decreases in PaO₂/FiO₂ and pH and concurrent increases in PaCO₂ were found in LVt animals after three hours of MV (Figure 1). pH in animals spontaneously breathing was slightly higher than in those receiving MV, and remained stable during the experiment.

Figure 2 shows representative images of lungs in each experimental group. Lung neutrophilic infiltration and LIS were significantly higher in MV rats than in unventilated rats, but no differences between LVt and HVt were found.

Neuronal activation evidenced by an increased number of c-fos immunopositive cells was observed in the RS (Figure 3) and thalamus (Figure 4) of HVt rats, but not in LVt or basal rats. c-fos expression was also observed in the CeA (Figure 5), PVN (Figure 6), and SON (data not shown) of MV rats, although activation did not differ between HVt and LVt animals. Similarly, no differences in c-fos activation in other cortical areas or in the hippocampus were observed between the experimental groups (data not shown).

Animals breathing spontaneously showed similar levels of activation in CeA and PVN than those observed in the basal group (Figures 5 and 6). Conversely, the c-fos signal in RS and Thalamus was higher than those found in BAS and LVt groups (Figures 3 and 4).

Figure 7 shows plasma and lung levels of pro- and anti-inflammatory mediators. MV increased plasma levels of IL-6, IL-10, IL-1β, MCP-1, and MIP-2, irrespective of the Vt level (LVt or HVt) (P < 0.05). However, plasma TNFα levels increased significantly after three hours of HVt ventilation (P = 0.005) but remained unaltered in the LVt group.

In the lungs, irrespective of the Vt level, MV increased IL-6, IL-10, IL-1β, and MIP-2 levels (Figure 7). Lung TNFα levels were similar in MV and unventilated animals. We also observed a trend to higher MCP-1 in HVt compared to LVt (Figure 7). Taken all together, the inflammatory response was higher (but also more variable) in the HVt group than in the LVt group.

Animal models of complex diseases are potentially limited by interspecies differences and have no immediate applicability to humans; nevertheless, animal models are accepted as a valid strategy for the initial approach to multifactorial conditions. In this sense the HVt level applied in this work is not clinically relevant, but has been used to simulate those areas exposed to locally high pressures in injured lungs. Nevertheless, HVt application in our model of healthy lungs resulted only in a moderated increase in Pplat, compared with other more aggressive models in the literature [39, 40]. Whereas this approach did preclude functional lung alterations, the short duration of MV also limited the analysis of brain alterations to early events like neuronal activation detected by c-fos immunostaining. Furthermore, our study design does not allow us to conclude whether inflammation and c-fos increased expression are mechanistically linked, let alone the nature of this possible link. In fact, as mentioned, hypercapnia would contribute to brain activation by a different mechanism. Moreover, the tight control of anesthesia and neuromuscular paralysis used in MV groups precluded any differences in this regard. Spontaneously breathing animals served to explore the effect of instrumentation and anesthesia, as they were not paralyzed.

Due to the novelty of this issue (brain activation and MV) and the limitations of the study, we can only speculate about the translation of these results to the clinical setting. The etiology of cognitive impairment in critically ill patients is undoubtedly multifactorial and is the subject of ongoing discussion [2, 3]. Nevertheless, crosstalk between the lung and brain is poorly understood [22], and although many randomized controlled clinical trials have evaluated the efficacy and safety of various methods of MV in ARDS and ALI patients, few studies have explored the influence of MV patterns at the neuronal level. Our findings about regional brain activation during MV could help define particular areas susceptible to be activated by mechanoreceptors in the lung. Those areas might play a crucial role in regulating early events occurring during the application of non-adequate MV patterns. Our findings might have implications for understanding how the brain senses incoming signals or insults from the lungs in anesthetized and paralyzed subjects.