Ischemic stroke is the second leading cause of death worldwide and will affect at least one-sixth of the population [
], with significant morbidity and mortality [
]. Stroke patients require intensive care unit admission for close monitoring [
]. Survivors experience long-term disability and significantly impaired quality of life [
Ischemic stroke leads to cerebral cell death, causing brain inflammation and neurological deficits [
]. In addition to local immunoinflammatory responses in the brain, stroke also triggers systemic responses, which may have important distal organ effects.
Notably, the lungs are particularly vulnerable in the event of severe brain damage, including ischemic and hemorrhagic strokes [
]. Bai et al. [
] reported that 15.6% of stroke patients developed acute lung injury within 36 h of hospital admission, and 7.8% of patients had pneumonia or bronchitis during hospitalization. Studies in experimental models of stroke have focused on stroke-induced immunosuppression and lung infection [
], whereas lung injury and inflammation, and their effects on the phagocytic capability of alveolar macrophages, have been overlooked. This is in contrast to the literature on traumatic brain injury, where lung inflammation has been well characterized in a variety of animal models [
We hypothesized that stroke might induce acute systemic and lung inflammation affecting respiratory parameters, pulmonary histological features, and alveolar macrophage phagocytic behavior. In this context, we investigated the impact of focal ischemic stroke on the respiratory pattern, lung histology, behavior of innate immune cells that reside in the lung, and activation of the innate immune system in the brain, lung, and circulatory compartments. Additionally, complementary in-vitro studies were conducted to elucidate the possible mechanisms underlying changes in the behavior of professional phagocytes, alveolar macrophages, following stimulation with bronchoalveolar lavage fluid (BALF) or serum isolated from naïve (unstimulated) or stroke rats. The expression of proinflammatory mediators was also evaluated in alveolar macrophages and epithelial and endothelial cells isolated from Sham and Stroke animals.
In the rat model of focal ischemic stroke used herein, V
, and V
increased while RR and T
decreased during spontaneous breathing, but lung mechanics and gas exchange did not differ during mechanical ventilation. Lungs from stroke rats showed evidence of increased diffuse alveolar damage mainly due to augmented edema and inflammation as shown by increases in total protein levels in BALF. Ultrastructural changes were observed in lung parenchyma, with damage to type 2 pneumocytes and endothelial cells, increased number of macrophages, as well as enlarged basement membrane thickness. Expression of TNF-α and IL-6 in the brain, the TNF-α level in plasma and BALF, as well as the IL-6 level in plasma alone were increased. Additionally, macrophages and endothelial cells, but not epithelial cells, from Stroke animals exhibited increased IL-6 gene expression. In parallel, the phagocytic capability of alveolar macrophages was decreased. Remarkably, all of these changes occurred within 24 h after the induction of cerebral ischemia, which is in agreement with the rapid onset of ALI in stroke patients [
]. In-vitro experiments revealed that the phagocytic capability of alveolar macrophages was reduced and the expression of TNF-α and IL-6 was increased in alveolar macrophages isolated from naïve rats only after exposure to serum from rats that had experienced ischemic stroke.
This is the first experimental study to investigate lung function and histology, systemic inflammation, and the phagocytic capability of alveolar macrophages in an experimental model of focal ischemic stroke. We hypothesized that brain damage would induce systemic inflammation, stimulating alveolar macrophages and triggering a phenotype shift, thus reducing phagocytic capability (Additional file
: Figure S9). From our data, it seems that specific factors contained in the circulatory compartment following an acute stroke may in turn modulate the innate immune system in the lung. Identification of the factor(s) responsible for these effects (cytokines, alarmins, or other inflammatory mediators) warrants further investigation.
We chose a model of focal, not global, ischemia because focal ischemic stroke has a higher incidence, accounting for approximately 80% of all strokes worldwide [
]. In keeping with previous studies, stroke was induced through thermocoagulation of pial blood vessels over the primary sensorimotor cortices, leading to sensorimotor dysfunction 24 h postoperatively [
]. The fact that peak systolic velocity in the carotid ipsilateral to the lesion decreased after stroke suggests that the regional blood flow in the right cerebral hemisphere was reduced. Following a focal ischemic insult, reperfusion is known to follow a biphasic pattern, with a transient increase (post ischemic hyperperfusion) [
] followed by a more sustained hypoperfusion [
]. As we measured blood flow 24 h after ischemic stroke, our findings likely correspond to the hypoperfusion phase.
Alterations in respiratory pattern following ischemic stroke are common both clinically [
] and experimentally [
], and have been attributed to autonomic dysfunction [
]. In mice, the coefficient of variation for V
and RR increased after an ischemic insult, while the mean values of these variables decreased, leading to reduced minute ventilation [
]. The fact that Stroke animals in our study exhibited increased V
and decreased RR might be due not only to interspecies differences, but also to the severity and location of brain injury as well as central or neurogenic hyperventilation.
Our observation that ischemic strokes are associated with a significant increase in DAD score and lung ultrastructural changes might be explained by higher levels of TNF-α and IL-6 in the plasma and lungs, and of TNF-α in BALF. These data are consistent with experimental [
] and clinical [
] reports of increased systemic inflammation after stroke. The fact that total protein levels in BALF were higher in Stroke animals likely reflects increased permeability of the pulmonary capillary membrane due to inflammation [
]. In turn, elevated inflammatory marker levels in the lungs could have resulted from decompartmentalization of the inflammatory response in the brain, where overexpression of cytokines was detected. Stroke has indeed been shown to disrupt the blood–brain barrier (BBB) [
]. Alternatively, lung inflammation may have been a result of focal ischemia-induced parasympathetic nervous system impairment, leading to loss of the protective cholinergic anti-inflammatory pathway [
]. We cannot rule out that higher V
contributed, at least partly, to increased lung damage and inflammation in Stroke animals. Increased V
has been associated with development of lung injury even in the absence of a first hit [
Despite the presence of alveolar edema and bronchoconstriction, these alterations were not sufficient to impair blood gases and lung mechanics. Bronchoconstriction might be explained by increased airway narrowing due to circulating proinflammatory cytokines [
]. In patients with brain injury, bronchoconstriction with increased airway resistance is common [
]. Despite histologic evidence of edema and bronchoconstriction, lung mechanics did not differ between Stroke and Sham animals. One possible explanation is that functional changes are observed only after a certain threshold of pulmonary damage has been exceeded [
]. Additionally, animals were sedated, anesthetized, paralyzed, and mechanically ventilated with PEEP = 3 cmH
O, which prevented development of possible lung mechanical changes. Interestingly, in patients with severe brain damage, respiratory mechanics and arterial blood gases differed at ZEEP but not at PEEP = 8 cmH
], which is consistent with our results.
Although we are unable to claim increased lung infection susceptibility based on our data, it is possible to speculate that the reduction of alveolar macrophage phagocytic capability that we observed could increase the risk of pneumonia after stroke in the clinical setting. Pneumonia is a common complication of stroke, affecting up to 22% of patients after stroke, and is known to worsen clinical and neurological outcomes [
]. The negative results of two randomized trials of prophylactic antibiotics in stroke [
] indicate that the mechanisms leading to lower respiratory tract infections after stroke need to be elucidated [
]. The pathophysiological processes that result in immunosuppression and gut translocation of bacteria after ischemic stroke have been studied in mice, and are partially dependent on sympathetic nervous system activation [
]. Catecholamine release has also been implicated in neurogenic pulmonary edema after severe brain injury, although it has become evident that other mechanisms might contribute to lung dysfunction, including systemic release of proinflammatory mediators, alarmins, and extracellular vesicles, which may induce lung inflammation and injury [
]. The cholinergic anti-inflammatory pathway also seems to be involved in brain–lung crosstalk [
], but the mechanisms remain unclear.
Possible clinical implications
Our results suggest that focal ischemic stroke may lead to brain–lung crosstalk resulting in increased inflammation in pulmonary tissue. As many stroke patients ultimately require mechanical ventilation, this inflammation might serve as a first hit, and particular attention should be paid to avoid ventilator-induced lung injury.
The reduced alveolar macrophage function associated with focal ischemic stroke can expose patients to a higher risk of pneumonia. In patients who need mechanical ventilation, the risk of ventilator-associated pneumonia might be increased, and prophylactic measures should be considered judiciously.
Importantly, the lack of gas exchange and lung mechanics impairment in focal ischemic stroke should not be interpreted as evidence for an absence of brain–lung crosstalk.
This study has several limitations. First, anesthesia was achieved with ketamine, which is a known bronchodilator and might have masked bronchoconstriction during measurements of lung mechanics. Second, we used a model of focal ischemia, and cannot rule out the possibility that results would differ in other models of brain damage. Third, it must be kept in mind that our model does not reproduce the more complex clinical scenario, and our findings cannot be directly extrapolated to human patients. Fourth, the breathing pattern might be influenced by animal species and size [
], as well as the location and intensity of brain damage [
]; however, we cannot rule out the influence of these factors on lung injury or macrophage function. Fifth, controversial results have been reported concerning the effects of ketamine on intracranial pressure [
], but ketamine is still used, even in the clinical setting, due to its protective effects on hemodynamic responses and lung function [
]. Finally, mechanisms of immunosuppression, which were previously evaluated in experimental models of stroke [
], were not analyzed in our study, but are a future line of investigation in the laboratory.