Background
Respiratory dysfunction is among the leading causes of death in Parkinson’s disease (PD) [
1]. The spectrum of respiratory dysfunction associated with PD includes the obstructive, restrictive, and mixed types of respiratory defects, decreased respiratory muscle strength and upper airway obstruction [
2]. However, pulmonary functional impairment usually goes unnoticed until the advanced stages of the disease [
3]. Therefore, the early identification and prevention of respiratory dysfunction are clinically important. Relatedly, there is accumulating evidence showing that the effects of PD on respiration occur via central and peripheral mechanisms. For example, one recent study suggested the involvement of the brain stem in the development of abnormal ventilatory control and sleep-related breathing dysfunction in PD [
4], while another study found that cervical arthrosis and motor disturbances lead to upper airway obstruction and respiratory muscle impairment [
3]. Despite these findings regarding respiratory disturbances, the exact pathophysiologic and anatomical brain alterations underlying respiratory dysfunction in PD have remained unclear.
The involvement of systemic oxidative stress and apoptosis in the degradation of dopaminergic neurons is emphasized in PD patients [
5]. These afflicted neurons cause the deafferentiation of striatal dopamine, which eventually leads, in turn, to degeneration of the extrapyramidal system [
6]. Furthermore, systemic oxidative stress has also been found to be associated with impaired pulmonary function and various lung diseases such as central obstructive pulmonary disease (COPD), asthma, and obstructive sleep apnea (OSA) [
7‐
9]. The inflammatory markers previously found to be related to respiratory dysfunction include thiobarbituric acid reactive substances (TBARS), thiol [
9], endothelial progenitor cells (EPCs) [
8], soluble platelet-selectin (sP-selectin), soluble endothelial-selectin (sE-selectin), soluble intercellular (sICAM-1) and soluble vascular cell adhesion molecules (sVCAM-1) [
10]. Taken together, these findings suggested a potential direct or indirect role of systemic oxidative stress in the pathogenesis of respiratory dysfunction in PD. However, the impact of systemic oxidative stress on PD patients with and without respiratory dysfunction is not yet fully elucidated.
Recent studies have demonstrated brain structural damage in certain pulmonary diseases, such as COPD and OSA [
11,
12]. Meanwhile, a rat PD model revealed that brain structural deficits are related to dysfunction in respiratory rhythm generation [
13]. That said, while the presence of regional gray matter changes associated with the autonomic dysfunction in PD has been well demonstrated [
14], there has been little exploration of brain structural changes associated with respiratory dysfunction in PD. It is clear that the central nervous system plays a role in ventilatory regulation and the coordination of upper airway musculature [
15]. A functional neuroimaging study revealed the involvement of cortical, limbic, and paralimbic (cortico-limbic) brain regions and the cerebellum in respiratory control and perception [
15]. Brain structural damage in this respiratory-related cortico-limbic circuit has been found to be associated with air hunger [
16] and impaired respiratory response to hypercapnia [
17]. Additionally, some authors have suggested that autonomic failure might contribute to respiratory dysfunction in PD [
3], while recent research has established the concept of a central autonomic network [
18]. Gray matter atrophy involving part of this network has been associated with autonomic dysfunction in PD [
14]. On the basis of these results and concepts, it appears that respiratory dysfunction in PD may be linked to the impairment of the cortico-limbic respiratory circuit and central autonomic network. To date, however, the relationship between respiratory dysfunction and the cortico-limbic respiratory circuit and central autonomic network has not been well demonstrated.
In this study, we aimed to evaluate the relationship between brain structural alterations, as determined by voxel-based morphometry (VBM), and respiratory dysfunction in patients with PD and to correlate them with systemic inflammatory markers. We hypothesized that PD patients with respiratory dysfunction would exhibit increased oxidative stress and regional gray matter atrophy which might be associated with impaired pulmonary function.
Discussion
Using VBM analysis, this study demonstrated the presence of brain structural alteration in PD patients with respiratory dysfunction. Consistent with our hypothesis, the APF group showed smaller GMVs in several brain regions than the NPF group. Furthermore, GMV loss in these brain regions was specifically correlated with poor pulmonary parameters as evaluated by pulmonary function tests. This is the first study to establish an association between respiratory dysfunction and regional cortical deficits in PD.
According to the Braak staging hypothesis [
26], the damage to the brainstem occurs earliest in PD. The early involvement of the medullary respiratory center might account for the autonomic dysfunction and respiratory dysfunction in the early stage of the disease [
4,
27]. However, the various phenotypes of respiratory dysfunction in PD cannot be attributed solely to the involvement of the brainstem. Interestingly, our present study showed that the APF group had smaller GMVs in some parts of the central autonomic network, including the cerebellum crus, parahippocampal gyrus, and fusiform gyrus. It has previously been recognized that the autonomic nervous system plays a pivotal role in the regulation of lung ventilation, gas exchange, and airway smooth muscle [
28]. Moreover, growing evidence suggests that the central autonomic network participates in regulatory control of sympathetic, parasympathetic, and respiratory motor neurons [
29]. In the proposed concept of the central autonomic network [
18], sympathetic regulation is primarily associated with the cingulate gyrus, cerebellum, insula, and supplementary motor area, whereas parasympathetic regulation is associated with the hippocampal formation, amygdala, insula, and cerebellum. The association of regional cortical deficits with autonomic dysfunction [
14] supported our hypothesis that interference with the central autonomic network in PD might contribute to the development of respiratory dysfunction.
Our results also showed that the smaller GMVs in several brain regions of the APF group were positively correlated with pulmonary functional parameters, particularly FEV1 and MEF50. FEV1 and MEF50 are markers used to assess airway obstruction and detect early small airway diseases such as COPD and asthma in clinical practice [
30]. These alterations might be explained by the complex sympathetic and vagal innervation of the lung. The autonomic nerve supply is one of the main mechanisms controlling the contraction of the airway smooth muscle [
28]. The vagus causes bronchoconstriction and increased airway resistance, while the sympathetic nerves are commonly thought to cause the relaxation of bronchial caliber and diminished airflow resistance [
28]. Although the mechanism remains unclear, it seems that gray matter deficits in PD would interfere with the autonomic innervation of the airways, leading to increased airflow resistance and airway obstruction.
Furthermore, compared to the NPF group, the APF group had smaller GMVs in the cerebellum, parahippocampal gyrus, and fusiform gyrus, regions which are considered to be principal components of respiratory-related cortico-limbic circuitry. In the proposed model for respiratory sensorimotor neural circuitry [
15], the motor division comprises the motor cortex, supplementary motor area, basal ganglia, cerebellum, and brainstem, while the sensory division includes the insula, amygdala, sensory cortex, and cerebellum. The motor division is responsible for volitional breathing via cortico-spinal circuitry, whereas the sensory division mediates the integration of processing of dyspnea stimuli. The neuronal activation within the parahippocampal gyrus, fusiform gyrus, and parietal cortex during CO2-stimulated breathing demonstrate the regulatory capacity of supra-brainstem structures in sensory and motor respiratory responses to hypercapnia [
17]. Our results are highly consistent with the physiological findings of the previous studies.
The cerebellum is classically associated with the regulation of motor coordination. However, the roles of cerebellar structures in respiratory patterns [
16] and autonomic dysfunction [
14] are also well recognized. The involvement of the cerebellum in the experience of hypercapnia and air hunger has been established [
16]. Moreover, the parahippocampal gyrus and fusiform gyrus, traditionally thought to be involved in memory and cognitive functions, also participate in normal respiratory responses to chemosensory stimuli [
17]. Thus, cortical deficits in the cerebellum, fusiform gyrus, and parahippocampus may explain abnormal ventilatory responses to CO2 in PD [
27]. These findings further support the assumption that GMV loss might play contributory roles in the development of respiratory dysfunction in PD.
Elevated systemic oxidative stress has been demonstrated to be the main etiology of PD [
5]. The imbalance between oxidative stress and antioxidative capacity results in elevated neuroinflammation, the death of dopaminergic neurons, and a cascade of cell repairing processes. In recent studies, alterations to systemic oxidative stress have also been found to have profound associations with various lung diseases [
7‐
9] and impaired pulmonary functional parameters [
9,
10]. Although previous studies have yielded relevant associations, in this study, the APF group showed no significant differences in systemic oxidative stress with the NPF group, presumably because of the small sample of recruited patients. However, while the results of the study cannot clarify the influences of oxidative stress on respiratory dysfunction in PD, they do allow us to draw a direct association between brain structural alterations and pulmonary function impairments with less concern over potential confounding factors such as oxidative stress and clinical severity.
Limitations
While significant results were found with respect to brain structural alterations, several limitations of our study should be acknowledged. First, our results may have been influenced by the lack of a healthy control group and the relatively small sample size, which may explain the insignificant differences for GMV among the APF subgroups. Second, the complex mechanism underlying the oxidant/antioxidant balance of PD patients may be affected by various factors, including individual genetic variations and physical exercise, and these confounding factors were not comprehensively addressed in our study. Finally, the final sample of 25 PD patients in the APF group did not allow division of the group into subgroups in order to further examine relevant variables. Although no significant GMV differences were found among the different types of respiratory dysfunction in this study, we cannot exclude the possibility that brain structural alterations associated with specific types of respiratory dysfunction could be detected with a larger sample. Future longitudinal studies utilizing larger sample sizes and including healthy controls could help to better characterize the different phenotypes and clarify the causal relationship underlying respiratory dysfunction in PD.