Background
Parkinson’s disease (PD) is a neurodegenerative disorder characterized by significant loss of dopaminergic neurons in the substantia nigra. Recently, there has been a large body of research searching for biomarkers useful in the early stages and for the differential diagnosis of PD and many studies are underway to identify disease-specific biomarkers in cerebrospinal fluid (CSF), peripheral plasma and other bodily fluids. Growing evidence has suggested that inflammation plays an important role in the pathologic features and symptoms of PD [
1,
2] and inflammatory biomarkers could be potential early indicators of neurodegenerative disease.
Abundant inflammation has been found in several animal models of dopaminergic neuron degeneration and in PD patients, resulting in the impaired regulation of several cytokines and chemokines secreted by activated microglia, which precedes the progression of PD [
3]. Inflammatory factors in blood are more readily accessible than those in CSF and the passage of some of these factors through the blood–brain barrier allows them to be measured in the periphery [
4]. Moreover, some peripheral inflammatory markers can cross the blood–brain barrier [
5]; therefore, peripheral cytokines and chemokines may be useful biomarkers for the detection of PD progression.
The inflammatory cytokine and chemokine levels in the plasma and serum of PD patients vary significantly. Several studies have shown that PD pathophysiology is associated with increased inflammatory status [
6]. However, discordant results in the literature and a lack of information regarding the stability of inflammatory factors in different cohorts have hampered progress. To fill this void, we measured 10 promising cytokines and chemokines in a discovery cohort, validated them in four validation sets and further verified them in an exploratory set of patients with idiopathic REM sleep behaviour disorder (iRBD). Moreover, we evaluated the potential of longitudinal factors as prognostic biomarkers for monitoring PD progression. We hypothesized that the combination of certain biomarkers would help with the early diagnosis of PD and tracking its progression.
Methods
Study population
For the discovery cohort, 76 PD patients and 76 age- and sex-matched healthy controls (HCs) were recruited from the Department of Neurology, Ruijin Hospital Affiliated with Shanghai Jiaotong University School of Medicine from 2017 to 2019. The diagnosis of PD was performed according to the International Parkinson and Movement Disorder Society (MDS) diagnostic criteria by at least two neurologists skilled in movement disorders [
7]. For the validation set, 80 PD patients and 80 matched controls from four medical centres, including Tianjin Huanhu Hospital, the Affiliated Hospital of Qingdao University, the Affiliated Hospital of Jining Medical University and the First Affiliated Hospital of Zhengzhou University, were enrolled from 2018 to 2019. All of the participants enrolled form five centers were Han Chinese people without family history, which reduced the ethnical genetic variation between groups. In addition, 39 PD patients in the early stage from the discovery cohort were chosen for follow-up assessments. This study was approved by the ethics committee of Ruijin Hospital, Shanghai Jiaotong University School of Medicine. Subjects with any peripheral inflammatory diseases or taking anti-inflammatory medicines were excluded from the cohort.
As an exploratory set, 45 iRBD patients with video-polysomnography confirmation were subsequently recruited based on the consensus criteria of the International RBD Study Group [
8]. All iRBD patients were examined by neurologists to exclude motor signs of parkinsonism or secondary causes. Thirty-five controls without neurological disorders were also enrolled. The ages of all participants ranged from 50 to 80 years, and patients taking anti-inflammatory medicines were excluded. All participants provided written informed consent to the research protocol, which was approved by the local medical ethical committee and conformed to the Helsinki declaration.
Demographics and clinical assessment of PD patients
For the PD patients, the disease stage was determined using the modified Hoehn and Yahr (H-Y) staging scale (range 0–5) [
9]. The motor subscale of the Unified Parkinson's Disease Rating Scale (UPDRS) was used to evaluate motor symptoms in the ON medication state. The Non-Motor Symptom Questionnaire (NMSQ), Scale for Outcomes in PD-Autonomic (SCOPA-AUT), Sniffin’ Sticks 16-item test (SS-16), 17-item Hamilton Depression Rating Scale (HAMD-17), Mini-Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA) and Sleep Behavior Disorder Screening Questionnaire (RBDSQ) were also completed.
Sampling and biomarker selection
Whole blood samples (5 mL) were collected in EDTA tubes and centrifuged at 3000 rpm for 15 min at room temperature. Plasma samples were frozen within 2 h of collection and stored at − 80 °C until assayed. All the blood from five centres was obtained simultaneously with the clinical information accessed. Ten inflammatory cytokines and chemokines (IL-8, TNF-α, TGF-β, IL-10, IL-6, CXCL12, CX3CL1, CCL15, CCL3 and CCL20) were included based on a review of the existing literature [
10,
11] and their levels were measured using a Meso Scale Discovery (MSD) assay. Selection was based on the following criteria: (1) inflammatory chemokines or cytokines had a plausible association with PD-related phenotypes as previously reported [
11,
12] or based on our previous work [
13]; (2) all of the measurements were within the limit of detection in a pilot project using MSD U-PLEX Human Biomarkers kits.
MSD assay
The inflammatory markers from each of the participating centres were all measured using a human multiple-chemokine panel kit (MSD, Rockville, MD, USA) in accordance with the manufacturer’s instructions. For each marker, the concentrations were calculated with reference to a standard curve derived using various concentrations of the standards assayed in the same manner as the cell culture supernatants. The lower limit of detection (LLOD) was calculated as a signal concentration that was 2.5 SD greater than the zero calibrator. The upper limit of detection was calculated as a signal concentration that was 2.5 SD below the upper plateau of the standard curve. To avoid potential cross-interactions among different antibodies, not all ten markers were measured simultaneously. CX3CL1, IL-10, IL-6, IL-8, CCL3, CCL20 and TNF-α were measured simultaneously in one MSD assay, while TGF-β, CXCL12 and CCL15 were measured individually in three separate assays according to the manufacturer's guidance. All plasma samples were assayed in duplicate, and the results were averaged for analysis.
Statistical analysis
The data were analysed by SPSS software version 20.0 (SPSS Inc., Chicago, IL, USA). Owing to the non-normal distribution of the ten markers and clinical characteristics, Mann–Whitney U test was performed to analyse the difference between HC and PD participants. Statistical significance was established as a two-sided p value < 0.05, and multiple testing was controlled by Bonferroni correction (p < 0.05/10 biomarkers). Conditional logistic regression analysis was used to quantify the association of inflammatory biomarkers with PD adjusted for age, sex and sites, and the effect-size estimates were expressed as odds ratio (OR) and 95% confidence intervals (CIs). In addition, the association of inflammatory biomarkers with sex and age (≤ 60 or > 60) was separately analysed. Inflammatory biomarker relationships were assessed by Pearson correlation analyses. A receiver-operating characteristic (ROC) curve was created to evaluate the role of inflammatory biomarkers in the diagnosis of PD. For the longitudinal analysis, the increase or decrease of cytokines were defined by the changes of the value between baseline and follow-up, the “increase” group indicates difference above 0, while “decrease” means less or equal to 0. To assess associations between longitudinal chemokine levels and progression outcomes, Cox proportional hazards regression models were used to estimate hazard ratio (HR) with 95% CIs adjusted for age and sex.
Discussion
In the present study, we found a strong correlation between a panel of immune-cytokines and chemokine in the peripheral blood and the risk of PD. Furthermore, in multivariate analyses, increased levels of CXCL12, CX3CL1 and IL-8 and a decreased level of CCL15 were all independent diagnostic biomarkers of PD. In addition, the levels of CX3CL1 and IL-8 showed a higher distribution in the exploratory set of iRBD patients. Furthermore, in a prospective follow-up of early-stage PD patients, we found that decreased CX3CL1 levels were associated with motor progression after a mean interval of 43 months.
Several studies have shown that inflammatory biomarkers, including cytokines and chemokines, could be used to help diagnose PD [
11,
17,
18]. The levels of proinflammatory cytokines, mainly TNF-α, IL-1β, IL-6, IL-10 and IL-8, and growth factors (EGF and TGF-β1) have been demonstrated by many investigators to be markedly altered in parkinsonian patients when compared with control subjects. Chemokines, such as CXCL12, CCL2, CCL5, CX3CL1 and CCL5, play well-established roles in the immune system, and several recent reports have suggested that chemokines and their receptors may also play a role in the central nervous system (CNS) [
19,
20]. Although a number of previous studies have examined the distribution differences of inflammatory biomarkers between PD and HC subjects, no study has focused on the association between these candidate markers and PD susceptibility in a multicentre cohort. Comparing the previous studies, we identified in an exploration cohort out of a bigger panel of cytokines/chemokine, which are validated in a validation cohort using MSD assay. Therefore, the present study increases the strength of the evidence.
CX3CL1 is one of the most important mediators of the communication between neurons and microglia, and its emerging role in PD has been increasingly recognized. Studies show that CX3CL1 plays a neuroprotective role in 6-OHDA-induced dopaminergic lesion, thus it might be an effective therapeutic target for PD [
21]. In a PD mouse model, decreased levels of CX3CL1 in the brains of A53T mice were observed, but elevated levels were observed in the peripheral blood [
22]. Moreover, Jing Zhang’s team measured the levels of CX3CL1 in the CSF, and no difference was found between PD and control subjects; however, CSF CX3CL1/Aβ
1-42 was found to be positively correlated with PD severity and progression [
23]. In our current cohort, the levels of CX3CL1 were also markedly elevated in PD in independent groups. However, decreased levels of CX3CL1 were associated with the motor progression of PD patients. We supposed this may be due to the fact that CX3CL1 levels are increased in the explorative cohort as a reflection of an anti-inflammatory protective effect, thus it is reasonable that there is an elevated trend in the early stages of PD and a decrease as the disease progresses. This is similar to the distribution in AD studies, as higher CX3CL1 plasma levels were observed in patients with mild to moderate AD than in patients with severe AD [
24]. Larger studies are needed to investigate the biomarker potential of CX3CL1 for the diagnosis and prognosis of PD.
The relationship between CXCL12 and PD was first reported by Mika Shimoji based on the data from the postmortem brains of PD patients and control individuals, and it was suggested that CXCL12 may participate in the aetiology of PD [
25]. In our previous study, CXCL12 was involved in α-synuclein-triggered neuroinflammation and could be a novel target for the prevention of α-synuclein-triggered ongoing microglial responses [
13]. In a small sample, Vahid Bagheri’s team demonstrated that CXCL12 serum levels were significantly elevated in patients with PD compared to controls (30 PD patients and 40 controls) [
26]. In our current study, our findings also support the hypothesis that CXCL12 is a candidate biomarker for the diagnosis of PD. Patients in the discovery and validation sets with high levels of CXCL12 were more than 9 and 2.4 times more likely to experience PD, respectively, than those with low CXCL12 levels. The biological mechanisms underlying the role of CXCL12 are not fully understood, and further functional characterization is encouraged.
The trend of IL-8 levels in the peripheral blood of PD patients has not been confirmed. Reale et al. found that the level of IL-8 in PD patients was significantly higher than that in HCs [
27]; however, Vineeta’s team found that the level of IL-8 was significantly reduced in the serum of patients with PD [
28]. Extending the results of previous studies, our findings demonstrated a strong correlation between the levels of IL-8 and PD. Moreover, regarding CCL15, Hochstrasser et al. showed that in monocytes, CCL15 levels were significantly lower in AD patients than in healthy subjects, but they were increased in AD patient plasma [
29]. However, CCL15 plasma levels have been shown to be reduced in AD patients as compared to controls [
30]. In our current study, reduced CCL15 levels were strongly associated with PD, which first showed evidence of the value of CCL15 in diagnosing PD.
Dopaminergic therapy influences plasma protein levels; however, according to a recent study that analysed > 1000 plasma proteins in ~ 500 individuals, CXCL12, CX3CL1, CCL15 and IL-8 are not affected by dopaminergic medications [
31]. Moreover, no correlations were found between LEDD and levels of these four biomarkers as correlation analysis shown, which strength their predict value.
In summary, we used a combined statistical analysis of CXCL12, CX3CL1 and IL-8 levels to differentiate PD patients from HCs with high sensitivity and specificity. Considering inconsistent results have been reported for individual cytokines and between studies in the literature, our results provide evidence of a heightened proinflammatory cytokine profile in PD patients, strengthening the clinical evidence that patients with PD have an increased inflammatory response.
Our study is not without limitations. On the one hand, our findings need to be replicated in a larger population to further confirm the efficacy of CXCL12, CX3CL1, IL-8 and CCL15 in diagnosing PD. On the other hand, patients enrolled in the present study were not classified by different disease stages or phenotypes, and future studies with large samples specifically designed to examine the association of the four inflammatory biomarkers with different phenotypes of PD are warranted.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.