Imaging biomarker correlates with oxidative stress in Parkinson's disease
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/neuroindia.NI_981_15
Source of Support: None, Conflict of Interest: None
Background: While oxidative stress (OS) may be one of the crucial factors determining the initiation and progression of Parkinson's disease (PD), its correlation with gray matter (GM) atrophy is not known.
Keywords: Magnetic resonance imaging, malondialdehyde, oxidative stress, Parkinson's disease, voxel-based morphometry
Parkinson's disease (PD) is a progressive, neurodegenerative disorder, which affects more than 10 million people worldwide. Even though the disease onset varies between young, middle-aged, or elderly patients, it is more common observed in the elderly population. Clinical and pathological hallmarks of PD are characterized by dopaminergic cell loss in the substantia nigra (SN), which in turn leads to decreased level of dopamine in the striatum. Even after several decades of research on PD, the etiology of the disease is not completely understood. Environmental exposure, genetic susceptibility, and aging are some of the factors which play a crucial role in the pathogenesis of PD., Numerous cellular mechanisms such as oxidative damage, neuroinflammation, mitochondrial defect, formation of protein aggregates, and lysosomal dysfunction have been described in the pathogenesis of PD.,
Oxidative stress (OS) is thought to play a pivotal role in both the genetic and sporadic forms of PD. Reactive oxygen species damage biomolecules and result in protein aggregation, nucleic acid damage, and lipid peroxidation. This creates a perilous state within the neuronal cell that leads to a gradual decline in physiological functions., Peroxidation of cellular membrane lipids leading to the production of toxic end products, such as malondialdehyde (MDA), poses a serious threat to membrane integrity resulting in cell death. Postmortem studies on PD brains have reported increased levels of 4-hydroxynonenal, MDA, protein carbonyl changes, and DNA and RNA oxidation within the SN. MDA is known to be mutagenic and carcinogenic in nature., Changes in the plasma MDA level can be used as a potential peripheral biomarker to study OS in many pathological conditions. MDA serves as a principal aldehyde in thiobarbituric acid reactive substances (TBARS) assay.
Voxel-based morphometry (VBM) is an automated, quantitative, operator independent, and unbiased neuroimaging technique. VBM has been used to study various neurodegenerative disorders, movement disorders, epilepsy, multiple sclerosis, and schizophrenia. One of the advantages of this technique is that it allows the study of significant gray matter volume (GMV) changes, in-vivo (local, endemic) at group level. VBM analysis in PD patients has shown widespread GM loss in the frontal, parietal, and temporal cortical areas of the brain as a direct function of disease or as a consequence to it. VBM has been widely used to study the association between GM atrophy with various motor and nonmotor features of PD.
However, none of the studies have attempted to correlate serum oxidative markers with brain volume. The aim of this study is to investigate GM changes using VBM and correlate it with OS marker. We have correlated GMV in 72 patients with PD and 72 healthy controls with serum MDA levels. Our hypothesis is that PD patients may have raised MDA levels in comparison with healthy controls (HC) and MDA levels may inversely correlate with GMV in the patient group. Understanding molecular mechanisms of PD and correlating them with magnetic resonance imaging (MRI) findings are crucial for the better management of PD.
Seventy-two patients with PD fulfilling the UK Parkinson's Disease Society Brain Bank Clinical Diagnostic Criteria and 72 healthy age and gender-matched controls were prospectively recruited for the study over a period of 2 years (2012–2014). All patients were recruited from the Neurology Outpatient Department and Movement Disorder Clinic at a tertiary level neurology hospital. All patients were clinically examined and evaluated by a single movement disorder specialist. The stage of PD was assessed using the modified Hoehn and Yahr (H and Y) scale, and the degree of motor disability was assessed in the “ON” state using unified Parkinson's disease rating scale (UPDRS-III motor score). The motor scores of the right and left sides were calculated using the items 20–26 of the UPDRS-III score of each side. The patients were classified as 'Young onset' if the age of onset was <45 years. Controls with a family history of PD or Parkinsonian syndromes were excluded from the study. Information regarding the history of alcohol consumption, tobacco chewing, smoking, and diseases such as hypertension and diabetes were also obtained. All the participants were recruited after obtaining written informed consent. The study was approved by the institutional ethics committee.
Serum MDA, a measure of lipid peroxidation, was used as a marker of OS. Blood samples were collected into 5 ml vacutainers, without any anticoagulant, from all the participants by venipuncture after consumption of food. Blood samples were centrifuged after an hour at 2500 rpm for 10 minutes for separation of serum samples and stored immediately at −80°C until analysis. MDA levels were measured by a modified spectrophotometric method. In brief, 50 μl of serum sample or MDA working standard solution was added to 50 µl of trichloroacetic acid. Reaction mixture was centrifuged at 4000 rpm for 10 minutes at 4°C. Fifty microliter of supernatant was taken into a 10 ml test tube, and 1 ml of thiobarbituric acid (TBA) was added. Then, the samples were heated for 1 hr in a boiling water bath and cooled. This resulted in the formation of MDA-TBA adduct. Reaction mixture was extracted by the addition of n-butanol. Finally, the butanol phase was separated by centrifugation at 1500 rpm for 10 minutes. Supernatant was read at wavelengths of 540 nm for excitation and 590 nm for emission. Readings were compared with known concentration of MDA standards. All the samples were assayed in duplicates. The MDA levels were expressed in µmol/l.
MRI data were acquired on a 3T MRI scanner (Philips Achieva 3T, Philips Medical Systems, Netherlands) using a 32-channel head coil. MRI scan was performed in the ON-state of the patients to avoid movement artifacts. The same imaging protocol was used for all patients and healthy controls. T1-weighted images were acquired using a magnetization prepared rapid acquisition gradient echo sequence (repetition time = 8.2 ms, echo time = 3.8 ms, flip angle = 8°, field of view = 256 mm × 256 mm × 165 mm, 256 sagittal slices, voxel size = 1 mm × 1 mm × 1 mm). All the images were checked by a neuroradiologist for structural abnormalities before analysis.
Regional GM differences between the patient and control group were assessed using VBM. Data processing was performed with Statistical Parametric Mapping 8 (Wellcome Department of Cognitive Neurology, www.fil.ion.ucl.ac.uk/spm/software/spm8/) in MATLAB (Math Works, R2013a) setup using VBM8 toolbox. Images of two patients were excluded from the analysis due to poor quality. Before processing the data, each image was oriented into the standard anterior commissure-posterior commissure line. The preprocessing steps included spatial normalization into the Montreal Neurological Institute template and segmentation into GM, white matter, and cerebrospinal fluid. GM images were smoothened with an isotropic Gaussian kernel of 8 mm.
A two-sample t-test was performed between the PD group and control group to investigate the GM differences. Differences in the two groups were considered statistically significant if P was < 0.001 (uncorrected). Multiple regression analysis was performed to correlate GMV with MDA levels. The MRI result was presented uncorrected with a minimum cluster size of 100 voxels. Age, gender, and total intracranial volume (TIV) levels were considered covariates for group comparison. Age, gender, TIV, smoking, diabetes, hypertension, and alcoholism were used as covariates in both the groups for correlation analysis. In the patient group, along with these parameters, levodopa equivalent dose and duration of illness were also added as covariates.
The mean age of the patients was 51.4 ± 10.6 (22–71) years and that of controls was 50.6 ± 10.5 (20–73) years. The mean duration of illness was 5.1 ± 10.6 years in the patient group. The mean H and Y score in the patient group was 1.71 + 0.62, with 27 patients in stage 1 (37.5%), 38 patients in stage 2 (52.8%), and 7 patients in stage 3 (9.7%). The mean age of onset of symptoms in the PD group was 45.2 ± 11.4 years, with late-onset PD occurring in 72.2% (n = 52) and YOPD being 27.8% (n = 20) in the study group. There was a significant difference in the right and left UPDRS scores of patients with PD with right being more than left [7.9 ± 4.4 vs. 6.3 ± 5] (P < 0.0001) [Table 1]. Among the 72 patients, 44 (61.1%) patients reported that the tremor started on the right side, 26 (36.1%) patients reported it on the left side, and 2 (2.8%) patients reported bilateral tremors.
In PD patients, serum MDA levels were significantly higher than in the controls. The mean MDA level was significantly higher in patients compared to controls (0.56 ± 0.1 µmol/l vs. 0.427 ± 0.055 µmol/l, P < 0.0001). To confirm repeatability and consistency of assay results, intraassay coefficient of variability (CV) and interassay CV was calculated (intraassay CV: 4% and intraassay CV: 5.7%).
Magnetic resonance imaging
There was no significant difference in total GMV between patients and controls (553.8 ± 55.6 vs. 562.25 ± 47.92, P = 0.18). However, regional GM analysis revealed focal decreases in volume in the patients involving all the major brain lobes including frontal, parietal, occipital, and temporal. In addition to this, GM atrophy was also observed in the anterior cingulate gyrus, parahippocampal gyrus, cingulate gyrus, and the cerebellar culmen and declive (P = 0.0001, uncorrected) [Table 2] and [Figure 1]. No significant correlations were found between GMV and the UPDRS-III score, H and Y score, or sidedness in the patient group.
In both the patient and control groups, there was no significant correlation between MDA level and age or GMV. Our study did not find any correlation between MDA levels and UPDRS-III score, type of PD, and H and Y staging in the patient group.
Multiple regression analysis was performed between GMV and MDA levels in both groups. Age, gender, TIV, diabetes, hypertension, smoking, and alcoholism were considered to be covariates since these factors are known to affect the OS status. Only in the patient group, MDA levels showed a positive correlation with the volume of CN and the PC ( P< 0.000006, number of voxels = 2241) [Table 3].. There was no significant negative correlation between the MDA levels and VBM findings. In the control group, the MDA levels did not correlate significantly, either positively or negatively, with GMV of any brain region.
OS has been known to play a crucial role in nigrostriatal degeneration in many neurodegenerative disorders such as PD. Brain is the second organ next to adipose tissue with the highest content of lipids. Polyunsaturated fatty acids in the brain are more susceptible to oxidative damage. Various markers of lipid peroxidation such as MDA, cholesterol, lipid hydroperoxide, isofurans, and acrolein are also found to be significantly elevated in PD. MDA is a marker of lipid peroxidation and OS. It was significantly increased in our study, which supports the previous findings., A case-control study on 80 sporadic PD patients investigated by a thiobarbituric acid reactive substances (TBARS) assay showed increased lipid peroxidation level when compared to 29 healthy controls. This was confirmed by more studies in the Indian population.,, High performance liquid chromatography (HPLC) analysis performed by Chen et al., among 211 PD patients and 135 controls also confirmed elevated levels of MDA in blood plasma. Increased level of MDA suggests chronic OS in the brain. Even though MDA is a peripheral marker of OS, it does not have any diagnostic significance because many neurodegenerative disorders have elevated levels of MDA.
While the pathology of PD is typically linked with the SN, the degeneration affects the entire brain. In our study, GM atrophy was noted in bilateral medial frontal gyrus, right superior temporal gyrus, right cingulate gyrus, bilateral fusiform gyrus, right parahippocampal gyrus, left lentiform nucleus, cerebellar culmen and declive, right cuneus, left anterior cingulate, and right precentral gyrus in the patient group compared to the control group. Many studies have shown an associated atrophy of these regions with executive dysfunction, impaired social cognition, dementia, apathy, memory impairment, and psychosis in PD patients.,,
It is established that SN is more prone to oxidative damage in PD. However, the oxidative status of non-SN regions is not adequately studied. Interestingly, in our study, there was significant positive correlation between OS and non-SN regions such as the CN and PC. CN being part of striatum is the site for several pharmacologic interventions in PD. It is known to be associated with cognition, working memory, and multimodal information processing., 3D volume analysis using MRI documented decreased CN volume in the early stages of PD patients when compared to 15 age and gender-matched controls. Apostolova et al., have also confirmed CN atrophy in PD patients with dementia. Interestingly, some studies have reported increased GMV in CN of PD patients after practicing meditation for a long time.
Postmortem studies of human PD brains have shown relatively lower level of OS and increased antioxidant levels in various non-SN regions such as CN, putamen, and frontal cortex. In CN, increased protein oxidation and reduced lipid peroxidation have been reported. This study also documented increased level of total glutathione and astrocytic proliferation in these non-SN areas. Another study from the same group demonstrated the effect of ageing on neurodegeneration. According to this study, oxidative damage, decreased levels of antioxidants, and mitochondrial function in SN with physiological aging make it more vulnerable to neurodegeneration in PD when compared to the non-SN regions. Increase in carbonyl level, which is commonly used as a marker of oxidative protein injury has been observed in postmortem PD studies in CN, SN, putamen, and globus pallidus. The cause of the variability among these findings is not known. The probable explanation could be oxidative status varying with duration and stages of PD. Only a few autopsy studies have documented changes in the antioxidant level in various regions of PD patients because in-vivo estimations are difficult. The protective role played by CN in our study could indicate increased levels of antioxidants in non-SN regions. In-vivo analyses of such markers could help to prognosticate disease and predict drug response. In the present study, MDA level did not show significant negative correlation with any of the brain areas and, hence, OS may not be the only mechanism involved in widespread atrophy in PD.
First, usage of peripheral serum biomarkers for correlation analysis in structural imaging has not been validated widely. More such studies would help in bridging the gap between imaging findings and clinical research. Second, though we have attempted to regress majority of factors, there exists a statistical concern because direct application of these findings to patient care will require randomized control studies. The use of region of interest-based analysis of SN, red nucleus, and other basal ganglia may provide complimentary information to VBM in addressing direct correlation. Finally, the uncorrected threshold was used for VBM analysis, which may include false positives; however, the significant cluster size obtained in our study reduces the possibility of chance findings.
We observed an increased level of MDA in patients with PD compared to matched controls which positively correlated with the volume of CN and PC, suggesting that, even though the whole brain is affected in PD, some of the non-SN regions of the brain may have differential compensatory mechanisms. This could preserve them from oxidative damage.
Financial support and sponsorship
This work was financially supported by Department of Biotechnology (DBT), Government of India. [Grant Number: NO. BT/PR14315/MED/30/474/2010].
Conflicts of interest
There are no conflicts of interest.
[Table 1], [Table 2], [Table 3]