Brain mediators of systemic oxidative stress on perceptual impairments in Parkinson’s disease

Participants

The study was approved by the Local Ethics Committee on Human Research of Kaohsiung Chang Gung Memorial Hospital in Taiwan. All participants or their guardians provided written informed consent prior to participation in the study. Twenty-seven right-handed PD patients (11 men and 16 women, mean age: 54.6 ± 9.3 years) with no previous history of neurological or psychiatric illnesses, psychotropic medication, or contraindication to magnetic resonance imaging (MRI) were prospectively enrolled at the Neurology Department of Kaohsiung Chang Gung Memorial Hospital. Patients were included if they had been diagnosed with idiopathic PD by an experienced neurologist according to the Parkinson’s Disease Society’s criteria for idiopathic Parkinson’s disease [19]. Of the 28 PD patients, ten had never used any anti-Parkinson’s medication, whereas the others used dopaminergic medication (levodopa and dopamine agonists). The disease severity and functional status of each patient were evaluated with the Unified Parkinson’s Disease Rating Scale (UPDRS) [20], the modified Hoehn and Yahr stages (HY-stage) [21], and the Schwab and England activities of daily living scale (SE-ADL) [22] in the “OFF” state, which refers to when levodopa appears to become less effective in eliminating motor symptoms and the patients are unable to function properly. The patients’ mean disease duration, defined as the time since the given patient subjectively noticed his or her first symptoms, was 4.1 ± 3.7 years. For comparison, 25 sex- and age-matched healthy subjects (11 men and 14 women, mean age: 50.9 ± 10.5 years) with no medical history of neurologic diseases or psychiatric illnesses, alcohol/substance abuse, or head injury, and with similar levels of education, were recruited.

Laboratory measurements for oxidative stress factors

We evaluated the oxidative stress condition in all subjects by measuring the serum concentration of TBARS and thiol. All subjects received blood sampling on the same day as the MRI study and neuro-psychological testing. Sera were isolated from peripheral blood samples drawn from each subject before and after the examination.

The serum TBARS were measured based on a well-established method for detecting lipid peroxidation [23]. The ability of anti-oxidative defense in response to increased oxidative damage was evaluated by measuring the serum level of total reduced thiols since serum thiols are physiologic free radical scavengers [24].

Neuropsychological assessment of perceptual function

A clinical psychologist performed the neuro-psychological (NP) battery of tests from the Wechsler Adult Intelligence Scale-III (WAIS-III) [25]. The WAIS-III, with 14 subtests, allows us to summarize index scores and convert to different large ability areas reliably, including three IQs (Verbal IQ, Performance IQ, and Full Scale IQ) and four Indexes (Verbal Comprehension, Working Memory, Perceptual Organization, and Processing Speed).

In the current study, we focused on perceptual organization and processing speed functions in PD. These batteries comprise three tests for perceptual organization function and one test for processing speed function. The tests for perception organization function include picture completion, block design, and matrix reasoning tests. We sum the results from these tests to come up with an index score which represents the aggregate abilities of perceptual organization function.

The test for processing speed was a digit symbol-coding test which examines several cognitive processes of participants, such as detecting a test digit, searching for the corresponding symbol, and maintaining the symbol representation. We then further added scaled scores of perceptual organization and processing speed functions and converted them into the performance IQ.

MR image acquisition

Volumetric structural MRI scans were acquired on a GE Signa 3T whole-body MRI scanner (General Electric Healthcare, Milwaukee, WI, USA) using an eight-channel phase array head coil at the Kaohsiung Chang Gung Memorial Hospital in Taiwan. Whole brain three-dimensional T1 weighted images were collected for each participant using an inversion-recovery fluid-attenuated fast spoiled gradient-recalled echo pulse sequence with the following imaging parameters: repetition time (TR)/echo time (TE)/inversion time (TI) = 9.5/3.9/450 ms; flip angle = 15 degrees; number of excitations (NEX) = 1; field of view (FOV) = 240 × 240 mm²; matrix size = 512 × 512; voxel size = 0.47 × 0.47 × 1.3 mm³; and slice number = 110 axial slices (without interslice gaps). In order to identify any brain abnormalities, an additional volumetric axial T2-weighted fast spin-echo sequence (TR/TE = 4200/102 ms; echo train length = 18; NEX = 2; FOV = 240 mm²; slice thickness = 5 mm; matrix size = 320 × 256 and 18 slices) and axial T2-weighted inversion-recovery fluid-attenuated sequence (TR/TE/TI = 8000/100/2000 ms; NEX = 1; FOV = 240 mm²; slice thickness = 5 mm; matrix size = 320*256 and 18 slices) were used in the same imaging session.

T1-weighted structural MRI scans were analyzed with voxel-based morphometry (VBM) [26] using the VBM8 toolbox (http://dbm.neuro.uni-jena.de), which was implemented using the Statistical Parametric Mapping software program (SPM8, Wellcome Institute of Neurology, University College London, UK, http://www.fil.ion.ucl.ac.uk/spm/) with default settings. All the image processing procedures, including tissue segmentation, study-specific template construction, and spatial normalization, were conducted according to a previous study [27] and are summarized in the Additional file 1 for the present study.

Statistical analysis

Mediation analysis

A single-level three-variable mediation model [28], illustrated in Fig. 1, was used to investigate the causal relationships between oxidative stress, regional GM volume, and perceptual function. Mediation analysis tests whether the direct effect of an independent variable on a dependent variable can be explained by the indirect influence of the mediating variable. The primary hypothesis of this analysis asks whether the effect of the oxidative stress level (independent variable) on the perceptual function (dependent variable) was explained indirectly by the regional gray matter volume changes (mediator) with significant group main effect. In order to examine this indirect effect, the Mediation toolbox (http://wagerlab.colorado.edu/tools) with an accelerated bias-corrected bootstrap test of statistical significance was used (10,000 bootstrap samples). The path model jointly tested three effects of interest that are required if a regional GM volume links oxidative stress with the perceptual function: (a) the effect of the independent variable (TBARS level) on the mediator (regional gray matter volume) (indirect effect, path a); (b) the effect of the mediator on the dependent variable (perceptual performances) by controlling the effect for the TBARS (indirect effect, path b); and (c) the mediation effect a × b which is defined as the reduction of the relationship between the independent and dependent variables (TBARS—perceptual performance) (total relationship, path c) by including the mediator into the model (direct path, path c′). We set the statistical significance threshold at 0.05 for all the relevant paths [29].

Demographic data, perceptual scores, and serum concentrations of TBARS and thiol of PD patients and healthy controls

The mean serum concentration of TBARS was significantly higher in the PD patients than in the controls (P < 0.001). The mean serum concentration of free thiol was lower in PD patients than in the controls, but the difference failed to achieve statistical significance (P = 0.191). Since only the serum TBARS level revealed a significant group difference, the following correlation and mediation analyses were primarily focused on and carried out with regard to the TBARS levels.

Gray matter volume reduction in patients with Parkinson’s disease

Compared to healthy controls, PD patients showed a significant GM volume reduction in the following brain regions: (1) bilateral temporal gyri (middle temporal gyrus and fusiform gyrus), (2) bilateral frontal gyri, (3) limbic lobe (hippocampus and uncus), (4) left inferior parietal lobule, (5) right caudate nucleus, and (6) insula. We did not find increased GM volume in PD patients compared with the healthy controls (Additional file 2: Table S2; Fig. 2).

Correlations between serum TBARS and perceptual functions

Brain mediators of oxidative stress on perceptual functions

TBARS-digit symbol relationship mediator

Single-level three-variable mediation analysis revealed atrophy of middle temporal gyrus that negatively mediated the TBARS-digit symbol relationship. The middle temporal gyrus volume reduction was negatively associated with TBARS levels, but predicted impaired digit symbol test performance (Coef_a = −0.007, P_a < 0.001; Coef_b = 15.55, P_b = 0.003; Coef_ab = −0.10, P_ab = 0.005). This finding suggests that the serum level of TBARS plays a role in contributing negative impact on digit symbol test scores through the middle temporal gyrus. Figure 3A shows path diagrams and scatterplots for path analyses in middle temporal gyrus. We also found that volume reductions in the inferior parietal lobule (Coef_ab = −0.10, P_ab = 0.042), hippocampus (Coef_ab = −0.06, P_ab = 0.042), and insula (Coef_ab = −0.12, P_ab = 0.025) can negatively mediate the TBARS-digit symbol relationship (Table 3).

Clinical variable	Anatomy name	Path a			Path b			a × b				Path c’
		pcoef	z	p	pcoef	z	p	pcoef	z	p	pcoef	z	p
Digit symbol	MTG	−0.007	−3.97	<0.001	15.55	2.94	0.003	−0.10	−2.82	0.005	−0.11	−2.12	0.034
	IPL	−0.009	−3.30	<0.001	11.29	2.28	0.023	−0.10	−2.04	0.042	−0.11	−2.20	0.028
	HIPP	−0.004	−2.49	0.013	18.25	−1.58	0.030	−0.06	−2.03	0.042	−0.15	−2.25	0.025
	INS	−0.005	−2.84	0.005	24.73	2.38	0.017	−0.12	−2.24	0.025	−0.09	−2.17	0.030
Block design	MTG	−0.007	−3.40	<0.001	19.21	4.02	<0.001	−0.13	−2.91	0.004	−0.09	−2.67	0.008
	HIPP	−0.004	−2.46	0.014	16.85	2.53	0.011	−0.06	−2.03	0.042	−0.15	−2.55	0.011
	INS	−0.005	2.86	0.004	29.32	3.03	0.002	−0.14	−2.47	0.014	−0.07	−2.56	0.011
POI	MTG	−0.007	−3.64	<0.001	92.59	2.84	0.005	−0.63	−2.74	0.006	−0.71	−2.72	0.007
	HIPP	−0.004	−2.44	0.015	99.80	2.42	0.015	−0.34	−2.15	0.031	−1.01	−2.71	0.007
	INS	−0.005	−2.99	0.003	142.70	2.88	0.004	−0.70	−2.39	0.017	−0.65	−2.75	0.006
PIQ	MTG	−0.007	−3.59	<0.001	82.38	2.60	0.009	−0.56	−2.26	0.024	−0.68	−2.62	−0.009
	INS	−0.005	−2.96	0.003	139.79	2.83	0.005	−0.68	−2.43	0.015	−0.57	−2.56	0.010

The corresponding statistical results of each path and mediation effect between the oxidative stress (independent variable), perceptual functions (dependent variable), and anatomical regions with significant disease effect (mediators) are described in terms of corresponding path coefficient, z value, and p value. The anatomical regions were considered as potential brain mediators between oxidative stress level and perceptual functions if statistical significance, as indicated by a p of less than 0.05, was found in each of the following three effects in corresponding mediation analysis. (1) Indirect Path a: the relationship between oxidative stress level and the regional gray matter volume; (2) indirect Path b: the relationship between the regional gray matter volume and perceptual functions as determined by controlling the oxidative stress level and (3) Mediation a × b effect. All mediation analyses were conducted using the Mediation toolbox

P _coef path coefficient, HIPP hippocampus, INS insula, IPL inferior parietal lobule, MTG middle temporal gyrus, PIQ performance intelligence quotient; POI, perceptual organization index

Summary

Consistent with our hypothesis and in line with the extant literature, patients with PD experienced higher serum oxidative stress and worse perceptual functions. Furthermore, we identified gray matter atrophy throughout much of the cortex, including anatomical locations that typically appear to be particularly sensitive to the effects of inflammation, such as the temporal lobe, hippocampus, and insula [30]. We further demonstrated, for the first time, that gray matter volume atrophy in the temporal lobe, hippocampus, insula, and parietal lobe mediate the relationships between TBARS levels and general perceptual function scores. This suggests that higher oxidative stress levels are associated with poor perceptual functions by means of smaller temporal and mesolimbic volumes and that those regions support the retention of visual memories, the processing of sensory inputs, the comprehension of language, the storing of new memories and emotions in a manner that may differ from the way in which the prefrontal cortex supports executive functioning in PD.

Pathophysiology of elevated oxidative stress in PD

Studies have demonstrated that an imbalance in pro-oxidant/antioxidant homeostasis can enhance the generation of toxic reactive oxygen species [31] and may be related to disease severity in PD [32]. The TBARS level is one of the markers of oxidative stress, and has been found to be elevated not only in the substantia nigra but also in the CSF and plasma in PD [33–35]. In addition, many cytokines, as well as TBARS, are also markers of reactive oxygen species (ROS) activity, which can worsen central neuron cell injury through the elevation of reactivated microglia in the substantia nigra [36]. Cytokines produced in the brain can freely diffuse past the blood–brain barrier (BBB), a phenomenon which has been demonstrated to induce memory impairments [37]. Our study supports the previous hypothesis that oxidative stress is implicated in the pathogenesis of PD.

In addition to the nigrostriatal region, inflammation activated microglial cells in PD were also observed in various brain regions, such as the hippocampus, temporal lobe, and other regions of the cerebral cortex [30]. Actually, increased oxidative stress/inflammation has also been found to be associated with temporal lobe damage in aging [38] due to atherosclerosis and Alzheimer’s disease [39], with the pathological accumulation of amyloid and neurofibrillary tangles. Two leading factors might contribute this temporal lobe vulnerability: relatively high receptor and messenger RNA expression for proinflammatory cytokines [40] in those regions and higher links to portions of the salience network, such as the insula, which has a role in regulating the immune system and can enhance the peripheral inflammatory conditions [41] and subsequently affect the neural organization of semantic memory [42]. From animal studies, it has been demonstrated that injections of inflammatory cytokines into the hippocampus and the overexpression of IL-1 messenger RNA in the hippocampus selectively deteriorate spatial and contextual memory processes [43]. In addition, elevated peripheral inflammation can stimulate IL-1 expression within the central nervous system, including the medial temporal lobe, and generates increased cytokine expression with subsequent impacts on temporal lobe-dependent memory [44]. The vagus nerve, which serves as a neurally mediated immune-brain pathway, can also enhance hippocampal activity by peripheral inflammatory challenge and electrical stimulation [45, 46]. Taken together, these data demonstrate the sensitivity of human medial temporal lobe structures to systemic oxidative stress/inflammation and provide mechanistic insights relevant to the broader literature linking severe or chronic inflammation to the attrition of human memory.

Gray matter atrophy serves as a mediator of oxidative stress on perceptual functions

In a previous cortical thickness analysis of PD with mild cognitive impairment, patient groups revealed widespread cortical thinning compared with controls, including thinning of the right inferior temporal gyrus, left superior parietal cortex, precuneus, lateral occipital, temporal gyri anterior cingulate, and superior frontal gyri [1]. The same study also found that memory and visuospatial performance, the core feature of perceptual function, were associated with temporoparietal and superior frontal gyri thinning. Our results replicated these previous findings and also consistently demonstrated the interaction between perceptual functional impairment and volume atrophy in the middle temporal gyrus, inferior parietal lobe, hippocampus, and insula. Although fewer cortical regions were found to be involved in this study, the relatively small sample size and younger patient group might have led to these slightly different results. Furthermore, different gray matter quantitative morphometrical evaluation methods which are sensitive to different properties of gray matter tissue (i.e., volume vs. cortical thickness) might also have led to the different outcomes. We do believe, however, that the results of the current study, together with those of the previous investigation, indicate that particular cortical atrophy can serve as a marker for perceptual function decline in early PD.

The test of mediation amounts to a test of whether controlling for each brain mediator explains a significant amount of the covariance between the independent variable and the dependent variable. In the initial analyses, we found that higher serum TBARS levels were associated with worse perceptual functions. We further examined whether the relationship between serum TBARS levels and perceptual functions was explained indirectly by greater damage to perceptual functional specific anatomies in patients with PD using mediation analyses. We found that the middle temporal lobe and insula, as well as the hippocampus, play the central roles in mediating most of the TBARS-perceptual function relationships. Interestingly, the caudate nucleus and frontal lobe volume reduction, which is a hallmark of PD, failed to mediate the association between oxidative stress and perceptual function. Actually, perceptual impairment in PD occurs not only on visuomotor tasks requiring a complex motor response such as drawing but also on visuoperceptual tasks that require a limited motor response. Furthermore, this perceptual impairment is not related to a decrement in intellectual abilities and can be clearly distinguished from dementia [4]. In the present study, though the results could not exclude the influence of TBARS on other aspects of systemic status, they suggest that TBARS may play a particular role in the perceptual deficits observed in PD. Another explanation might be the ceiling effect, which suggests that inflammation occurs earlier in the striatum and then is followed by inflammation in the temporal lobe and hippocampus, resulting in the sequential clinical presentations progressing from movement disorders to cognitive impairments. These results, however, should be interpreted with caution.

The interpretation of the findings here must be tempered by some of the limitations of the present study. We do not know the temporal relationships between peripheral inflammation, symptomatology, and anatomical integrity. Further task design or longitudinal evaluation with manually controlled systemic oxidative stress levels in animal models should be carried out in the future. In addition, the measurement of only a few biomarkers of oxidative damage cannot be considered a valid tool for exploring the multifaceted, complex oxidant/antioxidant imbalance in PD. Furthermore, the oxidant/antioxidant balance of PD patients may be influenced by a multitude of parameters. The interactions between the CNS and systemic oxidative stress and the effect of individual genetic variations and physical exercise could also affect either the systemic oxidative stress or the brain volume, and these interactions were not well evaluated in the present study. While systemic oxidative stress is also likely to have widespread effects on the brain tissue, the specific anatomical volumes that mediate a given cognitive function are likely to differ depending on the examinations being assessed [47]. In addition, recent research suggests that network-based rather than regional anatomic involvements contribute to the neurodegeneration and behavioral differences seen in PD, such that assessing structural or functional connectivity between brain regions associated with these phenotypes would clarify the complex neural network involved in oxidative stress–cognition relationships.