Brain Atrophy Mediates the Relationship between Misfolded Proteins Deposition and Cognitive Impairment in Parkinson’s Disease

Parkinson’s disease is associated with cognitive decline, misfolded protein deposition and brain atrophy. We herein hypothesized that structural abnormalities may be mediators between plasma misfolded proteins and cognitive functions. Neuropsychological assessments including five domains (attention, executive, speech and language, memory and visuospatial functions), ultra-sensitive immunomagnetic reduction-based immunoassay (IMR) measured misfolded protein levels (phosphorylated-Tau, Amyloidβ-42 and 40, α-synuclein and neurofilament light chain) and auto-segmented brain volumetry using FreeSurfur were performed for 54 Parkinson’s disease (PD) patients and 37 normal participants. Our results revealed that PD patients have higher plasma misfolded protein levels. Phosphorylated-Tau (p-Tau) and Amyloidβ-42 (Aβ-42) were correlated with atrophy of bilateral cerebellum, right caudate nucleus, and right accumbens area (RAA). In mediation analysis, RAA atrophy completely mediated the relationship between p-Tau and digit symbol coding (DSC). RAA and bilateral cerebellar cortex atrophy partially mediated the Aβ-42 and executive function (DSC and abstract thinking) relationship. Our study concluded that, in PD, p-Tau deposition adversely impacts DSC by causing RAA atrophy. Aβ-42 deposition adversely impacts executive functions by causing RAA and bilateral cerebellum atrophy.


Introduction
Parkinson's disease (PD) is currently the second most common neurodegenerative disease [1], lacking both efficient methods of early diagnosis and a mechanism-based cure [2]. In addition to motor deficits in PD patients, cognitive decline is increasingly emphasized and investigated. Individuals with PD have increased risk of dementia compared to the general population [3]. Such cognitive impairments in PD patients have significant impacts on quality of life and provide daily burdens for caregivers [4]. Detailed neuropsychological assessments are lengthy and cost time, while also being heavily dependent on the adherence of patients. The lack of a long-term follow-up regarding cognitive function, early medical or psychological interventions remains challenging and potentially affects the prognosis. Previous studies have demonstrated that plasma misfolded proteins are correlated with cognitive decline and are potential biomarkers for the detection and prediction of cognitive impairment [5].
Compelling evidence suggests that misfolded protein aggregation plays an important role in cellular dysfunction and tissue damage, leading to development of PD [6]. Less invasive biomarkers moving from cerebrospinal fluid (CSF) to blood are a noted clinical trend and may act as a potential large-scale screening tool in longitudinal follow-up. Among them, α-synuclein (α-syn), total and p-Tau, Aβ-40 and 42, and neurofilament light chain (NfL) are most commonly investigated and may contribute to brain atrophy and cognitive impairment [7].
In this study, we aim to clarify the associations between misfolded protein deposition, brain atrophy, and cognitive impairment. Our hypotheses are as follows: (1) PD patients have higher neurotoxic plasma misfolded proteins than normal participants; (2) misfolded proteins are correlated with brain atrophy and cognitive decline; and (3) brain atrophy mediates the relationship between misfolded protein deposition and cognitive decline.

Subjects
In our study, 54 PD patients and 37 sex-and age-matched healthy volunteers were enrolled in the Neurology Department of Chang Gung Memorial Hospital. The diagnosis of PD was clinically established by a qualified neurologist in accordance with the UK Parkinson's disease society brain bank clinical diagnostic criteria [8]. Healthy volunteers without signs and symptoms of neurological disease, alcohol or substance abuse, psychiatric illness, nor head injury were enrolled as the control group.
Age, sex, and education levels were obtained for all participants. Disease duration, disease severity, and functional assessments were obtained for PD patients. Disease severity were assessed by the modified Hoehn and Yahr staging scale (HY-stage) [9] and Unified Parkinson's Disease Rating Scale (UPDRS) [10]. Three areas were assessed in our study including: (1) nonmotor experiences of daily living (13 items); (2) motor experiences of daily living (13 items); and (3) motor examination (18 items). Functional assessment was performed with the Schwab and England ADL Scale [11].

Image Acquisition
Each participant's head was immobilized to diminish motion artifacts. MR images were acquired using a 3.0 T whole body GE Signa MRI system (General Electric Healthcare). A 3D-FSPGR sequence (repetition time (TR) = 9.492 ms, echo time (TE) = 3.888 ms, flip angle = 20 • , field of view (FOV) = 24 cm × 24 cm, matrix size = 512 × 512, 110 continuous slices with a slice thickness of 1.3 mm and in-plane spatial resolution of 0.47 mm × 0.47 mm) was used for acquiring T1-weighted structured images through the whole head parallel to the anterior-posterior commissure.

Image Data Processing
Auto-segmentation of ROIs (region of interests) was carried out using FreeSurfer image analysis suite and each brain area was divided by estimated total intracranial volumes (eICVs). FreeSurfer is a powerful tool providing extensive and automated analysis of ROIs in human brains, freely available for download online (http://surfer.nmr.mgh. harvard.edu/, v7.1.0, accessed on 11 May 2020). It is based on a subject-independent probabilistic brain atlas in combination with nonlinear image registration of individual images to obtain subject-specific measurements. FreeSurfer morphometric procedures show good test-retest reliability across scanner manufacturers and across field strengths. The technical details of these procedures are described in prior publications [17].

Demographic Data
IBM SPSS Statistics version 20 (IBM, Armonk, NY, USA) was utilized for statistical analysis. All demographic data were compared between the normal controls (NCs) and the PD group. Sex was compared with Pearson's Chi Square test. Age and education levels were analyzed with 2-sample independent t-test. MMSE and neuropsychological assessments were analyzed using ANCOVA (Analysis of Covariance) with age, sex, and education covariates. Plasma biomarkers and ROI volumes were compared with ANCOVA after controlling for age and sex. All results are reported as the mean ± standard deviation (SD). For each ROI's volume, the left and right brain area volumes were analyzed, respectively, after correction via dividing by each intracranial volume. The Benjamini-Hochberg adjusted p value (Q value) < 0.05 was considered significant.

Partial Correlation
The variables with significant differences between the NC and PD groups were enrolled into the analysis of partial correlation, which was used to identify the relationship between plasma biomarkers, brain volumes, and cognitive functions. Plasma biomarkers included α-syn, Aβ-40 and 42, and NfL. Brain ROIs included bilateral cerebellum, thalamus, caudate, putamen, hippocampus, amygdala, nucleus accumbens, and brainstem. Neuropsychological assessments included digit symbol coding, arithmetic, abstract thinking, information, semantic fluency, picture completion, block design and drawing. Bonferroni adjusted p < 0.05 was considered significant. The correlation coefficient was also reported.

Mediation Analysis
In this mediation analysis, we aimed to investigate whether the effect of the plasma misfolded proteins (independent variable) on the neuropsychological assessments (dependent variable) was explained indirectly by the brain volume atrophy (mediator) with a significant group main effect. An indirect dialogue PROCESS v3.5 (https://www. processmacro.org/download.html, accessed on 2 May 2020) was downloaded into IBM SPSS Statistics v.20 for bootstrap test (statistical significance with 5000 bootstrap samples was used).
The path model examined three effects of interest required if brain atrophy links misfolded proteins with neuropsychological assessments: (a) The effect of plasma misfolded protein level on brain volume (indirect effect, path a); (b) The effect of brain volume on neuropsychological assessments by controlling the effect of plasma misfolded proteins (indirect effect, path b); (c) The mediation effects a × b are defined as the reduction in the relationship between the plasma misfolded proteins and the neuropsychological assessments (total relationship, path c) by including the brain volume in the model (direct path, path c).
Statistical significance threshold was set at 0.05 for all the relevant paths.

Basic Characteristics
The baseline clinical demographics, neuropsychological assessments, plasma misfolded protein levels, and brain volumes of all participants are listed in Table 1. Of the 54 PD patients and 37 normal controls, statistical analysis of clinical characteristics showed no significant difference between age and sex (age: p = 0.093, sex: p = 0.199). The mean of disease duration of the PD group was 4.97 years. The NC group performed better in MMSE than PD patients (p < 0.001).

Neuropsychological Assessments
WAIS-III and CASI neuropsychological assessments were performed in our study ( Table 2) and were applied to the five cognitive domains. The NC group performed better in four of the domains (execution, memory, speech and language, visuospatial). In the executive functions, DSC (p < 0.001), arithmetic (p = 0.008), and abstract thinking (p < 0.001) were significant. In the memory functions, only information (p = 0.012) was significant. In speech and language, comprehension (p = 0.001) and semantic fluency (p = 0.025) were significant. In visuospatial functions, picture completion (p = 0.001), block design (p < 0.001), and drawing (p = 0.037) were significant.

Plasma Biomarkers
Four plasma biomarkers were noted as significantly higher in the PD group, including p-Tau (p = 0.0012), Aβ-42 (p = 0.0004), α-syn (p = 0.0390), and NfL (p = 0.0016). The lower Aβ-40 level in the PD group (p = 0.0520) indicates that less neuroprotective proteins are present in PD patients ( Figure 1). The ROC curve analysis of plasma misfolded proteins was performed and attached in the supplementary file ( Figure S1).

Brain Volume Analysis
Our results showed significant brain atrophy of PD patients. The ROI analyses revealed significant atrophy of bilateral cerebellum (L: p < 0.001; R: p < 0.

. Misfolded Proteins and ROI Volumes in PD Patients
The correlation between plasma biomarkers and ROI volumes after controlling age, sex, and education level were examined (Table 3)

Neuropsychological Assessments and Brain Volumes in PD Patients
We examined the correlation between MMSE score, neuropsychological assessments, and ROI volumes after controlling for age, sex, and education (Table 3)

Brain Mediators of Plasma Misfolded Proteins on Cognitive Functions
• p-Tau-digit symbol coding relationship mediator Single-level three-variable mediation analysis revealed atrophy of RAA negatively and completely mediated the p-Tau-DSC relationship (Coef a = −0.318, P a = 0.002; Coef b = 0.329, P b = 0.002; Coef ab = −0.193, P ab = 0.07) (Figure 2A). The results indicate that, with regard to the effects of RAA atrophy on DSC, the effects from p-Tau to DSC disappeared, revealing that it is RAA that causes the poor performance of DSC. A bootstrap test revealed the significance (effect = −0.836, CI: −0.1669; −0.027) of the path analysis (Table 4).

•
Aβ42-digit symbol coding relationship mediator Table 4. Potential brain mediators of plasma misfolded proteins on the cognitive functions. The corresponding statistical results of each path and mediation effect between the misfolded proteins (independent variable), cognitive functions (dependent variable), and brain areas (mediators) are described in terms of corresponding path coefficient, z value, and p value. The RAA volume was negatively associated with the Aβ-42 level and predicted impaired DSC. Specifically, RAA partially mediated the Aβ42-DSC relationship (Coef a = −0.361, P a < 0.001; Coef b = 0.329, P b = 0.002; Coef ab = −0.243, P ab = 0.023) ( Figure 2B). This finding suggests that plasma Aβ-42 contributes to the negative impact on DSC via RAA atrophy. The bootstrap test achieved significance (effect = −0.3649, CI: −0.7552; −0.1123).

Discussion
Consistent with our hypothesis, PD patients experienced brain atrophy (bilateral cerebellar cortex, caudate nuclei, putamen, hippocampus, amygdala, RAA, and brainstem) as well as elevated misfolded proteins (p-Tau, Aβ-42, α-syn, NfL). The patients achieved lower scores in the neuropsychological assessments, including information, comprehension, arithmetic, semantic fluency, picture completion, block design, DSC, abstract thinking, and drawing. Furthermore, we determined that elevated plasma misfolded proteins (p-Tau and Aβ-42) are correlated with lower scores on neuropsychological assessments (MMSE, arithmetic, DSC and ABS) and brain atrophy (RAA, cerebellum and caudate nuclei), indicating the effects of misfolded proteins on neurotoxicity, neuron damage, and brain volume loss. Further mediation analysis revealed that RAA completely mediates the p-Tau-DSC relationship. Meanwhile, RAA and the bilateral cerebellum partially mediates the Aβ42-DSC and Aβ42-ABS relationships.

Pathophysiology of Misfolded Protein Deposition and Cortical Atrophy in PD Patients
The concentration of the plasma misfolded protein level is similarly and correlated to CSF level in Tau protein [18], Aβ-40 and Aβ-42 [19] in neurodegenerative disease. Plasma levels of misfolded proteins can be used as biomarkers to predictive cognitive impairment in PD [5]. Our results showed significantly higher plasma p-Tau in PD patients, which correlated with atrophy of the bilateral cerebellar cortex, right caudate nucleus, and RAA. In a previous study, Tau proteins became insoluble neurofibrillary tangles, deposited in several brain areas, including the frontal cortex [20], temporal cortex, striatum [21], accumbens area [22], putamen, caudate, and the dentate nucleus of cerebellum [21], with a notable association with cognitive decline and neuronal damage [23].
In addition, misfolded amyloid-β proteins gradually turn into insoluble neurotoxic amyloid plaques [24], causing synapse dysfunction, and promote tau phosphorylation into neurofibrillary tangles [25]. Pathologic studies have also revealed extensive Aβ deposits in the frontal lobe, leptomeningeal vessels, and accumbens area [26]. In this study, elevated Aβ-42 is correlated with atrophy of the bilateral cerebellar cortex, bilateral caudate nucleus, and RAA, indicating that Aβ-42 may have been deposited in those brain areas, causing atrophy. On the other hand, a negative correlation with Aβ-42 and Aβ-40 was observed in our study, which may be interpreted as increased deposition of Aβ-42 reducing the level of neuroprotective Aβ-40 [27].

Brain Atrophy Involved in Several Neuronal Circuits in Cognitively Impaired PD Patients
Cortical thinning and gray matter atrophy are associated with cognitive decline in PDD patients [28]. The caudate, putamen, globus pallidum, accumbens area, and thalamus are related to the frontal cortex through basal ganglia-thalamocortical circuits [29]. The nucleus accumbens serves as a motivation-to-movement interface by integrating information from the limbic drive to prefrontal cortex motor planning, based on glutamatergic afferents from the ventral hippocampus, basolateral amygdala, and the medial prefrontal cortex (PFC), and projects to the ventral pallidum. These brain regions are dedicated to executive functions [30]. Nucleus accumbens lesions cause disruptions in cognitive functions and affected processes [31]. Our results demonstrate that atrophy of the nucleus accumbens contributes to executive function impairment, including to DSC and ABS.
Another significant atrophic region is the cerebellum, which correlated with impaired executive functions (DSC and ABS). The cerebellum mediates cortical information processing via closed cortico-cerebellar loops, with strong connections to the lateral prefrontal cortex (PFC) [32]. As the PFC is connected with many different brain areas, PFC-cerebellar circuits are likely to be involved in a variety of processing activities, including cognitive domains such as memory or spatial perception [33].

Gray Matter Atrophy Contributions in Misfolded Proteins and Cognitive Impairment Correlation: Mediation Analysis
Previous studies have reported on correlations between elevated misfolded proteins and cognitive decline in PD patients. Cortical and subcortical atrophy has also been observed to be correlated with cognition impairments [34]. To our knowledge, this is the first study to investigate the mediation relationship between plasma misfolded proteins, brain structure abnormalities, and cognitive dysfunctions. Summarized and noted in Figure 3, misfolded protein depositions in RAA and bilateral cerebellum cortex cause underlying vulnerable brain gray matter atrophy, leading to subsequent circuit damage and executive dysfunction in PD patients. Figure 3. Misfolded deposition contributes to gray matter atrophy and cognitive dysfunction. We hypothesized that misfolded protein deposition in specific ROIs causes mediation effects. Plasma misfolded protein deposition is indicated and has a positive correlation with misfolded protein level in the brain. Deposition of p-Tau in RAA causes brain gray matter atrophy and is correlated with DSC and Aβ-42 in RAA and the bilateral cerebellum, mediated by executive function including ABS and DSC. This function includes the striato-thalamocortical circuit and cerebellum (cortico-basal ganglia-cerebellar circuit). Interconnections between cognitive regions of the basal ganglia, the cerebellum, and the cerebral cortex could provide part of the foundation for the executive control network. Misfolded protein accumulation in brain tissues with neuron toxicity leads to neuron death and causes gross brain structural changes such as gray matter atrophy, which are in turn correlated with clinical disease progression and cognitive function impairments. 1 RAA: right accumbens area, 2 Aβ-42: Beta-Amyloid-42.
Our results revealed that p-Tau and Aβ-42 are the primary proteins significant in the mediation of relationships. The mediation analysis revealed RAA atrophy completely mediates the higher plasma p-Tau and impaired DSC relationship. Neurofibrillary tangles deposit in the RAA and along the involved neuronal circuits, resulting in transportation dysfunction, neuronal death with volume loss, and pathway dysfunction [23]. In this study, we found elevated Aβ-42 correlated with atrophy in the RAA and executive functions (DSC and ABS) impairment. The possible clinical consequences of nucleus accumbens atrophy were recently suggested to include both neuropsychiatric and motor symptoms of PD due to the nucleus accumbens functioning as a "limbic-motor interface" and significantly decreasing in the volume in PD MCI patient [35]. Nucleus accumbens serves as an integrator of the mesolimbic dopaminergic circuit [36] and the atrophy of human nucleus accumbens in PD, called Mavridis atrophy (MA), which begins in early stage PD and is correlated with psychiatric symptoms [37]. Furthermore, reduced volumes of the nucleus accumbens are significantly associated with the performances of the attention, working memory and language domains in PD MCI patients [38].
In previous pathologic studies, Aβ-42 was also found in the accumbens area in PD patients and to be more aggravated in PDD patients [39]. The mediation analysis showed partial mediation of RAA atrophy in the Aβ-42-executive functions relationship. These data may indicate that elevated Aβ-42 as well as RAA atrophy together cause executive function impairments. The possible pathophysiology may involve deposition of Aβ-42 in RAA, which has a highly neurotoxic effect, resulting in neuronal loss and atrophy.
We further identified elevated Aβ-42 correlated with atrophy of the bilateral cerebellar cortex and executive functions (DSC and ABS) impairment. These data may indicate the deposition of Aβ-42 in the bilateral cerebellum having a highly neurotoxic effect, resulting in volume atrophy. The cerebellum is involved in movement and sensorimotor processing, cognitive control and executive function [40,41]. The interconnected circuit, namely the cortico-basal ganglia-cerebellar network [42], with involvement of prefrontal loops, thereby becomes dysfunctional. This pathway is associated with cognitive functions, connecting the prefrontal cortex with the dentate nucleus. Furthermore, the limbic domain of the basal ganglia, namely ventral striatum (including the nucleus accumbens), interconnected with the cerebellum, can be proved by functional neuroimaging [43].
Interestingly, α-syn, which is considered to be the hallmark of PD, was found to be insignificant in the mediation analysis. This finding may represent the direct effect of α-syn on cognitive impairment (data not shown). The onset and severity of dementia in Lewy body diseases is also associated with pathologies beyond changes in α-syn. Copathologies have been reported in postmortem studies, in vivo PET evaluations, and in animal models. Animal and cellular models suggest a synergistic relationship between α-syn and Aβ-42 and p-tau, which accelerates cognitive decline [44]. Postmortem studies show that pathological tau co-localizes to α-syn in Lewy bodies, further revealing that cognitive decline in PD patients is related to the combined influence of α-syn, Aβ-42, and p-tau [45]. The combination of α-syn with Aβ plaque and p-tau pathology is strongly correlated with dementia in PD patients [46,47]. In the present study, elevated p-Tau and Aβ-42 misfolded protein levels mediated with RAA brain atrophy and cognitive function impairment may indicate an underlying co-pathology of misfolded protein deposition with neurotoxic effects.

Limitations
There are several limitations of this study. First, the case number in our study is relatively small and limited to a single tertiary institution and thus may not be representative of the wider PD population; furthermore, only two cohorts were involved. Second, this is a cross-sectional study, so readers should thus interpret the results with caution, while future longitudinal studies are recommended. Third, plasma measurement is only an indirect index of cortical deposits [47]. The observed differences in peripheral blood may be due to multiple external factors, including the abundant plasma proteins, collection time, and storage conditions of the blood samples. Fourth, not all brain areas are discussed in our study; we only take deep gray matter regions into consideration. In addition, no grading of PD patients was classified in our study.

Conclusions
This study reveals that elevated plasma p-Tau and Aβ-42 may indicate RAA and bilateral cerebellum atrophy and executive function impairments via mediation analysis. We highlight the possibility that changes in some vulnerable anatomies, such as the RAA and bilateral cerebellum, may be correlated with elevated plasma p-Tau and Aβ-42 and the involvement of the underlying cortico-basal ganglia-cerebellar network, as seen in certain phenotypes of PD. This study offers valuable insights into the role of misfolded proteins in cognitive decline as associated with atrophy of certain brain regions and could provide novel targets for future neuroprotective therapies. Institutional Review Board Statement: The protocols of our study were approved by the ethics committee of Chang Gung Memorial Hospital. Written informed consent was provided by all study participants or their guardians prior to enrollment in the study.
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. We declare that all human and animal studies have been approved by the Institutional Review Board of Chang Gung Memorial Hospital and have therefore been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.
Data Availability Statement: Data available on request due to restrictions. The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical.