Genetic Influence on Microstructure Integrity and Motor Progression in Parkinson’s Disease

Abstract

Background:

Up to 10% of Parkinson’s disease (PD) populations carry a genetic risk variant, which may not only increase one’s chance of developing PD but also affect disease presentation and progression. We hypothesize motor impairment in genetic carriers of PD correlate to different patterns of microstructural changes over time.

Design/Methods:

Data were accessed from the Parkinson’s Progression Markers Initiative (PPMI) project. Connectometry analyses were performed for GBA1+ PD, LRRK2+ PD, and sporadic PD correlating white matter structural changes, as measured by quantitative anisotropy (QA), with motor impairment, as measured by MDS-UPDRS III.

Results:

There was a negative correlation between QA and MDS-UPDRS III in all 3 cohorts at 48 months. In GBA1+ PD (n=12), the white matter tracts identified were cortical and subcortical, while in the LRRK2+ PD (n=18) and sporadic PD (n=45) cohorts white tracts identified were primarily subcortical and within the brainstem.

Conclusions:

Our findings highlight the association between motor symptom progrerssion and structural connectivity in individuals with GBA1+ PD, LRRK2+ PD, and sporadic PD. Due to the small sample size, larger studies are needed in the future to confirm the findings.

Introduction

Approximately 10% of patients with Parkinson’s disease (PD) have an identifiable genetic risk variant, with variants in leucine rich repeat kinase (LRRK2) and glucocerebrosidase (GBA1) being the most common [1]. While the clinical presentation may be indistinguishable from sporadic PD, genetic carriers of PD may have different incidence of motor and non-motor symptoms [1,2]. People with LRRK2-PD tend to have milder motor symptoms at baseline, slower progression, and lower incidence of non-motor symptoms including REM-sleep behavior disorder (RBD), hyposmia and cognitive decline; patients with GBA1 variants are more likely to experience a faster rate of progression with a higher incidence of cognitive impairment.[1,2] Simuni et al confirmed similar findings but demonstrated that GBA1-PD patients with the N370S variant present similarly to that of sporadic PD early in the disease course. Interestingly, both LRRK2 and GBA1 variant carriers demonstrated better dopaminergic striatal binding compared to that of sporadic PD. [3] The mechanisms underlying these differences remain undetermined.

Understanding the evolution of structural changes in the brain throughout the course of PD can potentially inform diagnosis, mechanisms of motor and non-motor symptoms, and rate of progression amongst PD subtypes. Little is known regarding white matter integrity and its role in PD pathogenesis and progression [4,5].

Diffusion tensor imaging (DTI) has been increasingly utilized to detect microstructural integrity and structural connectivity. Studies have demonstrated a positive correlation between PD motor progression and increased mean diffusivity (MD) or decreased fractional anisotrophy (FA) within various brain regions, but results have been variable [6,7]. Traditional DTI analyses employs an end-to-end fiber tracking paradigm, which are prone to artifacts especially when evaluating for complex fiber patterns such as crossing or branching fibers [8]. In this study, we used connectometry, or correlational tractography, a novel DTI approach that tracks the correlation of adjacent white matter fascicles with a specific outcome variable.

The objective of this study was to evaluate changes in white matter microstructural integrity in sporadic, and two common genetic variants of PD (LRRK2 and GBA1 variants) and their correlation to progression of motor symptoms.

Methods

Study design

We performed a retrospective study using clinical and imaging data obtained from the Parkinson Progression Markers Initiative (PPMI) database. The detailed methodology of the PPMI database was previously published (http://www.ppmi-info.org/data). Data from this study were accessed from PPMI on 6/10/2020.

Participants

A total of 2,252 subjects, including healthy control, at-risk individuals, and subjects with PD were surveyed. Of the 2,252 subjects, 1,462 were excluded for being healthy controls or at risk-individuals without a clinical diagnosis of PD. The remainder 790 of PD subjects were screened for presence of known PD risk alleles at the time of enrollment. Four-hundred-ninety-one subjects were included in the PPMI “non-genetic” cohort, without known PD risk alleles at the time of enrollment. Seventy-four individuals were later identified as having a LRRK2 variant or a GBA1 variant based on the genetic testing performed as part of the PPMI protocol and recharacterized into the genetic cohort. From the genetic cohort, 271 subjects were identified as having one of the GBA1 or LRRK2 risk variants at the time of enrollment and included in our study. Subjects who tested positive for both GBA1 and LRRK2 variants were eliminated (n = 3). Two different time points were evaluated for longitudinal structural changes: baseline to 24 months, and baseline to 48 months. Not all subjects with imaging at 48 months also had 24 month imaging, therefore we only included subjects with baseline and 48 month DTI. Subjects without DTI or with poor imaging quality were excluded. Subjects were further characterized by the genetic risk variants they carry into sporadic PD, LRRK2+ PD, and GBA1+ PD.

Outcome measures

Motor progression was measured by changes in MDS-UPDRS III score from baseline to 48 months. Most participants were not on dopaminergic medications at time of enrollment. The difference between MDS-UPDRS III scores were calculated using off scores at baseline, and on-medication scores at 48 months. On-medication score at the 48 months time point was chosen to assess for changes in quantitative anisotropy (QA) associated with MDS-UPDRS III changes driven by structural differences, not dopaminergic deficit.

Data acquisition

Raw diffusion weighted imaging (DWI) data used in this study were obtained from the PPMI neuroimaging database (www.ppmi-info.org/data/). Imaging parameters can be found in the PPMI imaging protocol (https://www.ppmi-info.org/study-design/research-documents-and-sops).

Diffusion MRI processing and local connectometry database creation

Raw DWI NIFTI files were imported into publicly available software, DSI Studio (http://dsi-studio.labsolver.org) and converted into SRC files. The b-table was checked by an automatic quality control routine to ensure its accuracy [9]. FIB files were created using Q-Space Diffemorphic Reconstruction (QSDR) [8] to obtain the spin distribution function and bring images into ICBM152_adult template space. A diffusion sampling length ratio of 1.25 was used and the output resolution in diffeomorphic reconstruction was 2 mm isotropic. A connectommetry database was created by aggregating the FIB files across participants. The diffusion metric used in connectometry analysis was quantitative anisotropy (QA), for which decreases are associated with axonal loss [8]. QA is a more robust measure to the free water effect and partial volume of crossing fibers than traditional DTI metrics including FA and MD [10].

Statistical analysis

Statistical analyses were performed using R (R version 4.2.2) and DSI Studio. Group differences for sporadic PD, LRRK+ PD and GBA1+ PD at baseline in demographics and clinical data were assessed using one-way ANOVA and post-hoc Tukey tests. A p-value of less than 0.05 was considered significant.

The connectometry model uses nonparametric permutation testing to identify and track voxels that have high correlation to the outcome of interest, which was the change in MDS-UPDRS III between baseline and 48 months, and the change in QA. A multiple regression model was used to determine the correlation with covariates age, sex, and mean DaT striatal binding volumes at 48 months. A T-score threshold of 2 was assigned and tracked using a deterministic fiber tracking algorithm [10] to obtain correlational tractography. QA values were normalized. Topology-informed pruning was conducted with one iteration to remove false connections. All tracts generated from bootstrap resampling were included. A length threshold of 20 voxel distance was used to select tracts. The seeding number for each permutation was 50,000. To estimate the false discovery rate (FDR), a total of 2,000 randomized permutations were applied to the group label to obtain the null distribution of the track length. The tracking algorithm eliminated cerebellar regions to avoid false results from partial structures. FDR values less than 0.05 were considered significant.

Results

The baseline demographics and clinical data of the cohort are summarized in Table 1.

Table 1:

Clinical characteristics of the three cohorts at baseline and 48 months.

	Sporadic PD	GBA1+ PD	LRRK2+ PD	P
N	45	12	18
Age, years	61.56±8.86	58.14±12.75	55.37±8.62	0.127
Gender, %Male (N)	67 (30)	75 (9)	67 (12)
Race, %White	100	100	100
Education, years	15.09±3.04	16.25±1.86	16.17±3.19	0.279
Disease Duration at screening, months	6.27±6.84	8.07±9.57	8.22±12.36	0.665
LEDD at baseline, mg	0	0	340±56.57	<0.001*
LEDD at 48 months, mg	544.48±240.94	633.76±233.08	563.58±307.59	0.590
MoCA at screening	27.76±2.08	26.25±2.60	27.89±2.11	0.0839
MoCA at 48 months	27.11±3.28	26.58±4.08	28.22±1.63	0.312
MDS-UPDRS III at screening	20.29±9.43	22.08±10.67	19.78±9.33	0.800
MDS-UPDRS III at 48 months	20.09±13.85	17±8.73	17.28±9.80	0.603
DaT striatal binding volumes at screening, mm3	1.30±0.37	1.39±0.46	1.37±0.36	0.637

Data are presented as mean± SD. LEDD = levodopa equivalent daily dose; MoCA = Montreal Cognitive Assessment; MDS-UPDRS III = Movement Disorders Society Unified Parkinson’s Disease Rating Scale part III; DaT = dopamine transporter scan.

Connectometry

GBA1+ PD

At 48 months, motor progression was associated with QA in the following tracts: corpus callosum, inferior fronto-occipital fasciculus, corticostraital pathway, anterior commissure, inferior longitudinal fasciculus, extreme capsule, super longitudinal fasciculus, cingulum, spinothalamic tract, and superior cerebellar peduncle (FDR = 0.017).

LRRK2+ PD

From baseline to 48 months, motor progression was associated with a QA in the following tracts: corticospinal tract, frontopontine, parietopontine, corpus callosum, fornix, anterior commissure, optic radiation, superior cerebellar peduncle, right inferior-fronto-occipital fasciculus (FDR = 0.0146).

Sporadic PD

From baseline to 48 months, motor progression was associated with QA in corpus callosum, corticospinal, and parietopontine tracts (FDR = 0.005).

Figure 1 highlights the white matter tracts associated with motor progression in GBA1+ PD, LRRK2+ PD, and sporadic PD from baseline to 48 months.

Figure 1:

Select white matter tracts structural decline over 48 months in GBA1+ PD (red), LRRK2+ PD (green), and sporadic PD (yellow) cohorts in axial (left), coronal (middle) and sagittal planes (right). GBA1+ PD demonstrated more widespread cortical and subcortical white matter structural decline, whereas LRRK+ PD and sporadic PD demonstrated structural decline more limited to the corpus callosum and brainstem regions.

Discussion

This study shows structural changes in white matter tracts correlate to motor impairment progression in sporadic, GBA1+ and LRRK2+ PD. Our analysis showed that in GBA1+ PD, the white matter tracts identified were cortical and subcortical, while in the LRRK2+ PD and sporadic PD cohorts the white tracts identified were primarily subcortical and within brainstem. There was no significant difference between the change in MDS-UPDRS III from baseline to 48 months between sporadic, LRRK+ PD and GBA1+ PD. In fact, as subjects added dopaminergic treatment, MDS-UPDRS III at 48 months tended to decrease compared to that of baseline, suggesting any structural changes occurred independently from dopaminergic treatment. Our analysis showed that motor progression in subjects with GBA1 variants is associated with more widespread cortical white matter tract degeneration compared to sporadic or LRRK2+ PD. In sporadic and LRRK2+ PD, the association between motor progression and changes in microstructural integrity was ultimately seen in the corpus callosum and more caudal tracts in brainstem regions.

Prior DTI studies focused on examining group differences between PD and healthy controls, or within different phenotypes of PD [6,7]. To our knowledge, our study was the first to evaluate the potential role of genetic factors on microstructural integrity. Within the GBA1+ PD cohort, microstructural changes were observed within the corpus callosum, corticostriatal tract, inferior and superior longitudinal fasciculi, cingulum, inferior fronto-occipital fasciculus, extreme capsule, spinothalamic tract, superior cerebellar peduncle, and anterior commissure. Of these white matter tracts, the corpus callosum, anterior commissure, superior cerebellar peduncle, and inferior-fronto-occipital fasciculus were also associated with motor progression in the LRRK2+ PD cohort, suggesting these white matter pathways are relevant to PD motor symptoms. The corpus callosum has previously been associated with the progression of akinetic-rigid symptoms in sporadic PD symptoms [7], a predominant motor symptom in GBA1+ PD [1,2]. A decline in microstructural integrity in the superior longitudinal fasciculus, along with other large white matter tracts (inferior longitudinal fasciculis, inferior-occipital fasciculus, anterior thalamic radiation, and corpus callosum genu and body) has been associated with gait symptoms in sporadic PD [11], which are also prominent motor symptoms in GBA1+ PD [1,2]. Our data also highlights the significant overlap in brainstem white matter structural changes within LRRK2+ and sporadic PD cohorts, in keeping with their clinically near-identical presentations [1,3].

The findings from this analysis highlight the importance of white matter structural integrity in its relationship to motor symptoms in PD, with additional insights into the genetic influence on PD motor progression. Connectometry increases sensitivity and reliability compared to conventional end-to-end fiber tracking paradigm by determining the strongest association along a fiber pathway [8]. Using the same methodology, Ashraf-Ganjouei and authors found an association in PD dysautonomic symptoms and structural decline within the CC and cingulate gyri. [12] By using a longitudinal approach and pairing each subject with their own baseline scan, we were able to limit the issue of individual brain anatomical differences. In addition, the bottom-up analysis in a common stereotactic space of connectometry allows for an unbiased and statistically robust tracking of specific fiber bundles across the entire brain.

Our study has limitations that should be considered when interpreting these results. The primary focus was on motor symptoms, as measured by the MDS-UPDRS III score. MDS-UPDRS III scores are routinely used as a marker of motor progression in PD research, but it has significant weaknesses, including overweighting tremor symptoms and the strong effect of dopamine medications on the results. A second limitation was that relatively few subjects had diffusion imaging at least two time points at the time our data was assessed. We plan to reassess in hopes of increasing sample size and length of follow up. Finally, the issue of representation of diverse populations in PD research applies to the PPMI cohort, limiting the applicability of our conclusions on other populations. Future studies including additional non-motor clinical features will help provide a more in-depth understanding on how GBA1 and LRRK2 risk alleles affect PD disease progression

In conclusion, the results of this study suggest that motor progression in all three PD cohorts is associated with loss of microstructural integrity and that GBA1+ PD is associated with more widespread cortical and subcortical degeneration of select white matter tracts when compared to LRRK2+ and sporadic PD.

Highlights.

• Connectometry is a novel method of tracking microstructural integrity

• GBA+ PD have more extensive subcortical structural correlates to changes in MDS-UPDRS III

• LRRK2+ PD and sporadic PD have similar patterns of structural correlates to motor progression involving the corpus callosum and brainstem white matter tracts

• Further studies investigating non-motor symptoms are needed