BST1 rs4698412 allelic variant increases the risk of gait or balance deficits in patients with Parkinson’s disease

Summary Aims We aimed to explore effects of bone marrow stromal cell antigen‐1 (BST1) rs4698412 allelic variant on brain activation and associative clinical symptoms in Parkinson’s disease (PD). Methods A total of 49 PD patients and 47 healthy control (HC) subjects were recruited for clinical evaluations, blood samples collection for genotypes, and resting‐state functional MRI (rs‐fMRI) scans. Based on BST1 rs4698412 allelic variant (G → A), participants were further divided into 18 PD‐GG, 31 PD‐GA/AA, 20 HC‐GG, and 27 HC‐GA/AA carriers, which respectively indicated subjects carrying ancestral or risk allele in that locus in PD or HC. Two‐way analysis of covariance (ANCOVA) was applied to investigate main effects and interactions between PD and BST1 rs4698412 allelic variant on brain function via amplitude of low‐frequency fluctuations (ALFF). Spearman’s correlations were then utilized to detect associations between interactive brain regions and clinical symptoms. Results Compared to HC subjects, PD patients exhibited increased ALFF values in left cerebellum_8 and cerebellum_9. Significant interaction was in right lingual gyrus, where there were the lowest ALFF values and ALFF values were only negatively associated with Timed Up and Go (TUG) test time in PD‐GA/AA subgroup. Conclusion BST1 rs4698412‐modulated lingual gyrus functional alterations could be related to gait and balance dysfunction in PD.

4p15/bone marrow stromal cell antigen-1 (BST1) was identified to participate in the PD progression. 7 BST1, also known as CD157, plays a role in immune responses. 8 In addition, BST1 could produce cyclic ADP-ribose involved in regulating calcium homeostasis. 9,10 Either abnormal immune responses 5 or imbalance of calcium homeostasis 11,12 could cause the development of PD. Moreover, accumulating evidence demonstrated the most investigated single nucleotide polymorphism (SNP) in BST1 gene was rs4698412 variant (G → A).
A meta-analysis 13 and genomewide association studies (GWAS) 7,14 in different populations have affirmed that BST1 rs4698412 was relevant to the increased risk of sporadic PD. Particularly, a study in Chinese population also manifested the association between BST1 rs4698412 and PD. 15 Taken together, these results prompted the potential role of BST1 rs4698412 allele in the development of PD.
However, these studies indicated little or no information regarding the association between BST1 rs4698412 allelic variant and clinical features, especially about how to mediate neural function or brain deficits in the pathogenesis of PD. Thus, in the current study, we aimed to explore the probable impact of BST1 rs4698412 allelic variant on brain functions and clinical symptoms in PD patients.
Imaging genetics could provide a more thorough and precise insight into the influence of gene variants on the brain. 16 Meanwhile, imaging genetics has been applied in research of multiple diseases including Alzheimer's disease (AD) 17 and attention-deficit/hyperactivity disorder. 18 Resting-state functional magnetic resonance imaging (rs-fMRI) can indicate the phenomenon of spontaneous neuronal activity at rest by examining spontaneous fluctuations in the blood oxygen level dependent (BOLD) signal of brain regions without any explicit stimulation. 19 Therefore, it could provide a platform to explore the probable gene function in the brain.
Amplitude of low-frequency fluctuations (ALFF) reflects the level of regional brain activation by detecting the spontaneous amplitude of low-frequency (0.01-0.08 Hz) BOLD signal. 20 A series of studies have applied ALFF approach to focus on possible pathogenesis in PD previously. For example, a study revealed that ALFF analysis could sensitively and specifically distinguish individuals with PD in the off medication state from healthy controls. 21 Another study also showed apathy, depression, and motor severity in PD could be predicted by ALFF signals in some brain regions. 22 Nonetheless, these researchers didn't consider the possible role of related genetic variations in brain region dysfunction. Hence, in order to identify the gene-brain-behavior relationships, we investigated the underlying regulations of BST1 rs4698412 allelic variant in the PD progression by examining ALFF signals along with the correlations between BST1 rs4698412-modulated brain alterations and clinical symptoms.

| Participants
According to UK Parkinson's Disease Society Brain Bank criteria for idiopathic PD, 23 we recruited 49 right-handed sporadic PD patients from outpatients and inpatients in the First Affiliated Hospital of Nanjing Medical University. PD patients who were combined with other severe acute or chronic neurological diseases including stroke, brain tumor or psychiatric diseases, were ruled out from our study. Besides, we excluded PD patients with cognitive impairment (Mini-Mental State Examination (MMSE) score <24). Furthermore, 47 healthy control (HC) subjects were enrolled in our study by advertising in the community. Participants with contraindications for MRI scans or intakes of antidepressant, anxiolytic, or antipsychotic drugs recently were also excluded from this study. Additionally, to minimize conceivable pharmacological impacts on neural activity, PD patients underwent MRI scans and clinical examinations during off-state (at least 12-hour withdrawal of pharmacologic treatment for PD). All participants had given their informed consent before the study began, which was approved by the ethics committee of the First Affiliated Hospital of Nanjing Medical University.

| Clinical evaluations
All PD patients were assessed their cognitive condition via MMSE 24 and Frontal Assessment Battery (FAB). 25 Emotional states were evaluated by means of 17-item Hamilton Depression Rating Scale (HDRS-17), 26 Hamilton Anxiety Rating Scale (HAMA) 27 and Apathy scale (AS). 28 Moreover, we employed the motor component of PIGD score was the sum of UPDRS-III items 27-30. Gait and balance were quantified with Timed Up and Go (TUG) test. 30 Besides, we assessed the Epworth Sleeping Scale (ESS) 31 and Fatigue Severity Scale (FSS) 32 for each PD patient. We also calculated total levodopa equivalent daily dose (LEDD), LEDD of levodopa preparations and LEDD of dopamine receptor agonists in each PD patient of two subgroups, respectively. 33

| DNA isolation and SNP genotyping
Peripheral venous blood samples were obtained from 49 PD pa-  After that, the PCR product was disposed of by shrimp alkaline phosphatase (SAP) to remove the dissociative dNTPs that were out of the system. The SpectroCHIP chip after dotting in MassARRAY Nanodispenser RS1000 array machine system (Agena Bioscience) was analyzed by matrix-assisted laser desorption/ionization timeof-flight (MALDI-TOF) mass spectrometer (Agena Bioscience) and followed by TYPER4.0 software to get the raw data and the genetic diagram.
Afterward, the subjects were classified into different subgroups (A-allelic variant carriers and GG homozygote carriers) based on previous studies. 7,13,14,34

| MRI acquisition
Participants were scanned via a 3.0 T Siemens MAGNETOM Verio whole-body MRI system (Siemens Medical Solutions, Erlangen, Germany). Tight foam padding, along with ear-plugs, was used to minimize head movement and diminish noise. In addition, participants were instructed to hold motionless as much as possible, close their eyes but remain awake, and attempt not to think about anything particularly during the whole scanning procedure. Prior to functional running, three-dimensional T1-weighted anatomical images were obtained using the following volumetric 3D magnetization-

| Preprocessing of fMRI data
The images were preprocessed and analyzed via DPARSF software (http://www.restfmri.net/forum/dparsf). 35 First 10 time points were abandoned. The remaining 230 volumes were kept for slice timing correction and head motion correction. None of the participants with head motions exceeded 2 mm or 2° of translation or rotation in any direction was excluded in our study. We also calculated the mean head translation, mean head rotation, and frame-wise displacement (FD). 36 Analysis of these head motion parameters did not reveal any difference between the four subgroups (P > 0.05).
Afterward, high-resolution T1 structural images were coregistrated to functional images via a nonlinear image registration approach and segmented using a new segment algorithm with diffeomorphic anatomical registration through exponentiated lie algebra (DARTEL), followed by a 24 parameter Volterra expansion. Finally, fMRI images were spatially normalized into the Montreal Neurological Institute (MNI) template, resampled into a spatial resolution of 3 × 3 × 3 mm 3 and spatially smoothened with a 4 mm full width at half-maximum Gaussian kernel.

| ALFF processing
The procedure of ALFF calculation was performed using the REST software (http://restfmri.net/forum/REST). In brief, as introduced previously, 37 all voxels were converted from the time domain to the frequency domain via Fast Fourier Transform (FFT) after image preprocessing. Subsequently, the square root of the power spectrum across 0.01 Hz to 0.08 Hz for each voxel was calculated. The average square root was regarded as ALFF, which was then converted into zscores for standardization purposes. Afterward, band-pass filtering (0.01 < f < 0.08 Hz) was performed and linear trend was removed.

| Statistical Analysis
Demographic and clinical data analysis was performed via IBM SPSS statistics v20.0.0 software (SPSS, Chicago, IL, USA). The comparisons were conducted between PD and HC groups, within BST1 rs4698412 GG carriers and GA/AA carriers. A Kolmogorov-Smirnov (K-S) test was employed to test for normality. Two-sample t test or one-way analysis of variance (ANOVA) was used for normally distributed data. Asymmetrically distributed variables were tested with Mann-Whitney U or Kruskal-Wallis test. Besides, chi-square test was for gender. The Hardy-Weinberg Equilibrium (HWE) of the genotype frequencies were also tested via Chi-square test. A significant threshold was set at P < 0.05. Two-way factorial analysis of covariance (ANCOVA: groups × genotypes; groups: PD and HC, genotypes: GG carriers and GA/AA carriers) was performed adjusting for age, gender, and education years, followed by post hoc tests to further explore the main effects and interactions. All statistical thresholds were set at a corrected P < 0.001, determined by Monte Carlo simulation for multiple comparisons (http://afni.nimh.nih.gov/pub/dist/doc/manual/AlphaSim.pdf). In addition, Spearman's correlative analyses with significant thresholds set at P < 0.05 were performed between clinical test scores and ALFF values of the clusters showing significant interactions between groups and genotypes. Demographic characteristics and clinical evaluations of participants among four subgroups were shown in Table 2. No significant effect of diagnosis, genotype, or interaction between diagnosis and genotype was observed for gender and education years. However, age of PD patients was significantly different from HC subjects, which was then taken as a covariant during the following analysis. Additionally, post hoc analysis revealed that no significant difference was found  there was a significant difference in PIGD scores between PD-GG and PD-GA/AA subgroup (P = 0.009). But tremor scores, rigidity scores, and akinesia scores in PD-GG subgroup were not significantly different from PD-GA/AA subgroup.

| ALFF analysis
The influence of diagnosis and genotypes on ALFF between GG and GA/AA carriers in PD and HC groups was shown in Table 3 and

| D ISCUSS I ON
To our knowledge, this study was the first to investigate the potential effect of BST1 rs4698412 allelic variant on modulating restingstate brain function in sporadic PD.
BST1 rs4698412 was considered as a candidate SNP for risk of PD supported by several GWAS 7,14 and a meta-analysis, 13 but some other studies also indicated that there was no relationship between BST1 rs4698412 and PD. 34,38 We hypothesized that the discrepancy might be attributed to heterogeneity of individ- Nonetheless, there was an interaction between diagnostic groups (PD, HC) and genotypes (GG, GA/AA) in right lingual gyrus, which could manifest the differential effects of BST1 rs4698412 genotypes on brain functions in PD and HC. The exact mechanism underlying will be explored in future studies.
Lingual gyrus, located in occipital cortex (also known as visual association cortex), participated in spatial orientation 39 and visuospatial information processing. 40,41 Moreover, changes in the flow of visual information mainly dominated by right hemisphere, 42,43 had an influence on balance disorders 44 48 and predict fall risks in PD. 49,50 Besides, PIGD scores from UPDRS-III scores can also reflect gait or balance function in PD patients to some extent. Taken together, these results could help to interpret the impact of allele A of BST1 rs4698412 on a pathological process leading to gait or balance difficulties in PD patients.
Nevertheless, some limitations should be considered when interpreting our results. Firstly, the sample size of our study was relatively small, so GA carriers and AA carriers were merged into one group, which failed us to explore the respective effects of GA and AA allele on brain function. A larger sample size is necessary based on these preliminary findings in the future. Secondly, only BST1 rs4698412 was investigated in our study while brain spontaneous activities might be affected by various genes and SNPs. Therefore, gene-gene interactions or more complex haplotype analysis should be further investigated. Thirdly, in this study, we only chose TUG test time to assess gait or balance performance in PD patients. To be rigorous, we should adopt some more tests to further verify our conclusion in future studies. Finally, our present study was a cross-sectional design focusing on regional brain activation and could not show the causal inference, which needed to be explored in the longitudinal research with multimodal techniques deeply.

| CON CLUS ION
The present study demonstrated that functional alternations in right lingual gyrus caused by BST1 rs4698412 allelic variant (allele A) could be associated with gait or balance deficits in PD patients.
It might provide a new perspective for further studies on PD for more effective disease intervention. Furthermore, imaging genetics as a new promising approach has the potential to explore relevant neurobiological mechanisms underlying gene polymorphisms on the complicated clinical symptoms in PD.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.