You are here

Article Text

Article menu

Download PDFPDF

Movement disorders
Original research
Impact of GBA1 variants on long-term clinical progression and mortality in incident Parkinson’s disease
  1. Thomas B Stoker1,2,
  2. Marta Camacho1,
  3. Sophie Winder-Rhodes1,
  4. Ganqiang Liu3,4,
  5. Clemens R Scherzer4,5,
  6. Thomas Foltynie6,
  7. Jonathan Evans7,
  8. David P Breen8,9,10,
  9. Roger A Barker1,2,
  10. Caroline H Williams-Gray1
  1. 1 John van Geest Centre for Brain Repair, Department of Clinical Neurosciences, University of Cambridge, Cambridge, Cambridgeshire, UK
  2. 2 Wellcome Trust Medical Research Council - Cambridge Stem Cell Institute, Cambridge, UK
  3. 3 School of Medicine, Sun Yat-Sen University, Guangzhou, Guangdong, China
  4. 4 Advanced Center for Parkinson's Disease Research, Brigham and Women's Hospital, Boston, Massachusetts, USA
  5. 5 Precision Neurology Program, Brigham and Women's Hospital, Boston, Massachusetts, USA
  6. 6 Department of Clinical and Movement Neurosciences, UCL Institute of Neurology, London, UK
  7. 7 Department of Neurology, Nottingham University Hospitals NHS Trust, Nottingham, UK
  8. 8 Centre for Clinical Brain Sciences, University of Edinburgh, Edinburgh, UK
  9. 9 Anne Rowling Regenerative Neurology Clinic, University of Edinburgh, Edinburgh, UK
  10. 10 Usher Institute of Population Health Sciences and Informatics, University of Edinburgh, Edinburgh, UK
  1. Correspondence to Dr Thomas B Stoker, John van Geest Centre for Brain Repair, Department of Clinical Neurosciences, University of Cambridge, Cambridge, CB2 0PY, UK; tbs26{at}cam.ac.uk

Abstract

Introduction Variants in the GBA1 gene have been identified as a common risk factor for Parkinson’s disease (PD). In addition to pathogenic mutations (those associated with Gaucher disease), a number of ‘non-pathogenic’ variants also occur at increased frequency in PD. Previous studies have reported that pathogenic variants adversely affect the clinical course of PD. The role of ‘non-pathogenic’ GBA1 variants on PD course is less clear. In this study, we report the effect of GBA1 variants in incident PD patients with long-term follow-up.

Methods The study population consisted of patients in the Cambridgeshire Incidence of Parkinson’s disease from General Practice to Neurologist and Parkinsonism: Incidence, Cognition and Non-motor heterogeneity in Cambridgeshire cohorts. Patients were grouped into non-carriers, carriers of ‘non-pathogenic’ GBA1 variants and carriers of pathogenic GBA1 mutations. Survival analyses for time to development of dementia, postural instability and death were carried out. Cox regression analysis controlling for potential confounders were used to determine the impact of GBA1 variants on these outcome measures.

Results GBA1 variants were identified in 14.4% of patients. Pathogenic and ‘non-pathogenic’ GBA1 variants were associated with the accelerated development of dementia and a more aggressive motor course. Pathogenic GBA1 variants were associated with earlier mortality in comparison with non-carriers, independent of the development of dementia.

Discussion GBA1 variants, including those not associated with Gaucher disease, are common in PD and result in a more aggressive disease course.

https://creativecommons.org/licenses/by/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.

https://doi.org/10.1136/jnnp-2020-322857

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Heterozygous mutations in the GBA1 gene have been identified as a common risk factor for Parkinson’s disease (PD), being found in approximately 5% of patients.1–4 Biallelic pathogenic mutations result in the lysosomal storage disorder, Gaucher disease (GD).5 In addition to the known pathogenic mutations associated with GD, there are a number of ‘non-pathogenic’ GBA1 variants that are not associated with GD but that are found at increased frequency in patients with PD, including the common E326K and T369M variants.6–9 Carrying a pathogenic GBA1 mutation conveys an earlier age of onset10 11 and adversely affects the clinical course of PD, manifested by an increased risk of cognitive decline and dementia1 10 12–18 and increased rate of motor progression.1 10 12 19 20 However, some studies have found no effect on motor progression in GBA1 mutation-associated PD.14 15 17 Few studies have commented on the effect of GBA1 mutations on mortality in PD, though some have suggested that survival time is reduced in GBA1 mutation carriers.10 16 The clinical significance of carrying a ‘non-pathogenic’ GBA1 variant is disputed14 21 though the E326K variant in particular has been associated with an adverse clinical course in PD in some studies.20 22

    We previously reported on the epidemiology of GBA1 mutations in incident PD.1 Here we present long-term follow-up data of incident PD cases carrying pathogenic and ‘non-pathogenic’ variants in the GBA1 gene from two community-based cohorts: the ‘Cambridgeshire Incidence of Parkinson’s disease from General Practice to Neurologist’ (CamPaIGN) and ‘Parkinsonism: Incidence, Cognition and Non-motor heterogeneity in Cambridgeshire’ (PICNICS) cohorts. Advantages of this study over some previously published studies include the long duration of follow-up, along with the fact that both cohorts consisted of community-based patients with newly diagnosed PD.

    Methods

    Patients

    The study population consisted of patients in the CamPaIGN and PICNICS cohorts—both community-based incident PD cohorts table 1 and online supplementary table 3).23 PD diagnosis was determined by the UK Parkinson’s Disease Society Brain Bank Criteria. The CamPaIGN cohort was recruited between 2000 and 2002, and subjects were followed up at 2 yearly intervals for up to 18 years. The PICNICS cohort was recruited between 2006 and 2013, and subjects have been followed up at 18 monthly intervals, so far for up to 11.4 years. Inclusion and exclusion criteria for these cohorts are summarised in online supplementary table 1.Written informed consent was obtained from all subjects.

    Supplemental material

    View this table:
    Table 1

    Patient demographics

    Motor, cognitive and neuropsychiatric assessments were completed at each time point including the Unified Parkinson’s Disease Rating Scale (UPDRS) in the CamPaIGN cohort, Movement Disorders Society UPDRS (MDS-UPDRS) in the PICNICS cohort, Mini-Mental State Examination (MMSE), Hoehn and Yahr (HY) stage, levodopa equivalent dose, Beck Depression Inventory, National Adult Reading Test and Parkinson’s Disease Questionnaire-39. For the CamPaIGN cohort, estimated MDS-UPDRS scores were calculated as described previously.24 Postural instability and gait disability (PIGD) and tremor scores were calculated for the CamPaIGN cohort as described here.25 We adapted this method to allow calculation of similar motor phenotypes in the PICNICS cohort, using the mean value of the UPDRS-MDS scores for questions 2.10 (tremor history), 3.15 (postural tremor), 3.16 (kinetic tremor) and 3.17 (rest tremor) for tremor score, and questions 2.12 (walking and balance history), 2.13 (freezing history), 3.10 (gait examination) and 3.13 (posture examination) for PIGD score. GBA1 variants were identified through sequencing of the GBA1 gene targeted at exons 1 to 11 in 250 patients, as described previously.1 A further 127 patients underwent targeted screening using an Illumina Multi-Ethnic Genotyping Array (MEGA) ChIP for specific GBA1 variants (online supplementary figure 1).26

    Survival and regression analyses

    Subjects were grouped into non-carriers and carriers of any GBA1 variant (GBA1-PD). The GBA1-PD group was further stratified into carriers of pathogenic mutations and carriers of ‘non-pathogenic’ variants, as indicated below. Hereafter, variants referred to as ‘pathogenic’ are those associated with the development of GD in the homozygous or compound heterozygous state, and variants referred to as ‘non-pathogenic’ are those that are not associated with the development of GD. All patients in whom sequencing had been undertaken were included in survival analyses, as well as additional patients in whom any GBA1 variant was detected on the MEGA ChIP array. Patients in which only MEGA ChIP screening was performed that were not found to carry a GBA1 variant were excluded, given the possibility that they carried an undetected GBA1 variant that had not been screened for. For subgroup analysis, subjects that only had genotyping with the Mega ChIP array in which no GBA1 variant was identified, or in which a ‘non-pathogenic’ variant was identified (n=6), were excluded to ensure that there no patients with missed pathogenic mutations were included in the non-carrier or ‘non-pathogenic’ variant carrier groups.

    Time to mortality, postural instability and dementia were selected as outcome measures, as these represent important clinical milestones in the progression of PD. Kaplan-Meier curves and Cox Regression analyses controlling for potential confounders as detailed below were used for comparison between the groups. Development of postural instability was defined as time to reach HY stage three (HY3) or PIGD score of 2 or more, and dementia was defined as a MMSE score of 24 or less, with fulfilment of Diagnostic and Statistical Manual of Mental Disorders IV criteria (functional impairment (as defined by Schwab and England score of 70% or less) with cognitive deficits in at least two domains and exclusion of other causes of cognitive dysfunction). The CamPaIGN and PICNICS cohort studies predated establishment of the Movement Disorder Society PD dementia criteria. Patients were only considered to have reached these outcomes if these criteria were fulfilled at all subsequent visits.

    Statistical analysis

    All statistical tests were performed on SPSS V.25 software. Independent samples t-tests were used to compare between non-carriers and carriers of all GBA1 variants. Levene’s test was first performed to ensure that equality of variance could be assumed. One-way analysis of variance tests were used for comparison between non-carriers and carriers of ‘non-pathogenic’ or pathogenic GBA1 variants. χ2 tests were used to compare sex distribution. For survival analysis, log-rank values are reported, with comparison with the non-carrier group.

    Results

    Incidence and spectrum of GBA1 abnormalities

    GBA1 variants were present in 36 out of the 250 patients in which sequencing was performed (14.4%). Pathogenic mutations were present in 11 of these patients (4.4 %), which included N370S (n=3), L444P (n=3), R463C (n=1), G10S (n=1), N426K (n=1), R48W (n=1) and R257Q (n=1). Isolated ‘non-pathogenic’ variants were present in 25 of these patients (10.0 %). These included c.762 18T>A (n=8), E326K (n=7), T369M (n=7), E388K (n=1), L119L (n=1) and c.589 86A>G) (n=1). The carrier of the pathogenic R48W variant carried a concomitant E326K variant, and the carrier of the pathogenic N426K variant carried a concomitant c.762 18T>A variant. These were both included in the pathogenic variant group for subgroup analyses. Variants were identified in 12 of the 127 patients (9.4%) in which MEGA ChIP analysis was performed. These included the pathogenic N370S (n=4), R463C (n=1) and G10S (n=1) variants and the ‘non-pathogenic’ T369M (n=3), E326K (n=2) and E388K (n=1) variants.

    After excluding those patients in which only MEGA ChIP genotyping was performed that were not identified to carry a GBA1 variant, these combined cohorts contained 214 non-carriers and 48 carriers of GBA1 variants. Mean follow-up times were 8.0 years (maximum 18.0 years) for the CamPaIGN cohort (n=115) and 5.8 years (maximum 11.4 years) for the PICNICS cohort (n=147).

    Cognitive decline and risk of dementia

    Baseline cognitive function did not differ between variant carriers and non-carriers, with no significant differences in MMSE scores at initial assessment (table 1). Carriers of any GBA1 variant had an increased risk of progression to dementia compared with non-carriers (table 2 and figure 1). Cox regression analysis controlling for age at diagnosis, sex and baseline MMSE score revealed that development of dementia was significantly more likely in carriers of any GBA1 variant, compared with non-carriers, with an HR of 2.1 (95% CI 1.2 to 3.9, p=0.013). At the 10-year time point, fewer than half of the variant carriers remained dementia free, in comparison with 68.4% of non-carriers. At 15 years, 64.2% of surviving non-carriers remained dementia-free. All patients in the GBA1 carrier group had developed dementia or died prior to the 15-year time point. The estimated mean time to dementia was approximately 5 years earlier in variant carriers, compared with non-carriers (table 2).

    View this table:
    Table 2

    Time to dementia

    Figure 1
    Figure 1

    Effect of GBA1 variants on development of dementia in PD. (A) Survival analysis for time to dementia in carriers of any GBA1 variant compared with non-carriers. (B) Subgroup survival analysis for time to dementia in carriers of pathogenic GBA1 mutations, ‘non-pathogenic’ variants and non-carriers. Statistical significance as determined by log-rank tests indicated by asterisk. PD, Parkinson’s disease.

    Subclassification of the variant carrier group into those with pathogenic mutations and those with ‘non-pathogenic’ variants revealed a significantly increased risk of progression to dementia in carriers of pathogenic variants, with an HR of 3.6 (95% CI 1.5 to 8.3, p=0.003). In carriers of ‘non-pathogenic’ variants, there was a non-significant trend towards increased risk of dementia (HR 1.8 (95% CI 0.9 to 3.8, p=0.11).

    Motor progression

    The baseline HY stage was lower in non-carriers compared with carriers (table 1), though there were no significant differences between the subgroups. There were no differences in baseline UPDRS motor score, PIGD score or levodopa equivalent dose between any of the groups. The risk of progression to postural instability was increased in the GBA1-PD group in comparison with the non-carrier group (table 3 and figure 2). At 5 years from diagnosis, 67.5% of the GBA1-PD group had reached HY3, compared with 43% of non-carriers. After controlling for potential confounders (age at diagnosis, sex, baseline HY score, baseline UPDRS motor score and baseline PIGD score), the risk of progression to HY3 was significantly greater in both carriers of ‘non-pathogenic’ variants and pathogenic mutations, with an HR of 1.7 (95% CI 1.0 to 2.9, p=0.035) and 1.9 (95% CI 1.0 to 3.4, p=0.036), respectively (figure 2 and table 3).

    View this table:
    Table 3

    Time to HY3

    Figure 2
    Figure 2

    Effect of GBA1 variants on development of postural instability in PD. (A and C) Survival analysis for time to HY3 and PIGD score of 2 or more, respectively, in carriers of any GBA1 variant compared with non-carriers. (B and D) Subgroup survival analysis for time to HY3 and PIGD score of 2 or more, respectively, in carriers of pathogenic GBA1 mutations, ‘non-pathogenic’ variants and non-carriers. Statistical significance as determined by log-rank tests indicated by asterisk. HY3, Hoehn and Yahr stage three; PIGD, postural instability and gait disability.

    Time to PIGD score of 2 or more was also determined as a measure of motor progression. Pathogenic variants were associated with an increased risk of progression to PIGD score of 2 or more, with a HR of 2.8 (95% CI 1.4 to 5.7, p=0.005). There was no significant increased risk of progression to PIGD score of two or more in carriers of ‘non-pathogenic’ variants (table 4).

    View this table:
    Table 4

    Time to PIGD score of 2 or more

    Mortality

    Mean time to mortality was approximately 1 year earlier in carriers of any GBA1 variant compared with non-carriers (table 5 and figure 3). At 15 years from diagnosis, only 5.3% of GBA1 variant carriers were surviving, compared with 20.1% of non-carriers. When controlling for age and sex, the risk of mortality was increased in carriers of pathogenic mutations in comparison with non-carriers with an HR of 2.1 (95% CI 1.0 to 4.1, p=0.043) but not in those carrying ‘non-pathogenic’ variants (table 5).

    Figure 3
    Figure 3

    Effect of GBA1 variants on time to death in PD. (A) Survival analysis for time to death in carriers of any GBA1 variant compared with non-carriers. (B) Subgroup survival analysis for time to death in carriers of pathogenic GBA1 mutations, ‘non-pathogenic’ variants and non-carriers. PD, Parkinson’s disease.

    View this table:
    Table 5

    Time to mortality

    To explore whether the increase in mortality associated with pathogenic GBA1 variants was attributable to the development of dementia, the dementia outcome was added into the Cox regression model, along with age and sex. The increased risk of mortality associated with carrying any GBA1 variant persisted after controlling for dementia status (HR 2.0 (95% CI 1.0 to 4.1, p=0.046)), suggesting that the relationship was independent of the development of dementia. Causes of death in each group are shown in online supplementary table 2.

    Discussion

    In this study, we have described the clinical course of incident GBA1-PD for up to 18 years from the time of diagnosis, which is a longer follow-up period than any previous study. We have demonstrated that GBA1 variants confer an increased risk of dementia, faster motor progression and mortality in PD, in keeping with previous reports. Importantly, in addition to pathogenic GBA1 mutations, variants generally considered to be ‘non-pathogenic’ were found to adversely affect disease course in GBA1-PD.

    By studying patients with incident PD for up to 18 years, this study has uniquely described the natural history of GBA1-PD. Most prior longitudinal studies of GBA1-PD have employed prevalent cohorts in which variation in disease duration may confound the results, and many other studies have used clinic-based cohorts that may be disproportionately represented by atypical or severe cases. The two cohorts analysed in this study were community based and thus more representative of the general PD population. Another strength of this study was the inclusion of patients with GBA1 variants that do not cause GD, which occur at increased incidence in PD. These variants were termed ‘non-pathogenic’ to differentiate them from variants that cause GD, though taking into account results from this and previous studies, they should not be considered to be non-deleterious.

    Of the patients that had undergone genetic sequencing, 14% carried a GBA1 variant, with 4.4% carrying a pathogenic mutation2–4 27 previously associated with GD. The frequency of pathogenic mutations was similar to that seen in previous studies. The prevalence of any GBA1 variant was higher than that seen in other studies, probably due to the fact that we included non-coding variants such as c.762 18T>A, which has been reported as a potential risk factor for PD, though its importance is not clear.28 The most common variants were c.762 18T>A, E326K and T369M, which accounted for almost two-thirds of all variants. Carriers of these variants are a particularly interesting group to study, as their clinical significance is much less clear than that of so-called ‘pathogenic’ GBA1 mutations.

    We found that GBA1 variants were generally associated with a more aggressive disease course. Diagnosis of dementia and development of postural instability were chosen as outcome measures, since these represent important milestones in the clinical progression of PD.29 Consistent with previous studies, carriers of pathogenic mutations had an increased risk of reaching both of these milestones. Carriers of any GBA1 variant had a projected mean time to dementia that was approximately 5 years earlier than in non-carriers, while the estimated mean time to development of postural instability was approximately 2 years earlier than that seen in non-carriers. Carriers of pathogenic mutations had a 3.6-fold greater risk of development of dementia than in non-carriers and a 1.9-fold greater risk of progression to HY3. Importantly, carriers of variants not associated with GD conveyed a similar risk of development of postural instability in comparison with non-carriers, with a trend towards greater risk of dementia. These variants have previously been reported to have little effect on clinical course in PD,14 though it is possible that such effects would be missed due to short follow-up times and heterogeneous populations. Carriers of the common E326K variant have been shown to have an increased risk of dementia and motor progression in comparison with non-carriers in one US-based study.20 However, in a large meta-analysis of GBA1 variants in PD, there was no association between development of PD and the E326K variant.4 This study population was enriched for individuals from Jewish, Portuguese and Asian populations, in which the E326K accounts for a lower proportion of GBA1 variants than in other populations, which may account for the lack of association in this analysis.3 30

    We also found a significantly increased risk of mortality over the study period in carriers of pathogenic GBA1 variants, in which the risk of death was approximately twofold greater than that in non-carriers, with a mean time to mortality approximately 1 year earlier. This observation was not attributable to the increased incidence of dementia in GBA1 variant carriers, since the effect persisted after controlling for dementia status. Other factors (such as immobility, adverse drug effects and non-motor features) may contribute to the observed increase in mortality in GBA1-PD. ‘Non-pathogenic’ variants were not associated with an increased risk of mortality, though the time to mortality was non-significantly reduced in this group. The impact of GBA1 variants on PD progression may lie on a spectrum, with pathogenic mutations conveying a greater biological effect than ‘non-pathogenic’ variants. As such, our study may have been underpowered to detect a small reduction in time to mortality associated with ‘non-pathogenic’ variants, and it would be interesting to explore this in a larger cohort.

    Age of disease onset has previously been reported to be reduced in PD associated with GBA1 mutations.10 13 27 31 In this study, age of onset did not differ based on GBA1 status. The mean age of onset in both carriers of pathogenic mutations and carriers of ‘non-pathogenic’ variants in GBA1 was approximately 2 years earlier than in non-carriers, and it may be that this study was underpowered to detect such a difference in age at onset.

    The observation that ‘non-pathogenic’ GBA1 variants adversely affect the prognosis of PD patients is important. This significantly increases the proportion of patients with PD in which GBA1 abnormalities play a potentially pathogenic role, in comparison with considering GD-causative mutations as the only variants that alter PD course. Given that GBA1 mutations are associated with the specific pathology of lysosome–autophagy system dysfunction,32 it is possible that the optimal treatment approach in this group of patients will involve drugs that are able to restore activity of this system. As such treatments emerge, it may become important to screen patients with PD for variants in the GBA1 gene, so that precision medicine can be applied, and our data suggest that it will be important to screen for a much broader range of variants than those previously identified as ‘pathogenic’.

    The ‘pathogenic’ mutations associated with GD result in significant reduction in glucocerebrosidase enzyme activity, which underlies its clinical manifestations.5 33 However, polymorphisms such as E326K result in a relatively minimal reduction in enzyme activity,22 and the fact that these predispose to PD without being associated with GD, supports the concept that the pathogenic mechanisms of GBA1-PD are distinct from those of GD. This means that putative treatments for GBA1-PD may need to correct glucocerebrosidase folding and structure, or restore activity of dysfunctional pathways (such as the lysosome-autophagy system), rather than simply restoring glucocerebrosidase enzyme activity.

    We acknowledge that there are a number of limitations with our study. While we have controlled for relevant potential confounders in our survival analyses, data regarding prodromal features of PD that may influence prognosis (eg, rapid eye movement sleep-behaviour disorder) were not available. We have reported on the time from diagnosis, rather than time from symptom onset, to the development of clinical milestones, which may not truly reflect disease duration. However, the time of symptom onset is difficult to determine accurately, since many prodromal features may not be attributed to PD, which would potentially introduce bias.

    These cohorts predated the development of the MDS criteria for PD dementia. Cognitive status was measured in these cohorts with MMSE, though other cognitive examinations are now favoured. In order to maximise the accuracy of dementia diagnosis in these cohorts, we employed strict criteria that reflect the MDS criteria for PD dementia, including a requirement for irreversible deficits in multiple cognitive domains accompanied by functional impairment. Though we included baseline MMSE as a covariate in our Cox regression model for time to dementia, it is not known how many patients had mild cognitive impairment at baseline, which may be expected to influence prognosis. Because assessments were performed every 2 years for the CamPaIGN cohort and every 18 months for the PICNICS cohort, it was not possible to determine the exact time of the onset of dementia or progression to HY3 or PIGD score of 2 or more, which was determined as the midpoint between the assessment at which the patient had reached the outcome measure and the prior assessment.

    Another limitation of our study is the relatively limited sample size. We have attempted to address this by using two homogeneous incident cohorts from the same population base but separated in time.23 However, 115 patients were excluded from survival analyses as they had only undergone limited genetic screening, meaning that it was possible that they harboured unidentified, potentially clinically significant GBA1 variants. The sequencing of GBA1 conferred major benefits in this study, as a more comprehensive method for identifying both presence of absence of GBA1 variants. While we acknowledge that complete sequencing reads could not be achieved in all patients (as previously described in Winder-Rhodes et al 1 2013), we feel that the use of sequencing data allowed us to assign GBA1 variant carrier status with maximum possible accuracy. Further strengths of this study are the long duration of follow-up of up to 18 years, and the fact that our cohorts consisted of incident PD patients recruited from the community, allowing follow-up from diagnosis to the development of key clinical milestones in a representative population. As such we have presented the first long-term data on the effect of GBA1 variants on mortality in an incident PD population and detected significant differences in the natural history of PD associated with ‘non-pathogenic’ GBA1 variants, which may have been missed in previous studies, probably due to their limited follow-up durations.

    In conclusion, long-term follow-up of the CamPaIGN and PICNICS cohorts confirms that GBA1 mutations adversely affect clinical course. Importantly, carriers of GBA1 variants that are often considered to be ‘non-pathogenic’ had a significantly worse clinical course than non-carriers. GBA1 variants therefore contribute to the pathogenesis and clinical features of more than 10% of patients with PD, making them a particularly important genetic subgroup. Since GBA1 abnormalities appear to be associated with specific pathogenic mechanisms, it may be that the optimal disease-modifying approach differs in this group than in the wider PD population, with important implications for future clinical trials.

    References

    1. ↵
      2. Winder-Rhodes SE ,
      3. Evans JR ,
      4. Ban M , et al
      . Glucocerebrosidase mutations influence the natural history of Parkinson's disease in a community-based incident cohort. Brain 2013;136:392–9.doi:10.1093/brain/aws318 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23413260
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      2. Kumar KR ,
      3. Ramirez A ,
      4. Göbel A , et al
      . Glucocerebrosidase mutations in a Serbian Parkinson's disease population. Eur J Neurol 2013;20:402–5.doi:10.1111/j.1468-1331.2012.03817.x pmid:http://www.ncbi.nlm.nih.gov/pubmed/22812582
      OpenUrlCrossRefPubMed
    3. ↵
      2. Bras J ,
      3. Paisan-Ruiz C ,
      4. Guerreiro R , et al
      . Complete screening for glucocerebrosidase mutations in Parkinson disease patients from Portugal. Neurobiol Aging 2009;30:1515–7.doi:10.1016/j.neurobiolaging.2007.11.016 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18160183
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      2. Sidransky E ,
      3. Nalls MA ,
      4. Aasly JO , et al
      . Multicenter analysis of glucocerebrosidase mutations in Parkinson's disease. N Engl J Med Overseas Ed 2009;361:1651–61.doi:10.1056/NEJMoa0901281 pmid:http://www.ncbi.nlm.nih.gov/pubmed/19846850
      OpenUrlPubMed
    5. ↵
      2. Hruska KS ,
      3. LaMarca ME ,
      4. Scott CR , et al
      . Gaucher disease: mutation and polymorphism spectrum in the glucocerebrosidase gene (GBA). Hum Mutat 2008;29:567–83.doi:10.1002/humu.20676 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18338393
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      2. Lesage S ,
      3. Anheim M ,
      4. Condroyer C , et al
      . Large-scale screening of the Gaucher's disease-related glucocerebrosidase gene in Europeans with Parkinson's disease. Hum Mol Genet 2011;20:202–10.doi:10.1093/hmg/ddq454 pmid:http://www.ncbi.nlm.nih.gov/pubmed/20947659
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      2. Park JK ,
      3. Tayebi N ,
      4. Stubblefield BK , et al
      . The E326K mutation and Gaucher disease: mutation or polymorphism? Clin Genet 2002;61:32–4.doi:10.1034/j.1399-0004.2002.610106.x pmid:http://www.ncbi.nlm.nih.gov/pubmed/11903352
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      2. Mata IF ,
      3. Leverenz JB ,
      4. Weintraub D , et al
      . Gba variants are associated with a distinct pattern of cognitive deficits in Parkinson's disease. Mov Disord 2016;31:95–102.doi:10.1002/mds.26359 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26296077
      OpenUrlCrossRefPubMed
    9. ↵
      2. Berge-Seidl V ,
      3. Pihlstrøm L ,
      4. Maple-Grødem J , et al
      . The GBA variant E326K is associated with Parkinson's disease and explains a genome-wide association signal. Neurosci Lett 2017;658:48–52.doi:10.1016/j.neulet.2017.08.040 pmid:http://www.ncbi.nlm.nih.gov/pubmed/28830825
      OpenUrlCrossRefPubMed
    10. ↵
      2. Brockmann K ,
      3. Srulijes K ,
      4. Pflederer S , et al
      . GBA-associated Parkinson's disease: reduced survival and more rapid progression in a prospective longitudinal study. Mov Disord 2015;30:407–11.doi:10.1002/mds.26071 pmid:http://www.ncbi.nlm.nih.gov/pubmed/25448271
      OpenUrlCrossRefPubMed
    11. ↵
      2. Clark LN ,
      3. Ross BM ,
      4. Wang Y , et al
      . Mutations in the glucocerebrosidase gene are associated with early-onset Parkinson disease. Neurology 2007;69:1270–7.doi:10.1212/01.wnl.0000276989.17578.02 pmid:http://www.ncbi.nlm.nih.gov/pubmed/17875915
      OpenUrlCrossRefPubMed
    12. ↵
      2. Iwaki H ,
      3. Blauwendraat C ,
      4. Leonard HL , et al
      . Genetic risk of Parkinson disease and progression:: an analysis of 13 longitudinal cohorts. Neurol Genet 2019;5:e348.doi:10.1212/NXG.0000000000000348 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31404238
      OpenUrlAbstract/FREE Full Text
    13. ↵
      2. Iwaki H ,
      3. Blauwendraat C ,
      4. Leonard HL , et al
      . Genomewide association study of Parkinson's disease clinical biomarkers in 12 longitudinal patients' cohorts. Mov Disord 2019;34:1839–50.doi:10.1002/mds.27845 pmid:http://www.ncbi.nlm.nih.gov/pubmed/31505070
      OpenUrlCrossRefPubMed
    14. ↵
      2. Liu G ,
      3. Boot B ,
      4. Locascio JJ , et al
      . Specifically neuropathic Gaucher's mutations accelerate cognitive decline in Parkinson's. Ann Neurol 2016;80:674–85.doi:10.1002/ana.24781 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27717005
      OpenUrlCrossRefPubMed
    15. ↵
      2. Setó-Salvia N ,
      3. Pagonabarraga J ,
      4. Houlden H , et al
      . Glucocerebrosidase mutations confer a greater risk of dementia during Parkinson's disease course. Mov Disord 2012;27:393–9.doi:10.1002/mds.24045 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22173904
      OpenUrlCrossRefPubMed
    16. ↵
      2. Cilia R ,
      3. Tunesi S ,
      4. Marotta G , et al
      . Survival and dementia in GBA-associated Parkinson's disease: the mutation matters. Ann Neurol 2016;80:662–73.doi:10.1002/ana.24777 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27632223
      OpenUrlCrossRefPubMed
    17. ↵
      2. Lunde KA ,
      3. Chung J ,
      4. Dalen I , et al
      . Association of glucocerebrosidase polymorphisms and mutations with dementia in incident Parkinson's disease. Alzheimers Dement 2018;14:1293–301.doi:10.1016/j.jalz.2018.04.006 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29792872
      OpenUrlCrossRefPubMed
    18. ↵
      2. Oeda T ,
      3. Umemura A ,
      4. Mori Y , et al
      . Impact of glucocerebrosidase mutations on motor and nonmotor complications in Parkinson's disease. Neurobiol Aging 2015;36:3306–13.doi:10.1016/j.neurobiolaging.2015.08.027 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26422360
      OpenUrlCrossRefPubMed
    19. ↵
      2. Davis AA ,
      3. Andruska KM ,
      4. Benitez BA , et al
      . Variants in GBA, SNCA, and MAPT influence Parkinson disease risk, age at onset, and progression. Neurobiol Aging 2016;37:e1-09.e7 doi:10.1016/j.neurobiolaging.2015.09.014 pmid:26601739
      OpenUrlPubMed
    20. ↵
      2. Davis MY ,
      3. Johnson CO ,
      4. Leverenz JB , et al
      . Association of GBA mutations and the E326K polymorphism with motor and cognitive progression in Parkinson disease. JAMA Neurol 2016;73:1217–24.doi:10.1001/jamaneurol.2016.2245 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27571329
      OpenUrlPubMed
    21. ↵
      2. Mallett V ,
      3. Ross JP ,
      4. Alcalay RN , et al
      . Gba p.T369M substitution in Parkinson disease: polymorphism or association? A meta-analysis. Neurol Genet 2016;2:e104.doi:10.1212/NXG.0000000000000104 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27648471
      OpenUrlAbstract/FREE Full Text
    22. ↵
      2. Duran R ,
      3. Mencacci NE ,
      4. Angeli AV , et al
      . The glucocerobrosidase E326K variant predisposes to Parkinson's disease, but does not cause Gaucher's disease. Mov Disord 2013;28:232–6.doi:10.1002/mds.25248 pmid:http://www.ncbi.nlm.nih.gov/pubmed/23225227
      OpenUrlCrossRefPubMed
    23. ↵
      2. Evans JR ,
      3. Cummins G ,
      4. Breen DP , et al
      . Comparative epidemiology of incident Parkinson's disease in Cambridgeshire, UK. J Neurol Neurosurg Psychiatry 2016;87:1034–6.doi:10.1136/jnnp-2015-312581 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26800710
      OpenUrlFREE Full Text
    24. ↵
      2. Goetz CG ,
      3. Stebbins GT ,
      4. Tilley BC
      . Calibration of unified Parkinson's disease rating scale scores to movement disorder Society-unified Parkinson's disease rating scale scores. Mov Disord 2012;27:1239–42.doi:10.1002/mds.25122 pmid:http://www.ncbi.nlm.nih.gov/pubmed/22886777
      OpenUrlCrossRefPubMed
    25. ↵
      2. Jankovic J ,
      3. McDermott M ,
      4. Carter J , et al
      . Variable expression of Parkinson's disease: a base-line analysis of the DATATOP cohort. The Parkinson Study Group. Neurology 1990;40:1529–34.doi:10.1212/WNL.40.10.1529 pmid:http://www.ncbi.nlm.nih.gov/pubmed/2215943
      OpenUrlCrossRefPubMed
    26. ↵
      2. Bien SA ,
      3. Wojcik GL ,
      4. Zubair N , et al
      . Strategies for enriching variant coverage in candidate disease loci on a multiethnic genotyping array. PLoS One 2016;11:e0167758.doi:10.1371/journal.pone.0167758 pmid:http://www.ncbi.nlm.nih.gov/pubmed/27973554
      OpenUrlCrossRefPubMed
    27. ↵
      2. Malek N ,
      3. Weil RS ,
      4. Bresner C , et al
      . Features of GBA-associated Parkinson's disease at presentation in the UK Tracking Parkinson's study. J Neurol Neurosurg Psychiatry 2018;89:702–9.doi:10.1136/jnnp-2017-317348 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29378790
      OpenUrlAbstract/FREE Full Text
    28. ↵
      2. Moors TE ,
      3. Paciotti S ,
      4. Ingrassia A , et al
      . Characterization of brain lysosomal activities in GBA-Related and sporadic Parkinson's disease and dementia with Lewy bodies. Mol Neurobiol 2019;56:1344–55.doi:10.1007/s12035-018-1090-0 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29948939
      OpenUrlCrossRefPubMed
    29. ↵
      2. Evans JR ,
      3. Mason SL ,
      4. Williams-Gray CH , et al
      . The natural history of treated Parkinson's disease in an incident, community based cohort. J Neurol Neurosurg Psychiatry 2011;82:1112–8.doi:10.1136/jnnp.2011.240366 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21593513
      OpenUrlAbstract/FREE Full Text
    30. ↵
      2. Ruskey JA ,
      3. Greenbaum L ,
      4. Roncière L , et al
      . Increased yield of full GBA sequencing in Ashkenazi Jews with Parkinson's disease. Eur J Med Genet 2019;62:65–9.doi:10.1016/j.ejmg.2018.05.005 pmid:http://www.ncbi.nlm.nih.gov/pubmed/29842932
      OpenUrlCrossRefPubMed
    31. ↵
      2. Nichols WC ,
      3. Pankratz N ,
      4. Marek DK , et al
      . Mutations in GBA are associated with familial Parkinson disease susceptibility and age at onset. Neurology 2009;72:310–6.doi:10.1212/01.wnl.0000327823.81237.d1 pmid:http://www.ncbi.nlm.nih.gov/pubmed/18987351
      OpenUrlCrossRefPubMed
    32. ↵
      2. Mazzulli JR ,
      3. Xu Y-H ,
      4. Sun Y , et al
      . Gaucher disease glucocerebrosidase and α-synuclein form a bidirectional pathogenic loop in synucleinopathies. Cell 2011;146:37–52.doi:10.1016/j.cell.2011.06.001 pmid:http://www.ncbi.nlm.nih.gov/pubmed/21700325
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      2. Fernandes HJR ,
      3. Hartfield EM ,
      4. Christian HC , et al
      . Er stress and autophagic perturbations lead to elevated extracellular α-synuclein in GBA-N370S Parkinson's iPSC-derived dopamine neurons. Stem Cell Reports 2016;6:342–56.doi:10.1016/j.stemcr.2016.01.013 pmid:http://www.ncbi.nlm.nih.gov/pubmed/26905200
      OpenUrlPubMed

    Footnotes

    • RAB and CHW-G are joint senior authors.

    • Twitter @foltynie

    • Contributors TBS wrote the manuscript. TBS, MC, SW-R, GL, CS, TF, JE and DPB performed the research. RAB and CHW-G conceived the ideas for this study and provided critical review of the manuscript.

    • Funding The Cambridgeshire Incidence of Parkinson’s disease from General Practice to Neurologist (CamPaIGN) study has been supported by funding from the Wellcome Trust, the Medical Research Council, Parkinson’s UK and the Patrick Berthoud Trust. The Parkinsonism: Incidence, Cognition and Non-motor heterogeneity in Cambridgeshire (PICNICS) study has been supported by funding from the Cure Parkinson’s Trust, the Van Geest Foundation, the MRC and Parkinson’s UK. This work has also been supported by the National Institute for Health Research (NIHR) Cambridge Biomedical Research Centre Dementia and Neurodegeneration Theme (Grant Reference Number 146281). TBS received financial support from the Cure Parkinson’s Trust. CS's work is supported in part by National Institutes of Health grants R01AG057331, U01NS100603 and R01AG057331 and the American Parkinson Disease Association. Illumina MEGA ChIP genotyping was made possible by a philanthropic investment from Dooley LLC (to Brigham & Women's Hospital and CRS). TF has received grants from NIHR, Michael J Fox Foundation, Innovate UK, John Black Charitable Foundation, Van Andel Research Institute and Cure Parkinson’s Trust. DPB is supported by a Wellcome Clinical Research Career Development Fellowship. RAB is an National Institute for Health Research Senior Investigator (NF-SI-0616-10011) and is supported by the Wellcome Trust Stem Cell Institute (Cambridge 203151/Z/16/Z). CHW-G is supported by a RCUK/UKRI Research Innovation Fellowship awarded by the Medical Research Council (MR/R007446/1) and by the Cambridge Centre for Parkinson-Plus.

    • Competing interests None declared.

    • Patient consent for publication Not required.

    • Ethics approval The CamPaIGN and PICNICS studies were both approved by the Cambridgeshire Local Research Ethics Committee.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. All data relating to participants in this study is stored securely at the John van Geest Centre for Brain Repair, University of Cambridge.