Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Oct 1.
Published in final edited form as: Mov Disord. 2020 Jul 14;35(10):1755–1764. doi: 10.1002/mds.28189

NON-STEROIDAL ANTI-INFLAMMATORY USE AND LRRK2 PARKINSON’S DISEASE PENETRANCE

Marta San Luciano 1, Caroline M Tanner 1,2, Cheryl Meng 1, Connie Marras 3, Samuel M Goldman 1,4, Anthony E Lang 3, Eduardo Tolosa 5, Birgitt Schüle 6, J William Langston 6,7, Alexis Brice 8, Jean-Christophe Corvol 8, Stefano Goldwurm 9, Christine Klein 10, Simone Brockman 11, Daniela Berg 12, Kathrin Brockmann 12, Joachim J Ferreira 13, Meriem Tazir 14, George D Mellick 15, Carolyn M Sue 16, Kazuko Hasegawa 17, Eng King Tan 18, Susan Bressman 19, Rachel Saunders-Pullman 19; Michael J Fox Foundation LRRK2 Cohort Consortium
PMCID: PMC7572560  NIHMSID: NIHMS1611810  PMID: 32662532

Abstract

Background

The penetrance of Leucine Rich Repeat Kinase 2 (LRRK2) mutations is incomplete and may be influenced by environmental and/or other genetic factors. Non-steroidal anti-inflammatory drugs (NSAIDs) are known to reduce inflammation and may lower Parkinson’s disease (PD) risk, but their role in LRRK2-associated PD is unknown.

Objectives

To evaluate the association of regular NSAID use and LRRK2 associated PD.

Methods

Symptomatic (“LRRK2-PD”) and asymptomatic (“LRRK2-nonPD”) participants with LRRK2 G2019S, R1441X or I2020T variants (“definitely pathogenic variant carriers”), or G2385R or R1628P variants (“risk variant carriers”) from two international cohorts provided information on regular ibuprofen and/or aspirin use (≥2 pills/week for ≥6 months) prior to index date (diagnosis date for PD, interview date for non-PD). Multivariate logistic regression was used to evaluate the relationship between regular NSAID use and PD for any NSAID, separately for ibuprofen and aspirin in all carriers and separately in pathogenic and risk variant groups.

Results

259 LRRK2-PD and 318 LRRK2-non-PD participants were enrolled. Regular NSAID use was associated with reduced odds of PD in the overall cohort (OR=0.34, 95%CI 0.21–0.57), and in both pathogenic and risk variant carriers (ORPathogenic=0.38, 95%CI 0.21–0.67; ORRiskVariant=0.19, 95%CI 0.04–0.99). Similar associations were observed for ibuprofen and aspirin separately (ORIbuprofen=0.19, 95%CI 0.07–0.50, ORAspirin=0.51, 95%CI 0.28–0.91).

Conclusions

Regular NSAID use may be associated with reduced penetrance in LRRK2-associated PD. The LRRK2 protein is involved in inflammatory pathways and appears to be modulated by regular anti-inflammatory use. Longitudinal observational and interventional studies of NSAID exposure and LRRK2-PD are needed to confirm this association.

INTRODUCTION

Parkinson’s disease (PD) represents the second most common neurodegenerative illness after Alzheimer’s disease and is characterized clinically by bradykinesia, rest tremor, rigidity and postural instability as well as a variety of non-motor features1, 2. The pathogenesis of PD is unknown although both genetic and environmental factors are known to play a role3, 4. The most common Mendelian genetic form of PD is caused by pathogenic variants in the Leucine Rich Repeat Kinase 2 (LRRK2) gene, which phenotypically closely resembles idiopathic PD5, 6. Pathogenic variants in this gene are inherited in an autosomal dominant fashion with reduced age-dependent lifetime penetrance. Although still debated and highly variable across different populations, penetrance is estimated at approximately 30% by age 807–9. The factors responsible for the decreased penetrance of these mutations are largely unknown. Several other variants in LRRK2 have also been associated with modestly increased risk of PD (odds ratio, OR,1.5–3) and are predominantly found in Asian populations10.

Inflammation has long been thought to have a role in the pathogenesis of PD at multiple levels in both sporadic and familial disease11. Neuroinflammatory responses may perpetuate degenerative neuronal injury, ultimately leading to PD. In large epidemiological studies, non-steroidal anti-inflammatory medication (NSAIDs) use, including both ibuprofen and aspirin, has been associated with lower PD risk12–14, although this has not been universally observed15–18. Since the LRRK2 protein is expressed in most immune cells, including microglia, and may influence inflammatory pathways19–21, we investigated whether regular NSAID use affected risk of PD in two well-characterized international LRRK2 cohorts.

METHODS

Participants

This study utilized phenotypic and environmental data collected from members of two large international cohorts of LRRK2 pathogenic or risk variant carriers: the Parkinson’s Disease Genetic and Environmental Modifiers (PD-GEM, 26 sites) and the Michael J Fox Foundation (MJFF) LRRK2 Cohort Consortium (MJFF-LCC cross-sectional cohort, 14 sites). For both cohorts, demographic information, PD diagnosis and age of diagnosis, clinical features, response to therapy, and family history of PD are available. Case report forms and standard operating procedures for MJFF-LCC can be found at https://www.michaeljfox.org/data-sets. Description of study cohorts and determination of LRRK2 status in PD-GEM are described in Marras et al22. Information regarding relatedness of participants belonging to the cohorts was available for approximately 30% of participants. Participants with LRRK2 G2019S, R1441X and I2020T mutations were defined as “Pathogenic Variant Carriers”23–25, and those with G2385R or R1628P as “Risk Variant Carriers”26. Those participants meeting criteria for PD were defined as symptomatic (“LRRK2-PD”) and those without PD as asymptomatic (“LRRK2-nonPD”). The breakdown of participant totals, LRRK2-PD and LRRK2-nonPD, age at exam, variant group, cohort and region of origin are listed in Supplementary Table 1. Only cohort sites that provided information on NSAID use were included for this analysis.

Definition of PD

PD status was defined in accordance with standard diagnostic criteria for PD (UK Brain Bank, NIH)1 using available investigator assessments and clinical information.

Environmental Exposure Data

In both study cohorts, environmental exposures were assessed by using the NSAID module of the Parkinson’s Disease Risk Factor Questionnaire (PD-RFQ), a validated set of questionnaires developed to provide a standard tool for general use in epidemiologic studies of PD (RFQ-U CRFs, Version 1.0 Epidemiology Working Group of the Collaborative Centers for PD Environmental Research, https://commondataelements.ninds.nih.gov/report-viewer/23354/Risk%20Factor%20Questionnaire%20(RFQ)%20-%20Physical%20Activity%20and%20Sleep). The PD-RFQ assesses life-long exposures including amount, duration and time of use in freestanding questionnaires, each covering a specific domain, such as smoking, caffeine and alcohol consumption. For this specific analysis, history of regular NSAID use was determined using the following questions:

  1. Have you ever regularly taken ibuprofen-based non-aspirin medications, that is, at least two pills per week for 6 months or longer? These include ibuprofen, Advil, Motrin, Nuprin, and others.

  2. Have you ever regularly taken aspirin, that is, at least two pills per week for 6 months or longer?

  3. Have you ever regularly taken other anti-inflammatory medications for pain, inflammation, or swelling, that is, at least two pills per week for 6 months or longer? [Please do NOT include use of Tylenol or acetaminophen, or narcotic pain relievers such as Vicodin or codeine or Demerol]

--IF YES: a) At what age (or in what year) did you start regularly taking X; b) At what age (or in what year) did you stop regularly taking X; c) On average, about how many pills per week did you take?

Regular use of any NSAID prior to index date was defined as ≥2 pills/week for ≥6 months. Smoking status, also evaluated with the PD-RFQ, was included in all the models as ever (≥one cigarette per day for ≥6 months) vs. never smoker.

Statistical Analysis

All statistical data analysis was performed using SPSS v21 (IBM SPSS Statistics, Armonk, NY). Mean and standard deviation are presented for normally distributed continuous variables. Two-sample t-tests or non-parametric equivalent and Chi-square tests were used in unadjusted analyses. The relationship between NSAID use and PD was determined using logistic regression. Models were calculated for any regular NSAID use, and separately for ibuprofen or aspirin use. We considered use of NSAIDs prior to the diagnosis date in LRRK2-PD and prior to interview date in non-PD (index date). Analyses were performed for all genotypes combined and separately for pathogenic and risk variant groups. All logistic models were adjusted for index age (as a continuous variable), gender, and smoking history. Both race/ethnicity and region of origin in our dataset were highly correlated with specific LRRK2 variants, and adjusting for race/ethnicity and/or region of origin did not significantly alter the results of the models. For those models that included all genotypes combined, we adjusted for gene group (pathogenic or risk variant carriers) and included a product term for the interaction between smoking and gene group, in order to evaluate whether the effect of smoking would be greater in one genetic subpopulation than in another. To control for potential reverse causality (i.e., early undiagnosed disease causing increased usage of NSAIDs), exposure-lagging analysis was performed at 5 and 10 years prior to index date27.

RESULTS

Of the 577 participants enrolled, 259 (44.9%) were LRRK2-PD and 318 (55.1%) were LRRK2-nonPD. Demographic and clinical characteristics are summarized in Table 1. Participants with PD were older than those without PD (mean 66.2 years, standard deviation, SD, 11.5 vs. 60.9 years, (SD 14.9), p<0.001) and more likely to be men (51% vs. 38.7%, p=0.003). More than half of the participants carried the common G2019S pathogenic variant, including 189 (76.5%) LRRK2-PD and 126 (39.6%) LRRK2-nonPD. The specific region of origin of all carriers is listed in Supplementary Table 1. Approximately 70% were derived from the MJFF-LCC cohort and 30% from PD-GEM, and as expected, all risk variant carriers were from Asian sites (Supplementary Table 1). Among those with PD, age at onset did not differ between pathogenic and risk variant carriers, and clinical features at time of exam were comparable with the exception of presence of disease asymmetry and having a positive response to levodopa therapy, both of which were more common among pathogenic variant carriers (95.5 vs. 81.5% and 95.4 vs. 90.9%, respectively, p=0.001 for both, Supplementary Table 2).

Table 1.

Subject characteristics

Total LRRK2_PD LRRK2_nonPD
N N (%) % male N (%) % male
Subject totals* 577 259 51.0 318 38.7
Age (mean, SD, range)** 577 66.2 (11.5, 26–94) 60.9 (14.9, 18–89)
Age groups
 <50 142 71 (50) 53.5 71 (50.0) 40.8
 50–59 148 90 (60.8) 41.1 58 (39.2) 39.7
 60–69 154 69 (44.8) 50.7 85 (55.2) 36.5
 70–79 105 27 (25.7) 40.7 78 (74.3) 37.2
 >79 28 2 (7.1) 50.0 26 (92.9) 42.3
Gene group***
Pathogenic variants age** 401 66.3 (11.7, 26–94) 52.0 (14.1, 19–87)
Pathogenic variants
 Overall 401 232 (57.9) 49.1 169 (42.1) 41.4
 G2019S 315 189 (60.0) 49.7 126 (40.0) 40.5
 R1441 80 37 (46.3) 48.6 42 (53.8) 44.2
 I2020T 7 7 (100.0) 42.9 0 (0) 0.0
Risk variants age** 176 65.1 (9.7, 45–81) 70.9 (7.6, 52–89)
Risk variants
 Overall 176*** 27 (15.3) 63.0 149 (84.7) 35.6
 G2385R 114 24 (21.1) 66.7 90 (78.9) 38.9
 R1628P 64 5 (7.8) 40.0 59 (92.2) 30.5
Region
 Africaa 11 11 (100.0) 27.3 0 (0.0) 0.0
 Asiab 181 32 (17.7) 59.4 149 (82.3) 35.6
 Australia 10 8 (80.0) 62.5 2 (20.0) 0.0
 Europec 273 167 (61.2) 50.9 106 (38.8) 45.3
 North Americad 102 41 (40.2) 46.3 61 (59.8) 36.1
Open in a new tab
*

Test of gender distribution between PD and non-PD, chi square=8.579, p=0.03

**

Age at enrollment; test of age difference, p<0.05

***

Due to overlapping mutations the total count of variant carriers exceeds the total count of gene group

a:

all from Algeria

b:

China n=175, Japan n=4, Singapore n=2; b: Spain n=162, France n=30, Germany n=20, Italy n=23, Norway n=22, Portugal n=16

d:

USA n=76, Canada=26.

In the overall sample, rates of regular NSAID use were higher among LRRK2-nonPD than in LRRK2-PD (31.8% vs. 10.9%), including for both ibuprofen (7.3% vs. 2.4%) and aspirin (24.2% vs. 8.2%, Table 2). A similar pattern was observed when pathogenic and risk variants were analyzed separately, and when men were studied individually. Among women, rates of NSAID use were higher in LRRK2-nonPD in all groups with the exception of aspirin use in pathogenic variant carriers, which was higher among those with PD, although the sample size was small in this subgroup (Table 3).

Table 2.

Regular NSAID use by age, gene group and region

Any NSAID Ibuprofen Aspirin
LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD
Overall (n, %) 28 (10.9) 101 (31.8) 6 (2.37) 23 (7.26) 21 (8.20) 77 (24.21)
Age
 <50 3 (4.2) 9 (12.7) 0 (0.0) 8 (11.27) 2 (2.9) 3 (4.23)
 50–59 6 (6.7) 14 (24.1) 1 (1.12) 6 (10.34) 5 (5.6) 7 (12.07)
 60–69 14 (20.3) 32 (37.6) 4 (5.97) 4 (4.71) 10 (14.7) 25 (29.41)
 70–79 5 (18.5) 35 (44.9) 1 (4.0) 3 (3.85) 4 (14.8) 32 (41.03)
 >79 0 (0.0) 11 (42.3) 0 (0.0) 2 (8.0) 0 (0.0) 10 (38.46)
Gene group
Pathogenic variants carriers
 All 26 (11.2) 38 (22.5) 6 (2.7) 20 (11.9) 19 (8.3) 15 (8.9)
 G2019S 24 (12.7) 31 (24.6) 6 (3.3) 18 (14.4) 17 (9.1) 12 (9.5)
 R1441 2 (5.4) 7 (16.3) 0 (0.0) 2 (4.7) 2 (5.6) 3 (7.0)
 I2020T 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Risk variant carriers
 All 2 (7.1) 63 (42.3) 0 (0.0) 3 (2.0) 2 (7.1) 62 (41.6)
 G2385R 2 (8.3) 33 (36.7) 0 (0.0) 1 (1.1) 2 (8.3) 33 (36.7)
 R1628P 0 (0.0) 30 (50.8) 0 (0.0) 2 (3.4) 0 (0.0) 29 (49.2)
Region
 Asia 2 (6.3) 63 (42.3) 0 (0.0) 3 (2.0) 2 (6.3) 62 (41.6)
 Australia 1 (12.5) 0 (0.0) 0 (0.0) 0 (0.0) 1 (12.5) 0 (0.0)
 Europe 10 (6.0) 18 (17.0) 1 (0.6) 11 (10.5) 9 (5.4) 4 (3.8)
 North America 15 (36.6) 20 (32.8) 5 (12.8) 9 (14.8) 9 (22.5) 11 (18.0)
Open in a new tab

Table 3.

NSAID use by age and genetic mutation, by gender.

Any NSAID Ibuprofen Aspirin
Analysis restricted to men
LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD
Overall (n, %) 12 (9.5) 35 (28.5) 1 (0.8) 7 (5.7) 11 (8.53) 29 (23.6)
Gene Group
Pathogenic variants carriers
 All 11 (9.6) 17 (24.3) 1 (0.9) 6 (8.7) 10 (8.9) 11 (15.7)
 G2019S 10 (10.6) 12 (23.5) 1 (1.1) 5 (10.0) 9 (9.8) 8 (15.7)
 R1441X 1 (5.6) 5 (26.3) 0 (0.0) 1 (5.3) 1 (5.6) 3 (15.8)
 I2020T 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Risk variant carriers
 All 1 (5.9) 18 (34.0) 0 (0.0) 1 (1.9) 1 (5.9) 18 (34.0)
 G2385R 1 (6.3) 10 (28.6) 0 (0.0) 0 (0.0) 1 (6.3) 10 (28.6)
 R1628P 0 (0.0) 8 (44.4) 0 (0.0) 1 (5.6) 0 (0.0) 8 (44.4)
Analysis restricted to women
LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD LRRK2-PD LRRK2-nonPD
Overall (n, %) 16 (12.4) 66 (33.8) 5 (4.07) 16 (8.2) 10 (7.87) 48 (24.6)
Gene Group
Pathogenic variants carriers
 All 15 (12.7) 21 (21.2) 5 (4.4) 14 (14.1) 9 (7.7) 4 (4.0)
 G2019S 14 (14.7) 19 (25.3) 5 (5.6) 13 (17.3) 8 (8.4) 4 (5.3)
 R1441X 1 (5.3) 2 (8.3) 0 (0.0) 1 (4.2) 1 (5.6) 0 (0.0)
 I2020T 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0)
Risk variant carriers
 All 1 (9.1) 45 (46.9) 0 (0.0) 2 (2.1) 1 (9.0) 44 (45.8)
 G2385R 1 (12.5) 23 (41.8) 0 (0.0) 1 (1.8) 1 (12.5) 23 (41.8)
 R1628P 0 (0.0) 22 (53.7) 0 (0.0) 1 (2.4) 0 (0.0) 21 (51.2)
Open in a new tab

In the adjusted logistic regression models, regular NSAID use was associated with lower risk for PD (ORany NSAID=0.34, 95%CI: 0.21–0.57, Table 4). Both aspirin and ibuprofen were independently associated with lower risk of PD (ORIbuprofen=0.19, 95% CI: 0.07–0.50, ORASA=0.51, 95%CI: 0.28–0.91) and similar findings were seen in both pathogenic and risk variant carriers when analyzed separately. In gender-specific models adjusted for smoking and gene group (pathogenic or risk carrier), overall NSAID use was associated with lower PD risk, however, when broken down by NSAID type, the confidence intervals for the results for aspirin in women were wide and no longer statistically significant. In analyses stratified by age groups (<50, 50–69, >69 and ≤64, ≥65), overall NSAID use was associated with lower PD risk in all age groups, but when studied separately by NSAID, the results were only significant for ibuprofen in the younger group, likely reflecting increased uncertainty from small sample size (Table 5). The analysis was repeated separately in each cohort and the results were similar for each (MJFF-LCC ORany NSAID=0.26, 95%CI: 0.13–0.52; PD-GEM ORany NSAID=0.44, 95%CI: 0.19–1.02, Mantel-Haenszel test of homogeneity p-value= 0.59)

Table 4.

Logistic Regression Models

Overall Model OR (95%CI) 5y Lag Model OR (95%CI) 10y Lag Model OR (95%CI) Model for Men** OR (95%CI) Model for Women** OR (95%CI)
Overall*
Any NSAID 0.34 (0.21–0.57) 0.48 (0.27, 0.85) 0.90 (0.47,1.74) 0.28 (0.13–0.61) 0.42 (0.21–0.82)
Ibuprofen 0.19 (0.07–0.50) 0.25 (0.08, 0.72) 0.52 (0.16,1.71) 0.06 (0.01–0.50) 0.27 (0.09–0.78)
Aspirin 0.51 (0.28–0.91) 0.72 (0.37, 1.43) 1.57 (0.69,3.56) 0.41 (0.18–0.93) 0.77 (0.32–1.85)
Pathogenic variant carriers*
Any NSAID 0.38 (0.21–0.67) 0.51 (0.27, 0.98) 1.05 (0.49,2.28) 0.27 (0.11,0.65) 0.48 (0.23,1.03)
Ibuprofen 0.19 (0.07–0.50) 0.23 (0.08, 0.69) 0.49 (0.14,1.66) 0.07 (0.01,0.66) 0.25 (0.08,0.74)
Aspirin 0.79 (0.38–1.67) 1.11 (0.46, 2.69) 5.14 (1.11,23.82) 0.45 (0.17,1.20) 1.86 (0.54,6.42)
Risk variant carriers*
Any NSAID 0.19 (0.04–0.99) 0.17 (0.02, 1.66) 0.50 (0.04,5.85) 0.09 (0.01,1.07) 0.26 (0.03,2.75)
Ibuprofen 0.00 0.00 0.00 0.00 0.00
Aspirin 0.20 (0.04–1.00) 0.18 (0.02, 1.70) 0.53 (0.04,6.26) 0.09 (0.01,1.07) 0.27 (0.03,2.84)
Open in a new tab
*

Model adjusted for index age, gender, smoking and gene group

**

Model adjusted for index age and smoking

Table 5.

Logistic Regression Models by Age Group

Age Group <50* OR (95%CI) Age Group 50–69* OR (95%CI) Age Group >69* OR (95%CI) Age Group ≤ 64* OR (95%CI) Age Group ≥65* OR (95%CI)
Any NSAID 0.37 (0.09–1.48) 0.42 (0.22–0.81) 0.38 (0.00–0.31) 0.36 (0.19–0.67) 0.28 (0.10–0.76)
Ibuprofen 0 0.26 (0.08–0.86) 0.11 (0.00–1.87) 0.21 (0.7–0.60) 0.15 (0.01–1.68)
Aspirin 0.90 (0.14–5.95) 0.70 (0.33–1.52) 0.09 (0.01–0.60) 0.56 (0.26–1.22) 0.47 (0.16–1.38)
Open in a new tab
*

Model adjusted for gender and smoking

Results of a 5-year lagged model (Table 4) were very similar to the primary analysis (ORAny NSAID=0.48, 95%CI: 0.27–0.85). In the 10-year lag, confidence intervals were very wide and no longer significant, with the exception of aspirin, which suggested an increased risk (10-year lag ORASA=5.14, 95%CI: 1.11–23.82). However, the much smaller sample size in this subgroup precludes confident interpretations.

DISCUSSION

Results of this study suggest that regular use of NSAID medication may lower risk for PD among LRRK2 variant carriers, including pathogenic and risk variants, and for both ibuprofen and aspirin. This finding supports a growing body of research evidence involving inflammatory pathways in PD pathogenesis and data from prior epidemiological studies supporting a potential role for NSAIDs in lowering risk of sporadic PD.

Animal models as well as human epidemiological data support the importance of inflammatory mechanisms in PD at multiple levels. Activated microglia, inflammatory cytokines, and abnormalities in the complement system all appear to play significant roles28. Reactive microglia are observed in nigral pathology specimens in PD29, in humans and animals exposed to 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)29, 30 and in other PD animal models29, 31, 32. Such activated microglia may underlie the oxidative stress thought to be largely responsible for dopaminergic cell loss in PD31. Increased levels of the inflammatory cytokines IL-1b, IL-6 and TNFα, known to sustain inflammation and immune responses, as well as the complement component C-reactive protein, have been found in basal ganglia and cerebrospinal fluid of PD subjects33, and genetic variants affecting expression of certain cytokines have been linked to increased risk for neurodegenerative disorders, including PD34. Endogenous alpha-synuclein is involved in normal immune function35, and pathogenic alpha-synuclein directly induces both innate and adaptive immune responses in the central nervous system and in the periphery36. Pharmacologic immunosuppression and use of NSAIDs in rodent PD models also reduce neurodegeneration, highlighting aberrant neuroinflammation as a viable target for disease modification in PD37.

Epidemiological studies have shown that self-reported aspirin and non-aspirin NSAID use including ibuprofen may decrease risk for PD12–17. However, the findings from such studies are heterogeneous and not universal, since some studies did not find associations for aspirin14, or the protective effect of aspirin was only present in women13, or no association was found at all16. Several meta-analyses have evaluated these associations. A meta-analysis from 2010 suggested a protective effect for non-aspirin NSAID use on risk for PD estimated at ~15%12, although a Cochrane review in 2011 did not demonstrate sufficient evidence for the use of NSAIDs in PD prevention18. Among two more recent meta-analyses, one supported the association of non-aspirin NSAID on PD risk (pooled RR=0.91, 95%CI 0.84–0.99)38, but could not demonstrate a dose-response, and the other had a pooled RR in the same direction, but did not reach statistical significance (pooled RR=0.95, 95%CI 0.86–1.05)39. While it is likely that differences in methodology from the different studies underlie some of the heterogeneity in these results, pathophysiologic heterogeneity may also account for some of the variability. Studying genetically homogeneous sub-groups such as LRRK2 may provide additional insight.

The role of LRRK2 in inflammation raises the possibility that inflammation may be even more important in LRRK2-associated PD than in idiopathic PD21. Variants in the LRRK2 gene have been associated with Crohn’s disease and other chronic inflammatory conditions19, 20, suggesting potentially common genetic, molecular and cellular pathological pathways with PD. The LRRK2 protein is involved in inflammatory pathways and is highly expressed in monocytes, macrophages and microglia40. Within monocytes, LRRK2 may have a role in their maturation41. Exposure to certain antigens results in LRRK2 up-regulation and impaired autophagy in macrophages42. Lastly, the magnitude of the association between NSAID exposure and PD risk appears to be larger for LRRK2-PD. Pooled RR from meta-analyses of idiopathic PD range from 0.90–0.9512, 38, 39, and our observed OR of 0.34 would correspond to an RR of 0.42 (assuming a PD prevalence of 0.30 in non-exposed7–9), although the heterogeneity of study designs and populations make this comparison difficult.

The physiologic pathways on which NSAIDs exert their potentially neuroprotective effects have not been definitively determined. NSAIDs inhibit cyclooxygenase enzymes and hence may protect against reactive oxidative species and glutamate induced toxicity43. In this specific genetic population, since the LRRK2 protein is known to be involved in inflammatory pathways 19, 20, 40, 42 and impaired autophagy may result from LRRK2 protein upregulation, it is possible that modulation of this neuroinflammatory response by NSAIDs may modify the neurodegenerative effect of these mutations and variants. More recently, specific peripheral immune profiles were able to discriminate between three different clinical severity phenotypes in LRRK2-PD, suggesting that certain pro-inflammatory proteins could predict disease progression44. Peripheral inflammation has also been detected in asymptomatic G2019S carriers and inflammatory markers can discriminate between idiopathic PD and LRRK2-PD45.

PD is a complex disease believed to result from the combined effect of environmental and genetic determinants and while largely unknown, concrete evidence of this interaction is starting to arise46, 47. While gene-gene and gene-environment interactions are thought to underlie the incomplete and relatively low penetrance of LRRK2 mutations and variants, as well as its phenotypic heterogeneity, to date very few genetic or environmental factors have been associated with LRRK2-PD disease expression. Using data from a GWAS replication genotyping in a large case-control study with prospectively collected environmental information, Gao et al. examined the potential interaction between PD SNPs from several genes and HLA loci and smoking and caffeine intake47. They found a single significant interaction between a SNP near the LRRK2 gene and combined smoking and caffeine intake, however, known LRRK2 variants were not genotyped. In a recent case control study from Israel, cigarette smoking and increase coffee and tea intake were associated with older age of onset of LRRK2-PD48. Lastly, unaffected LRRK2 mutation carriers with higher urate levels have been found to have a reduced risk of developing PD 49, adding to prior evidence suggesting a protective effect of urate in idiopathic PD50. While still unknown, the modulation of urate levels on LRRK2-PD could be related to neuroinflammation, since the urate-activated Nrf2 antioxidant pathway likely plays a role in LRRK2 pathogenesis51–53.

There are several limitations to this study. First, there are innate limitations of a case-control study design, including the potential for selection and recall bias. NSAID dosage information used in this analysis was obtained retrospectively via questionnaires, and the variables necessary to perform a reliable dose-response analysis could not be analyzed separately due to insufficient data. Since the questionnaires did not capture the indication for NSAID use, we cannot explore the possibility of confounding by indication. However, if an indication for NSAID use (such as chronic inflammatory conditions19 or metabolic syndrome/cardiovascular risk factors54 for aspirin) were also a cause of PD, one would expect the association with NSAIDs to be in the opposite direction (that is, higher NSAID use among manifesting carriers).

Since pain is a common indication for NSAID use, hyperuricemia is associated with gouty arthritis, and higher urate levels are associated with reduced PD risk, there is a question as to whether urate levels may be confounding the NSAID-PD association. However, the majority (>75%) of individuals with hyperuricemia (levels >7–8mg/dL) are entirely asymptomatic55, and the incidence of gouty symptoms among the many hyperuricemic individuals with levels between 7–8mg/dL is very low56. Urate levels reported in unaffected LRRK2 individuals had means below 7mg/dL and many below 6mg/dL. It would hence be unlikely that asymptomatic individuals within those urate levels would have higher rates of NSAID usage. Another potential confounder that could explain the findings would be related to stoicism as part of the ‘premorbid personality’ described in PD leading to unwillingness to take NSAIDs for pain; however, while there is some evidence for higher rates of stoicism, harm avoidance and less novelty seeking behavior among PD patients57,58, there is little data to support that such personality traits would necessarily result in lessened analgesic intake, especially preceding a diagnosis of Parkinson’s disease. Results in lagged analyses additionally suggest that observed results were not due to differential use of NSAIDs by persons with early undiagnosed PD, since pain is commonly reported prior to PD motor symptom onset59. Also, as stated above, a reverse causation would be more likely in the opposite direction than the inverse relationship we observed. Another potential limitation is that case and control selection and recruitment methods differed between the centers, and not all centers recruited non-PD carrier controls (Supplementary Table 1). However, when the analysis was limited to each cohort separately, the results were similar for both cohorts even though the results limited to the smaller PD-GEM cohort were no longer statistically significant. Unaffected LRRK2 variant carriers represent first and second-degree relatives of cases in many centers. Very limited information was available on relatedness of subjects from the cohorts, which precluded analytic adjustment. However, when limiting the analysis to those participants who were known to be unrelated (a subset of PD-GEM), this subcohort showed an association in the same direction as our main analysis, though the confidence intervals were wide (unrelated cohort ORAny NSAID=0.11, 95%CI: 0.01–1.14).

Our study has significant strengths, including the large numbers of manifesting and non-manifesting LRRK2 carriers of both known pathogenic mutations and common risk variants from well-characterized international cohorts, as well as the use of validated questionnaires on NSAIDs exposure.

In summary, the results of this investigation in two large and well-characterized LRRK2 cohorts supports the hypothesis that both aspirin and ibuprofen may reduce the risk of PD manifestation. The optimal dose and duration of exposure, and the exact mechanisms by which protection may be exerted are however unknown. Anti-inflammatory drugs may potentially be useful as disease-modifying treatments in LRRK2 PD, and the ability to identify an at-risk population makes interventions in this subgroup particularly feasible. Lastly, since studies of LRRK2 kinase inhibitors are already underway, the results of this study highlight the potential need to control for the use of concomitant anti-inflammatory medications. However, prospective observational and interventional studies are needed to establish definitively whether or not this relationship is causal.

Supplementary Material

Supplementary Tables

ACKNOWLEDGEMENT

Study supported by the Michael J Fox Foundation (CT, CM, CLM, SG, BS), the Smart Family Foundation (MS) and NIH/NINDS K23NS0099441-O1A (MSL=). The LRRK2 Cohort Consortium is coordinated and funded by the Michael J. Fox Foundation for Parkinson’s Research. CK is supported by the DFG (FOR2488). We thank Aaron Viser for his assistance in reviewing and preparing the manuscript and tables for publication.

Funding sources for study: Study supported by the Michael J Fox Foundation (CT, CM, CLM, SG, BS), the Smart Family Foundation (MLS) and NIH/NINDS K23NS0099441-O1A (MS). The LRRK2 Cohort Consortium is coordinated and funded by the Michael J. Fox Foundation for Parkinson’s Research. CK is supported by the DFG (FOR2488).

Footnotes

CONFLICTS OF INTEREST

None of the authors had conflicts of interest to report in regard to this manuscript.

Financial Disclosure/Conflict of Interest: None of the authors had conflicts of interest to report in regard to this manuscript

REFERENCES

  • 1.Hughes AJ, Daniel SE, Blankson S, Lees AJ. A clinicopathologic study of 100 cases of Parkinson’s disease. Arch Neurol 1993;50:140–148. [DOI] [PubMed] [Google Scholar]
  • 2.Poewe W, Seppi K, Tanner CM, et al. Parkinson disease. Nat Rev Dis Primers 2017;3:17013. [DOI] [PubMed] [Google Scholar]
  • 3.Goldman SM. Environmental toxins and Parkinson’s disease. Annu Rev Pharmacol Toxicol 2014;54:141–164. [DOI] [PubMed] [Google Scholar]
  • 4.Lill CM, Roehr JT, McQueen MB, et al. Comprehensive research synopsis and systematic meta-analyses in Parkinson’s disease genetics: The PDGene database. PLoS Genet 2012;8:e1002548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Klein C, Schlossmacher MG. The genetics of Parkinson disease: Implications for neurological care. Nat Clin Pract Neurol 2006;2:136–146. [DOI] [PubMed] [Google Scholar]
  • 6.Trinh J, Zeldenrust FMJ, Huang J, et al. Genotype-phenotype relations for the Parkinson’s disease genes SNCA, LRRK2, VPS35: MDSGene systematic review. Mov Disord 2018;33:1857–1870. [DOI] [PubMed] [Google Scholar]
  • 7.Healy DG, Falchi M, O’Sullivan SS, et al. Phenotype, genotype, and worldwide genetic penetrance of LRRK2-associated Parkinson’s disease: a case-control study. Lancet Neurol 2008;7:583–590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Clark LN, Wang Y, Karlins E, et al. Frequency of LRRK2 mutations in early- and late-onset Parkinson disease. Neurology 2006;67:1786–1791. [DOI] [PubMed] [Google Scholar]
  • 9.Troiano AR, Elbaz A, Lohmann E, et al. Low disease risk in relatives of north african lrrk2 Parkinson disease patients. Neurology 2010;75:1118–1119. [DOI] [PubMed] [Google Scholar]
  • 10.Tan EK. The role of common genetic risk variants in Parkinson disease. Clin Genet 2007;72:387–393. [DOI] [PubMed] [Google Scholar]
  • 11.Deleidi M, Gasser T. The role of inflammation in sporadic and familial Parkinson’s disease. Cell Mol Life Sci 2013;70:4259–4273. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Gagne JJ, Power MC. Anti-inflammatory drugs and risk of Parkinson disease: a meta-analysis. Neurology 2010;74:995–1002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Wahner AD, Bronstein JM, Bordelon YM, Ritz B. Nonsteroidal anti-inflammatory drugs may protect against Parkinson disease. Neurology 2007;69:1836–1842. [DOI] [PubMed] [Google Scholar]
  • 14.Gao X, Chen H, Schwarzschild MA, Ascherio A. Use of ibuprofen and risk of Parkinson disease. Neurology 2011;76:863–869. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Chen H, Jacobs E, Schwarzschild MA, et al. Nonsteroidal antiinflammatory drug use and the risk for Parkinson’s disease. Ann Neurol 2005;58:963–967. [DOI] [PubMed] [Google Scholar]
  • 16.Ton TG, Heckbert SR, Longstreth WT Jr., et al. Nonsteroidal anti-inflammatory drugs and risk of Parkinson’s disease. Mov Disord 2006;21:964–969. [DOI] [PubMed] [Google Scholar]
  • 17.Bornebroek M, de Lau LM, Haag MD, et al. Nonsteroidal anti-inflammatory drugs and the risk of Parkinson disease. Neuroepidemiology 2007;28:193–196. [DOI] [PubMed] [Google Scholar]
  • 18.Rees K, Stowe R, Patel S, et al. Non-steroidal anti-inflammatory drugs as disease-modifying agents for Parkinson’s disease: evidence from observational studies. Cochrane Database Syst Rev 2011:CD008454. [DOI] [PubMed] [Google Scholar]
  • 19.Dzamko NL. LRRK2 and the Immune System. Adv Neurobiol 2017;14:123–143. [DOI] [PubMed] [Google Scholar]
  • 20.Wallings RL, Tansey MG. LRRK2 regulation of immune-pathways and inflammatory disease. Biochem Soc Trans 2019;47:1581–1595. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Kluss JH, Mamais A, Cookson MR. LRRK2 links genetic and sporadic Parkinson’s disease. Biochem Soc Trans 2019;47:651–661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Marras C, Alcalay RN, Caspell-Garcia C, et al. Motor and nonmotor heterogeneity of LRRK2-related and idiopathic Parkinson’s disease. Mov Disord 2016;31:1192–1202. [DOI] [PubMed] [Google Scholar]
  • 23.Simon-Sanchez J, Marti-Masso JF, Sanchez-Mut JV, et al. Parkinson’s disease due to the R1441G mutation in Dardarin: a founder effect in the Basques. Mov Disord 2006;21:1954–1959. [DOI] [PubMed] [Google Scholar]
  • 24.Lewis PA, Greggio E, Beilina A, Jain S, Baker A, Cookson MR. The R1441C mutation of LRRK2 disrupts GTP hydrolysis. Biochem Biophys Res Commun 2007;357:668–671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Gloeckner CJ, Kinkl N, Schumacher A, et al. The Parkinson disease causing LRRK2 mutation I2020T is associated with increased kinase activity. Hum Mol Genet 2006;15:223–232. [DOI] [PubMed] [Google Scholar]
  • 26.Tan EK, Peng R, Teo YY, et al. Multiple LRRK2 variants modulate risk of Parkinson disease: a Chinese multicenter study. Hum Mutat 2010;31:561–568. [DOI] [PubMed] [Google Scholar]
  • 27.Checkoway H, Pearce N, Hickey JL, Dement JM. Latency analysis in occupational epidemiology. Arch Environ Health 1990;45:95–100. [DOI] [PubMed] [Google Scholar]
  • 28.McGeer PL, McGeer EG. Inflammation and neurodegeneration in Parkinson’s disease. Parkinsonism Relat Disord 2004;10 Suppl 1:S3–7. [DOI] [PubMed] [Google Scholar]
  • 29.Langston JW, Forno LS, Tetrud J, Reeves AG, Kaplan JA, Karluk D. Evidence of active nerve cell degeneration in the substantia nigra of humans years after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine exposure. Ann Neurol 1999;46:598–605. [DOI] [PubMed] [Google Scholar]
  • 30.Gao H-M, Liu B, Zhang W, Hong J-S. Critical role of microglial NADPH oxidase-derived free radicals in the in vitro MPTP model of Parkinson’s disease. The FASEB Journal 2003;17:1–22. [DOI] [PubMed] [Google Scholar]
  • 31.McGeer PL, Schwab C, Parent A, Doudet D. Presence of reactive microglia in monkey substantia nigra years after 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine administration. Ann Neurol 2003;54:599–604. [DOI] [PubMed] [Google Scholar]
  • 32.Cicchetti F, Brownell AL, Williams K, Chen YI, Livni E, Isacson O. Neuroinflammation of the nigrostriatal pathway during progressive 6-OHDA dopamine degeneration in rats monitored by immunohistochemistry and PET imaging. European Journal of Neuroscience 2002;15:991–998. [DOI] [PubMed] [Google Scholar]
  • 33.McGeer PL, Yasojima K, McGeer EG. Inflammation in Parkinson’s disease. Adv Neurol 2001;86:83–89. [PubMed] [Google Scholar]
  • 34.Schulte T, Schols L, Muller T, Woitalla D, Berger K, Kruger R. Polymorphisms in the interleukin-1 alpha and beta genes and the risk for Parkinson’s disease. Neurosci Lett 2002;326:70–72. [DOI] [PubMed] [Google Scholar]
  • 35.Xiao W, Shameli A, Harding CV, Meyerson HJ, Maitta RW. Late stages of hematopoiesis and B cell lymphopoiesis are regulated by α-synuclein, a key player in Parkinson’s disease. Immunobiology 2014;219:836–844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Allen Reish HE, Standaert DG. Role of α-synuclein in inducing innate and adaptive immunity in Parkinson disease. Journal of Parkinson’s disease U6 - ctx_ver=Z3988–2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummonserialssolutionscom&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rftgenre=article&rftatitle=Role+of+%CE%B1-synuclein+in+inducing+innate+and+adaptive+immunity+in+Parkinson+disease&rftjtitle=Journal+of+Parkinson%27s+disease&rftau=Allen+Reish%2C+Heather+E&rftau=Standaert%2C+David+G&rftdate=2015–01-01&rfteissn=1877–718X&rftvolume=5&rftissue=1&rftspage=1&rft_id=info%3Apmid%2F25588354&rft_id=info%3Apmid%2F25588354&rftexternalDocID=25588354&paramdict=en-US U7 - Journal Article 2015;5:1. [Google Scholar]
  • 37.Hirsch EC, Hunot S. Neuroinflammation in Parkinson’s disease: a target for neuroprotection? The Lancet Neurology 2009;8:382–397. [DOI] [PubMed] [Google Scholar]
  • 38.Ren L, Yi J, Yang J, Li P, Cheng X, Mao P. Nonsteroidal anti-inflammatory drugs use and risk of Parkinson disease: A dose-response meta-analysis. Medicine (Baltimore) 2018;97:e12172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Poly TN, Islam MMR, Yang HC, Li YJ. Non-steroidal anti-inflammatory drugs and risk of Parkinson’s disease in the elderly population: a meta-analysis. Eur J Clin Pharmacol 2019;75:99–108. [DOI] [PubMed] [Google Scholar]
  • 40.Hakimi M, Selvanantham T, Swinton E, et al. Parkinson’s disease-linked LRRK2 is expressed in circulating and tissue immune cells and upregulated following recognition of microbial structures. J Neural Transm (Vienna) 2011;118:795–808. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Thevenet J, Pescini Gobert R, Hooft van Huijsduijnen R, Wiessner C, Sagot YJ. Regulation of LRRK2 expression points to a functional role in human monocyte maturation. PLoS One 2011;6:e21519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Schapansky J, Nardozzi JD, LaVoie MJ. The complex relationships between microglia, alpha-synuclein, and LRRK2 in Parkinson’s disease. Neuroscience 2015;302:74–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Hirsch EC, Hunot S. Neuroinflammation in Parkinson’s disease: a target for neuroprotection? Lancet Neurol 2009;8:382–397. [DOI] [PubMed] [Google Scholar]
  • 44.Brockmann K, Schulte C, Schneiderhan-Marra N, et al. Inflammatory profile discriminates clinical subtypes in LRRK2-associated Parkinson’s disease. Eur J Neurol 2017;24:427–e426. [DOI] [PubMed] [Google Scholar]
  • 45.Dzamko N, Rowe DB, Halliday GM. Increased peripheral inflammation in asymptomatic leucine-rich repeat kinase 2 mutation carriers. Mov Disord 2016;31:889–897. [DOI] [PubMed] [Google Scholar]
  • 46.Tanner CM, Goldman SM, Ross GW, Grate SJ. The disease intersection of susceptibility and exposure: chemical exposures and neurodegenerative disease risk. Alzheimers Dement 2014;10:S213–225. [DOI] [PubMed] [Google Scholar]
  • 47.Gao J, Nalls MA, Shi M, et al. An exploratory analysis on gene-environment interactions for Parkinson disease. Neurobiol Aging 2012;33:2528 e2521–2526. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Yahalom G, Rigbi A, Israeli-Korn S, et al. Age at Onset of Parkinson’s Disease Among Ashkenazi Jewish Patients: Contribution of Environmental Factors, LRRK2 p.G2019S and GBA p.N370S Mutations. J Parkinsons Dis 2020. [DOI] [PubMed] [Google Scholar]
  • 49.Bakshi R, Macklin EA, Logan R, et al. Higher urate in LRRK2 mutation carriers resistant to Parkinson disease. Ann Neurol 2019;85:593–599. [DOI] [PubMed] [Google Scholar]
  • 50.Ascherio A, Schwarzschild MA. The epidemiology of Parkinson’s disease: risk factors and prevention. Lancet Neurol 2016;15:1257–1272. [DOI] [PubMed] [Google Scholar]
  • 51.Angeles DC, Ho P, Dymock BW, Lim KL, Zhou ZD, Tan EK. Antioxidants inhibit neuronal toxicity in Parkinson’s disease-linked LRRK2. Ann Clin Transl Neurol 2016;3:288–294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Bakshi R, Zhang H, Logan R, et al. Neuroprotective effects of urate are mediated by augmenting astrocytic glutathione synthesis and release. Neurobiol Dis 2015;82:574–579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Skibinski G, Hwang V, Ando DM, et al. Nrf2 mitigates LRRK2- and alpha-synuclein-induced neurodegeneration by modulating proteostasis. Proc Natl Acad Sci U S A 2017;114:1165–1170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54.Abbott RD, Ross GW, White LR, et al. Midlife adiposity and the future risk of Parkinson’s disease. Neurology 2002;59:1051–1057. [DOI] [PubMed] [Google Scholar]
  • 55.Campion EW, Glynn RJ, DeLabry LO. Asymptomatic hyperuricemia. Risks and consequences in the Normative Aging Study. Am J Med 1987;82:421–426. [DOI] [PubMed] [Google Scholar]
  • 56.Choi HK, Mount DB, Reginato AM, American College of P, American Physiological S. Pathogenesis of gout. Ann Intern Med 2005;143:499–516. [DOI] [PubMed] [Google Scholar]
  • 57.Menza MA, Golbe LI, Cody RA, Forman NE. Dopamine-related personality traits in Parkinson’s disease. Neurology 1993;43:505–508. [DOI] [PubMed] [Google Scholar]
  • 58.Santangelo G, Garramone F, Baiano C, et al. Personality and Parkinson’s disease: A meta-analysis. Parkinsonism and Related Disorders 2018;49:67–74. [DOI] [PubMed] [Google Scholar]
  • 59.Ha AD, Jankovic J. Pain in Parkinson’s disease. Movement Disorders 2012;27:485–491. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Tables
NIHMS1611810-supplement-Supplementary_Tables.docx (19.7KB, docx)

RESOURCES