Fig 2: The LRRK2 protective variant reduces LRRK2 levels and activity in cellular models. (A) Schematic overview of LRRK2 risk and protective variants associated with PD and Crohn’s disease (CD). N551K R1398H: protective variant associated with both PD and CD (in blue). N2081D: a risk variant associated with CD (in orange). (B,C) In HEK293T cells overexpressing different LRRK2 variants and Rab10, total LRRK2 and pT73 Rab10 levels were reduced with the LRRK2 N551K R1398H variant on its own or with PD-linked LRRK2 variants in the kinase and WD40 domains; n = 5–6. Data were normalized for protein concentration and then normalized to the median within the batch and to the wildtype group; shown as mean ± SEM with p values based on paired t-tests on log transformed data, without adjustment for multiple comparisons. (D,E) LRRK2 and pT73 Rab10 levels were reduced in LRRK2 N551K R1398H KI A549 cells. Three wildtype pooled cells and three clones of LRRK2 N551K R1398H KI A549 cells were used; n = 4 experiments. Data are displayed as MSD signals normalized for protein concentration and then normalized to the median within the batch and to the wildtype group and are shown as mean ± SEM with p values: one-way ANOVA with Sidak’s multiple comparison test. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

Fig 3: Development of LRRK2 and pT73 Rab10 assays for human blood-based biomarker analysis. (A) pS935 LRRK2 and total LRRK2 levels were measured with serial dilutions of lysates from human PBMCs. Each run denotes an independent experiment with a pooled sample from 4 to 5 donors. (B) pT73 Rab10 and total Rab10 levels were measured with serial dilutions of lysates from human PBMCs. (C) Inhibition of LRRK2 kinase activity decreased pS935 LRRK2 and pT73 Rab10 levels in a dose-dependent manner in PBMCs with ex vivo treatment of MLi-2 (1 h). PBMCs were from n = 3 donors. The MSD signals were normalized to the response to the lowest concentration of MLi2 (0.003 nM) as 100% within each donor. (D) pS935 and total LRRK2 levels measured with serial dilutions of samples of human whole blood. Each run denotes an independent experiment with a pooled sample from 4 to 5 donors. (E,F) Whole blood specimens from 12 Parkinson’s Disease patients (3 Female, 9 Male) and 6 Non-Parkinson’s controls (2 Female, 4 Male) from the 24-h Biofluid Sampling Study (provided by the Michael J. Fox Foundation) were analyzed. Biospecimens were donated by each subject at 11 time points over 26 h. (E) Total LRRK2 levels in whole blood from healthy subjects and PD patients were stable within subjects over the course of 24 h (average coefficient of variation (CV) = 7%; average maximum to minimum value within-subject fold change of 1.25), while displaying high variability between subjects (66% CV, 11.6-fold difference in maximum to minimum value). (F) pS935 LRRK2 levels measured in human whole blood samples varied to a greater extent than total LRRK2 within subjects (average CV = 17%, average within-subject maximum to minimum value fold change of 1.76), with less variability between subjects than that observed for total LRRK2 (36% CV, 3.8-fold difference in maximum to minimum value).

Fig 4: LRRK2 and pT73 Rab10 levels assessed in PD patients and healthy subjects. LRRK2 levels and activities were measured in PBMCs from PD patients and healthy subjects with and without LRRK2 G2019S variant provided by the LRRK2 Detection in PBMC Consortium32; n = 16 non-PD-LRRK2 G2019S carriers, n = 22 non-PD non-G2019S carriers, n = 46 PD non-G2019S carriers, n = 33 PD and G2019S carriers. pS935 and total LRRK2 levels were shown as ng/ml and were calculated based on a standard curve of recombinant LRRK2 protein that was analyzed along with the PBMC lysates as in-plate reference material. For each variable, an ANCOVA model was fit, with log2 (value) as response and terms for Cohort, G2019S status (and their interaction), Gender and Age. For each variable, a forward model selection step was performed to assess the usefulness of adjusting for Age or including an interaction term; p values were based on pairwise comparison between groups. (A) For pS935 LRRK2, p = 0.0002 between G2019S carriers and non-carriers in PD (Relative Geometric mean ± 95% confidence interval = 0.651 (0.524–0.810)); p = 0.0134 between PD and non-PD in G2019S carriers (relative geometric mean ± 95% confidence interval = 0.673 (0.492–0.920)); p = 0.0233 between G2019S carriers and non-carriers (Relative Geometric mean ± 95% confidence interval = 0.805 (0.660–0.970)). (B) For LRRK2, p = 0.0028 between G2019S carriers and non-carriers in PD (relative geometric mean ± 95% confidence interval = 0.688 (0.540–0.877)); p = 0.0396 between PD and non-PD in G2019S carriers (relative geometric mean ± 95% confidence interval = 0.693 (0.489–0.982)). (C,D) For pT73 Rab10/LRRK2, p = 0.0171 between G2019S carriers and non-carriers in PD and in non-PD (relative geometric mean ± 95% confidence interval = 1.44 (1.07–1.95)). Relative geometric means given are adjusted for age and sex. Data are shown as geometric mean ± 95% confidence interval (CI). *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.

Fig 5: LRRK2 levels and activity are abundant in mouse glia cells and human iPSC-derived microglia cells and are regulated by lysosomal stress and inflammatory stimuli. (A) LRRK2-dependent phosphorylation of Rab10 occurs in primary cultured cortical neurons, astrocytes, and microglia and is highest in astrocytes. Inhibition of LRRK2 kinase activity by MLi-2 (500 nM, 2 h) significantly reduced phosphorylation of Rab10 across primary cultured cortical brain cells as assessed by western blot analysis; n = 3. Data are shown as mean ± SEM with p values: two-way ANOVA with Sidak’s multiple comparison test. (B) pT73 Rab10 levels were increased and pS935 LRRK2 levels were decreased in primary mouse astrocytes from G2019S KI mice; n = 3. Data were normalized for protein concentration and normalized to the median within each batch and then to the control group, shown as mean ± SEM with p values based on paired t-test. (C) Human iPSC-derived microglia expressed high levels of LRRK2 and pT73 Rab10, and LRRK2 inhibition (MLi-2, 500 nM, 2 h) significantly reduced pT73 Rab10 levels; n = 4. Data are shown as MSD signals normalized for protein concentration and displayed as mean ± SEM with p values: paired t-test. (D,E) iMicroglia treated with pharmacological agents that induce lysosomal stress (50 µM chloroquine (CQ) for 24 h and 100 nM Bafilomycin A1 (Baf) for 24 h) resulted in increased pT73 Rab10 levels. LRRK2 levels were largely not affected, with only a mild increase observed with Bafilomycin A1 treatment; n = 5–8. (F,G) IFN-\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\gamma$$\end{document}γ treatment (20 ng/mL, 24 h) induced a significant increase in both LRRK2 and pT73 Rab10 levels in iMicroglia cells; n = 5–8. (D–G) MSD signals were normalized for protein concentration, and data were then normalized to the median within each batch and to the control group, shown as mean ± SEM with p values based on one-way ANOVA with Sidak’ multiple comparison test on log transformed data. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.