Research ArticleParkinson’s Disease

LRRK2 activation in idiopathic Parkinson’s disease

See allHide authors and affiliations

Science Translational Medicine  25 Jul 2018:
Vol. 10, Issue 451, eaar5429
DOI: 10.1126/scitranslmed.aar5429

Proximity matters

Abnormally increased kinase activity due to mutations in the leucine-rich repeat kinase 2 (LRRK2) gene is the cause of Parkinson’s disease (PD) in about 3 to 4% of patients with this neurodegenerative disorder. Di Maio and colleagues now show that in postmortem brain tissue from individuals with idiopathic PD (iPD), LRRK2 kinase activity is aberrantly increased in vulnerable nigrostriatal dopamine neurons. The authors propose that oxidative stress involving α-synuclein and mitochondrial impairment activated LRRK2 kinase activity and that this caused endolysosomal dysfunction and accumulation of phosphorylated α-synuclein in the brains of these individuals. Thus, activation of LRRK2 kinase activity, independent of mutations, contributes to pathogenesis in iPD, suggesting that LRRK2 kinase inhibitors may be useful for treating patients with iPD and PD patients carrying LRRK2 mutations.

Abstract

Missense mutations in leucine-rich repeat kinase 2 (LRRK2) cause familial Parkinson’s disease (PD). However, a potential role of wild-type LRRK2 in idiopathic PD (iPD) remains unclear. Here, we developed proximity ligation assays to assess Ser1292 phosphorylation of LRRK2 and, separately, the dissociation of 14-3-3 proteins from LRRK2. Using these proximity ligation assays, we show that wild-type LRRK2 kinase activity was selectively enhanced in substantia nigra dopamine neurons in postmortem brain tissue from patients with iPD and in two different rat models of the disease. We show that this occurred through an oxidative mechanism, resulting in phosphorylation of the LRRK2 substrate Rab10 and other downstream consequences including abnormalities in mitochondrial protein import and lysosomal function. Our study suggests that, independent of mutations, wild-type LRRK2 plays a role in iPD. LRRK2 kinase inhibitors may therefore be useful for treating patients with iPD who do not carry LRRK2 mutations.

INTRODUCTION

Parkinson’s disease (PD) is a common neurodegenerative disorder that results in motor impairment, cognitive and psychiatric symptoms, and autonomic dysfunction (1). Whereas a number of gene mutations are known to cause familial PD, about 90% of PD cases are of unknown cause, that is, idiopathic PD (iPD). Missense mutations in the gene encoding leucine-rich repeat kinase 2 (LRRK2) are the most common cause of autosomal dominant PD and may account for about 3% of cases overall (2, 3). The LRRK2 locus also contains a risk factor for iPD (4), but the role of LRRK2 in typical iPD is unknown. It is generally believed that the common missense mutations in LRRK2 confer a toxic gain of function, and increased LRRK2 kinase activity has been strongly implicated in PD pathogenesis (5). Among the kinase substrates of LRRK2 are a subset of the Rab GTPases (guanosine triphosphatases), including Rab10, which has been implicated in the maintenance of endoplasmic reticulum, vesicle trafficking, and autophagy (6). LRRK2-induced phosphorylation of Rab10 inhibits its function by preventing binding to Rab GDP (guanosine diphosphate) dissociation inhibitor factors necessary for membrane delivery and recycling. Hence, aberrantly enhanced LRRK2 kinase activity is likely to be associated with reduced activity of Rab10 and its effectors.

Assessment of the kinase activity state of LRRK2 under various conditions has been somewhat cumbersome, although there appears to be a growing consensus that autophosphorylation of LRRK2 at Ser1292 correlates with activity (7, 8). Phosphoserine1292 (pSer1292) has generally been detected by Western blotting rather than immunocytochemistry, which limits anatomical or cellular resolution. LRRK2 is believed to be a rather low-abundance protein, and efforts to detect it immunocytochemically sometimes rely on high antibody concentrations, which may reduce specificity. This problem may be compounded for pSer1292 when only a fraction of the total (small) pool of LRRK2 is phosphorylated. The activity of LRRK2 is also regulated by its interaction with 14-3-3 proteins, whose binding to LRRK2 requires phosphorylation at LRRK2 serine residues 910 and 935 (9, 10), which are not autophosphorylation sites. Although the binding of LRRK2 to 14-3-3 proteins is associated with reduced kinase activity (11), this interaction can be disrupted by oxidative stressors including hydrogen peroxide (H2O2 ) (12). The interaction between LRRK2 and 14-3-3 proteins has generally been assessed by coimmunoprecipitation. A critical barrier to understanding the role of LRRK2 in iPD has been the absence of a practical, sensitive, high-resolution assay for its activation state.

We have developed and validated proximity ligation assays with excellent anatomical resolution that can rapidly provide information regarding activation state, cellular localization, and physiological regulators of LRRK2. The assay is based on (i) Ser1292 phosphorylation and (ii) dissociation of 14-3-3 proteins from LRRK2.

RESULTS

New proximity ligation assays were developed and validated

As a LRRK2 autophosphorylation site, pSer1292 reflects the activity of LRRK2 per se. We developed a proximity ligation assay to amplify the signal and increase the specificity of an antibody recognizing the pSer1292 epitope of the LRRK2 protein. For the proximity ligation assay, we paired the anti-pSer1292 antibody with a validated antibody directed against an epitope in the C terminus of Roc (COR) domain of the LRRK2 protein. In this way, the signal of the anti-pSer1292 antibody was amplified and detected only if it was in proximity to the anti-COR domain antibody (that is, specific LRRK2 pSer1292 signals were amplified, whereas potential signals from nonspecific or off-target antibody binding were filtered out by the proximity ligation assay). We developed a second proximity ligation assay that assessed the interaction between LRRK2 and 14-3-3 proteins, whose binding to LRRK2 is associated with reduced kinase activity. The proximity ligation assays were designed such that LRRK2 activity would be associated with a strong pSer1292 signal and a weak 14-3-3 signal; conversely, when LRRK2 is inactive, there would be little pSer1292 signal and a robust 14-3-3 signal.

To validate the assays, we used wild-type LRRK2, mutant LRRK2 (LRRK2G2019S/G2019S), and LRRK2-deficient (LRRK2−/−) human embryonic kidney–293 (HEK-293) cells, obtained by CRISPR/Cas9 gene editing. In wild-type cells, there was little pSer1292 proximity ligation assay signal and a strong 14-3-3 signal (Fig. 1, A, B, D, and E). The G2019S mutation is known to cause increased LRRK2 kinase activity (5, 6). Accordingly, in LRRK2G2019S/G2019S HEK-293 cells, there was a bright pSer1292 signal and loss of the 14-3-3 interaction [P < 0.0001 for both compared to wild-type cells; analysis of variance (ANOVA) with Bonferroni correction]. No signal in either proximity ligation assay was seen in LRRK2−/− HEK-293 cells lacking LRRK2 (Fig. 1, A, B, D, and E), further establishing the specificity of the assays. The small GTPase, Rab10, has recently been shown to be directly phosphorylated at Thr73 by LRRK2 (6). Using a specific pThr73-Rab10 antibody, we found low amounts of phosphorylated Rab10 in wild-type cells and much higher amounts in LRRK2G2019S/G2019S cells (P < 0.0001 compared to wild-type cells), in keeping with increased kinase activity of the mutant protein (Fig. 1, A, C, and D). After assay development, validation included blinded analysis and correct identification of all three cell lines with these assays alone. Using the selective LRRK2 kinase inhibitors, GNE-7915 and MLi-2 (13), in dose-response studies, we found that LRRK2 activation state, assessed by the pSer1292 signal in the proximity ligation assay, correlated closely with phosphorylation of its substrate Rab10 (Fig. 1, F and G). We next looked at LRRK2 activation state in patient-derived lymphoblastoid cell lines. Relative to cells derived from a healthy age-matched control, there was marked elevation of the pSer1292 signal in a LRRK2 G2019S mutation carrier; titration with the selective LRRK2 kinase inhibitor, GNE-7915, dose-dependently reduced the pSer1292 proximity ligation signal and pThr73-Rab10 signal in parallel (Fig. 1H).

Fig. 1 Validation of proximity ligation assays using CRISPR/Cas9 gene-edited HEK-293 cells and LRRK2 kinase inhibitors.

(A) A proximity ligation (PL) assay showing LRRK2 kinase activation by means of phosphorylation of the autophosphorylation site Ser1292 (red signal) and immunofluorescence of phosphorylation of the LRRK2 substrate Rab10 (green signal). In wild-type HEK-293 cells (HEK wild type; top row), there was little proximity ligation signal or pThr73-Rab10 immunofluorescence. HEK-293 cells carrying a homozygous G2019S mutation in LRRK2 (HEK G2019S; middle row) showed elevated LRRK2 kinase activity, indicated by a bright pSer1292 proximity ligation signal and strong pThr73-Rab10 immunofluorescence. In HEK-293 cells lacking LRRK2 [HEK LRRK2 knockout (KO; bottom row)], there was no pSer1292 proximity ligation signal and very little pThr73-Rab10 signal. DAPI (4′,6-diamidino-2-phenylindole; blue) was used as a nuclear stain. (B) Quantification of the pSer1292 proximity ligation signal in wild-type, G2019S mutant, and knockout HEK-293 cells. Results reflect three independent experiments. Each symbol represents signal from a single cell. Statistical testing by ANOVA with post hoc Bonferroni correction. (C) Quantification of the pThr73-Rab10 signal in wild type, G2019S mutant, and knockout HEK-293 cells. Results reflect three independent experiments. Each symbol represents signal from a single cell. Statistical testing by ANOVA with post hoc Bonferroni correction. (D) Proximity ligation assay of 14-3-3 binding to LRRK2 and immunofluorescence of Rab10 phosphorylation at Thr73. In wild-type HEK-293 cells (top row), there was a strong 14-3-3–LRRK2 proximity ligation signal (red) and little pThr73-Rab10 immunofluorescence (green). In HEK-293 cells carrying a homozygous G2019S mutation in LRRK2 (middle row), there was loss of 14-3-3 binding and a marked increase in pThr73-Rab10 signal. In HEK-293 LRRK2 knockout cells (bottom row), there was no 14-3-3–LRRK2 signal and little pThr73-Rab10 signal. (E) Quantification of the 14-3-3–LRRK2 proximity ligation signal in HEK-293 wild type, G2019S mutant, and LRRK2 knockout cells. Results reflect three independent experiments. Each symbol represents signal from a single cell. Statistical testing by ANOVA with post hoc Bonferroni correction. (F) Dose-response curves for the LRRK2 kinase inhibitor GNE-7915 against the pSer1292 proximity ligation signal (filled circles) and the pThr73-Rab10 signal (open circles) in HEK-293 G2019S mutant cells. Cells were cultured for 24 hours with various LRRK2 kinase inhibitor concentrations. Results are from three independent experiments. Symbols show means ± SEM. IC50 values were calculated by GraphPad Prism software. (G) Dose-response curves for the LRRK2 kinase inhibitor MLi-2 against the pSer1292 proximity ligation signal (filled circles) and the pThr73-Rab10 signal (open circles) in HEK-293 G2019S mutant cells. Cells were cultured for 24 hours with various LRRK2 kinase inhibitor concentrations. (H) Dose-response curves for the LRRK2 kinase inhibitor GNE-7915 against the pSer1292 proximity ligation signal (filled circles) and the pThr73-Rab10 signal (open circles) in lymphoblastoid cells derived from an individual carrying the G2019S LRRK2 mutation. Cells were cultured for 24 hours with various LRRK2 kinase inhibitor concentrations.

Endogenous wild-type LRRK2 is activated in iPD

Conventional assays of LRRK2 activity often rely on overexpression, and they require substantial amounts of tissue, lack cellular/anatomical resolution, and cannot be performed in previously fixed tissue. In contrast, our proximity ligation assays could assess the activation state of endogenous LRRK2 on a cell-by-cell basis using fixed cells or tissue. In this context, we measured pSer1292 proximity ligation and pThr73-Rab10 by quantitative confocal immunofluorescence in sections of substantia nigra from postmortem brain tissue from seven individuals who had died with iPD and from eight controls matched for age and postmortem interval. In nigrostriatal dopamine neurons from healthy controls, there were very low basal levels of pSer1292 signal and pThr73-Rab10 immunoreactivity and a strong 14-3-3 proximity ligation signal (Fig. 2). In contrast, the remaining nigrostriatal dopamine neurons of the iPD cases showed about a sixfold increase in pSer1292 proximity ligation (P < 0.0002, two-tailed, unpaired t test; P < 0.002 with Welch’s correction for unequal variances), and this was associated with a fourfold increase in phosphorylation (pThr73) of the LRRK2 substrate Rab10 (P < 0.0002). The increase in LRRK2 activation state in iPD dopamine neurons corresponded to a fivefold decrease in the 14-3-3 proximity ligation signal (P < 0.0001; P < 0.0004 with Welch’s correction). This suggested that endogenous wild-type LRRK2 may be activated in dopamine neurons in iPD and that this activation was associated with increased substrate phosphorylation.

Fig. 2 Activation of LRRK2 kinase in nigrostriatal dopamine neurons in human iPD postmortem brain tissue.

(A) Shown are the pSer1292 proximity ligation signal (red) and pThr73-Rab10 immunofluorescence signal (gray) in sections of substantia nigra from a healthy, age-matched control human brain (top row) and a brain from an individual with iPD (bottom row). In the control brain, there was little pSer1292 or pThr73-Rab10 signal, but in the iPD brain, there were strong signals for both. TH, tyrosine hydroxylase, a marker of dopamine neurons (blue). (B) Quantification of pSer1292 proximity ligation signal in eight control brains and seven iPD brains. Statistical comparison by unpaired two-tailed t test. (C) Quantification of pThr73-Rab10 signal in eight control brains and seven iPD brains. Statistical comparison by unpaired two-tailed t test. (D) Shown are 14-3-3–LRRK2 proximity ligation signal (red) and pThr73-Rab10 immunofluorescence signal (gray) in sections of substantia nigra from a control human brain (top row) and a brain from an individual with iPD (bottom row). In the control brain, there was a strong 14-3-3–LRRK2 proximity ligation signal and little pThr73-Rab10 signal, but in the iPD brain, the opposite pattern was seen. (E) Quantification of 14-3-3–LRRK2 proximity ligation signal in eight control brains and seven iPD brains. Statistical comparison by unpaired two-tailed t test.

In addition to neuronal expression, LRRK2 is expressed by microglia (14). We found detectable levels of pSer1292 proximity ligation signal in nigral microglia in controls (fig. S1, A and B), and the signal was more than doubled in microglia from iPD cases (P < 0.0005, two-tailed, unpaired t test).

Endogenous wild-type LRRK2 activation is found in the rotenone (mitochondrial) and α-synuclein rat models of PD

Mitochondrial impairment and α-synuclein aggregation and accumulation have been strongly implicated in PD pathogenesis (15). Therefore, we tested whether LRRK2 was activated in two rat models of PD. First, we used the rotenone model of PD, which has been shown to reproduce or even predict pathological and pathogenic features of the disease (16). Using substantia nigra sections from the brains of rotenone-treated rats that had reached behavioral endpoint (severe parkinsonism after 10 to 14 days), we found a 10-fold increase in pSer1292 proximity ligation signal in nigrostriatal dopamine neurons compared to vehicle-treated control rat nigrostriatal dopamine neurons (P < 0.0001, unpaired, two-tailed t test) (Fig. 3, A to C). In these animals, there was a marked loss of the 14-3-3 proximity ligation signal (P < 0.0001, unpaired, two-tailed t test) similar to the changes we observed in human postmortem brain tissue from iPD patients. We found detectable pSer1292 proximity ligation signal in microglia in vehicle-treated control rat brain (fig. S1, C and D); the signal was more than doubled after rotenone treatment of rats (P < 0.05, unpaired, two-tailed t test). To determine whether LRRK2 activation occurs before neurodegeneration, we examined tissue from animals that had received only 1 or 5 days of rotenone treatment, time points at which we detect no degeneration of the nigrostriatal dopamine neurons (fig. S2). After a single dose of rotenone, the pSer1292 proximity ligation signal was increased fivefold relative to vehicle treatment (P < 0.0001, ANOVA with Bonferroni correction), and after five daily doses, the proximity ligation signal was increased sixfold (P < 0.0001).

Fig. 3 LRRK2 activation in nigrostriatal dopamine neurons in two rat models of PD.

(A) Shown are pSer1292 and 14-3-3–LRRK2 proximity ligation signals in the substantia nigra of the brains of rats treated with vehicle (control, top row) or the pesticide rotenone (bottom row). In the rotenone-treated rats, there was increased pSer1292 proximity ligation signal and loss of 14-3-3–LRRK2 proximity ligation signal, indicating LRRK2 activation. TH, tyrosine hydroxylase, a marker of dopamine neurons (blue). (B) Quantification of pSer1292 proximity ligation signal in nigrostriatal dopamine neurons from control vehicle- and rotenone-treated rats. Symbols represent individual animals. Statistical comparison by unpaired two-tailed t test. (C) Quantification of 14-3-3–LRRK2 proximity ligation signal in nigrostriatal dopamine neurons from control vehicle- and rotenone-treated rats. Symbols represent individual animals. Statistical comparison by unpaired two-tailed t test. (D) Shown are pSer1292 proximity ligation signal and 14-3-3–LRRK2 proximity ligation signal in the substantia nigra of the brains of rats that received a unilateral injection of AAV2-hSNCA into one brain hemisphere. In the hemisphere overexpressing α-synuclein (bottom row), there was increased pSer1292 proximity ligation signal and loss of 14-3-3–LRRK2 proximity ligation signal, indicating LRRK2 activation in nigrostriatal neurons compared to the hemisphere that was not injected (top row). (E) Quantification of pSer1292 proximity ligation signal in nigrostriatal dopamine neurons from the control and AAV-hSNCA–injected rat brain hemispheres. Symbols represent mean values from each hemisphere. Statistical comparison by paired two-tailed t test. (F) Quantification of 14-3-3–LRRK2 proximity ligation signal in nigrostriatal dopamine neurons from the control and AAV-hSNCA–injected rat brain hemispheres. Symbols represent mean values from each hemisphere. Statistical comparison by paired two-tailed t test.

α-Synuclein accumulation and Lewy body pathology are hallmarks of PD. Elevated wild-type α-synuclein may cause PD, and several groups have used viral vector–mediated overexpression of α-synuclein as a model of PD (17). Here, we used adeno-associated virus type 2 (AAV2)–mediated overexpression of wild-type human α-synuclein (hSNCA) injected unilaterally into the substantia nigra pars compacta of rats to induce slowly progressive neurodegeneration. Six weeks after vector injection when neurodegeneration was ongoing, the remaining nigrostriatal dopamine neurons showed a marked 10-fold increase in pSer1292 proximity ligation signal compared to the contralateral, uninjected hemisphere (P < 0.0001, paired two-tailed t test) (Fig. 3, D to F); there was a concomitant loss of the 14-3-3 proximity ligation signal (P < 0.0001).

Both rotenone treatment and AAV2-mediated α-synuclein overexpression lead to oligomerization of α-synuclein, as well as accumulation of Ser129-phosphorylated α-synuclein (18, 19). We recently identified soluble oligomers and pSer129-α-synuclein as specific forms of α-synuclein that have deleterious effects on mitochondrial protein import machinery and that cause mitochondrial impairment (15). In analogous fashion, when SNCA−/− HEK-293 cells were treated with exogenous soluble oligomers of α-synuclein (400 nM monomer equivalent) for 24 hours, there was a marked activation of endogenous wild-type LRRK2 with increased pSer1292 and loss of the 14-3-3 proximity ligation signal (P < 0.0001, ANOVA with Bonferroni correction; fig. S3). Treatment with monomeric α-synuclein did not activate LRRK2.

Both rotenone treatment and elevated α-synuclein increase formation of reactive oxygen species (ROS), and both insults activate wild-type LRRK2, which raises the possibility that it is secondary generation of ROS that actually activates LRRK2. To test directly whether ROS can activate LRRK2, we treated wild-type HEK-293 cells with H2O2 (Fig. 4, A and B). Treatment with H2O2 dose-dependently (50 nM to 5 μM) activated the pSer1292 proximity ligation signal (P < 0.0001 versus control for all H2O2 doses; one-way ANOVA with Bonferroni correction) and increased phosphorylation of its substrate Rab10 (P < 0.0001 for all doses above 50 nM H2O2). The antioxidant α-tocopherol blocked H2O2 activation of the pSer1292 signal (P < 0.0001) and Rab10 phosphorylation (P < 0.0001).

Fig. 4 LRRK2 is activated in HEK-293 cells by ROS.

(A) The pSer1292 proximity ligation signal is increased dose-dependently by H2O2 (blue symbols) in wild-type HEK-293 cells. This H2O2-induced increase was blocked by the antioxidant α-tocopherol (5 μM) (red symbols). Results represent three independent experiments. Symbols represent measurements from individual cells. Red asterisks denote P < 0.0001 versus no H2O2, ANOVA with Bonferroni correction; blue asterisks denote P < 0.0001 versus H2O2 alone at the same concentration. (B) pThr73-Rab10 signal was increased dose-dependently by H2O2 (blue symbols) in wild-type HEK-293 cells, and the H2O2-induced increase was blocked by the antioxidant α-tocopherol (5 μM) (red symbols). Results represent three independent experiments. Symbols represent measurements from individual cells. Red asterisks denote P < 0.0001 versus no H2O2, ANOVA with Bonferroni correction; blue asterisks denote P < 0.0001 versus H2O2 alone at the same concentration. ns, not significant. (C) In wild-type HEK-293 cells, rotenone treatment increased the pSer1292 proximity ligation signal and pThr73-Rab10 immunoreactivity. Both effects were blocked by the specific NOX2 inhibitor Nox2ds-tat. (D) Quantification of the pSer1292 proximity ligation signal in vehicle- and rotenone-treated cells. Results represent three independent experiments. Symbols represent measurements from individual cells. Comparison by ANOVA with Bonferroni correction. (E) Quantification of the pThr73-Rab10 immunofluorescence signal in vehicle- and rotenone-treated cells. Results represent three independent experiments. Symbols represent measurements from individual cells. Comparison by ANOVA with Bonferroni correction.

Further evidence of oxidative activation of LRRK2 came from the study of endogenous NADPH oxidase 2 (NOX2). We found that rotenone treatment of wild-type HEK-293 cells caused an increase in the pSer1292 proximity ligation signal and Rab10 phosphorylation (Fig. 4, C to E). Although rotenone may cause mitochondrial ROS formation, mitochondrially derived ROS may also activate NOX2 in a process known as ROS-induced ROS release, which can feed forward to amplify ROS production (20, 21). We found that cotreatment with rotenone plus the specific NOX2 inhibitor peptide, Nox2ds-tat (22), blocked rotenone’s effects on LRRK2 activation and phosphorylation of its substrate (P < 0.0001, one-way ANOVA with Bonferroni correction). Thus, NOX2-generated superoxide appears to be important in activating LRRK2.

A LRRK2 kinase inhibitor prevents rotenone-induced activation of nigrostriatal LRRK2 and its downstream effects in rats

The rotenone model of PD reproduces many features of the human disease, including accumulation of pSer129-α-synuclein, impairment of autophagy, and reduced mitochondrial protein import (15). To determine whether systemic treatment with a brain-penetrant LRRK2 inhibitor could block rotenone-induced LRRK2 activation and to survey some of the potential downstream effects of LRRK2 activation, we treated rats for 5 days with rotenone (2.8 mg/kg per day, i.p.) with or without concomitant PF-360 (10 mg/kg, p.o., twice daily), a highly selective LRRK2 kinase inhibitor (23, 24). This PF-360 dosing regimen resulted in a pharmacokinetic profile in which an IC90 concentration in rat brain was achieved for 15 hours daily, and an IC50 concentration was achieved for a full 24 hours.

In a new cohort of rats treated with rotenone for 5 days, there was a marked increase in pSer1292 proximity ligation signal in nigrostriatal dopamine neurons, which was associated with an increase in phosphorylation of Rab10 (Fig. 5, A to C). Cotreatment with PF-360 effectively blocked the rotenone-induced activation of LRRK2 (P < 0.0001, two-way ANOVA with Sidak correction) and phosphorylation of Rab10 (P < 0.0001). Thus, the pSer1292 proximity ligation assay provided an ex vivo assay of target (LRRK2) engagement by PF-360, which was corroborated by measurement of pThr73-Rab10.

Fig. 5 LRRK2 activation and pSer129-α-synuclein accumulation in rat nigrostriatal dopamine neurons can be blocked by a brain penetrant LRRK2 kinase inhibitor.

(A) Shown are the pSer1292 proximity ligation signal (red) and the pThr73-Rab10 signal (gray) in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. TH, tyrosine hydroxylase, a marker of dopamine neurons (blue). (B) Quantification of pSer1292 proximity ligation signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats. Comparison by ANOVA with Bonferroni correction. (C) Quantification of pThr73-Rab10 signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats. Comparison by ANOVA with Bonferroni correction. (D) Shown is pSer129-α-synuclein immunoreactivity in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. (E) Quantification of pSer129-α-synuclein signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats from a single experiment. Comparison by ANOVA with Bonferroni correction.

We reported previously that chronic rotenone treatment (10 to 14 days) leads to elevated pSer129-α-synuclein (15). Here, we found that the rats treated for only 5 days also accumulated pSer129-α-synuclein, and cotreatment with PF-360 prevented this accumulation (Fig. 5, D and E). The mechanism by which pSer129-α-synuclein accrues in response to rotenone is uncertain, but it has been suggested that phosphorylation of α-synuclein at Ser129 targets the protein for degradation by autophagy (25, 26). Both chaperone-mediated autophagy (CMA) and macroautophagy play roles in α-synuclein degradation (27, 28). Therefore, we assessed a marker for CMA, Lamp2A, which is located on lysosomes, and another marker, Lamp1, which may label late endosomes, autolysosomes, or lysosomes. There were abundant Lamp2A and Lamp1 punctae in nigrostriatal dopamine neurons from the brains of vehicle-treated rats, which were markedly lost after rotenone treatment and preserved by cotreatment with PF-360 (Fig. 6, A to E). Together, these results suggest that there may be early impairment of CMA and lysosomal function, which is downstream of LRRK2 kinase activity. To complement these pharmacological studies, we examined the effects of rotenone on pSer129-α-synuclein in wild-type and LRRK2−/− HEK-293 cells. We found that rotenone treatment caused accumulation of pSer129-α-synuclein in wild-type cells; however, there was no such accumulation in the LRRK2 null cells (fig. S4), suggesting that buildup of pSer129-α-synuclein may be LRRK2-dependent. Moreover, the rotenone-induced increase in pSer129-α-synuclein in wild-type cells was effectively blocked by PF-360 to the same extent as in LRRK2−/− cells, confirming the specificity of the PF-360 effect.

Fig. 6 Rotenone induces lysosomal and CMA defects in rat nigrostriatal dopamine neurons that are prevented by cotreatment with a LRRK2 kinase inhibitor.

(A) Shown is Lamp1 and p62/SQSTM1 immunoreactivity in the nigrostriatal dopamine neurons of rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. TH, tyrosine hydroxylase, a marker of dopamine neurons (red). (B) Shown is Lamp2A immunoreactivity in the nigrostriatal dopamine neurons of rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. (C) Quantification of Lamp1 signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats from one experiment. Comparison by ANOVA with Bonferroni correction. (D) Quantification of p62/SQSTM1 signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats from one experiment. Comparison by ANOVA with Bonferroni correction. (E) Quantification of Lamp2A signal in rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats from one experiment. Comparison by ANOVA with Bonferroni correction. (F) Lamp1 and p62/SQSTM1 immunoreactivity in the substantia nigra of a postmortem human healthy, age-matched control brain and two postmortem brains from iPD patients (iPD-1 and iPD-2). In the control brain, nigrostriatal dopamine neurons contained many small punctae of Lamp1 immunoreactivity and little detectable p62/SQSTM1. In the two postmortem iPD brains, there was loss of Lamp1 puncta and accumulation of p62/SQSTM1 into large inclusions (Lewy bodies) in nigrostriatal dopamine neurons. (G) Quantification of Lamp1 in postmortem brain nigrostriatal dopamine neurons from three healthy age-matched control subjects and three patients with iPD. Symbols represent individual brains. Comparison by unpaired two-tailed t test. (H) Quantification of p62 in postmortem brain nigrostriatal dopamine neurons from three healthy age-matched control subjects and three patients with iPD. Symbols represent individual brains. Comparison by unpaired two-tailed t test.

Similar to rotenone-treated rats, there was a marked loss of Lamp1 puncta in human postmortem brain tissue from individuals with iPD (P < 0.001, unpaired two-tailed t test), and this was accompanied by an accumulation of the autophagy cargo receptor p62/SQSTM1 in Lewy bodies (P < 0.03), indicating autophagic and lysosomal dysfunction (Fig. 6, F and G). As reported by many other groups, and as seen in rotenone-treated rats, there was accumulation of pSer129-α-synuclein in the substantia nigra of postmortem brain tissue from iPD patients.

In vitro experiments have shown that pSer129-α-synuclein binds to TOM20 and inhibits mitochondrial protein import; however, this has not been examined directly in human brain or in the rotenone-treated rat model of PD. Examination of human iPD postmortem brain tissue (Fig. 7, A and B) revealed a marked increase in the pSer129-α-synuclein–TOM20 proximity ligation signal (P < 0.0001, unpaired, two-tailed t test), indicating that accumulation of this specific form of α-synuclein may have toxic consequences in terms of mitochondrial protein import. Similarly, in the rotenone-treated rats (Fig. 7, C to E), the increased pSer129-α-synuclein we found at 5 days was associated with its binding to TOM20, measured as a strong pSer129-α-synuclein–TOM20 proximity ligation signal (P < 0.0001, two-way ANOVA with Bonferroni correction), as well as reduced levels and redistribution of the imported complex I subunit, Ndufs3, from mitochondria to cytosol (P < 0.0001). Cotreatment with PF-360 prevented the elevation in pSer129-α-synuclein (Fig. 5E) and, as a result, there was little binding to TOM20 (P < 0.0001 versus rotenone alone), and there was preservation of normal levels and mitochondrial localization of Ndufs3 (Fig. 7, C and E). Thus, both the accumulation of pSer129-α-synuclein and its toxic consequences appear to be downstream of LRRK2 kinase activity.

Fig. 7 pSer129-α-synuclein binding to TOM20 in postmortem iPD brain tissue and in rotenone-treated rats is prevented by cotreatment with a LRRK2 kinase inhibitor.

(A) Shown is the pSer129-α-synuclein (pSer129syn)–TOM20 proximity ligation signal in the substantia nigra of postmortem brain tissue from a healthy, age-matched control individual and a patient with iPD. (B) Quantification of the pSer129syn-TOM20 proximity ligation signal in eight postmortem control brains and seven postmortem iPD brains. Comparison by unpaired two-tailed t test. Symbols represent individual brains. (C) Shown is pSer129syn-TOM20 proximity ligation signal (red) and Ndufs3 immunoreactivity (gray) in the substantia nigra of the brains of rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. In rotenone-treated rats, there was increased pSer129syn-TOM20 proximity ligation signal and a reduced amount and diffuse redistribution of the nuclear encoded and imported complex I subunit Ndufs3. These abnormalities were prevented by treatment with the LRRK2 kinase inhibitor PF-360. (D) Quantification of the pSer129syn-TOM20 proximity ligation signal in the substantia nigra of the brains of rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Symbols represent individual rats from a single experiment. Comparison by ANOVA with Bonferroni correction. (E) Graphical representation of the distribution and fluorescence intensity of Ndufs3 in nigrostriatal dopamine neurons in the brains of rats treated with vehicle, PF-360 alone, rotenone alone, or rotenone + PF-360. Note the loss of punctate, high-intensity staining in the rotenone-treated animals that was preserved by cotreatment with PF-360.

DISCUSSION

Development of new proximity ligation assays for detecting the pSer1292-LRRK2 autophosphorylation site and for LRRK2 binding to 14-3-3 proteins allowed us to show that endogenous wild-type LRRK2 was activated in nigrostriatal dopamine neurons in postmortem brain tissue from patients with iPD and that this finding could be reproduced in rodent models of the disease. Whereas our new assay does not measure LRRK2 activity per se, it provides a snapshot of relative LRRK2 activation state and does so with a cellular level of resolution (for example, in dopamine neurons and microglia). The assay was validated (i) using CRISPR/Cas9-edited HEK-293 cells, (ii) by demonstrating that the readout of LRRK2 activation state (pSer1292 proximity ligation signal) correlated with phosphorylation of Rab10 substrate, and (iii) by showing that both the pSer1292 proximity ligation signal and pThr73-Rab10 signal responded appropriately and coordinately to three structurally distinct LRRK2 kinase inhibitors (GNE-7915, MLi-2, and PF-360). Consistent with our results, others showed recently, by means of immunoblotting, that the phosphorylation state of LRRK2 (at Ser935) correlates with Rab10 phosphorylation in dose-response studies using PF-360 (23).

The finding that treatment of HEK-293 LRRK2G2019S/G2019S mutant cells, or patient-derived LRRK2wildtype/G2019S lymphoblastoid cells, with LRRK2 kinase inhibitors blocks the pSer1292 proximity ligation signal, indicates that this readout is readily reversible. Thus, the relative degree of LRRK2 activation detected reflects the physiological state of LRRK2 in the cell or tissue at the specific time of fixation. In addition, the ease with which inhibitor dose-response relationships can be assessed suggests that the assay can provide a quantitative measure of target (LRRK2) engagement.

Recent work from West and colleagues showed that genetic ablation or pharmacological kinase inhibition of endogenous wild-type LRRK2 reduced the toxicity of AAV2-hSNCA injected into the substantia nigra of rats, suggesting a possible role of LRRK2 in α-synuclein toxicity (29, 30). Implicit in such a conclusion is the assumption that wild-type LRRK2 kinase must be active in nigrostriatal dopamine neurons under these experimental conditions; however, this has been difficult to demonstrate with conventional assays. Consistent with this supposition, we found that AAV2-hSNCA (as used by West and colleagues) activated LRRK2 in nigrostriatal dopamine neurons in rats. In this context, our findings indicate that LRRK2 is activated in the vulnerable dopamine neurons of the nigrostriatal pathway in rats and in human iPD postmortem brain tissue and suggest that endogenous wild-type LRRK2 may play a role in iPD pathogenesis.

Our assay allows relatively facile assessment of physiological regulators of LRRK2 activity. In cell culture, we showed that low concentrations of oligomeric, but not monomeric, α-synuclein activated LRRK2. We recently reported that oligomeric, but not monomeric, α-synuclein binds to TOM20, impairs mitochondrial protein import, and causes mitochondrial dysfunction and aberrant ROS production (15). Similarly, the mitochondrial toxin rotenone induces ROS production, and we showed that it caused LRRK2 activation and LRRK2 substrate phosphorylation. The rotenone-induced activation of LRRK2 was blocked by inhibition of NOX2, implicating ROS-induced ROS release and NOX2-derived superoxide in the phenomenon (20). To directly test whether ROS can activate LRRK2, HEK-293 cells were treated with physiological concentrations of H2O2, and this resulted in increased pSer1292 proximity ligation signal and concomitant Rab10 phosphorylation. These results suggest that oxidative stress, long implicated in PD pathogenesis, may up-regulate LRRK2 activity. Further, our results are consistent with the work of Mamais et al., which showed that H2O2 dissociates LRRK2 from 14-3-3 proteins (12), and the study by Li et al., which found a small activation of LRRK2 by H2O2 (31). The molecular mechanisms by which ROS stimulates LRRK2 kinase activity are unknown; whether this activation results from direct effects on the LRRK2 protein or from effects on interacting proteins, such as protein phosphatases, remains to be explored.

The consistent finding that LRRK2 is activated in vulnerable dopaminergic neurons in human iPD raises the possibility that LRRK2 kinase activity may play a central role in the pathogenesis of many cases of both familial and sporadic PD. To investigate experimentally the possibility that LRRK2 activation may have downstream deleterious effects on putative pathogenic mechanisms, we treated rats for 5 days with rotenone with or without cotreatment with the LRRK2 kinase inhibitor PF-360. We found that rotenone (i) activated LRRK2 in nigrostriatal dopamine neurons, (ii) increased phosphorylation of a LRRK2 substrate, Rab10, (iii) increased pSer129-α-synuclein and binding to TOM20, and (iv) disrupted markers of autophagy. All of these effects were blocked by the LRRK2 kinase inhibitor PF-360, and we therefore conclude that these mechanisms are downstream of LRRK2 activity (fig. S5).

Rab10 is a bona fide direct substrate of LRRK2 and has been implicated in the maintenance of endoplasmic reticulum, vesicle trafficking, and autophagy (3234). LRRK2-induced phosphorylation of Rab10 inhibits its function by preventing binding to Rab GDP dissociation inhibitor factors necessary for membrane delivery and recycling (6). Thus, LRRK2 activation leads to Rab10 phosphorylation (and inhibition), and this likely impairs autophagic or lysosomal function. Phospho-Rab10–induced autophagic dysfunction, in turn, may account for the early accumulation of pSer129-α-synuclein, which is normally degraded by CMA and macroautophagy. Consistent with our findings, Volpicelli and colleagues reported that the G2019S mutation in LRRK2 was associated with accumulation of pSer129-α-synuclein–positive inclusions, which was prevented by LRRK2 kinase inhibitors, including MLi-2 (35). It is also noteworthy that pSer129-α-synuclein is one of the species of α-synuclein we reported that binds to TOM20 and impairs mitochondrial function (15). The fact that all of these abnormalities are prevented by treatment with a LRRK2 kinase inhibitor supports this hypothetical scheme (fig. S5). Further, our experimental approach of surveying potential pathogenic mechanisms in rodent and cellular models of PD, with and without concomitant treatment with a LRRK2 kinase inhibitor, may lead to new insights into how aberrant LRRK2 activity causes neurodegeneration.

Our study is not without limitations. As noted, the proximity ligation assays we used do not measure activity per se. However, several convergent lines of evidence confirm that we measured reliable surrogates of LRRK2 activity: The pSer1292 proximity ligation signal correlated with substrate (Rab10) phosphorylation and with loss of 14-3-3 binding signal, and both the pSer1292 proximity ligation signal and the pThr72-Rab10 signal were inhibited dose-dependently by selective LRRK2 kinase inhibitors of different chemical classes. In this context, we provided compelling evidence that wild-type LRRK2 is activated in nigrostriatal dopamine neurons in human iPD postmortem brain tissue and in two rat models of PD. Nevertheless, although LRRK2 activation occurred very early in the course of rotenone-induced parkinsonism, and early intervention with a LRRK2 kinase inhibitor was beneficial, the time course of LRRK2 activation in the human brain is unknown, and the clinical effects of LRRK2 inhibitors remain to be examined.

The fact that there are a variety of selective LRRK2 kinase inhibitors reflects the interest by the pharmaceutical industry in target-specific therapeutics for PD, even for the relatively small number of cases caused by LRRK2 mutations. The results presented here suggest that wild-type LRRK2 is activated by ROS in dopamine neurons in iPD and that this, in turn, may trigger a downstream pathological cascade of events culminating in neurodegeneration. We believe that LRRK2 inhibitors may be beneficial not only for the 3 to 4% of people with PD who carry LRRK2 mutations but also for iPD patients who do not carry LRRK2 mutations.

MATERIALS AND METHODS

Study design

This study was designed to assess the role of wild-type LRRK2 in iPD. To do so, we developed new proximity ligation assays to assess the phosphorylation state of the LRRK2 autophosphorylation site, Ser1292, and separately, the binding of LRRK2 to 14-3-3 proteins, which is associated with decreased LRRK2 kinase activity. These assays were validated using CRISPR/Cas9 genome-edited HEK-293 cells treated with three chemically distinct LRRK2 kinase inhibitors. The validated assays were then used to assess the LRRK2 activation state in iPD postmortem brain tissue and in two rat models of PD. Additional studies examined the role of oxidative stress in activating LRRK2, as well as downstream consequences of LRRK2 activity in vivo. All in vitro experiments were replicated at least three times, and key validation studies identifying CRISPR/Cas9-edited cell lines by means of proximity ligation assays were analyzed by blinded assessors. In vivo experiments using rotenone treatment, with or without concomitant PF-360 treatment, were performed in a single cohort of rats (n = 6 per active treatment group), and outcomes were analyzed by blinded assessors. Rats were randomized to the treatment group. There was no exclusion of outliers.

CRISPR/Cas9 genome editing of HEK-293 cells to produce LRRK2 knockout and knock-in cell lines

To generate a LRRK2 knockout cell line, a CRISPR/Cas9 genome editing approach was used. A guide RNA (gRNA) targeting exon 41 of the LRRK2 gene (5′-ATTGCAAAGATTGCTGACTAGTTTT-3′) was cloned into the GeneArt CRISPR Nuclease Vector (Invitrogen) as described by the manufacturer. HEK-293 cells were transfected by nucleoporation using a Nucleoporater II device (Amaxa). Transfected cells were collected and enriched by FACS (fluorescence-activated cell sorting). Sorted cells were grown and expanded for polymerase chain reaction (PCR) and DNA sequencing analyses. To confirm gene editing of the LRRK2 gene, a region of exon 41 was PCR-amplified using a forward primer (5′-TTTAAGGGACAAAGTGAGCA-3′) and reverse primer (5′-CACAATGTGATGCTTGCATTT-3′), and the resulting PCR product was sequenced. The LRRK2 G2019S knock-in cell line was generated using a similar approach but included transfection of a 120-mer single-stranded oligonucleotide containing a G to A substitution coding for a glycine-to-serine amino acid change.

An SNCA knockout cell line was generated using a similar approach but used a gRNA targeting exon 4 of the α-synuclein (SNCA) gene (5′-CTTTGTCAAAAAGGACCAGT-3′). To confirm gene editing of the SNCA gene, a region of exon 4 was PCR-amplified using a forward primer (5′-CCACCCTTTAATCTGTTGTTGC-3′) and reverse primer (5′-ATATAAAGGTAGCACTTTTTCACAAGG-3′), and the resulting PCR product was sequenced.

Proximity ligation assays

Proximity ligation was performed as described previously in 4% PFA (paraformaldehyde)–fixed tissue or cells (15). Samples were incubated with specific primary antibodies to the proteins to be detected. Secondary antibodies conjugated with oligonucleotides were added to the reaction and incubated. Ligation solution, consisting of two oligonucleotides and ligase, was added. In this assay, the oligonucleotides hybridize to the two proximity ligation probes and join to a closed loop if they are in close proximity. Amplification solution, consisting of nucleotides and fluorescently labeled oligonucleotides, was added together with polymerase. The oligonucleotide arm of one of the proximity ligation probes acts as a primer for “rolling-circle amplification” using the ligated circle as a template, and this generates a concatemeric product. Fluorescently labeled oligonucleotides hybridize to the rolling circle amplification product. The proximity ligation signal was visible as a distinct fluorescent spot and was analyzed by confocal microscopy (Duolink; Sigma-Aldrich). Control experiments included routine immunofluorescence staining of the proteins of interest under identical experimental conditions.

Statistical analyses

Each result presented here was derived from three to six independent experiments. For simple comparisons of two experimental conditions, two-tailed, unpaired t tests were used. Where variances were not equal, Welch’s correction was used. When AAV vector was injected into one hemisphere of the rat brain and the other hemisphere was used as a control, two-tailed paired t tests were used. For comparisons of multiple experimental conditions, one-way or two-way ANOVA was used, and if significant, overall post hoc corrections (with Bonferroni or Sidak tests) for multiple pairwise comparisons were made. P values less than 0.05 were considered significant.

SUPPLEMENTARY MATERIALS

www.sciencetranslationalmedicine.org/cgi/content/full/10/451/eaar5429/DC1

Materials and Methods

Fig. S1. Active LRRK2 is detected by proximity ligation in microglia in control brains and is increased in iPD and in rotenone-treated rats.

Fig. S2. Time course of in vivo rotenone-induced LRRK2 activation as assessed by pSer1292-LRRK2 proximity ligation signal.

Fig. S3. LRRK2 is activated by oligomeric but not monomeric α-synuclein.

Fig. S4. Rotenone-induced accumulation of pSer129-α-synuclein is LRRK2-dependent.

Fig. S5. Activation of LRRK2 kinase activity in iPD and its downstream consequences.

References (3638)

REFERENCES AND NOTES

Funding: This work was supported by research grants from the NIH (NS100744, R21ES027470, NS095387, and AG005133), the Blechman Foundation, the American Parkinson Disease Association, and friends and family of S. Logan. The brain bank has received support from the University of Pittsburgh Brain Institute. Work in the Alessi laboratory on LRRK2 is supported by the Michael J. Fox Foundation for Parkinson’s research (grant number 6986), the Medical Research Council (grant number MC_UU_12016/2), and the pharmaceutical companies supporting the Division of Signal Transduction Therapy Unit (Boehringer-Ingelheim, GlaxoSmithKline, and Merck KGaA to D.R.A.). Author contributions: R.D.M. designed, performed, and analyzed the proximity ligation experiments and edited the manuscript; E.K.H. was responsible for molecular biology and created and validated cell lines; E.M.R. performed and analyzed in vivo rotenone experiments; M.T.K. performed and analyzed many of the proximity ligation experiments; L.H.S. performed proof-of-concept experiments with lymphoblastoid cells; B.R.D.M. did ex vivo staining and in vitro cell culture experiments; A.V.L. helped to design experiments; A.Z. did in vivo α-synuclein overexpression work; A.F.S. and T.A.L. provided input on the design and analysis of PF-360 experiments; J.K.K. provided human neuropathology expertise and samples; E.A.B. provided input and collaboration on the in vivo α-synuclein overexpression experiments; D.R.A. provided guidance on LRRK2 biology and antibodies to LRRK2 substrates; T.G.H. provided guidance and helped to design experiments; and J.T.G. supervised the project, designed and analyzed the experiments, and wrote the paper. Competing interests: J.T.G. briefly held an advisory position at Pfizer. A.F.S. and T.A.L. were employees of Pfizer at the time this work was performed. The other authors declare that they have no competing interests. Data and materials availability: All data associated with this study can be found in the paper and the Supplementary Materials. J.T.G. received the PF-360 compound from Pfizer under an MTA.
View Abstract

Navigate This Article