Lrrk promotes tau neurotoxicity through dysregulation of actin and mitochondrial dynamics

Abstract

Mutations in leucine-rich repeat kinase 2 (LRRK2) are the most common cause of familial Parkinson disease. Genetics and neuropathology link Parkinson disease with the microtubule-binding protein tau, but the mechanism of action of LRRK2 mutations and the molecular connection between tau and Parkinson disease are unclear. Here, we investigate the interaction of LRRK and tau in Drosophila and mouse models of tauopathy. We find that either increasing or decreasing the level of fly Lrrk enhances tau neurotoxicity, which is further exacerbated by expressing Lrrk with dominantly acting Parkinson disease-associated mutations. At the cellular level, altering Lrrk expression promotes tau neurotoxicity via excess stabilization of filamentous actin (F-actin) and subsequent mislocalization of the critical mitochondrial fission protein dynamin-1-like protein (Drp1). Biochemically, monomeric LRRK2 exhibits actin-severing activity, which is reduced as increasing concentrations of wild-type LRRK2, or expression of mutant forms of LRRK2 promote oligomerization of the protein. Overall, our findings provide a potential mechanistic basis for a dominant negative mechanism in LRRK2-mediated Parkinson disease, suggest a common molecular pathway with other familial forms of Parkinson disease linked to abnormalities of mitochondrial dynamics and quality control, and raise the possibility of new therapeutic approaches to Parkinson disease and related disorders.

Fig 1. Manipulation of Lrrk enhances tau neurotoxicity.

(A) Increasing or decreasing Drosophila Lrrk enhances the toxicity of human tau^R406W, which is further enhanced by expressing mutant forms of Lrrk engineered to mimic Parkinson’s disease mutations. (B) Representative immunofluorescence images showing activated caspase in neurons (arrowheads, identified as neurons using anti-elav) of Drosophila brains. Scale bars represent 5 μm. (C) Western blot showing that manipulation of Lrrk does not change levels of transgenic tau. Quantification from three different blots is shown on the right panel. ns: not significant. Blots are reprobed for actin and GAPDH to illustrate equivalent protein loading. (D) Cell cycle activation in postmitotic neurons as monitored by PCNA staining mirrors the pattern observed for caspase activation. (E) Western blot showing levels of Drosophila Lrrk protein in the brains of Lrrk mutant and overexpressing flies. Quantification from three different blots is shown on the lower panel. Blots are reprobed for actin and GAPDH to illustrate equivalent protein loading. Y1383C, I1915T, and G1914S are Lrrk mutants homologous to Parkinson disease related human LRRK2 mutants Y1699C, I2020T, and G2019S, respectively. n = 6 per genotype and time point for histological assessments (A, D). *P < 0.01, ANOVA with supplementary Neuman—Keuls. Control is elav-GAL4/+; UAS-CD8-PARP-Venus/+ in A and B and elav-GAL4/+ in the remaining panels. Flies are 10 days old. See S1 Data for individual numerical values underlying the summary data displayed in A, C, D, and E. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Lrrk, leucine-rich repeat kinase; Lrrk^-, Lrrke⁰³⁶⁸⁰; Lrrk-OE, wild-type Lrrk overexpression; PCNA, proliferating cell nuclear antigen.

Fig 2. Manipulation of Lrrk enhances the excess stabilization of the actin cytoskeleton caused by tau.

(A) ELISA specific for F-actin shows an increase in F-actin levels in tau transgenic flies with further increase when Lrrk levels are increased or decreased or when Lrrk-GS is expressed. n = 3. (B) Western blot demonstrating equivalent levels of total actin and total protein, as illustrated by GAPDH, in the genotypes studied. Quantification of actin intensity from three blots is shown on the right panel. (C) Immunofluorescence microscopy images of sections of the indicated genotypes flies stained for actin illustrating actin rods. Images shown are of cortical Kenyon neurons in the mushroom bodies. Scale bars represent 10 μm. (D) The number of actin-rich rod-like structures (actin rods) present in sections from tau transgenic flies is increased when Lrrk levels are increased or decreased. (E) Western blot showing that Lrrk overexpression in a Lrrk mutant background results in Lrrk levels similar to those in wild-type control animals. The blot is reprobed for GAPDH to illustrate equivalent protein levels. Quantification from three different blots is shown on the right panel. (F) Expression of wild-type Lrrk in tau transgenic flies in the Lrrk mutant background does not increase the number of neurons with activated caspase, indicating no increase of neuronal toxicity. (G) There is no increase in cell cycle activation in postmitotic neurons when Lrrk is expressed with tau in flies in the Lrrk mutant background. (H) There is no incr ease in the number of actin rods when Lrrk is expressed with tau in flies with Lrrk mutant background. n = 6 per genotype and time point for histological analyses. *P < 0.01, ANOVA with supplementary Neuman—Keuls. Control is elav-GAL4/+; UAS-CD8-PARP-Venus/+ in F, and elav-GAL4/+ in the remainder of the panels. Flies are 10 days old. See S1 Data for individual numerical values underlying the summary data displayed in A, B, and D—H. ELISA, enzyme-linked immunosorbent assays; F-actin, filamentous actin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Lrrk, leucine-rich repeat kinase; Lrrk-GS, Lrrk carrying the G1914S mutation.

Fig 3. Mislocalization of Drp1, mitochondrial elongation, and mitochondrial dysfunction are enhanced by Lrrk manipulation.

(A) Representative confocal images of cortical Kenyon neurons of the Drosophila brain sections show mislocalization of the mitochondrial fission protein Drp1 and elongated mitochondria in tau transgenic flies. Mitochondrial elongation and Drp1 mislocalization are enhanced in flies lacking or overexpressing Lrrk. Scale bars represent 2 μm. (B) Quantification of mitochondrial length shows elongated mitochondria in tau transgenic flies with modulated levels of Lrrk. (C) Quantification of the number of mitochondria colocalized with Drp1 shows reduced mitochondrial localization of Drp1 in tau transgenic flies, with further reduction with manipulation of Lrrk. (D) Measurement of Pearson’s correlation coefficient reflects reduced colocalization with genetic manipulation of tau and Lrrk. (E) Mitochondrial interconnectivity is increased with expression of tau and further increased by manipulating Lrrk expression. (F) Western blot showing equal levels of Drp1 protein in experimental genotypes as indicated. The blot is reprobed for actin and GAPDH to illustrate equivalent protein loading. The lower panel shows quantification of the Drp1 band intensity from three different blots. (G) Confocal images of whole fly brains freshly dissected and stained with the mitochondrial superoxide dye MitoSOX. Two-dimensional projection of z-stacks showing the brightest section for each sample. (H) Quantification of the fluorescence intensity for the entire brain shows an increase in superoxide levels in the mitochondria of tau transgenic flies, which is further enhanced in tau transgenic flies with altered Lrrk expression. (I) Measurement of the ratio of red to green fluorescence in flies with the UAS-MitoTimer reporter transgene reflects increased reporter oxidation in tau transgenic flies, which is further increased with manipulation of Lrrk expression. (J) Measurement of red fluorescence indicating the levels of ROS in the brains of flies of the indicated genotypes shows increased ROS levels in flies expressing tau, with further increases with manipulation of Lrrk expression. n = 6 per genotype and time point in B, C, D, E, H, I, and J. *P < 0.01, ANOVA with supplementary Neuman—Keuls. Control is elav-GAL4/+; UAS-mito-GFP/+; HA-Drp1/+ in A—F, elav-GAL4/+ in G, H, and I, and elav-GAL4/+; UAS-MitoTimer/+ in J. Flies are 10 days old. See S1 Data for individual numerical values underlying the summary data displayed in B—F and H—J. Drp1, dynamin-1-like protein; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; Lrrk, leucine-rich repeat kinase; mito-GFP, mitochondrially directed GFP; ns, not significant; ROS, reactive oxygen species; HA, hemagglutinin.

Fig 4. Pharmacological inhibition of actin polymerization rescues neurodegeneration and mitochondrial deficits in tau transgenic flies.

(A) Dose-dependent improvement in tau neurotoxicity with feeding of the actin-depolymerizing compound LatA, as assessed by the number of activated caspase positive cells. (B) Cell cycle activation is reduced in flies treated with LatA. (C, D) Treatment with Cyto-B results in a dose-dependent reduction in caspase activation (C), and also in cell cycle activation (D). (E, F) Treatment with Cyto-D results in a significant reduction in caspase activation (E) and in cell cycle activation (F). (G) Western blot showing that drug treatments at 10 μM do not change levels of transgenic human tau. (H) Measurement of the ratio of red to green fluorescence in flies with the UAS-MitoTimer reporter transgene shows a reduction in reporter oxidation in tau transgenic flies treated with 10 μM of either LatA, Cyto-B, or Cyto-D. (I) Representative confocal images of either control or tau transgenic fly brains treated with actin depolymerization drugs and stained for superoxide indicator dye MitoSOX. Two-dimensional projections of z-stacks showing the brightest section for each sample. (J) Quantification of the fluorescence intensity of the entire brain shows a decrease in superoxide levels in tau transgenic flies treated with 10 μM of actin-depolymerizing drugs. n = 6 per genotype and treatment. *P < 0.01, ANOVA with supplementary Neuman—Keuls. Control is elav-GAL4/+; UAS-CD8-PARP-Venus/+ (A, C, and E), elav-GAL4/+ in B, D, F, G, I, and J, and elav-GAL4/+; UAS-MitoTimer in H. Flies are 10 days old. See S1 Data for individual numerical values underlying the summary data displayed in A—F, H, and J. Cyto-B, cytochalasin B; Cyto-D, cytochalasin D; LatA, latrunculin A.

Fig 5. Enhancement of neurodegeneration, actin rod formation, Drp1 mislocalization, and mitochondrial elongation in mice expressing human tau and human LRRK2.

(A) H&E staining of sagittal brain sections of mice shows neuronal loss in the hippocampus of transgenic mice expressing human tau (tau^P301L), which is enhanced in mice overexpressing wild-type human LRRK2 and human tau (tau—LRRK2). (B) Higher magnification of the CA1 region. Scale bars in A, B represent 100 μm. (C) Quantification of the number of neurons in the CA1 region in the brain sections of mice with the indicated genotypes. n = 10 per genotype. Overall P value (ANOVA) for difference is <0.0001. Bonferroni multiple comparison posthoc test shows that each group is significantly different from all others with a P value lower than 0.01 across the board. The single exception is the comparison of nontransgenic and LRRK2 groups, which are not significantly different from one another. (D) Quantification of actin-rich rod-like structures in brain sections of mice shows an increase in tau—LRRK2 mice. (E) Representative immunofluorescence images showing actin rods in the brains of transgenic mice. Scale bars represent 10 μm. (F) Immunofluorescence images of mouse hippocampal neurons stained for ATPVa to visualize the mitochondria and Drp1 show a decrease in Drp1 localization to mitochondria in human tau transgenic mice and a further reduction in tau—LRRK2 mice. G) Pearson’s correlation coefficient calculated as a measure of colocalization. (H) Quantification of mitochondrial length in the neuronal cell bodies reveals an increase in average mitochondrial length in human tau transgenic mice, which is further increased in tau-LRRK2 mice. (I) Mitochondrial interconnectedness is increased in human tau transgenic mice, and further increased in tau—LRRK2 mice. (J) Immunofluorescent images of mouse brain sections stained with NeuN to visualize the hippocampal pyramidal neurons and ATPVa to demonstrate mitochondrial morphology. Scale bars represent 10 μm. n = 5 per genotype. *P < 0.01, ANOVA with supplementary Neuman—Keuls for D, G, H, and I. Controls are age-matched nontransgenic mice. Mice are 5.5 months old. See S1 Data for individual numerical values underlying the summary data displayed in C, D, G, H, and I. H&E, hematoxylin and eosin; ATPVa, vacuolar protein-ATPase A-subunit; Drp1, dynamin-1-like protein; LRRK2, leucine-rich repeat kinase 2.

Fig 6. Lrrk enhances actin depolymerization by severing F-actin filaments.

(A) Actin depolymerization assay using pyrene-labeled actin (4 μM) and recombinant LRRK2 proteins, either wild type, LRRK2-GS, or 1:1 mixture of the two (mix) used at 1 nM shows an enhancement of actin depolymerization by LRRK2. (B) Initial 10 minutes of the actin depolymerization assay showing the linear phase of depolymerization. (C) Initial rates of depolymerization calculated from the liner portion of the depolymerization curves. (D) Fold increase in depolymerization of actin filaments in the presence of LRRK2, LRRK2-GS, or the mix at 60 minutes. (E) Native gel with silver stain showing oligomerization states of LRRK2 proteins with 0 or 60 minutes incubation at room temperature. (F) Graph showing the percentage increase in the oligomer/monomer ratio in 60 minutes for the specified samples after quantification of three different gels. (G) Representative pictures of the results of actin severing assays on in vitro polymerized actin filaments with either wild-type LRRK2, LRRK2-GS, or the mix at 1 nM. (H) Quantification of the actin severing assay showing severing activity by LRRK2, LRRK2-GS, or the mix after a two-minute incubation with polymerized actin filaments. (I) Representative pictures of the results of the actin severing assays on in vitro polymerized F-actin performed after incubating the indicated LRRK2 proteins for 1 hour. (J) Quantification of the actin severing assays performed after preincubation of the LRRK2 proteins. (K) Biotinylated phalloidin precipitation of F-actin from control (genotype: elav-GAL4/+) Drosophila heads shows that Lrrk interacts with F-actin in vivo. (L, M) Immunofluorescence images of freshly dissected brains from flies expressing HA-tagged Lrrk from the endogenous promoter (genotype: elav-GAL4/+; Lrrk^HA/+) and costained with an anti-HA antibody and fluorescent phalloidin show colocalization of Lrrk with F-actin. n = 3 for all experiments. *P < 0.01, ANOVA with supplementary Neuman—Keuls. Flies are 1–3 days of age. Scale bars represent 10 μm for G and I and 5 μm for L. See S1 Data for individual numerical values underlying the summary data displayed in A—D, F, H, J, and M. (N) Model of the proposed effects of Lrrk oligomerization state on actin severing activity. At normal levels, Lrrk exists primarily in the monomeric state and severs actin filaments efficiently. When Lrrk is overexpressed, higher concentrations of Lrrk lead to the formation of oligomeric species that have reduced or absent actin severing activity. In the presence of mutant Lrrk, oligomers form with increased efficiency, further reducing actin-severing activity. F-actin, filamentous actin; HA, hemagglutinin; LRRK, leucine-rich repeat kinase; LRRK2-GS, LRRK2-G2019S.