LRRK2 modifies α-syn pathology and spread in mouse models and human neurons

Abstract

Progressive aggregation of the protein alpha-synuclein (α-syn) and loss of dopaminergic neurons in the substantia nigra pars compacta (SNpc) are key histopathological hallmarks of Parkinson's disease (PD). Accruing evidence suggests that α-syn pathology can propagate through neuronal circuits in the brain, contributing to the progressive nature of the disease. Thus, it is therapeutically pertinent to identify modifiers of α-syn transmission and aggregation as potential targets to slow down disease progression. A growing number of genetic mutations and risk factors has been identified in studies of familial and sporadic forms of PD. However, how these genes affect α-syn aggregation and pathological transmission, and whether they can be targeted for therapeutic interventions, remains unclear. We performed a targeted genetic screen of risk genes associated with PD and parkinsonism for modifiers of α-syn aggregation, using an α-syn preformed-fibril (PFF) induction assay. We found that decreased expression of Lrrk2 and Gba modulated α-syn aggregation in mouse primary neurons. Conversely, α-syn aggregation increased in primary neurons from mice expressing the PD-linked LRRK2 G2019S mutation. In vivo, using LRRK2 G2019S transgenic mice, we observed acceleration of α-syn aggregation and degeneration of dopaminergic neurons in the SNpc, exacerbated degeneration-associated neuroinflammation and behavioral deficits. To validate our findings in a human context, we established a novel human α-syn transmission model using induced pluripotent stem cell (iPS)-derived neurons (iNs), where human α-syn PFFs triggered aggregation of endogenous α-syn in a time-dependent manner. In PD subject-derived iNs, the G2019S mutation enhanced α-syn aggregation, whereas loss of LRRK2 decreased aggregation. Collectively, these findings establish a strong interaction between the PD risk gene LRRK2 and α-syn transmission across mouse and human models. Since clinical trials of LRRK2 inhibitors in PD are currently underway, our findings raise the possibility that these may be effective in PD broadly, beyond cases caused by LRRK2 mutations.

Keywords: Aggregation; Alpha-synuclein; GBA; Genetic interaction; LRRK2; Parkinson’s disease.

Fig. 1

PFF-based screen in primary neuron identifies Lrrk2 and Gba as genetic modifiers of α-synuclein transmission and aggregation. a Schematic of experimental design: lentiviral knock down or overexpression of Parkinson’s disease (PD) risk genes in primary neuron cultures exposed to fragmented α-syn preformed fibrils (PFFs) at 10 days in vitro (DIV). Survival and aggregation of endogenously expressed α-syn was measured using immunolabeling of phosphorylation of α-syn at Serine 129 (pSer129) 10 days after PFF treatment. b Summary of targeted screening results. The aggregation score was calculated based on pSer129 immunolabeling and normalized to the neuronal viability for each shRNA. Neuronal viability was determined by the number of NeuN-positive cells/well in PFF compared to vehicle-treated wells. The dotted line marks the average of all non-targeting scramble shRNAs. A minimum of 2 shRNAs with a KD-efficiency of 50 percent were used per PD risk gene. α-syn (Snca) constructs are highlighted in yellow, Lrrk2 in red, scramble controls in gray and Gba in blue (n = 12–24 wells/shRNA). Data expressed as mean + SEM; *p < 0.05; **p < 0.01; ****p < 0.0001 compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons. c, d Representative images (c) and quantification of pSer129-positive (green) α-syn aggregation (d) following knock down of Snca, Gba, Lrrk2 and non-targeting scramble (SCR) control (n = 8–10 independent wells/treatment). pSer129 immunolabeling was normalized to neuron counts using the neuronal marker NeuN (red). e, f Experimental design (e) and quantification of PFF-based α-syn aggregation (f) in primary neuron cultures derived from LRRK2 G2019S-mutant (G2019S) and WT mouse embryos (n = 10 wells/genotype, 10 images/well). g Relative α-syn (Snca) mRNA expression level measured in WT and G2019S-LRRK2 mutant neurons at 10 DIV using quantitative RT-qPCR analysis (n = 5 samples/group). Data expressed as mean + SEM; *p < 0.05; **p < 0.01; compared by unpaired Student’s t test and one-way ANOVA with a Tukey’s post-test for multiple comparisons

Fig. 2

The LRRK2 G2019S mutation exacerbates α-syn pathology in a PFF-based in vivo mouse model of α-syn transmission and aggregation. a, b Schematic of mouse genotypes and experimental design. Adult LRRK2 G2019S BAC-transgenic mice (G2019S or GS) and wild-type (WT) littermates were injected with α-syn PFFs or vehicle control (Veh). Histological analysis and behavioral testing was performed at 1, 3, 6 months post injection (PI). c Representative images of pSer129 (green) immune-labeling in TH-positive neurons (red) in the substantia nigra pars compacta (SNpc) at 3 months PI. d Quantification of the percentage of TH-positive neurons that are also pSer129-positive in the SNpc of PFF-injected animals (n = 8–13 animals/group, 8–10 brain sections/animal). e Quantification of the number of TH-positive neurons in the SNpc of WT and LRRK2-G2019S animals at 1, 3, 6 months post injection with PFFs or vehicle control (n = 6–12 animals/group, 8–10 brain sections/animal). f Schematic and results of open field behavioral testing 6 months post injection with PFFs or vehicle control (11–13 animal/group). g Rotarod behavioral assay measuring the latency to fall from an accelerating rotating cylinder at 6 months post injection (n = 11–13 animals/group). Data expressed as mean + SEM; *p < 0.05; compared by Student’s t test and one-way ANOVA with a Tukey’s post-test for multiple comparisons

Fig. 3

The neuroinflammatory response is altered in PFF-injected LRRK2 G2019S mutant mice. a–c Representative images (a) and quantification (b, c) of microglial markers Iba1 (green) and Cd68 (red) in the dorsal striatum of LRRK2 G2019S (G2019S or GS) mice and wild-type (WT) littermate controls 6 months post injection (PI) with PFFs or Vehicle (Veh) control (n = 6–12 animals/group). d–f Experimental design and quantification of gene expression of microglial (e) and astrocyte-associated (f) activation markers. mRNA was isolated from the striatum and gene expression was analyzed using RT-qPCR analysis (n = 5–6 animals/group). g–h Representative images (g) and quantification (h) of intensity of C1q immunolabeling in the dorsal striatum of PFF and vehicle-injected mice 6 months post injection (n = 5–6 animals/group). Data expressed as mean + SEM; *p < 0.05, **p < 0.01; compared by Student’s t test, one-way ANOVA with a Tukey’s post-test for multiple comparisons or and two-way ANOVA with Bonferroni post hoc correction

Fig. 4

Recombinant human α-syn PFFs induce aggregation of endogenous α-syn in human iPS-derived induced neurons. a Experimental design: iPS-derived induced neurons (iNs) were differentiated for 3 weeks before the addition of human α-syn PFFs to the culture media. Aggregation of endogenously expressed α-syn was assessed using immunolabeling of pSer129. b, c Representative images (b) and quantification (c) of pSer129 immunolabeling (red) up to 14 days (D) post PFF treatment (n = 19–21 frames/time point). Nuclei labeled with DAPI (blue). d Representative images of pSer129 α-syn (red) co-localizing with the axonal marker BIII-tubulin (magenta) 7 days post PFF treatment. e Representative images of pSer129 α-syn (red) co-localizing with ubiquitin (green) 14 days post PFF treatment. Data expressed as mean + SEM; compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons

Fig. 5

LRRK2 mutations and level modulate aggregation of α-syn in human-induced neurons. a Experimental design of internalization of α-syn PFFs. 3 week-old induced neurons (iNs) were exposed to fluorescently labeled α-syn PFFs for 0–24 h. Fluorescence in the media was quenched with 0.1% Trypan blue, revealing the internalized labeled PFFs. b Representative image of internalized PFFs (green) in human iNs (brightfield) after 6 h of incubation. c Quantification of internalized PFFs in isogenic G2019S mutant, corrected and LRRK2 knock out (KO) iNs after 1–24 h of incubation (n = 300–450 cells from two biological replicates/condition). Data expressed as mean + SD, compared by two-way ANOVA with Bonferroni post hoc correction. d, e Representative images and quantification of pSer129-positive α-syn aggregates after 3 weeks of incubation with recombinant human PFFs (n = 10–12 wells/group). Data expressed as mean + SEM; *p < 0.05; **p < 0.01; compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons. f Representative Western blot of isogenic LRRK2 G2019S mutant, corrected and LRRK2 knock out (KO) iN lysates 2 weeks post PFF treatment, probed with anti-pSer129, anti-PSD95, anti-SYNAPTOPHYSIN (SYP) and anti-ACTB antibodies. Synaptic proteins and ACTB were used as loading controls. Proteins were sequentially extracted using 1% Triton-X followed by 2% SDS lysis buffers. g–i Quantification of pSer129 (g), PSD95 (h) and SYNAPTOPHYSIN (i) 14 days post PFF treatment (n = 4/group). Data expressed as mean + SEM; *p < 0.05; **p < 0.01; compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons

Fig. 6

α-syn expression is altered in human LRRK2 knock out-induced neurons. a Representative Western blot of lystates from isogenic LRRK2 G2019S mutant, corrected and knock out (KO) induced neurons (iNs) at 4 weeks of differentiation probed with anti-LRRK2, anti-α-syn (SNCA), anti-PSD95, anti-SYNAPTOPHYSIN (SYP), anti-HOMER1 and anti-ACTB antibodies. ACTB was used as a loading control. b–e Quantification of α-syn (f), PSD-95 (g), SYNAPTOPHYSIN (SYP) (d) and HOMER1 level at 4 weeks of differentiation (n = 3/group). Data expressed as mean + SD; compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons. f–i RT-qPCR expression analysis of mature neuron markers MAP2 (f), MAPT (g) and synaptic markers SNAP25 (h) and GRIA1 (i) during 4 weeks of differentiation. Data normalized to the 0 weeks of differentiation time point (n = 4/group and time point). Data expressed as mean + SEM; *p < 0.05; **p < 0.01; compared by one-way ANOVA with a Tukey’s post-test for multiple comparisons and two-way ANOVA with Bonferroni post hoc correction

Fig. 7

PD-linked mutations in LRRK2 exacerbate aggregation, glial activation and neuronal survival in mouse and human models of PFF-transmitted α-syn pathology. In primary neuron cultures and human iPS-derived induced neurons (iNs), PFFs are rapidly internalized and trigger the recruitment of soluble α-syn into cytoplasmic aggregates. The LRRK2 G2019S mutation is one of the most common genetic causes of Parkinson’s disease. Human and mouse neurons, derived from patient-derived iPS cells or LRRK2 G2019S BAC-transgenic mice, develop normally in vitro. While internalization of PFFs is unaltered, LRRK2 G2019S neurons display increased accumulation of aggregated endogenous α-syn upon PFF treatment. Further, following intrastriatal PFF delivery into the brains, LRRK2 G2019S mice display an altered neuroinflammatory response with increased expression of inflammatory markers and cytokines and accelerated degeneration of dopaminergic neurons