Research ArticleNeurodegeneration

Phosphorylation of amyloid precursor protein by mutant LRRK2 promotes AICD activity and neurotoxicity in Parkinson’s disease

See allHide authors and affiliations

Science Signaling  18 Jul 2017:
Vol. 10, Issue 488, eaam6790
DOI: 10.1126/scisignal.aam6790

The amyloid connection in Parkinson’s disease

Parkinson’s disease (PD) is a progressive neurodegenerative disorder marked by motor control impairments and, eventually, dementia due to the loss of dopaminergic neurons in the brain. Activating mutations in the kinase LRRK2 are among the most common genetic associations with the disease. Chen et al. discovered why activating LRRK2 mutations are toxic to neurons. Mutant LRRK2 phosphorylated amyloid precursor protein (APP) at a residue within a cleaved region called the APP intracellular domain (AICD). Analysis of a mouse model, patient-derived neurons, and postmortem brain tissue suggests that phosphorylation of APP increases its nuclear translocation and transcriptional activity, leading to the loss of dopaminergic neurons. Inhibiting LRRK2 or blocking APP cleavage might be therapeutic in patients. The findings also connect the pathologies of PD and Alzheimer’s disease, in which dementia is commonly associated with the production of another cleavage product of APP, amyloid-β.


Mutations in LRRK2, which encodes leucine-rich repeat kinase 2, are the most common genetic cause of familial and sporadic Parkinson’s disease (PD), a degenerative disease of the central nervous system that causes impaired motor function and, in advanced stages, dementia. Dementia is a common symptom of another neurodegenerative disease, Alzheimer’s disease, and research suggests that there may be pathophysiological and genetic links between the two diseases. Aggregates of β amyloid [a protein produced through cleavage of amyloid precursor protein (APP)] are seen in both diseases and in PD patients carrying G2019S-mutant LRRK2. Using patient-derived cells, brain tissue, and PD model mice, we found that LRRK2 interacted with and phosphorylated APP at Thr668 within its intracellular domain (AICD). Phosphorylation of APP at Thr668 promoted AICD transcriptional activity and correlated with increased nuclear abundance of AICD and decreased abundance of a dopaminergic neuron marker in cultures and brain tissue. The AICD regulates the transcription of genes involved in cytoskeletal dynamics and apoptosis. Overexpression of AICD, but not a phosphodeficient mutant (AICDT668A), increased the loss of dopaminergic neurons in older mice expressing LRRK2G2019S. Moreover, the amount of Thr668-phosphorylated APP was substantially greater in postmortem brain tissue and dopaminergic neurons (generated by reprogramming skin cells) from LRRK2G2019S patients than in those from healthy individuals. LRRK2 inhibitors reduced the phosphorylation of APP at Thr668 in the patient-derived dopaminergic neurons and in the midbrains of LRRK2G2019S mice. Thus, APP is a substrate of LRRK2, and its phosphorylation promotes AICD function and neurotoxicity in PD.


Parkinson’s disease (PD) is a prevalent neurodegenerative disease that is characterized by the loss of midbrain dopaminergic (mDA) neurons, which leads to a reduction of dopamine in the striatum (1). Among the known genetic contributors to PD, pathogenic leucine-rich repeat kinase 2 (LRRK2) mutations are currently recognized to be the most common, accounting for up to 40% of familial cases in certain populations (2). Increasing evidence suggests that some LRRK2 mutations exert their toxicity through an enhanced LRRK2 kinase activity (3). In particular, the common G2019S mutation has been shown to elevate LRRK2 kinase activity toward generic kinase substrates and to induce neuronal loss by enhancing the phosphorylation of putative downstream targets (4). To improve our understanding on how enhanced LRRK2 kinase activity contributes to PD development, it is important to identify specific LRRK2 substrates that mediate its neurotoxicity.

Alzheimer’s disease (AD) is the most common cause of dementia among the elderly. Although AD and PD have distinct genetic risk factors that contribute to susceptibility to each disorder, previous research revealed that there might be common pathophysiological and genetic links between these two diseases (58). LRRK2 has been found to phosphorylate Tau by promoting the activity of glycogen synthase kinase–3β (GSK-3β), which induces the mislocalization of Tau (9, 10). Altered Tau phosphorylation is closely related to AD pathogenesis (11). Similarly, hyperphosphorylated amyloid precursor protein (APP) has been observed in the brains of patients with AD (12), and the phosphorylation of APP at its tyrosine or serine or threonine residues can promote APP processing and result in neuronal loss (13). For example, phosphorylation of APP at Thr668 generates amyloid depositions (12). Moreover, amyloid depositions are seen to accumulate in postmortem specimens derived from carriers of LRRK2 mutations (14), suggesting a link between the two pathogenic proteins. However, little is known about the potential pathophysiological interplay between LRRK2 and APP, particularly whether phosphorylation of APP might predispose neurons to LRRK2-mediated toxicity in PD. Here, we used LRRK2G2019S mouse models and induced pluripotent stem cell (iPSC)–derived neurons from PD patients to investigate the potential pathophysiological interplay between LRRK2 and APP in PD.


LRRK2 interacts with APP

Previous studies show that LRRK2 phosphorylates Tau and induces its mislocalization (9, 10), suggesting a link between the neurodegenerative processes in PD and AD. Given this, we speculated that LRRK2 might also interact with APP, a pathogenic protein of AD. To determine whether LRRK2 interacts with APP, we cotransfected Flag-tagged APP with LRRK2 complementary DNAs (cDNAs) into human embryonic kidney (HEK) 293T cells. Upon confirmation of overexpressed APP and LRRK2 in HEK293T cells, we performed immunoprecipitation and subsequent Western blot analysis. The assays revealed that LRRK2 and Flag-APP reciprocally coimmunoprecipitated with APP and LRRK2 (Fig. 1, A and B). When APP was coexpressed with PD-associated mutant LRRK2 (G2019S) or kinase-deficient LRRK2 (D1994A), we found that Flag-APP bound to LRRK2G2019S or LRRK2D1994A with similar efficiency relative to its interaction with wild-type LRRK2 (Fig. 1C) and vice versa (Fig. 1D). To examine the interaction on an endogenous level, we immunoprecipitated LRRK2 or APP from adult mouse brain lysates, and LRRK2 was present in the APP immunoprecipitates and vice versa (Fig. 1E). Our data therefore suggest that LRRK2 may interact with APP.

Fig. 1 LRRK2 interacts with APP.

(A and B) LRRK2 and Flag-APP were cotransfected into HEK293T cells. Whole-cell lysates (WCL) were extracted and subjected to immunoprecipitation (IP) with antibody against LRRK2 (A) and antibody against Flag (B). Blots are representative of three independent experiments. IB, immunoblotting. (C and D) LRRK2, LRRK2G2019S, and kinase-deficient LRRK2D1994A were cotransfected with Flag-APP into HEK293T cells. Flag-APP–only transfection and LRRK2-only transfection were considered as negative controls. Whole-cell lysates were extracted and subjected to immunoprecipitation with antibody against LRRK2 (C) and antibody against Flag (D). Blots are representative of three independent experiments. WT, wild type. (E) Endogenous pull-down assay in brain tissue lysates from adult mice with antibody against APP and antibody against LRRK2. Rabbit immunoglobulin G (IgG) was used as a negative control.

LRRK2 phosphorylates APP at Thr668

We then explored whether LRRK2 phosphorylates APP. Full-length LRRK2 protein and recombinant human APP were mixed with adenosine triphosphate (ATP) in kinase assay buffer and then loaded onto a 4 to 12% SDS–polyacrylamide gel electrophoresis (SDS-PAGE) gel. After electrophoresis, in-gel digestion was performed, and liquid chromatography/tandem mass spectrometry (LC/MS/MS) analysis was carried out on an Orbitrap Fusion (Thermo Fisher Scientific). We found that the only identified phosphorylation site of APP was K.QYTSIHHGVVEVDAAVTPEER.H (Thr724, equivalent to Thr668 of APP695 protein sequence) (Fig. 2A). The phosphorylation degree [area of phosphorylated peptide/(area of unphosphorylated peptide + area of phosphorylated peptide)] was estimated as 2.6% (Fig. 2B), suggesting that LRRK2 directly phosphorylated APP at Thr668. This prediction was confirmed with in vitro kinase assays (fig. S1A) as well as Western blot analyses showing that expression of either wild-type or PD-associated mutant (G2019S), but neither ROC (“Ras of complex proteins”) domain–mutant [R1441G; (15)] nor kinase-deficient (D1994A) LRRK2, increased Thr668-phosphorylated APP in transfected HEK293T cells (Fig. 2, C and D, and fig. S1, B and C). To determine whether LRRK2 phosphorylates APP at Thr668 in neuronal cells, Western blot analysis of cortical neurons isolated from the brain tissue of nontransgenic (NTg) or transgenic (Tg) LRRK2G2019S mice showed an increased abundance of phosphorylated APP in Tg neurons compared to NTg control neurons (Fig. 2, E and F). Short hairpin RNA–1 (shRNA-1)–mediated knockdown of endogenous LRRK2 abundance in NTg primary cortical neurons (fig. S1, D and E) reduced APP phosphorylation at Thr668 (Fig. 2, G and H).

Fig. 2 LRRK2 phosphorylates APP at Thr668.

(A) The higher-energy collision dissociation (HCD) MS/MS spectrum of QYTSIHHGVVEVDAAVpTPEER [mass/charge ratio (m/z), 806.3766, 3+] from APP acquired by Orbitrap [30,000 full width at half maximum at m/z of 200 and mass accuracy of <5 parts per million (ppm)]. (B) The phosphorylation degree was calculated as 2.6% at T724 (equivalent to Thr668 of APP695 protein sequence) based on the area of the extracted chromatography of the triply charged unmodified peptide (top: QYTSIHHGVVEVDAAVTPEER; MA, 106906932; retention time, 21.97 min) and phosphorylated peptide (bottom: QYTSIHHGVVEVDAAVpTPEER; MA, 2810456; retention time, 21.92 min). (C and D) Western blot analysis of phosphorylated APP at Thr668 in HEK293T cells transfected with control, LRRK2, or LRRK2G2019S constructs. Blots are representative of three independent experiments. Data are mean ± SD, n = 3 experiments. *P < 0.05, ***P < 0.001 by one-way analysis of variance (ANOVA) with Tukey’s honest significant difference (HSD) test. (E and F) Western blot analysis of phospho-APP protein at Thr668 from Tg-LRRK2G2019S and control mouse cortical neurons. (G and H) Western blot analysis for phospho-APP, total APP, and LRRK2 abundance in cortical neurons derived from NTg mice and electroporated with control and LRRK2-targeted shRNA-1. Data in (E) to (H) are mean ± SD, n = 3 experiments. **P < 0.01 by Student’s t test. Blots are representative of three independent experiments.

LRRK2 stimulates the transcriptional activity of the APP intracellular domain

Phosphorylation of APP at Thr668 affects APP processing and is known to stimulate APP intracellular domain (AICD) production through nonamyloidogenic metabolism of the protein (16). Given our demonstrated relationship between LRRK2 and APP phosphorylation, we evaluated the effect of LRRK2 on AICD transcriptional activity using the APP-Gal4 and C99-Gal4 luciferase reporter systems (Fig. 3A), in which AICD fused to Gal4 interacts with the Gal4 response element and activates the expression of a luciferase reporter (17, 18). We found that coexpression of APP-Gal4 with LRRK2 or LRRK2G2019S in HEK293T cells led to enhanced reporter activity by up to two- to fourfold compared to vector control. Moreover, LRRK2G2019S induced more AICD transcriptional activity than did wild-type LRRK2. In contrast, kinase-deficient LRRK2D1994A abolished AICD transcriptional activity when compared to LRRK2 or LRRK2G2019S (Fig. 3B). Similar to APP-Gal4, coexpression of C99-Gal4 led to a two- to threefold increase of AICD transcriptional activity in the presence of LRRK2 or, to a greater extent, LRRK2G2019S (Fig. 3C). To further confirm the role of Thr668 phosphorylation in APP processing, we mutated the APP-Gal4 and C99-Gal4 luciferase constructs at this residue to alanine (T668A). The luciferase assay showed that AICD reporter activity was inhibited significantly in cells expressing APPT668A (Fig. 3, B and C). In addition, Western blotting showed that AICD protein abundance was greater in LRRK2G2019S-transfected cells than in wild-type or kinase-deficient mutant LRRK2–transfected cells (fig. S2, A and B). Together, the data indicate that Thr668 phosphorylation promotes Gal4/reporter interaction, thus suggesting that Thr668 phosphorylation may promote AICD function.

Fig. 3 LRRK2 stimulates the transcriptional activity of the AICD.

(A) Schematic diagram of the Gal4-reporter assay system: Gal4 luciferase reporter, APP-Gal4, and C99-Gal4 constructs. (B and C) Luciferase assays were performed to show AICD reporter activity by overexpression of LRRK2 or LRRK2G2019S or LRRK2D1994A through APP-Gal4 (B) and C99-Gal4 (C) reporter systems in HEK293T cells. APP-Gal4 and C99-Gal4 luciferase constructs mutated at T668 (T668A) served as negative controls. Data are mean ± SD, n = 3 experiments. **P < 0.01, ***P < 0.001 by one-way ANOVA with Tukey’s HSD test.

AICD translocates to the nucleus and is associated with dopaminergic neuron loss in 20-month-old Tg LRRK2G2019S mice

Given the role of LRRK2G2019S in dopaminergic (DA) neurodegeneration (1, 3) and our findings that it phosphorylated APP at Thr668, we hypothesized that the two observations were linked and that LRRK2G2019S may induce DA neuronal loss through the phosphorylation of APP. To investigate this in vivo, we isolated the midbrain from 12- and 20-month-old NTg and LRRK2G2019S Tg mice, respectively. By Western blotting and immunohistochemical staining analysis, we found no significant difference in the abundance of Thr668-phosphorylated APP (fig. S3, A and B) or the amount of the DA neuron marker tyrosine hydroxylase (TH) in 12-month-old LRRK2G2019S Tg mice compared to NTg controls (fig. S3, A and C to E). This is consistent with the lack of DA neurotoxicity that is observed in 12-month-old LRRK2G2019S mice (19). In contrast, in midbrains from 20-month-old LRRK2G2019S mice, we saw increased abundance of Thr668-phosphorylated APP relative to controls (Fig. 4, A and B), which correlated with a decreased amount of TH (Fig. 4, A and C to E). To evaluate whether the neuron toxicity is selective to DA neuron in the substantia nigra (SN), we performed TH and neuronal nuclei immunofluorescence staining; the result showed that DA neuronal loss in 20-month-old LRRK2G2019S Tg mice was mainly selective to the loss of DA neurons in the SN, but not other types of neurons (fig. S7). These results suggest that phosphorylation of APP at Thr668 correlates with age-associated LRRK2G2019S-induced DA neuronal loss in vivo.

Fig. 4 APP phosphorylation at Thr668 and AICD nucleus translocation associate with DA neuronal loss in 20-month-old Tg LRRK2G2019S mice.

(A to C) Western blot analysis of phosphorylated APP (A and B) and TH (A and C) amounts in the midbrains of 20-month-old LRRK2G2019S Tg mice and NTg control mice. (D) Representative images of 3,3′-diaminobenzidine (DAB) staining of SN DA neurons with antibody against TH in 20-month-old NTg and Tg mice. Scale bars, 500 μm. (E) TH-positive cell counts are shown by stereological counting. (F) Representative images of immunohistochemical staining of SN DA neurons with phospho-APP Thr668 and TH from 20-month-old LRRK2 Tg and NTg mice. Scale bars, 50 μm (long length) and 5 μm (short length). (G) The immunostaining intensity of intranuclear phospho-APP in DA neurons was measured and compared between LRRK2 Tg and NTg mice. Data in (A) to (G) are mean ± SD from n = 5 mice each. Blots are representative of three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001 by Student’s t test.

Immunostaining of 20-month-old mouse midbrain tissues showed that phosphorylated AICD (hereafter referred to as “phospho-APP” per antibody used) was localized mainly to the nuclear compartment in LRRK2G2019S mice (Fig. 4F). We quantified the intranuclear phospho-APP immunostaining signal intensity and found that the abundance of phospho-APP per TH-positive cell in LRRK2G2019S Tg mice was significantly higher than that in NTg mice, but with less TH (cytoplasmic) in LRRK2G2019S than NTg control mice (Fig. 4, F and G). These results suggested that increased phospho-APP in the nuclear fraction was associated with a reduction in TH abundance. To investigate whether the phosphorylation of APP by LRRK2 promotes AICD nuclear translocation, we isolated the nuclear fraction from 20-month-old LRRK2G2019S midbrains and performed Western blot analysis to detect AICD using an antibody against the C terminus of APP. We found that the amount of AICD was increased in the nuclear fraction of LRRK2G2019S mouse midbrain tissues (fig. S3, F and G). We used the nuclear protein histone deacetylase 1 (HDAC1) to assess enrichment of the nuclear fraction (20). Collectively, these results indicated that the phosphorylation of APP leads to AICD translocation to the nucleus and triggers TH-positive cell loss in LRRK2G2019S mice.

AICD exacerbates LRRK2G2019S-mediated neurotoxicity in culture

Several studies demonstrate that nuclear translocation of the AICD induces neurotoxicity (16, 21). Given that phosphorylation of APP at Thr668 enhanced the nuclear translocation of the AICD and was correlated with LRRK2G2019S-induced neuronal loss (Fig. 4), we explored the direct regulation of AICD in LRRK2G2019S-mediated neurotoxicity in cultured cells using the lentiviral UbC promoter to drive AICD transgene expression and the P2A promoter to drive tdTomato (Fig. 5A). Western blot analysis demonstrated the expression of wild-type and phosphodeficient AICD constructs in HEK293T cells (Fig. 5B). Lentivirally expressed wild-type or phosphodeficient AICD (AICDT668A) was introduced into cortical neurons derived from LRRK2G2019S Tg and control (NTg) mice. We found that expression of AICD alone was sufficient to induce neuronal toxicity and exacerbated toxicity when introduced into LRRK2G2019S neurons (Fig. 5, C and D), whereas expression of phosphodeficient AICDT668A exhibited no neurotoxic effect in NTg control or LRRK2G2019S neurons (Fig. 5, C and D). Neural injury was assessed as a measure of dendritic length after AICD transduction into NTg control and LRRK2G2019S neurons. Healthy neurons were defined as having at least one smooth neurite extension that was twice the length of the cell body (22, 23). With this criterion, we found that AICD led to neuronal damage and further enhanced LRRK2G2019S-mediated neural injury (Fig. 5, E and F). Our study suggests that APP, as a substrate of LRRK2G2019S, promotes neuronal cell death and neurite shortening.

Fig. 5 AICD enhances LRRK2G2019S-mediated neurotoxicity in vitro.

(A) Schematic diagram of FUGW WT and mutant (T668A) AICD-P2A-tdTomato constructs. Lentiviral ubiquitin C (UbC) promoter drives transgene expression, and P2A promoter drives tdTomato. (B) Western blot analysis demonstrated the expression of WT AICD and phosphodeficient AICD constructs in HEK293T cells. Blots are representative of three independent experiments. (C) LRRK2G2019S-induced neurotoxicity was promoted by AICD overexpression in mouse cortical neurons, and phosphodeficient T668A AICD has no effect. Immunocytostaining was performed using antibody against TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling). Arrowheads indicate neurons that are double-positive for red fluorescent protein (RFP) and TUNEL. Scale bars, 50 μm. Images are representative of three independent experiments. (D) RFP and TUNEL double-positive cells were counted, and the results are shown. More than 1000 RFP-positive cells were counted for each condition. Data are mean ± SD, n = 3 experiments. ***P < 0.001 by one-way ANOVA with Tukey’s HSD test. (E) LRRK2G2019S-induced toxicity, indicated by neurite shortening, was promoted by AICD overexpression in mice cortical neurons, and phosphodeficient T668A AICD has no effect. Arrowheads indicate the injured neurons. Images are representative of three independent experiments. (F) The ratio of injured neurons to the total number of viable cells was assessed. Data are mean ± SD, n = 3 experiments (n > 1000 RFP-positive cells assessed for each condition). ***P < 0.001 by one-way ANOVA with Tukey’s HSD test. Scale bars, 20 μm.

AICD exacerbates LRRK2G2019S-mediated neurotoxicity in vivo

To confirm the effect of AICD on the regulation of LRRK2G2019S-mediated neurotoxicity in vivo, we performed intrastriatal administration of lentiviral-expressing wild-type and phosphodeficient AICD into 12-month-old LRRK2G2019S mice. These mice were used because we did not observe any TH loss at this age (fig. S3, A to E). Three weeks after stereotaxic injection of AICD or AICDT668A virus into the ipsilateral striatum of LRRK2G2019S mice, we found that lentivirally expressed AICD and AICDT668A were retrogradely transported to the TH-positive SN pars compacta region and infected TH-positive cells (fig. S4, A and B). Control virus was injected into the left hemisphere, whereas AICD or AICDT668A was injected into the right hemisphere of the same mouse. Immunostaining for TH revealed that wild-type but not phosphodeficient mutant AICD expression caused neuronal loss (TH cell viability), whereas phosphodeficient AICD had no significant effect in NTg mice (Fig. 6, A to C) or LRRK2G2019S Tg mice (Fig. 6, D to F); however, wild-type AICD expression enhanced the neurotoxicity of LRRK2G2019S, reducing the proportion of TH-positive cells in Tg mice further by ~20% (Fig. 6G). Thus, the cell culture and in vivo data suggest that the AICD promotes LRRK2G2019S-mediated DA neuronal loss. Together, these results indicated that AICD enhances LRRK2 neurotoxicity. Conversely, AICDT668A has no effect on neuronal cell loss.

Fig. 6 AICD enhances LRRK2G2019S-mediated neurotoxicity in vivo.

(A to C) Gain of function in AICD led to TH loss in NTg mouse in vivo. Representative images (A) and analysis (B and C) of DAB staining for TH in the SN postlentivirus-mediated delivery of control, AICD, and mutant AICD to 12-month-old NTg mice by intrastriatal administration are shown. Scale bars, 500 μm. Data are mean ± SD from n = 5 mice each. *P < 0.05 by paired t test. (D to F) Gain of function in AICD promoted LRRK2G2019S-induced neurotoxicity in Tg mice in vivo. Representative images (D) and analysis (E and F) of DAB staining for TH in the SN postlentivirus-mediated delivery of control, AICD, and mutant AICDT668A to 12-month-old Tg-LRRK2G2019S mice by intrastriatal administration are shown. Scale bars, 500 μm. Data are mean ± SD from n = 5 mice each. *P < 0.05 by paired t test. (G) AICD exacerbated neurotoxicity in Tg-LRRK2G2019S compared to NTg mice. Data are mean ± SD, n = 5 mice. **P < 0.01 by Student’s t test.

LRRK2G2019S phosphorylates APP at Thr668 in human DA neurons and postmortem tissues

To address the pathophysiological relevance of APP phosphorylation by LRRK2G2019S, we examined human DA neurons generated from human control and LRRK2G2019S-derived iPSCs. iPSC-derived neurons were characterized by immunostaining (fig. S5). LRRK2G2019S neurons displayed a significantly increased abundance of endogenous phospho-Thr668 APP (Fig. 7, A and B). We detected a significant decrease in the amount of TH in LRRK2G2019S neurons (Fig. 7, A and C), suggesting neurodegeneration. Next, we examined whether phospho-Thr668 APP was increased in LRRK2G2019S. Western blot analysis of patient tissue and age- and gender-matched healthy donor postmortem brain tissue revealed increased APP phosphorylation in the cytosolic fraction of the cortex from LRRK2G2019S patients (Fig. 7, D and E). The total abundance of LRRK2 was not significantly different (Fig. 7D).

Fig. 7 An LRRK2 inhibitor reduces APP phosphorylation and restores TH abundance in LRRK2G2019S patient–derived DA neurons and in 20-month-old LRRK2G2019S mice.

(A to C) Western blot analysis showed the amounts of phospho-APP and TH proteins in human iPSC-derived DA neurons in LRRK2G2019S and control tissues. Blots are representative of three independent experiments. Data are mean ± SD, n = 3 experiments. *P < 0.05, **P < 0.01 by Student’s t test. (D and E) Western blot analysis of the phospho-APP in cytosolic fractions of LRRK2G2019S human postmortem cortex was performed. Blots are representative of three independent experiments. Data are mean ± SD, n = 3 human samples. *P < 0.05 by Student’s t test. (F to H) Western blot analysis of phospho-APP and TH in human LRRK2G2019S iPSC-derived DA neurons treated with the LRRK2 kinase inhibitor LRRK2-IN-1 was performed at different concentrations. Blots are representative of three independent experiments. Data are mean ± SD, n = 3 experiments. **P < 0.01, ***P < 0.001 by one-way ANOVA with Tukey’s HSD test. (I to K) Western blot analysis of phosphorylated LRRK2 and APP and total TH in midbrain tissue from LRRK2G2019S mice treated with the LRRK2 kinase inhibitor HG-10-102-01 by intraperitoneal injection (50 mg/kg for 24 hours). Blots are representative of three independent experiments. Data are mean ± SD, n = 5 mice. *P < 0.05 by Student’s t test. (L) We propose that LRRK2 interacts with APP to trigger the C-terminal Thr668 phosphorylation of APP in the cytoplasm and to promote AICD translocation to the nucleus and regulate associated genes expression and then induce DA neuron loss, which contributes to LRRK2G2019S-mediated neurotoxicity.

An LRRK2 inhibitor reduces APP phosphorylation and restores TH abundance in LRRK2G2019S patient–derived DA neurons and in 20-month-old LRRK2G2019S mice

To confirm that the increased phosphorylation of APP was caused specifically by LRRK2, we applied the LRRK2 catalytic inhibitor LRRK2-IN-1 (24, 25) to cultured human iPSC-derived LRRK2G2019S DA neurons. Given that LRRK2 kinase activity is dependent on phosphorylation of Ser935 and Ser910, we evaluate the phosphorylation status of these sites to assess the efficiency of the LRRK2 inhibitor. Dose response curves showed that LRRK2-IN-1 blocked the phosphorylation of both LRRK2G2019S at Ser935 and Ser910 (fig. S6, A to C) and APP at Thr668 in a dose-dependent manner (Fig. 7, F and G). Decreased phosphorylation of APP correlated with increased abundance of TH (Fig. 7, F and H), suggesting a greater number of neurons in cultures treated with LRRK2-IN-1. To determine the pathological significance of direct phosphorylation of APP by LRRK2 in vivo, we used HG-10-102-01, a potent and selective LRRK2 inhibitor that is capable of crossing the blood-brain barrier (26). Twenty-four hours after treating mice with HG-10-102-01 [50 mg/kg by intraperitoneal injection (26)], Western blotting of harvested midbrain tissues revealed substantially less Ser935 phosphorylation of LRRK2 in 20-month-old LRRK2G2019S mice than in controls (Fig. 7I), as well as significantly less APP phosphorylation at Thr668 (Fig. 7, I and J) and less neurotoxicity as inferred from increased TH abundance (Fig. 7, I and K). Together, our results suggest that LRRK2G2019S phosphorylates APP in DA neurons, which promoted its intracellular domain translocation to the nucleus and further induce neurotoxicity (Fig. 7L).


It is interesting that LRRK2 mutations have been associated with AD-like pathology (14, 27), which supports the existence of overlapping pathways between the two diseases. APP is an important molecule in AD, and the phosphorylation of APP at tyrosine, serine, or threonine residues can promote APP processing and result in neuronal loss (16). Here, we show that LRRK2 phosphorylated APP at Thr668 and that it further promoted the transcriptional activity of AICD, an intracellular domain of APP, and its translocation to the nucleus to enhance LRRK2G2019S-induced neurotoxicity (Fig. 7L). In addition, we found that the phosphorylation of Thr668 was increased in LRRK2G2019S PD brains and in LRRK2G2019S iPSC-derived DA neurons. Conversely, treatment with LRRK2 kinase inhibitor reduced APP phosphorylation at Thr668. These observations suggest that the enhancement of Thr668-phosphorylated APP was a pathology that is specific to LRRK-related PD and that the transcriptional activity of AICD contributes to LRRK2-mediated neuronal cell loss.

By LC/MS/MS analysis of APP, we found that Thr668 was the unique site to be phosphorylated by LRRK2. Although various approaches have been performed to identify LRRK2 substrates, like SILAC (28) and phosphoproteomics (29), none of the approaches had identified APP as a potential substrate of LRRK2. Because these experiments have been set using different values and individual criteria, both false-negative and false-positive results can occur. Direct interaction experiments are still the most useful in validating specific interactions.

There is no doubt that screening of the most important LRRK2 substrates is one of the key steps in developing efficacious drugs. Steger et al. (29) identified only a subset of Rab guanosine triphosphatases as bona fide physiological LRRK2 substrates with a high-stringency assay, but they still did not rule out the existence of other functional substrates. Lin et al. (30) had explored the neuropathology of APP and G2019S double-Tg mice in their work about LRRK2 and α-synuclein and had not observed obvious APP-mediated astrocytosis and microgliosis. However, the mice they detected were only 1 month old, and the pathological change may not take place at this early stage. Therefore, the relationship between LRRK2 and APP was still elusive; our work has now uncovered APP as an LRRK2 substrate.

Phosphorylation of APP, seen in postmortem brain tissue from AD patients, promotes APP processing by β- and γ-secretase to generate Aβ and promote neuronal loss (12). In contrast, the AICD, which is also a product of this cleavage event, has received much less attention. However, previous studies have shown that AICD regulates a number of important cellular events (3133). Here, we describe a PD-related function of AICD, which promoted neuronal cell loss, reduced dendrite length, and promotes LRRK2G2019S-induced neurotoxicity in vitro and in vivo. Thus, we have discovered a new physiological role for AICD in PD pathogenesis.

Previous studies have suggested that the release of Aβ is not sufficient to explain all of the neuropathological consequences of APP cleavage and that AICD may contribute to cellular dysfunction and death in APP-associated neurodegenerative diseases (31). Mice overexpressing AICD do not show Aβ accumulation, but they do display AD-like pathologies, including Tau hyperphosphorylation, neuroinflammation, GSK-3β activation, and working memory deficits (33). Moreover, AICD is found to be increased in human AD brains (32). It has therefore been proposed that AICD may be causally involved in AD pathogenesis. Understanding AICD signaling is more important than ever because AICD is the pharmacological target of a compound under clinical investigation for AD treatment and prevention. The compound 1-(3′,4′-dichloro-2-fluoro[1,1′-biphenyl]-4-yl)-cyclopropanecarboxylic acid (CHF5074), which is in clinical trials aimed at treating mild cognitive impairment and AD (34, 35), has been shown to interact with the AICD and impair its nuclear translocation and activity (36). In the future, it will be interesting to investigate whether the inhibition of AICD will protect against LRRK2G2019S-induced neuronal loss in PD.

In conclusion, we have identified APP as a new substrate of LRRK2 that contributes to LRRK2-mediated cell loss in PD. Our study provides new insights into the mechanisms underlying LRRK2-induced neurotoxicity. It links LRRK2 and APP, two important proteins that are involved in two of the most common neurodegenerative diseases (PD and AD). Furthermore, the identification of targets of AICD can provide new therapeutic approaches to managing LRRK2-linked PD.



LRRK2G2019S Tg mice were generated using a bacterial artificial chromosome containing the entire mouse LRRK2G2019S mutation (19) and were purchased from the Jackson Laboratory (#012467). Genotypes were verified by polymerase chain reaction (PCR) using genomic DNA from tails. LRRK2G2019S Tg and NTg male mice (12 and 20 months old) were maintained in accordance with institutional guidelines, and all protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the National Neuroscience Institute (NNI) of Tan Tock Seng Hospital. The mice were maintained in a pathogen-free facility and exposed to a 12-hour light/dark cycle with food and water.

Postmortem brain tissues

We obtained LRRK2G2019S and control postmortem samples from UK Brain Bank. Postmortem study was approved by the SingHealth Institutional Review Board Committee.

Constructs and inhibitors

Human wild-type LRRK2 cDNA was synthesized by PCR using PfuUltra fusion HS DNA polymerase (#600670, Stratagene) and cloned into the pEGFPN1 vector containing GFP at the C terminus. A point mutation (G2019S) was introduced into the above LRRK2 by using an XL QuikChange site-directed mutagenesis kit (#200521, Stratagene). The APP751 (NM_201413.1) clone was purchased from OriGene. The C-terminal fragment of APP (AICD59) was amplified by PCR from human APP cDNA with the following primers: 5′-CTAGCTAGCGCCACCATGATAGCGACAGTGATCGTCATCACC-3′ (forward) and 5′-CCGCTCGAGGTTCTGCATCTGCTCAAAGAACTTG-3′ (reverse). The PCR product was cloned into the FUGW-P2A-tdTomato lentiviral vector. A point mutation (T668A) was introduced by site-directed mutagenesis according to the method described above (#200521, Stratagene). All constructs were verified by sequencing to ensure the integrity of the cloned open reading frames.To generate the APP-Gal4 and C99-Gal4 mutant constructs, the LRRK2 substrate site (Thr668, ACC) was mutated to Ala (GCC) based on the wild-type constructs using the mutagenesis kit described above with the following primers: 5′-GACGCCGCTGTCGCCCCAGAGGAGC-3′ (forward) and 5′-GCTCCTCTGGGGCGACAGCGGCGTC-3′ (reverse). The mutant constructs were verified by sequencing. The LRRK2 kinase inhibitor HG-10-102-01 was purchased from ApexBio (#B1262), and LRRK2-IN-1 was purchased from Tocris Bioscience (#4273).

Cell culture

The American Type Culture Collection cell line of HEK293T was maintained in Dulbecco’s modified Eagle’s medium (DMEM) (#D1152, Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS), 1% penicillin-streptomycin, 1% glutamine, 1% nonessential amino acid (NEAA), and 1% sodium pyruvate in an atmosphere of 5% CO2 at 37°C.


HEK293T cells were cotransfected with LRRK2 and APP plasmids using Lipofectamine 2000 (#P/N 52887, Invitrogen). Cells were collected 24 hours after transfection for Western blot analysis. Transfected HEK293T cells were washed with phosphate-buffered saline (PBS) and lysed in M-PER mammalian protein extraction reagent buffer (#78501, Thermo Fisher Scientific) supplemented with protease inhibitors (#11697498001, Roche). The lysates were then incubated with antibody against LRRK2 (ab133474, Abcam) overnight at 4°C. For reverse coimmunoprecipitation, lysates were incubated with antibody against Flag (F3165, Sigma-Aldrich) overnight at 4°C. Endogenous pull-down was performed using adult mouse brain lysates with antibody against LRRK2 (NB300-268, Novus Biologicals) and antibody against APP (#A8717, Sigma-Aldrich). The precipitates were then washed five times using NP-40 buffer [50 mM tris (pH 7.4), 300 mM NaCl, and 1% NP-40] and resuspended in 2× SDS loading buffer for Western blot analysis.

LC/MS/MS analysis

Wild-type, full-length LRRK2 protein (1 μl; cat. no. A15197, 0.2 mg/ml; Life Technologies) and recombinant human APP751 (10 μl; cat. no. 842601, 0.1 mg/ml; BioLegend) were mixed with kinase assay buffer [50 mM tris-HCl (pH 7.4) and 50 mM MgCl2] and heated to 30°C. We added 10 μl of ATP solution [1 mM in 10 mM Hepes buffer (pH 7.4)] and incubated for 30 min at 30°C. Laemmli buffer was added to stop the reaction and loaded to a 4 to 12% SDS-PAGE gel (NuPAGE, Thermo Fisher Scientific).

In-gel digestion was performed according to the standard protocol. The LC/MS/MS analysis was carried out on an Orbitrap Fusion (Thermo Fisher Scientific). The column was 20 cm × 75 μm (inner diameter) home-packed with 3-μm C18 (Reprosil, Dr. Maisch). The tryptic peptides were separated on a nano-UPLC (Easy-nLC 1000, Thermo Fisher Scientific) with gradient from 0% mobile phase A (water with 0.1% formic acid) to 30% mobile phase B (acetonitrile with 0.1% formic acid) in 30 min. The mass spectrometer was configured to perform MS/MS (HCD) on top abundant multiple-charged ions (400 to 1500 m/z) every 3 s.

The mass spectrometry and MS/MS spectra from the raw data were sent to the Mascot (Matrix Science) database search engine against the UniProt human protein database (90,441 sequences). The search parameters were peptide mass tolerance (7 ppm), fragment mass tolerance (0.3 Da), enzyme (trypsin), maximum missed cleavages (1), fixed modification [carbamidomethyl (C)], and variable modifications [oxidation (M) and phospho (ST)]. The quantification of phosphorylation was based on the extracted ion chromatograph of the triply charged nonphosphorylated and phosphorylated peptide containing Thr724.

Primary mouse cortical neuronal cultures and electroporation

Primary cortical neuronal cultures were prepared from embryonic day 16 (E16) Tg LRRK2G2019S mouse cortex. The cortex was dissected in Hepes-buffered HBSS, incubated in 0.025% trypsin for 30 min, and then washed with HBSS. We triturated brain tissues very gently with a fire-polished glass Pasteur pipette with a reduced diameter (10 to 15 times) in seeding medium (DMEM supplemented with 5% FBS). We allowed the cells to settle for 2 min and then removed the supernatant to a new tube. Dissociated cells were counted for subsequent experiments. Electroporation of the AICD, AICD mutant, and control plasmids into primary neurons was performed using a Neon electroporator (Invitrogen) according to the manufacturer’s instructions. Briefly, 1 to 2 × 106 cells were resuspended in Nucleofector solution and electroporated using a preestablished program. The electroporated cells were plated on poly-l-lysine–coated four-well coverslips at a density of 0.5 to 0.7 × 106 cells per well. After 4 hours, the medium was exchanged for feeding medium (Neurobasal containing GlutaMAX, B27, and antibiotics). After 48 hours, the transfected neurons were fixed, permeabilized, and immunostained.

iPSC-induced human DA neuron cultures

Human LRRK2G2019S and control iPSCs were purchased from Coriell Institute. All iPSCs were cultured in chemically defined mTeSR medium (#05850, STEMCELL Technologies) on Matrigel-coated tissue culture plates. Confluent cultures were passaged using Dispase (1 mg/ml; #07923, STEMCELL Technologies) at a split ratio of 1:6 every 7 days. For neuronal differentiation, iPSC colonies were detached using Dispase (1 mg/ml; #07923, STEMCELL Technologies) and were split 1:6 and seeded on Matrigel-coated six-well plates in the presence of 10 μM ROCK inhibitor. Neural induction from iPSCs into neural progenitor cells (NPCs) was initiated 1 day after passaging the iPSCs. At day 1, mTeSR medium was replaced with N2B27 medium (DMEM/F12 with 2 mM l-glutamine and 1× N2 supplement and Neurobasal medium with 0.2 mM NEAA and 1× B27 without vitamin A supplement, mixed at a 1:1 ratio) supplemented with 4 μM CHIR99021 (#C2447-2S, Cellagen Technology), 3 μM SB431542 (#C7243-5, Cellagen Technology), 0.1 μM Compound E (#565790, γ-Secretase Inhibitor XXI, EMD Chemicals Inc.), bovine serum albumin (5 μg/ml), and human leukemia inhibitory factor (10 ng/ml) (#L5283, Sigma-Aldrich) for 7 days. At day 8, cells were dissociated into single cells by Accutase (#SCR005, Merck Millipore) and split on cell culture plates that were coated with poly-l-ornithine hydrochloride and laminin. Starting from day 8, NPCs were treated with mDA patterning medium consisting of N2 and B27 media supplemented with fibroblast growth factor 8 (100 ng/ml) (#100-25, PeproTech), 2 μM purmorphamine (#540220, Merck Millipore), dibutyryl cyclic adenosine monophosphate (db-cAMP; 300 ng/ml) (#D0627, Sigma-Aldrich), and 200 μM l-ascorbic acid (l-AA; #A4403, Sigma-Aldrich) for 14 days. From day 22, cells were fed with maturation medium consisting of N2B27 medium supplemented with human glial cell line–derived neurotrophic factor (20 ng/ml; #212-GD, R&D Systems), human brain-derived neurotrophic factor (20 ng/ml; #248-BD, R&D Systems), 0.5 mM db-cAMP, and 200 μM l-AA. mDA neuron differentiation cultures were replated as single cells on days 36 to 40 on poly-l-ornithine hydrochloride and laminin-coated plates for further experiments.

Subcellular fractions and Western blot analysis

Tissues or cells were lysed in the lysis buffer supplemented with protease inhibitors [50 mM Hepes (pH 7.3), 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, 1 mM Na3VO4, 1 mM phenylmethylsulfonyl fluoride, and aprotinin (10 μg/ml)]. Harvested protein lysates were separated on 7 to 12% SDS-PAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes (#IPVH00010, Millipore). The blots were incubated with antibody against LRRK2 (1:1000; ab133474, Abcam), APP (1:1000; A8717, Sigma-Aldrich), phospho-APP Thr668 (1:1000; #3823S, Cell Signaling), TH (1:500; NB300-109, Novus Biologicals), or β-actin (1:3000; sc-69879, Santa Cruz Biotechnology) antibodies. For detection of the APP C-terminal fragment (AICD) in the mouse midbrain, cellular fractionation was performed. Nuclear and cytoplasmic fractions were extracted using the NE-PER Nuclear and Cytoplasmic Extraction Kit (#78835, Thermo Fisher Scientific) following the manufacturer’s protocol. Cytoplasmic and nuclear extracts were separated by 4 to 20% Mini-PROTEAN TGX Stain-Free Precast gels (#456-8094, Bio-Rad) and transferred to a Trans-Blot Turbo Transfer Pack with 0.2-μm PVDF (#170-4156, Bio-Rad). The blots were probed with antibody against β-tubulin (1:3000; #05-661, Millipore) to detect the cytosolic fraction and antibody against HDAC1 (1:1000; #06-720, Millipore) to detect the nuclear fraction. The membranes were incubated with primary antibodies overnight at 4°C in TBST (tris-buffered saline–Tween 20) supplemented with 5% skim milk, washed in TBST, and then incubated with secondary horseradish peroxidase (HRP)–conjugated antibody against mouse IgG (#NA931V, GE Healthcare) or HRP-conjugated antibody against rabbit IgG (#NA934V, GE Healthcare) and developed using an ECL detection kit (#RPN2106, GE Healthcare) or a SuperSignal West Femto detection kit (#34096, Thermo Fisher Scientific).

Immunohistochemistry and intranuclear intensity measurement

Animals were anesthetized and intracardially perfused with ice-cold PBS followed by 4% paraformaldehyde in 0.1 M phosphate buffer. The brains were removed and fixed in the same fixative. After cryoprotection with 30% sucrose, a complete set of 30-μm-thick serial sections was cut through the SN on a freezing microtome (#CM3050S model, Leica). Every sixth section was selected. Free-floating sections in one group were blocked in 10% goat serum plus 0.3% Triton X-100 and incubated with antibodies against phospho-APP (#3823S, Cell Signaling) and TH (#MAB318, Millipore), followed by incubation with streptavidin–Alexa 488 (1:500; #S-11223, Invitrogen) and donkey against mouse Alexa 555 (1:500; #A-31570, Invitrogen) secondary antibodies. Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) (1:200; #268298, Calbiochem). Finally, sections were mounted on glass slides. Images were acquired with a confocal microscope (Olympus). The fluorescence intensities of phosphorylated APP in nuclear fraction were measured by analyzing the images with ImageJ software.

Luciferase assay

The day before transfection, HEK293T cells were cultured in 24-well plates and then transiently cotransfected with the following constructs: 0.1 μg per well of pG5E1B and APP-Gal4 or C99-Gal4, luciferase internal control plasmid pCMV-β-Gal and 0.05 μg per well of LRRK2 or LRRK2G2019S or LRRK2D1994A overexpression constructs using Lipofectamine reagent (#18324-012, Life Technologies). Cells were harvested 24 to 48 hours after transfection, and luciferase activity was measured with the Steady-Glo Luciferase Assay Kit (E2510, Promega) using a GloMAX 20/20 luminometer (Promega). Luciferase activity was normalized to pCMV-β-Gal luciferase activity.

Lentivirus production

Lentivirus was prepared according to the established protocols. Briefly, lentivirus packaging was performed by cotransfecting lentivectors with pLP, VSV-G (Invitrogen), and psPAX2 (Addgene plasmid #12260) into 293FT cells cultured in Opti-MEM I medium using Lipofectamine 2000 (all from Invitrogen). Culture supernatant was collected on day 2 and 3 after transfection and passed through a 0.45-μm filter. Viral particles were concentrated from culture supernatants by ultracentrifugation through a 20% sucrose cushion at 20,000 rpm in an SW 28 rotor (Beckman Coulter) for 3 hours at 4°C. Viral pellets used for in vivo delivery were resuspended in a minimal volume of HBSS (Invitrogen), whereas pellets for neuronal transduction were resuspended in Neurobasal medium (Invitrogen).

Lentivirus in vivo delivery

For stereotaxic injection of lentivirus particles overexpressing control, AICD, and mutant AICD, experimental procedures were followed according to the guidelines of the Laboratory Animal Manual of the NNI Guide for the Care and Use of Animals. All procedures used in this study were approved by the NNI Animal Care Committee and the NNI IACUC. Twelve-month-old LRRK2G2019S Tg mice were anesthetized using 2% xylazine and 10% ketamine and placed in a Kopf stereotaxic frame. For nigrostriatal transduction, either lentivirally expressed control, AICD, or mutant AICDT668A infectious particles was injected bilaterally into the striatum (anterior/posterior, +1.2 mm; medial/lateral, −1.3 mm; and dorsal/ventral, −3.2 mm) (2 μl per site) at a rate of 0.25 μl/min with a 5-μl Hamilton syringe (25 gauge) driven by a Harvard pump. After infusion, the cannula remained in place for 4 min to prevent reflux. Each lentivirus particle was injected into least six sites in different mice brains.

Three weeks after injection, animals were perfused with 1× PBS followed by 4% paraformaldehyde. Brains were postfixed with 4% paraformaldehyde, cryoprotected in 30% sucrose, and processed for immunohistochemistry. Thirty-micrometer coronal sections were cut throughout the brain, including the striatum and SN. Every sixth section was taken for TH staining.

In vitro kinase assay

In vitro kinase assay was carried out using recombinant LRRK2 protein (cat. no. A15197, Life Technologies), ATP (Sigma-Aldrich), kinase buffer (Cell Signaling), and purified APP–glutathione S-transferase (GST) protein. The kinase assay was carried out for 2 hours at room temperature and was stopped by adding 2× SDS loading buffer and boiled for 5 min at 95°C. The protein was loaded on SDS-PAGE gel, and APP phosphorylation was detected by antibody against phospho-APP T668.

In vitro toxicity assay

Neuronal toxicity assay includes cell death and neurite shortening (neural injury) assays. For cell death assays, neurons were subject to TUNEL and DAPI staining following the manufacturer’s protocols and then visualized using a confocal microscope (Olympus). For neurite shortening assays, healthy neurons were defined as having at least one smooth neurite extension that was twice the length of the cell body (25, 26), whereas injured neurons were defined as those with neurite extensions that were less than twice the length of the cell body. The ratio of injured neurons to the total number of viable cells was assessed.

DAB staining and stereology counting

For TH DAB staining, brain sections were reacted with a 1:1000 dilution of mouse monoclonal antibody against TH (MAB318, Millipore) and visualized with biotinylated goat against mouse IgG, followed by streptavidin-conjugated HRP (Mouse on Mouse kit, PK-2200, Vector Laboratories). Positive immunostaining was visualized with DAB after reaction with hydrogen peroxide (DAB kit, SK-4100, Vector Laboratories). Total numbers of TH-stained neurons in substantia nigra pars compacta were counted using the stereological cell counting with the Optical Fractionator probe of the Stereologer 2000 software.

Statistical analyses

For each experiment, at least three independent experiments were performed. The images obtained from one representative experiment were presented. Statistical analysis was performed using SPSS. Data were presented as mean ± SD for Western blotting and cell counting quantification. For two-group comparisons, two-tailed Student’s t test was performed. Paired t test was used to evaluate TH viability after virus injection in 12-month-old mice. For multiple comparisons, data with a normal distribution were analyzed by one-way ANOVA followed by the Tukey’s HSD test. The investigators were blinded to genotypes during experiments involving immunohistochemical counting. Significance level was set at P < 0.05.


Fig. S1. LRRK2 phosphorylates APP at Thr668.

Fig. S2. G2019S mutation in LRRK2 stimulates AICD abundance.

Fig. S3. There was no TH loss in the 12-month-old LRRK2G2019S mice.

Fig. S4. Expression and wide diffusion of lentiviral AICD in striatum.

Fig. S5. Characterization of iPSC-derived human DA neurons.

Fig. S6. Dose-dependent study of LRRK2-IN-1 in human LRRK2G2019S neurons.

Fig. S7. AICD-associated TH loss in 20-month-old LRRK2G2019S mice is selective to DA neurons in the SN.


Acknowledgments: We thank L. Parsons for providing the LRRK2G2019S and control postmortem samples. Funding: This research was supported by the Singapore National Research Foundation under its Translational and Clinical Research Flagship Programme and administered by the Singapore Ministry of Health’s National Medical Research Council and by the National Medical Research Foundation’s StaR Award. Author contributions: Z.-C.C., W.Z., and L.-L.C. conceived and designed the study and acquired, analyzed, and interpreted the data. C.C., L.L., R.L., Z.-D.Z., Z.C., and D.C.A. acquired the data. L.W.S. and J.-H.P. analyzed and interpreted the data. K.-L.L. drafted the manuscript. L.Z. and E.-K.T. conceived and designed the study, acquired, analyzed, and interpreted the data, and drafted the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The raw mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium through the jPOST Repository [(; data set identifier JPST000263/PXD006399].

Stay Connected to Science Signaling

Navigate This Article