Antibodies against PARP (CST, 9532) and cleaved-PARP (CST, 9541) were from Cell Signaling, Inc. The anti-flag antibody, H₂O₂, and curcumin were from Sigma. Anti-LRRK2 [MJFF4 (c81-8)] was from MJFF. ECL goat anti-Rabbit IgG and ECL goat anti-Mouse IgG were from GE Healthcare (NA934V and NA931V, respectively). Mito-tracker and MitoSox were from Molecular Probe.

WT-LRRK2, G2385R-LRRK2, and G2019S-LRRK2 constructs contain human LRRK2 cDNA with Flag-tags as described previously (Yang et al., 2012; Thomas et al., 2016). Human embryonic kidney (HEK293) cells were from the American Type Culture Collection (ATCC; Manassas, VA, USA) and were grown in Dulbecco's modified Eagle medium (DMEM; Gibco) with 10% heat-inactive fetal bovine serum (FBS; Gibco) and 1% penicillin–streptomycin (Gibco) in a humidified 5% CO₂ atmosphere at 37°C. SH-SY5Y human neuroblastoma cells were from ATCC grown in Opti-Mem I media with 10% FBS and 1% penicillin-streptomycin. Cells were transfected with vectors, GFP, and various LRRK2 constructs using LipofectAMINE and a Plus reagent (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. For treatment experiments, curcumin (1 μM) was added 1 h prior to transfection and then transfected various constructs in transfection media without curcumin. Post-transfection lasted 4 h, and cells were incubated to normal medial with curcumin (1 μM in double-distilled H₂O) for 48 h. H₂O₂ (100 μM) was added 4 h post-transfection and until the end of the experiment.

SH-SY5Y cell viability assays were performed as described previously with slight modification (Smith et al., 2005; Yang et al., 2012; Thomas et al., 2016). Briefly, post-transfection lasted 46 h, and neurons expressing GFP were imaged for 2 h with 20 min intervals on a Zeiss Axiovert 200 inverted microscope. The healthy GFP-positive cells were counted (where at least one process is smooth with two folds of cell body length) by an automatic program to detect healthy and smooth neurites and DAPI nuclei as described previously (Smith et al., 2005; Yang et al., 2012).

Cells were transfected with vector or various LRRK2 constructs for 12–48 h. MitoSOX (5 μM) and MitoTracker (1 μM) were added 20 min before imaging as described previously (Wang et al., 2020). The unused free MitoSox was washed away using PBS. Hoechst 33342 (blue) was used for nuclei staining. Fluorescent images were captured under a fluorescent microscopy image system (Carl Zeiss) with identical exposure settings. NIH image software was used to quantify the density of MitoSOX (Red), which represents the levels of mitochondrial superoxide. The MitoTracker was used to confirm the localization of MitoSOX in mitochondria.

The Nexcelom ViaStain™ Live Caspase-3/7 Detection kit (CSK-V0003-1, Nexcelom) was employed to detect caspase-3/7 activities according to manufactural protocol. The assay is based on the NucViewTM reagent with a nucleic-acid-binding dye linked to a fluorescent probe that is attached to a four-amino-acid peptide sequence DEVD (Asp-Glu-Val-Asp) to form a cell-membrane-permeable DEVD-DNA complex. During apoptosis, activated caspase-3/7 cleaves the DEVD-DNA dye complex to release the high-affinity DNA dye, which translocates to the nucleus and binds to the DNA, producing a bright green, fluorescent signal. When cells are healthy, the nucleic-acid dye is linked to the DEVD peptide sequence, and the dye is unable to bind to DNA, resulting in a lack of flourescence. Cells were transfected with a vector or LRRK2 constructs for 48 h and then incubated with NucViewTM reagent for 30 min. The green fluorescence was read using a Celigo fluorescent reader.

Mouse primary cortical neurons were generated from E16 embryos as described previously (Nucifora et al., 2016), and they were cultured in Neurobasal medium with 2 mM Glutamax, 2% B27, and 1% PenStrep antibiotics. At 5 days in vitro (DIV), neurons were co-transfected with various LRRK2 plasmids and green fluorescent protein (GFP) at a ratio of 10:1 using Lipofectamine 2000 (Invitrogen) as described previously (16, 17). About 95% of GFP-positive cells expressed LRRK2 variants confirmed by co-immunostaining using anti-GFP and anti-LRRK2 antibodies. Curcumin was added 1 h prior to transfection, and H₂O₂ (100 μM) was added 4 h post-transfection.

Neuronal injury assays were performed as described previously (Smith et al., 2005; Liu et al., 2016; Wang et al., 2021). Neurons were imaged at end of experiments. Neurons with injury neurites were blindly quantified using NIH ImageJ software. Injured neurons were defined as displaying loss of neuronal projections, a disappearance of nerites, the presence of soma fragmentation, nerites with beading morphology, or a soma shape change from oval to completely circular. An arbitrary morphological score was assigned as either 100 for healthy cells or 0 for injured cells in each frame. The mean morphological score of all the cells was calculated. Data analysis was performed using a Gehan-Breslow test to determine the statistical difference between groups, followed by an All Pairwise Multiple Comparison (Holm-Sidak method) to identify the differences between groups.

Cells were collected using cell lysis buffer (Cell Signaling, Inc, CST-9803S) with 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton, 2.5 mM sodium pyrophosphate, 1 mM beta-glycerophosphate, 1 mM Na₃VO₄, 1 μg/ml leupeptin, and 1 mM PMSF. Cell lysates from each group were subjected to protein assays. An aliquot of 30 μg protein from each sample was resolved on 4–12% NuPAGE Bis/Tris gels and transferred onto polyvinylidene difluoride membranes (Millipore). The membranes were blocked with 4% milk for 1 h, then incubated with various primary antibodies as indicated in Reagents section: anti-Flag (1:1,000 dilution), rabbit anti-PARP (1:1,000 dilution), and rabbit anti-cleaved PARP (1:1,000 dilution) followed by HRP-conjugated secondary antibodies: goat anti-Rabbit IgG (1:5,000 dilution) or goat anti-Mouse IgG (1:5,000 dilution). Protein bands were detected by ECL enhanced chemiluminescence reagents (PerkinElmer) as described previously (Li et al., 2014; Yang et al., 2018; Arbez et al., 2020).

Quantitative data represent the means ± SEM of three separate experiments. There were three parallel reactions for each sample for all biochemistry experiments. Statistically significant differences among groups were analyzed by ANOVA by using the GraphPad Prism 5 software. The p value for significance was 0.05.

G2385R-LRRK2 was widely identified in the genetic studies of the Asian population (Zhang et al., 2018). Its neurobiological function is understudied compared with the most common PD-linked mutation, G2019S-LRRK2. We found that transient expression of G2385R-LRRK2 significantly induced neurotoxicity in human neuroblastoma SH-SY5Y cells compared with vector cells alone or cells expressing wild-type (WT)-LRRK2 (Figure 1). Compared to G2019S-LRRK2, G2385R-LRRK2 induced slightly less cell toxicity but exhibited no statistical difference. Thus, we only focused on G2385R-LRRK2 in this study.

To study whether G2385R-LRRK2 alters mitochondrial ROS, MitoSOX florescent assays were used. The MitoSOX™ Red reagent can selectively label the production of mitochondrial superoxide under fluorescence microscopy. Expression of either WT or G2385R-LRRK2 increased the ROS level in mitochondria compared with vector cells (Figures 2A,B), while cells expressing G2385R-LRRK2 had more mitochondrial ROS than those of WT-LRRK2. Increased ROS (oxidative stress) could induce mitochondrial dysfunction and result in the activation of a caspase cascade. In fact, we found that expression of WT-LRRK2 slightly increased caspase 3/7 activities (Figure 2C). However, G2385R-LRRK2 significantly increased caspase 3/7 activities compared with vector cells or cells expressing WT-LRRK2. Treatment with an antioxidant, curcumin, at 1 μM concentration (the most effective dose based on our pilot experiment) significantly protected against G2385R-induced neurodegeneration in SH-SY5Y cells compared with vehicle control (Figure 3).

To further validate G2385R-LRRK2-induced neurodegeneration, we used mouse primary cortical neurons. Consistent with our findings in SH-SY5Y cells, expression of G2385R-LRRK2 dramatically induced neurite injury in mouse primary neurons compared with vector or WT-LRRK2 groups (Figure 4). Curcumin significantly attenuated G2385R-induced neurite injury compared with the vehicle group (Figure 4).

To further study the interaction between environmental stress and G2385R mutation, we added subtoxic doses of H₂O₂ (100 μM) in neurons expressing vector or LRRK2 (WT or G2385R) variants. We found that H₂O₂ promoted both WT-LRRK2 and G2385R-LRRK2-induced neurite injury (Figure 4), but the neurons with G2385R were more vulnerable to H₂O₂ than those of WT-LRRK2 cells. There was about a 3.5-fold increase in neurons with neurite injury in the G2385R-LRRK2 plus H₂O₂ group compared to vector cells with H₂O₂ (Figure 4). Curcumin also protected against this combined neurotoxicity (H₂O₂ and G2385R), up to 50% (Figure 4).

Exposure of subtoxic doses of H₂O₂ (100 μM) in SH-SY5Y cells also dramatically increased both WT-LRRK2- and G2385R-LRRK2-induced mitochondrial ROS levels compared to no exposure control group (Figure 5A), while the ROS increased more in cells expressing the G2385R variant. Moreover, Curcumin treatment significantly attenuated the mitochondrial ROS induced by H₂O₂ and G2385R-LRRK2 compared with vehicle control (Figure 5A). Exposure of subtoxic doses of H₂O₂ also dramatically increased both WT-LRRK2 and G2385R-LRRK2-induced caspase-3/7 activation compared to those cells without H₂O₂ (Figure 5B). Curcumin treatment significantly reduced caspase-3/7 activation (Figure 5B).

PARP is an enzyme in the nucleus and can be cleaved (inactive) by active caspase-3 during apoptosis. To assess whether G2385R alters PARP cleavage, we used HEK293 cells transiently transfected with LRRK2 constructs as these cells have high transfected efficiency (over 80%) and extract the protein, then perform western blot analysis. We found that expression of mutant G2385R-LRRK2 significantly increased PARP cleavage compared to vector cells (Figure 6). Moreover, H₂O₂ dramatically increased G2385R-LRRK2-induced PARP cleavage compared to those cells without H₂O₂ (Figure 6). While curcumin treatment significantly attenuated PARP cleavage-induced by H₂O₂ and G2385R-LRRK2 compared with vehicle controls (Figure 6).