Abstract
GBA1 encodes the lysosomal enzyme β-glucocerebrosidase (GCase) which converts glucosylceramide into ceramide and glucose. Mutations in GBA1 lead to Gaucher’s disease and are a major risk factor for Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB), synucleinopathies characterized by accumulation of intracellular α-synuclein. In this study, we examined whether decreased ceramide that is observed in GCase-deficient cells contributes to α-synuclein accumulation. We demonstrated that deficiency of GCase leads to a reduction of C18-ceramide species and altered intracellular localization of Rab8a, a small GTPase implicated in secretory autophagy, that contributed to impaired secretion of α-synuclein and accumulation of intracellular α-synuclein. This secretory defect was rescued by exogenous C18-ceramide or chemical inhibition of lysosomal enzyme acid ceramidase that converts lysosomal ceramide into sphingosine. Inhibition of acid ceramidase by carmofur resulted in increased ceramide levels and decreased glucosylsphingosine levels in GCase-deficient cells, and also reduced oxidized α-synuclein and levels of ubiquitinated proteins in GBA1-PD patient-derived dopaminergic neurons. Together, these results suggest that decreased ceramide generation via the catabolic lysosomal salvage pathway in GCase mutant cells contributes to α-synuclein accumulation, potentially due to impaired secretory autophagy. We thus propose that acid ceramidase inhibition which restores ceramide levels may be a potential therapeutic strategy to target synucleinopathies linked to GBA1 mutations including PD and DLB.
Introduction
Parkinson’s disease (PD) is a progressive neurodegenerative disease resulting in dopaminergic neuronal loss in the substantia nigra pars compacta, and clinically characterized by bradykinesia, tremor, postural instability and non-motor symptoms including depression, hyposmia, gastrointestinal dysfunction and sleep disturbance (1,2). Lewy body inclusions containing α-synuclein are a pathological hallmark of multiple synucleinopathies including PD (3). Accumulating evidence suggests that multiple pathways including mitochondrial dysfunction, vesicle trafficking and autophagy/lysosomes are involved in the pathogenesis of synucleinopathies (4).
Mutations in GBA1 represent the greatest genetic risk factor for PD, and increase the risk of PD by 5–6-fold in the general population (5). Approximately 5–10% of PD patients carry a GBA1 mutation (6) and the penetrance rate of PD by age 85 in GBA1 mutation carriers is ∼10.9% (7). GBA1 encodes the lysosomal enzyme β-glucocerebrosidase (GCase), which is ubiquitously expressed in multiple cell types and converts glucosylceramide (GluCer) into ceramide and glucose within the lysosome. Homozygous or compound heterozygous mutations in GBA1 cause Gaucher’s disease (GD), the most common autosomal recessive lysosomal storage disease, and the severe loss of GCase activity (>75%) in GD patients results in elevated levels of GluCer and its deacylated form glucosylsphingosine (GluSph) primarily in macrophages but also in neurons (8,9).
In contrast, most GBA1-linked PD patients carry heterozygous GBA1 mutations and do not present with GD symptoms (10). Clinically, PD patients with GBA1 mutation are indistinguishable from sporadic PD patients and are positive for Lewy body pathology (11). GBA1 mutations also increase the risk of Dementia with Lewy Body (DLB) by ∼9-fold (12), suggesting that GBA1 mutations contribute to the pathogenesis of synucleinopathies.
Recent evidence has shown that loss of GCase activity is correlated with α-synuclein accumulation (13). In sporadic PD, reduced GCase activity is associated with increased α-synuclein levels (14,15), and PD and DLB patient brains show selective decreased activity of GCase, but not of multiple other lysosomal hydrolases (16). GBA1D409V/D409V knock-in mice lacking ∼80% of GCase activity accumulate α-synuclein (17,18), and chemical inhibition of GCase with conduritol beta-epoxide (CBE) also promotes α-synuclein accumulation (19). Conversely, adeno-associated virus-mediated expression of GBA1 ameliorates α-synuclein accumulation in GBA1D409V/D409V mice and in transgenic α-synuclein-overexpressing mice (17,20). Moreover, small chemical chaperone activators of GCase reduce intracellular α-synuclein levels in GBA1-linked PD patient dopaminergic neurons (21,22). We previously found that GluCer, the lipid substrate of GCase, promotes the formations of α-synuclein fibrils and oligomers in a pH-dependent manner (13), whereas in this study we examined whether decreased ceramide in GCase-deficient cells contributes to α-synuclein accumulation.
Ceramides are a lipid family consisting of sphingosine (Sph) and a fatty acid, which can be generated via de novo synthesis in the ER (23,24). However, ceramides can also be generated in the lysosome via the catabolic salvage pathway by several lysosomal enzymes including GCase, which converts GluCer into ceramide (25,26). Lysosomal ceramide is subsequently converted to Sph by acid ceramidase, a downstream enzyme in the ceramide pathway (27,28). Although recent efforts have focused on the role of GCase and its potential as a therapeutic target in PD (21,22,29), whether targeting the downstream activity of acid ceramidase is beneficial for decreasing α-synuclein levels in synucleinopathies has not been studied.
We hypothesized that impaired ceramide generation in GCase-deficient cells contributes to α-synuclein accumulation, and that restoring lysosomal ceramide levels by acid ceramidase inhibition promotes the clearance of α-synuclein. We demonstrated that loss of GCase activity leads to a reduction of C18-ceramide species and alters the intracellular localization of Rab8a, a small GTPase implicated in secretory autophagy, contributing to impaired Baf-A1-induced α-synuclein secretion and increased intracellular α-synuclein accumulation. We further show that exogenous C18-ceramide (C18-Cer) or chemical inhibition of acid ceramidase in GCase-deficient cells rescues defects in Baf-A1-induced α-synuclein secretion and secretory autophagy. Finally, we found that chemical inhibition of acid ceramidase decreased oxidized α-synuclein and ubiquitinated protein species in dopamine neurons derived from a PD patient harboring a heterozygous GBA1-c.84dupG frameshift mutation or in dopamine neurons derived from a PD patient carrying a heterozygous GBA1-N370S mutation, suggesting that acid ceramidase inhibition may be an important angle for therapeutically targeting α-synuclein toxicity in synucleinopathies such as PD.
Results
Secretory autophagy is disrupted by loss of GBA1
To examine cellular pathways affected by loss of GCase activity, we generated a GCase-deficient HEK293-FT cell line using CRISPR/Cas9-based genome editing. The GBA1-CRISPR/Cas9 construct targeted four out of five human GBA1 isoforms (Fig. 1A), and led to almost complete loss of GCase protein by immunoblot analysis using two independent GCase antibodies detecting either the N-terminal or C-terminal region of GCase (Fig. 1B). We further verified that this led to dramatically decreased GCase activity (Fig. 1C), and confirmed that the majority of GCase activity in wild-type cells was sensitive to CBE, an irreversible inhibitor of GCase (Fig. 1C). Immunostaining for GluCer, the lipid substrate of GCase, demonstrated that GCase-deficient cells exhibited increased GluCer compared with wild-type cells (Fig. 1D).
Characterization of GCase-deficient cells. (A) Schematic diagram of human GBA1 gene structure and target sequence of GBA1-CRISPR/Cas9 construct. The GBA1-CRISPR construct targets four out of five human GBA1 isoforms. (B) Cell lysates from wild-type (WT) and GCase-deficient (KO) HEK293-FT cells were subjected to immunoblot analysis using an N-terminal or C-terminal GCase antibody. (C) Triton X-100 soluble cell lysates were prepared from wild-type or GCase-deficient cells. GCase activity in 7.5 μg of cell lysates was measured in the presence or absence of CBE. The detailed GCase assay is described in the Materials and Methods section. GCase activity was measured in triplicate. (D) Cells were fixed with 4% formaldehyde in PBS and immuno-stained with mouse anti-GluCer antibody and DAPI. Representative images are shown. Data represent mean ± S.E.M. N = 30 microscopic fields for wild-type cells, n = 25 microscopic fields for GCase-deficient cells, two-tailed unpaired t-test, ***P < 0.001 compared with wild-type cells. n.s. = not significant.
Consistent with other models of GCase deficiency (13,17,18,30–34), we found elevated α-synuclein levels (Fig. 2A), as well as accumulation of autophagic substrates p62/SQSTMI and LC3 and lysosomal protein LAMP2, compared with wild-type cells (Fig. 2B–D). As reported previously (35,36), Rab7 and Lamp2 protein levels were also increased in GCase-deficient cells (Fig. 2D). We did not observe any significant changes in the rate of starvation-induced autophagic degradation of p62 (Fig. 3A and B), CCCP-induced autophagic degradation of p62 or CCCP-induced proteasomal mitofusin2 (Mfn2) degradation (Fig. 3C and D). However, we found decreased secretion of p62 in GCase-deficient cells (Fig. 3E and F), suggesting the involvement of secretory autophagy. To study secretory autophagy, we examined the intracellular and extracellular media from cells treated with bafilomycin-A1 (Baf-A1), a potent inhibitor of the vacuolar ATPase (V-ATPase), which induces extracellular secretion via various pathways including secretory autophagy (37,38) and lysosomal exocytosis (39). The Baf-A1-induced extracellular secretion of p62 is thought to be mediated by secretory autophagy (also known as exophagy) (37,38). In addition, the secretion of α-synuclein has been reported previously to be mediated in part by secretory autophagy (40,41), as well as via exosomal secretion (42), conventional Brefeldin A (BFA)-sensitive Golgi secretion (43) and BFA-insensitive non-conventional secretion (44). Upon Baf-A1 treatment of wild-type cells, we found that intracellular p62 accumulated and was robustly secreted into the extracellular media. In contrast, GCase-deficient cells showed more robust intracellular accumulation of p62 and impaired p62 extracellular secretion (Fig. 3E). We then calculated the secretion index for p62 (extracellular levels/intracellular levels after Baf-A1) and found that it was significantly reduced in GCase-deficient cells compared with wild-type cells (Fig. 3F). In addition, GCase-deficient cells demonstrated increased intracellular α-synuclein accumulation as observed previously, and impaired extracellular α-synuclein secretion upon Baf-A1 treatment (Fig. 3E), resulting in a similar reduction in α-synuclein secretion index (Fig. 3F). This impaired secretion appeared to be preferential for autophagic substrates, as the secretion index of the non-autophagic substrate sortilin was not disrupted in GCase-deficient cells (Fig. 3E and F). We also found impaired secretion of the mature lysosomal hydrolase Cathepsin-D (mCat-D), but not its immature form (iCat-D) in GCase-deficient cells (Fig. 3G and H), suggesting that GCase may also regulate lysosomal exocytosis in addition to secretory autophagy.
Loss of GBA1 leads to α-synuclein and autophagy substrate accumulation. (A) Cells were lysed with 2× SDS sample buffer and cell lysates were analyzed with immunoblot analysis using indicated antibodies. Blot band intensities were normalized to tubulin, and compared with wild-type cells. Graphs show normalized band intensities of intracellular α-synuclein. N = 3, two-tailed unpaired t-test, ***P < 0.001 compared with wild-type cells. (B) Cells were fixed with 4% formaldehyde in PBS for 10 min and subsequently fixed again with cold methanol for 5 min, and immuno-stained with anti-p62 antibody. Representative images are shown. Data represent mean ± S.E.M. n = 10 microscopic fields for wild-type, n = 7 microscopic fields for GCase-deficient cells. Two-tailed unpaired t-test, ***P < 0.001 compared with wild-type cells. (C) LAMP2 and p62 levels are increased in GCase-deficient cells. Cells were prepared as in (B), and immuno-stained with mouse anti-Lamp2 antibody and rabbit anti-LC3B antibody. Representative images are shown. Data represent mean ± S.E.M. n = 10 microscopic fields for wild-type, n = 10 microscopic fields for GCase-deficient cells. Two-tailed unpaired t-test, ***P < 0.001, **P < 0.01 compared with wild-type cells. (D, E) Cells were lysed with 2× SDS sample buffer, and cell lysate samples were analyzed with immunoblot analysis using indicated antibodies. Band intensities were normalized to tubulin, and compared with wild-type cells. Graphs show normalized band intensities. Data represent mean ± S.E.M. N = 4, two-tailed unpaired t-test, ***P < 0.001 compared with wild-type cells.
GCase-deficient cells exhibit defective Baf-A1-induced extracellular secretion of α-synuclein, p62 and mature cathepsin-D. (A–D) Stimulation (starvation or CCCP)-induced degradative autophagic process in GCase-deficient cells. (A) Representative immunoblot blot from cell lysates from wild-type and GCase-deficient cells starved for 5 h with Hanks’ balanced salt solution. (B) Quantification of tubulin normalized LC3 II levels (n = 4). (C) Representative immunoblot blot data from cell lysates from wild-type and GCase-deficient cells treated with 5 μm CCCP for indicated times. Cell lysates were immunoblotted with antibodies against GCase, p62, Mfn2, and tubulin. (D) Band intensities were normalized to tubulin and compared with 0 h wild-type sample. Data represent mean ± S.E.M. n = 3. (E, F) GCase-deficient cells exhibit defective extracellular secretion of α-synuclein and p62. Wild-type and GCase-deficient cells were treated with 300 nm Baf-A1 for the indicated times. Both intracellular fractions and extracellular media fractions from each timepoint were collected as described in the Materials and Methods section. Protein samples from intracellular lysates and extracellular media fractions were analyzed with immunoblot analysis using indicated antibodies. (E) Representative immunoblot data are shown. (F) Graph shows secretion indexes (extracellular level/intracellular level at 5 h) of α-synuclein, and p62 and sortilin. Data represent mean ± S.E.M. N=4 independent experiments. Two-tailed paired t-test, ***P < 0.001, compared with wild-type cells. (G, H) GCase-deficient cells show defective extracellular secretion of mature cathepsin-D. Wild-type and GCase-deficient cells were treated with 300 nm Baf-A1 for the indicated times. Both intracellular fractions and extracellular media fractions from each timepoint were collected. Protein samples from intracellular lysates and extracellular media fractions were analyzed with immunoblot analysis using cathepsin-D antibodies. (G) Representative immunoblot data are shown. (H) Graph shows secretion indexes (extracellular level/intracellular level at 5 h) of mature cathepsin-D (mCat-D) and immature cathepsin-D (iCat-D). Data represent mean ± S.E.M. N = 4 independent experiments. Two-tailed paired t-test, *P < 0.05, compared with wild-type cells. n.s. = not significant.
Exogenous C18-Cer treatment rescues defects in secretory autophagy and Rab8a localization
We then asked how loss of GCase might contribute to impaired secretory autophagy. The Rab family of small GTPases plays critical roles in various steps of vesicular trafficking (45), of which Rab8a is the only known Rab to regulate secretory autophagy (37,41,46). We found that Rab8a protein levels were elevated in GCase-deficient cells compared with wild-type cells (Fig. 4A). As shown by immunostaining, GCase-deficient cells also exhibited fewer Rab8a puncta than wild-type cells (Fig. 4B), suggesting that disruption of Rab8a localization may contribute to defective secretory autophagy in GCase-deficient cells.
Rab8a levels and distribution are altered by loss of GBA1. (A) Rab8a protein levels are elevated in GCase-deficient cells. Total cell lysates from wild-type and GCase-deficient cells were subjected with immunoblot analysis using indicated antibodies. Band intensities were normalized to tubulin, and compared with wild-type cells. Data represent mean ± S.E.M. N = 4, two-tailed unpaired t-test, ***P < 0.001, compared with wild-type cells. (B) Representative confocal images of Rab8a in wild-type and GCase-deficient cells. Cells were seeded onto PDL-coated glass coverslip with the cell density of ∼40 000 cells per coverslip. Three days later, cells were fixed with 4% formaldehyde in PBS for 12 min, and subsequently fixed with cold methanol for 5 min. The fixed cells were then stained for Rab8a and nuclear DNA. (C) Acid ceramidase expression in wild-type cells increases Rab8a protein levels. Wild-type HEK293-FT cells were transfected with either empty pCAG vector or pCAG-ASAH1 vector. Two days later, total cell lysates from transfected cells were subjected to immunoblot analysis using indicated antibodies. Band intensities were normalized to tubulin, and compared with empty pCAG-transfected cells. Data represent mean ± S.E.M. N = 6, two-tailed unpaired t-test, ***P < 0.001, compared with pCAG-transfected cells.
As GCase promotes catabolic ceramide generation in the lysosome, we next asked whether the ceramide pathway might contribute to the impaired secretory autophagy and Rab8a mislocalization. Using mass spectrometry-based lipid analysis, we found that GCase-deficient cells contained increased levels of the substrate GluCer compared with wild-type cells (see GluCer species in Fig. 7A), consistent with our observations by immunofluorescence (Fig. 1D). Although the levels of C16-Cer or longer acyl chain-containing ceramide species were not altered, C18-Cer was significantly decreased in GCase-deficient cells (see ceramide species in Fig. 7B). We thus investigated whether exogenous addition of C18-Cer to GCase-deficient cells could rescue defects in secretory autophagy and Rab8a localization. Cells were treated with either DMSO or C18-Cer for 20 h, and additionally with Baf-A1 to induce secretion. C18-Cer treatment led to decreased levels of intracellular and increased extracellular α-synuclein upon Baf-A1 treatment (Fig. 5A), resulting in increased α-synuclein secretion index (Fig. 5B). In addition, C18-Cer treatment also rescued impaired p62 secretion in GCase-deficient cells, by decreasing its intracellular and increasing its extracellular levels (Fig. 5A). Although C18-Cer treatment robustly increased the secretion indices of both α-synuclein and p62, it did not affect the secretion of the non-autophagic substrate sortilin (Fig. 5B). We then examined whether C18-Cer treatment could rescue Rab8a mislocalization found in GCase-deficient cells. Indeed, C18-Cer treatment for 20 h was able to restore Rab8a distribution similar to that observed in wild-type cells (Fig. 5C), further suggesting that reduced levels of C18-Cer contribute to defects in secretory autophagy in GCase-deficient cells.
Exogenous C18-Cer treatment rescues defective extracellular α-synuclein and p62 secretion and Rab8a mislocalization upon loss of GBA1. (A, B) Exogenous C18-Cer treatment corrects defective secretion of α-synuclein and p62 in GCase-deficient cells. GCase-deficient cells were pre-treated with either DMSO or 4 μm C18-Cer for ∼18 h. Cells were then treated with 300 nm Baf-A1 for the indicated times. Both intracellular fraction and extracellular media fraction from each timepoint were collected as described in the Materials and Method section. Protein samples from intracellular lysates and extracellular media fractions were analyzed with immunoblot analysis using indicated antibodies. (A) Representative immunoblot data are shown. (B) Graph shows secretion indexes (extracellular level/intracellular level at 5 h) of α-synuclein, and p62 and sortilin. Data represent mean ± S.E.M. N = 3, two-tailed paired t-test, **P < 0.01, *P < 0.05, compared with DMSO-treated cells. n.s = not significant. (C) Exogenous C18-Cer treatment reverts altered Rab8a localization in GCase-deficient cells. Wild-type and GCase-deficient cells were seeded onto poly-d-lysine coated glass coverslip with the cell density of ∼40 000 cells per coverslip. Two days later, cells were treated with either DMSO or 4 μm C18-Ceramide for ∼20 h. Cells were then fixed with 4% formaldehyde in PBS for 12 min, and subsequently fixed with cold methanol for 5 min. The fixed cells were then stained for Rab8a and nuclear DNA. Representative confocal images are shown.
Inhibition of acid ceramidase rescues lipid abnormalities observed in GCase-deficient cells
Lysosomal ceramide species generated by GCase and other lysosomal enzymes [acid sphingomyelinase and galactosylceramidase (GALC)] are directly converted into Sph and free fatty acids by acid ceramidase (encoded by the ASAH1 gene). We thus investigated if acid ceramidase might also play a role in regulating cellular dysfunction in GCase-deficient cells. Surprisingly, there was a significant increase in acid ceramidase protein levels in GCase-deficient cells, compared with wild-type cells (Fig. 6A and B). Of note, lysosomal β-hexosaminidase (m-HEX-B) levels were also increased, consistent with previous reports (47), whereas cathepsin-D levels were not altered (Fig. 6A and B).
Acid ceramidase expression phenocopies loss of GBA1. (A, B) Protein levels of acid ceramidase are elevated in GCase-deficient cells. Total cell lysates from wild-type and GCase-deficient cell were subjected to immunoblot analysis using indicated antibodies. Immunoblot data is shown in (A). (B) Band intensities were normalized to tubulin, and compared with wild-type cells. Graph represents mean ± S.E.M; n = 3, two-tailed unpaired t-test, **P < 0.01; compared with wild-type cells. (C, D) Acid ceramidase expression phenocopies GCase-deficient cells. HEK293-FT cells were transiently transfected either with empty pCAG vector or pCAG-ASAH1 vector. Two days after transfection, total cell lysates were prepared with 2× SDS sample buffer, and subjected to immunoblot analysis using indicated antibodies. Representative immunoblot data are shown in (C). (D) Protein levels were normalized to tubulin, and compared with empty vector transfected cells. Graph represents mean ± S.EM. n = 6, two-tailed unpaired t-test, **P < 0.01; ***P < 0.001; compared with wild-type cells.
We hypothesized that increased acid ceramidase in GCase-deficient cells may further lower the levels of lysosomal ceramide and contribute to cellular dysfunction. To test this, we first investigated whether acid ceramidase overexpression in wild-type cells might phenocopy cellular defects observed in GCase-deficient cells. Indeed, acid ceramidase expression resulted in the accumulation of α-synuclein, LC3, p62, Rab7, ubiquitinated proteins and Lamp2 (Fig. 6C and D), and increased Rab8a protein levels (Fig. 4C), similar to what we observed in GCase-deficient cells. Next, we examined if inhibiting acid ceramidase might rescue cellular defects observed in cells with low GCase activity. Carmofur is a highly potent brain-penetrant chemical inhibitor of acid ceramidase that can be administered orally and has been used in the treatment of breast and colorectal cancer (48). To confirm that carmofur was indeed inhibiting acid ceramidase activity, we performed mass-spectrometry lipidomic analysis of wild-type and GCase-deficient cells treated with either DMSO or carmofur for ∼20 h. We observed a selective change in the levels of abundant individual ceramide species (C18-Cer, C24-1Cer, C16-Cer), including a 0.4-fold decrease in C18-Cer, but no change in C16-Cer in GCase-deficient cells compared with wild-type cells (Fig. 7B). Importantly, carmofur could rescue this decrease in ceramide levels, leading to a dramatic increase of all ceramide species measured (Fig. 7B). Consistent with previous reports in GD patient brain and GD mouse models (17,49), GluSph levels were increased in GCase-deficient cells, and rescued by carmofur treatment (Fig. 7A). We also observed a significant increase of GluCer and reduction of GalCer species in GCase-deficient cells compared with wild-type cell that were not altered by carmofur (Figs 7A and 8A). Because the hydrolysis of GalCer is mediated by lysosomal GALC, which breaks down GalCer into ceramide and galactose in lysosomes, we tested whether GALC levels are increased in the GCase-deficient cells. Indeed, the level of ∼50 kDa mature forms of GALC (50) was increased by ∼1.4-fold in the GCase-deficient cells. On the other hand, the levels of ∼80 kDa immature precursor form of GALC (50) levels were not changed (Fig. 8B). Finally, Sph but not sphingosine-1-phosphate (Sph-1-P) levels were significantly increased in GCase-deficient cells compared with wild-type cells, and carmofur treatment increased both Sph and Sph-1-P levels (Fig. 8C). Taken together, these results demonstrate that carmofur partially rescues lipid abnormalities in GCase-deficient cells including reverting the decrease in ceramide and the accumulation of GluSph.
Carmofur treatment reverts reduced ceramide levels and increase GluSph levels in GCase-deficient cells. Wild-type and GCase-deficient cells were treated with either DMSO or 7.5 μm carmofur for ∼20 h. After drug treatments, total lipids were extracted and measured as described in the Materials and Methods section. Lipid analysis results were expressed as lipid levels of picomoles (pmol)/three million cells. Levels of GluSph, GluCer species and dihydroceramide (A), and ceramide species (B) are shown. For each group, three independent samples were analyzed. Data represent mean ± S.E.M. Two-tailed unpaired t-test, ***P < 0.001, **P < 0.01, *P < 0.05.
Galactosylceramide (GalCer) levels are reduced in and mature GALC levels are elevated in GCase-deficient cells. (A) Reduced galactosylceramide (GalCer) levels in in GCase-deficient cells. Wild-type and GCase-deficient cells were treated with either DMSO or 7.5 μm carmofur for ∼20 h. After drug treatments, total lipids were extracted and measured as described in the Materials and Methods section. Lipid analysis results were expressed as lipid levels of picomoles (pmol)/three million cells. Levels of galactosylceramide species. For each group, three independent samples were analyzed. Data represent mean ± S.E.M. Two-tailed unpaired t-test, ***P < 0.001. (B) Mature GALC levels are elevated in GCase-deficient cells. Total cell lysates from wild-type and GCase-deficient cell were subjected to immunoblot analysis using antibody against GALC, GCase or tubulin. Immunoblot data is shown in the left. Band intensities of the mature form of galactosylceramidase (m-GALC) and the immature form of galactosylceramidase (i-GALC) were normalized to tubulin, and compared with wild-type cells. Graph represents mean ± S.E.M. n = 3, two-tailed unpaired t-test, **P < 0.01; compared with wild-type cells. n.s. = not significant. (C) Levels of Sph and Sph-1-P. For each group, three independent samples were analyzed. Data represent mean ± S.E.M. Two-tailed unpaired t-test, **P < 0.01, *P < 0.05.
Inhibition of acid ceramidase rescues secretory defects and reduces levels of intracellular α-synuclein in GCase-deficient cells
Having shown that inhibition of acid ceramidase improves lipid abnormalities observed in GCase-deficient cells, we next investigated the effects on accumulation of α-synuclein. Carmofur treatment decreased intracellular α-synuclein levels in wild-type cells (Fig. 9A), and it reversed elevated α-synuclein, LC3II, ubiquitinated proteins and Rab7 in GCase-deficient cells (Fig. 9B and C). As GCase-deficient cells showed defects in secretory autophagy, we examined if carmofur might also rescue Baf-A1-induced extracellular secretion of α-synuclein and p62. In comparison with wild-type cells, GCase-deficient cells exhibited pronounced intracellular accumulation of p62, LC3-II and α-synuclein after Baf-A1 induction (Fig. 10A and B), which was robustly rescued by pre-treatment with carmofur (Fig. 10A and B).
Carmofur reduces intracellular α-synuclein and autophagic substrate accumulation in GCase-deficient cells. (A) Carmofur reduces intracellular α-synuclein levels in a dose-dependent manner. Wild-type HEK cells were treated with DMSO or the indicated concentration of carmofur for 18 or 45 h. Total cell lysates were analyzed with immunoblotting with α-synuclein and tubulin antibody. Band intensities of α-synuclein were normalized to tubulin levels. (B, C) Carmofur reduces α-synuclein, Rab7 and autophagic substrates in GCase-deficient cells. Wild-type cells were treated with DMSO, and GCase-deficient cells were treated with either DMSO or 7.5 μm carmofur for ∼18 h. Total cell lysates were analyzed with immunoblotting using indicated antibodies. (B) Representative immunoblot data are shown. (C) Blot band intensities were normalized to tubulin, and compared with DMSO-treated wild-type cells. Graphs show normalized band intensities of intracellular α-synuclein (n = 4), LC3B-II (n = 4), p62 (n = 4), ubiquitinated proteins (n = 4), LAMP2 (n = 4) and Rab7 (n = 4). Data represent mean ± S.E.M. Two-tailed unpaired t-test, *P < 0.05, compared with DMSO-treated GCase-deficient cells. n.s = not significant.
Carmofur promotes Baf-A1-induced α-synuclein and p62 extracellular secretion in GCase-deficient cells. (A, B) Carmofur in GCase-deficient cells ameliorates intracellular accumulation of α-synuclein and autophagic substrates (p62 and LC3II). Wild-type cells were pre-treated with DMSO for ∼18 h, and GCase-deficient cells were pre-treated with either DMSO or 7.5 μm carmofur for ∼18 h. Cells were then treated with 300 nm Baf-A1 for the indicated times. Total cell lysates were analyzed with immunoblot analysis using indicated antibodies. (A) Representative immunoblot data are shown. (B) Blot band intensities were normalized to tubulin, and compared with DMSO-treated GCase-deficient cells. Graphs show normalized band intensities of intracellular α-synuclein (N = 3), LC3B-II (N = 4) and p62 (N = 4). Data represent mean ± S.E.M. Two-tailed paired t-test, *P < 0.05; **P < 0.01; compared with DMSO-treated GCase-deficient cells. (C, D) Carmofur promotes extracellular secretion of α-synuclein and p62 in GCase-deficient cells. GCase-deficient cells were pre-treated with either DMSO or 7.5 μm carmofur for ∼18 h. Cells were then treated with 300 nm Baf-A1 for the indicated times. Both intracellular fractions and extracellular media fractions from each time point were collected as described in the Materials and Methods section. Protein samples from intracellular lysates and extracellular media fractions were analyzed with immunoblot analysis using indicated antibodies. (C) Representative immunoblot data are shown. (D) Graph shows secretion indexes (extracellular level/intracellular level at 5 h) of α-synuclein, and p62, and sortilin. N = 3 independent experiments. Data represent mean ± S.E.M. Two-tailed paired t-test, ***P < 0.001, *P < 0.05, compared with DMSO-treated cells. n.s = not significant.
In addition, carmofur was able to promote robust extracellular secretion of p62 and α-synuclein induced by Baf-A1, as compared with DMSO-treated GCase-deficient cells (Fig. 10C and D). This effect was accompanied by decreased intracellular accumulation of α-synuclein and p62. In contrast, carmofur treatment did not affect the extracellular secretion of sortilin (Fig. 10C and D), suggesting that inhibition of acid ceramidase preferentially promotes secretion of autophagic substrates α-synuclein and p62.
To further test carmofur’s ability to inhibit intracellular accumulation of α-synuclein in a relevant PD neuronal model, we generated human iPSC-derived dopaminergic neurons from a GBA1-linked PD patient carrying a heterozygous GBA1-c.84dupG frame-shift mutation (resulting in p.Leu29Ala fs*18). These iPSCs were differentiated into dopaminergic neurons (51), and treated with either 7.5 μm carmofur or DMSO. We found that carmofur potently reduced ubiquitinated protein species in mutant neurons (Fig. 11A and B). Although the total levels of α-synuclein were not reduced by a 2 days carmofur treatment, oxidized α-synuclein species were decreased (Fig. 11A and B) (52). Importantly, tyrosine hydroxylase (TH) level was not altered by carmofur treatment (Fig. 11A and B). Consistent with the findings from dopaminergic neurons carrying a heterozygous GBA1-c.84dupG frame-shift mutation, carmofur treatment of neurons derived from a PD patient carrying heterozygous GBA1-N370S mutation significantly reduced ubiquitinated proteins, total α-synuclein and oxidixed α-synuclein (Fig. 11C and D). Importantly, the same carmofur treatment of control neurons did not affect levels of ubiquitinated proteins, oxidized α-synuclein and total α-synuclein (Fig. 11E and F). TH levels were not altered by carmofur treatment (Fig. 11C–F).
Carmofur reduces oxidized α-synuclein and ubiquitinated proteins in PD patient-derived dopaminergic neurons harboring a heterozygous GBA1-c.84dupG mutation or heterozygous GBA1-N370S mutation. (A) iPSCs derived from a PD patient carrying heterozygous GBA1-c.84dupG mutation were differentiated into dopaminergic neurons for 40 days. Dopaminergic neurons were treated with either DMSO or 7.5 μm carmofur for 2 days. Total cellular lysates were analyzed with immunoblotting using antibodies against ubiquitinated proteins (Ub), tubulin, TH, oxidized α-synuclein (Syn303), or total α-synuclein (C20). Saturated pixels in blot images are shown in red. (B) Band intensities were normalized to tubulin, and compared with DMSO-treated control. Data represent mean ± S.E.M. N = 6 from two independent differentiations into dopaminergic neurons, two-tailed unpaired t-test, **P < 0.01, *P < 0.05 compared with DMSO control. (C) iPSCs (clone PD267) derived from a PD patient carrying heterozygous GBA1-N370S mutation were differentiated into dopaminergic neurons for 40 days and treated with either DMSO or 7.5 μm carmofur for 2 days. Total lysates were analyzed with immunoblotting using antibodies against ubiquitinated proteins (Ub), tubulin, TH, oxidized α-synuclein (Syn303), or α-synuclein (C20). Saturated pixels in blot images are shown in red. (D) Band intensities were normalized to tubulin, and compared with DMSO-treated controls. Data represent mean ± S.E.M. N = 6, two-tailed unpaired t-test, *P < 0.05, compared with DMSO control. (E) Control neurons were differentiated for 40 days and treated with either DMSO or 7.5 μm carmofur for 2 days. Total lysates were analyzed with immunoblotting using antibodies against ubiquitinated proteins (Ub), tubulin, TH, oxidized α-synuclein (Syn303), or α-synuclein (C20). Saturated pixels in blot images are shown in red. (F) Band intensities were normalized to tubulin, and compared with DMSO-treated control. Data represent mean ± S.E.M. N = 6, two-tailed unpaired t-test, n.s. = not significant, compared with DMSO control.
Taken together, these data demonstrate that carmofur eliminates oxidized α-synuclein and ubiquitinated protein species in dopaminergic neurons carrying GBA mutations, and further suggest that acid ceramidase inhibition may be an important therapeutic angle for GBA1-linked synucleinopathies such as PD.
Discussion
Acid ceramidase inhibition reverts ceramide deficiency in GCase-deficient cells
This study demonstrates that acid ceramidase inhibition by carmofur ameliorates cellular defects including α-synuclein accumulation, both in a GCase-deficient cell model and in GBA1-linked PD patient-derived dopaminergic neurons. Carmofur treatment also robustly decreases intracellular α-synuclein accumulation, potentially by rescuing defects in secretory trafficking. Our data further implicate increased acid ceramidase activity in the context of decreased GCase, as acid ceramidase levels were elevated upon loss of GBA1, and increased acid ceramidase expression led to accumulation of α-synuclein. Taken together, our results suggest that inhibition of acid ceramidase may be a relevant therapeutic target for GBA1-linked synucleinopathies, although potential systemic side effects of carmofur will have to be considered.
Acid ceramidase acts in the lysosome to convert ceramide into Sph. Although ceramide was originally implicated in apoptosis signaling, recent studies have revealed a role for ceramide species in diverse cellular processes including proliferation, differentiation, stress response and autophagy (23,53).
Several lines of evidence now point to reduced ceramide generation in the lysosome as a potential contributor to PD pathogenesis. First, loss of function of three different lysosomal hydrolases (GCase, acid sphingomyelinase and galactocylceramidase) linked to lysosomal storage disorders which contribute to the catabolic lysosomal salvage pathway of ceramide generation have been implicated in either increased PD risk or intracellular α-synuclein accumulation. GCase converts GluCer into ceramide and glucose, and heterozygous mutations in GCase are a major risk factor for PD (5), whereas loss of GBA1 leads to α-synuclein accumulation in multiple models as well as in GBA1-linked PD patients (13,15,17,18,30–34). Acid sphingomyelinase hydrolyzes sphingomyelin into ceramide and phosphorylcholine in lysosomes, and loss of its activity leads to Types A and B Niemann-Pick disease characterized by sphingomyelin accumulation. Interestingly, the L302P loss of function mutation in acid sphingomyelinase is associated with a 9-fold increased PD risk in Ashkenazi Jews (54). Finally, GALC breaks down galactosylceramide into ceramide and galactose in lysosomes, and a mouse model of Krabbe disease caused by GALC mutations was found to have brain accumulation of α-synuclein aggregates and ubiquitinated proteins (55). Further supporting the relationship between PD and ceramide levels, early stage PD patients exhibit reduced brain levels of ceramide (15). More extensive ceramide species analysis from PD brain by Abbot et al. (56) showed that in anterior cingulate cortex (affected brain area in PD) but not occipital cortex (unaffected brain area in PD), total ceramide including C18-ceramide was reduced from control levels by ∼53%. Additionally a genome-wide RNAi-screen in Caenorhabditiselegans demonstrated that knock-down of lagr-1, a ceramide synthase in C. elegans, could potently increase α-synuclein inclusion formation (57), further implicating ceramide in synucleinopathies.
Second, our data demonstrate that ceramide species with long acyl chain including C18-Cer are reduced in GCase-deficient cells. Inhibiting acid ceramidase with carmofur in GCase-deficient cells successfully rescues ceramide levels, and importantly, also reduced α-synuclein and ubiquitinated protein accumulation. As carmofur has been shown to elevate C18-Cer levels in mouse brain (48), such treatment may be a potential in vivo strategy for increasing ceramide levels in models of GCase deficiency.
Finally, we find that application of exogenous C18-Cer reverts secretory trafficking defects in GCase-deficient cells, leading to reduced intracellular accumulation of α-synuclein and the autophagic substrate p62. C18-Cer has been reported previously to be concentrated in Rab8a/Rab11-positive vesicles (58), and we show that exogenous C18-Cer restores punctate Rab8a localization in the cellular periphery of GCase-deficient cells. As it is increasingly recognized that ceramides with different fatty acid chain lengths play distinct roles in the cell (23,59,60), our data point to preferentially elevating C18-Cer levels as a potential therapeutic target for synucleinopathies.
Effect of acid ceramidase inhibition on levels of GluSph
Our mass-spectrometry lipidomic analysis also showed that GluSph levels were dramatically elevated in GCase-deficient cells and subsequently reduced by carmofur treatment. As it has been reported that GluCer can be deacylated to GluSph in lysosomes by acid ceramidase (61), we propose that the reduction of GluSph levels by carmofur is due to direct inhibition of acid ceramidase activity. Interestingly, another GBA1 model using 9-month old GBA1D409V/D409V knock-in mice also showed GluSph accumulation along with proteinase-resistant α-synuclein aggregates and ubiquitinated protein species in the brain. Importantly, these phenotypes in addition to elevated GluSph were reversed by long-term chronic administration of a brain-penetrant GluCer synthase inhibitor, GZ667161 (18), suggesting that GluSph levels may be positively correlated with the levels of aggregated α-synuclein and ubiquitinated protein species. Recently, it has been reported that GluSph binds to α-synuclein and potently accelerates pathogenic α-synuclein aggregation in vitro and in cells (62). We hypothesize that elevated GluSph levels may further contribute to cellular toxicity in GBA1-linked PD by promoting α-synuclein oligomerization or fibril formation, or by increasing oxidative stress as GluSph also inhibits cytochrome C oxidase activity in mitochondria (63). GluSph is increasingly being recognized as a key biomarker for GD (64). In addition, many clinical symptoms of GD are closely linked to extensive immune dysregulation (65) and patients have been found to develop antibodies against GluSph (66). Thus, the potent reduction of GluSph levels by carmofur additionally point to acid ceramidase inhibitors as a potential therapeutic candidate.
Secretory trafficking defects in Parkinson’s disease
GCase-deficient cells in our study exhibited impaired extracellular secretion, potentially leading to the intracellular accumulation of α-synuclein, which was partially rescued by either carmofur treatment or addition of exogenous C18-Cer. Consistent with the idea that impaired extracellular secretion may contribute to PD pathogenesis, acid sphingomyelinase regulates the ectosome release of intracellular proteins through plasma membrane shedding (67) whereas its loss-of-function L302P mutation in associated with a 9-fold increased risk of PD in Ashkenazi Jews (54). Additionally, loss-of-function mutations in ATP13A2 gene, encoding a late endosomal/lysosomal p-type ATPase, cause juvenile-onset Parkinsonism (68) and also lead to reduced α-synuclein secretion, whereas overexpression of wild-type ATP13A2 increases α-synuclein secretion (42). These studies thus suggest that correcting defects in secretory trafficking may be a potential therapeutic strategy for reducing intracellular α-synuclein accumulation. We further found that levels of three lysosomal hydrolases (HEX-B, GALC and acid ceramidase) were elevated in GCase-deficient cells, suggesting that lysosomal exocytosis might be disrupted in GCase-deficient cells.
The role of Rab8a GTPase in synucleinopathies
Among the Rab family of small GTPases, Rab8a is the only Rab currently known to regulate secretory autophagy (37,40,41,46). Interestingly, several studies have recently reported Rab8a functional connection to several PD genes. In particular, α-synuclein has been shown to interact with the switch region of Rab8a in a Ser129 phosphorylation-dependent manner (69) and Rab8a expression promotes extracellular α-synuclein secretion via secretory autophagy in PC12 cells (40). Rab8 expression also ameliorates α-synuclein-dependent behavioral defects in fruit flies (69) and suppresses α-synuclein-induced toxicity in C. elegans and rat primary dopaminergic neurons (70). More recently, Leucine-rich repeat kinase 2 (LRRK2) has also been reported to directly phosphorylate Rab8a at Thr-74, an evolutionary conserved residue in its switch II domain, whereas pathogenic LRRK2 variants lead to increased phosphorylation of Rab8a and its decreased binding to Rab-GDP dissociation inhibitors (GDIs) (71). In addition, another PD-linked gene PINK1 (PTEN-induced kinase 1) has also been reported to indirectly influence the phosphorylation of Rab8a at Ser-111 (72). Finally, we previously showed that both LRRK2 and the recently identified PD gene TMEM230 participate in Rab8a-mediated secretory autophagy (41). From our current study, we present evidence suggesting that reduced levels of ceramide species with long chain fatty acids in GCase-deficient cells may contribute to the intracellular accumulation of α-synuclein, partly due to altered Rab8a localization and defective extracellular secretion of α-synuclein. Together, these studies raise an intriguing possibility that the dysregulation of Rab8a function might be a pathogenic converging node in multiple genetic forms of PD.
In conclusion, we propose a model in which inhibition of acid ceramidase activity by carmofur partially restores sphingolipid metabolism in GCase-deficient cells by increasing ceramide and decreasing GluSph levels. This restoration of lipid levels rescues Rab8a-mediated secretory autophagy, and lowers the intracellular accumulation of α-synuclein. Further investigation of the relationship between these pathways will thus be important for our understanding of acid ceramidase inhibition as a viable therapeutic target in synucleinopathies linked to GCase dysfunction.
Materials and Methods
Antibodies and chemicals
The following antibodies were used in immunoblotting or immunostaining: mouse anti-GCase (N-terminal) (Santa Cruz, sc-365745); rabbit anti-GCase (C-terminal) (Sigma, G4171); rabbit anti-GluCer (Glycobiotech, RAS_0011); rabbit anti-α-synuclein (clone C20, Santa Cruz, sc-7011-R); mouse anti-α-tubulin (Sigma, T5168); mouse anti-LAMP2 (Developmental Studies Hybridoma Bank, H4B4-s); rabbit anti-LC3B (Cell Signaling, 3868S); rabbit anti-p62 (Sigma, P0067); rabbit anti-Rab7 (Sigma, R4779); mouse anti-Sortilin (clone G11, Santa Cruz, sc-376561); rabbit anti-Mfn2 (Cell Signaling, 11925); goat anti-cathepsin-D (Clone R20, Santa Cruz, sc-6478); rabbit anti-acid ceramidase (Proteintech); mouse anti-HEX-B (clone D9, Santa Cruz, sc-376781); mouse anti-ubiquitin (clone P4D1, Santa Cruz, sc-8017); mouse anti-Syn303 (Covance, MMS-5085-100); rabbit anti-TH (Calbiochem, 657012); rabbit anti-Rab8a (Proteintech, 55296-1-AP); mouse anti-GALC (clone 2D1, Santa Cruz SC-293200). CCCP, Baf-A1 and carmofur were purchased from Cayman Chemicals. C18-Ceramide (d18: 1/18: 0) were purchased from Avanti Polar Lipids (860518P). Other chemicals were purchased from Sigma, unless otherwise stated.
Plasmid, cell culture and transfection
The open reading frame of human acid ceramidase (ASAH1) was PCR amplified from MGC Human ASAH1 cDNA clone (Clone Id: 3923451, GE Healthcare Dharmacon). Primers used were: forward primer 5′-GTG GCG GCC GCA TGC CGG GCC GGA GTT GCG T-3′ and reverse primer 5′-GTC GGT ACC TCA CCA ACC TAT ACA AGG GT-3′. After restriction enzyme digestion with NotI/KpnI, the digested PCR product was subcloned into NotI/KpnI sites of pCAG mammalian expression vector, which drives gene expression under chicken β-actin promoter. The CRISPR/Cas9 construct targeting human GBA1 gene was constructed into pSpCas9(BB)-2A-Puro (PX459) V2.0 (a gift from Feng Zhang, Addgene plasmid, #62988). The web tool of CRISPR design (http://crispr.mit.edu/) was used to select gRNAs for GBA1 gene. The target sequence of human GBA1-CRISP/Cas-9 construct is 5′-GCGACGGATGGAGCTGAGTA-3′. Two oligonucleotides (5′-CAC CGC GAC GGA TGG AGC TGA GTA-3′ and 5′-AAA CTA CTC AGC TCC ATC CGT CGC-3′) were annealed and subcloned into BbSI-digested pSpCas9(BB)-2A-Puro (pX459) V2.0 vector. HEK293-FT (human embryonic kidney) cells were maintained in DMEM with 10% fetal bovine serum in 37°C CO2 incubator. Human embryonic kidney 293-FT (HEK293-FT) cells and GCase-deficient cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (Life Technologies), 10 U/ml penicillin, and 10 μg/ml streptomycin at 37°C with 5% CO2 incubation.
Generation of GCase-deficient cell line
HEK293-FT cells grown in 12-well plates were transfected with the CRISPR/Cas9 construct targeting human GBA1 gene. One day after transfection, cells were trypsinized, and re-seeded on a new 10 cm dish with low density (∼50 cells in 10 cm dish), and cultured until single cells formed cell colonies. Clonal colonies were picked from the 10 cm culture dish and expanded in new culture dishes. The GCase-deficient clone was identified by immunoblot analysis using anti-GCase antibody.
Cell lysis and immunoblotting
Cells were seeded ∼200 000 cells per well in 12-well plates. Two days later, cells were harvested for biochemical assays, unless otherwise stated. Cells were lysed with 2× SDS sample buffer [100 mm Tris–Cl (pH 6.8), 4% (w/v) SDS, 0.05% (w/v) bromophenol blue, 20% (v/v) glycerol, 200 mm dithiothreitol]. Protein samples from total cell lysates were subjected to SDS-PAGE. Proteins were then transferred to nitrocellulose membranes using Trans-Blot Turbo Transfer system (Bio-Rad). The membranes were incubated overnight with the indicated primary antibodies. All primary antibodies were diluted in an antibody dilution buffer (25 mm Tris, 0.15 m NaCl, 0.05% Tween-20, 5% BSA, 0.05% sodium azide). Primary antibodies were visualized using the appropriate horseradish peroxidase-conjugated secondary antibody (anti-mouse) (Jackson ImmunoResearch Laboratories, #115-035-146), anti-rabbit (Jackson ImmunoResearch Laboratories, #111-035-144), anti-goat (Jackson ImmunoResearch Laboratories, #805-035-180) and SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific, #34096). Chemiluminescence signals from blot were imaged on the ChemiDoc MP System (Bio-Rad) with a 16-bit CCD camera. Signal accumulation mode was used to acquire images at progressively longer exposure times. This allowed for acquisition of immunoblot images with band intensities within the linear range of the system. Quantification of protein levels was done using Bio-Rad ImageLab software and ImageJ (NIH) using non-saturated raw image files. Data were normalized to tubulin and expressed relative to control levels as indicated.
Immunostaining and imaging
Wild-type HEK293-FT cells or GCase-deficient cells were seeded onto poly-d-lysine coated glass coverslips with a density of ∼45 000 cells per coverslip in 12-well plates. Three days later, cells were washed once with PBS and fixed with 4% formaldehyde in PBS for 15 min at room temperature and subsequently formaldehyde-fixed cells were fixed again with cold methanol for 5 min. After PBS washing, fixed cells were incubated in blocking solution (1% BSA and 0.1% Tween-20 in PBS) for 1 h. Cells were then incubated with indicated antibodies in GDB buffer (30 mm phosphate buffer, pH 7.4, containing 0.1% gelatin, 0.3% Triton X-100, and 0.45 m NaCl) overnight at 4°C. Primary antibodies were visualized using goat Alexa-dye-conjugated secondary antibodies against rabbit or mouse. Immunofluorescence images were acquired with a Leica confocal microscope with a 63× oil objective. The confocal microscope settings were kept the same for all scans when fluorescence intensity was compared. Fluorescence intensity measurements were performed using ImageJ (NIH) using non-saturated raw image files. Data were expressed relative to wild-type levels as indicated.
GCase activity assay
Wild-type or GCase-deficient cells were washed once with PBS, harvested and lysed in 1% Triton lysis buffer (1% Triton X-100 in Tris-buffered saline). Protein concentrations were measured by BCA method. GCase activity in 7.5 μg of cleared cell lysates were assayed in 50 μl of GCase assay buffer [1 mm 4-methylumbelliferyl-β-d-glucopyranoside (4MU-Gluc, Sigma-Aldrich) 0.25% Triton X-100, 1% BSA, 0.25% (w/v) Taurocholic acid, 1 mm EDTA, in citrate/phosphate buffer, pH 5.4] in the presence or absence of Conduritol-B epoxide (CBE), an irreversible GCase inhibitor. Enzyme reaction mixtures were incubated for 30 min at 37°C, and the enzyme reaction was stopped by the addition of equal volume of 1 m glycine, pH 12.5. In endpoint mode, 4MU fluorescence (excitation wavelength = 355 nm, emission wavelength = 460 nm) was measured with SpectraMax i3 (Molecular Devices). The fluorescence value of control no-cell lysates were subtracted from 4MU fluorescence values. GCase activity was measured in triplicates.
Preparation of extracellular media fraction and cell lysates fraction
For the analysis of media fraction, cells were seeded at ∼200 000 cells per well in 12-well plates. One day after seeding, cells were then pre-treated with the indicated drugs. The next day, cells in 12-well plates were left untreated or treated with Baf-A1 (300 nm) for the indicated times at 37°C in 5% CO2 incubator. The cells were then cooled in ice for 5 min. About 0.8 ml of extracellular media samples were transferred from each well of 12 wells to a 1.5 ml tube, and centrifuged for 15 min at 550g at 4°C to remove the cell debris. Subsequently, 0.55 ml of the clear medium was transferred to a fresh 1.5 ml tube and the medium samples were then centrifuged for 30 min at 2500g at 4°C. After centrifugation, 0.2 ml of extracellular medium samples were transferred to fresh 1.5 ml tubes for immunoblot analysis. For immunoblot analysis of extracellular medium samples, 200 μl of extracellular media samples obtained from 2500g of centrifugation were mixed with 200 μl SDS 4× sample buffer [200 mm Tris–Cl (pH 6.8), 8% SDS 0.1% (w/v) bromophenol blue, 40% (v/v) glycerol, 400 mm DTT (dithiothreitol)] and boiled for 10 min. For the preparation of the cell lysate fraction, the cells of the culture dish were washed once with PBS and lysed with 2× SDS sample buffer. Protein samples were then analyzed with immunoblotting using indicated antibodies.
Carmofur treatment in iPSC-derived dopaminergic neurons
iPSCs (GBA1-c.84dupG, C14 clone) were generated from skin fibroblasts of a PD patient carrying a heterozygous GBA1-c.84dupG mutation. iPSCs (GBA1-N370S, clone PD267) were generated from skin fibroblasts of a PD patient carrying a heterozygous GBA1-N370S mutation. Control iPSC (clone 2132) was derived from skin fibroblasts of a healthy individual. iPSCs were differentiated into dopaminergic neurons for 40 days, according to the protocol by Kriks et al. (51). On day 40, we treated dopaminergic neurons with either 7.5 μm carmofur or DMSO for 2 days. On day 42, we analyzed total cell lysates with immunoblot analysis using indicated antibodies.
Mass spectrometry-based lipid analysis
Cells were grown on a 10 cm dishes at a cell density of ∼2.5 million cells per dish (six 10 cm dishes for wild-type cells; six 10 cm dishes for GCase-deficient cells). The next day, wild-type cells in three 10 cm dishes or GCase-deficient cells in three 10 cm dishes were treated with either DMSO or 7.5 μm carmofur for ∼20 h. Cells were washed with cold PBS, and were then counted. Three million cells from each 10 cm dishes were pelleted with spin at 300g for 5 min, and frozen at −80°C. At the University of South Carolina Medical Lipidomics Center, lipids were extracted from frozen cell pellets as a service provided by their core facilities. Levels of GluSph, GluCer, GalCer, ceramide and dihydroceramide were measured by a high-performance liquid chromatography/mass spectrometer (LC-MS/MS) method. Lipid analysis results were expressed as lipid levels of picomoles (pmol)/3 million cells. For each group, three independent samples were analyzed.
Statistical analysis
All values in figures and text refer to mean ± S.E.M. unless otherwise stated. N refers to the number of independent experiments unless otherwise indicated. Statistics and graphing were performed using Prism (GraphPad) software. Statistical analysis of data was performed with two-tailed paired t-test unless otherwise indicated.
Acknowledgements
We thank Yvette Wong for helpful suggestions, and Ming-Yi Chiang for technical assistance.
Conflict of Interest statement. None declared.
Funding
This work was supported by R01NS076054 (DK). Lipidomics were supported in part by the Lipidomics Shared Resource, Hollings Cancer Center, Medical University of South Carolina (P30 CA138313 and P30 GM103339).
Author Contributions
M.J.K. and D.K. conceived and designed the experiments. M.J.K. performed the experiments and analyzed data. S.J. performed in vitro GCase assay in Fig. 1C. S.J. and L.F.B. provided iPSC-derived dopaminergic neurons. M.J.K. and D.K. wrote the manuscript. All authors read the manuscript.
