Pfizer Neuroscience Research Unit, Cambridge, Massachusetts, USA
Address correspondence and reprint requests to Warren D. Hirst, Pfizer Global Research and Development, Neuroscience Research Unit, 610 Main Street, Cambridge, MA 02139, USA. E-mail: firstname.lastname@example.org
Aggregate-prone mutant proteins, such as α-synuclein and huntingtin, play a prominent role in the pathogenesis of various neurodegenerative disorders; thus, it has been hypothesized that reducing the aggregate-prone proteins may be a beneficial therapeutic strategy for these neurodegenerative disorders. Here, we identified two previously described glucosylceramide (GlcCer) synthase inhibitors, DL-threo-1-Phenyl-2-palmitoylamino-3-morpholino-1-propanol and Genz-123346(Genz), as enhancers of autophagy flux. We also demonstrate that GlcCer synthase inhibitors exert their effects on autophagy by inhibiting AKT-mammalian target of rapamycin (mTOR) signaling. More importantly, siRNA knock down of GlcCer synthase had the similar effect as pharmacological inhibition, confirming the on-target effect. In addition, we discovered that inhibition of GlcCer synthase increased the number and size of lysosomal/late endosomal structures. Although inhibition of GlcCer synthase decreases levels of mutant α-synuclein in neurons, it does so, according to our data, through autophagy-independent mechanisms. Our findings demonstrate a direct link between glycosphingolipid biosynthesis and autophagy in primary neurons, which may represent a novel pathway with potential therapeutic value for the treatment of Parkinson's disease.
Inhibition of GlcCer synthase enhances autophagy by inhibiting AKT-mTOR signaling, and increases the number and size of lysosomal/late endosomal structures. Furthermore, inhibition of GlcCer synthase decreased levels of mutant α-synuclein in neurons, which may represent a potential therapeutic target for Parkinson's disease.
Parkinson's disease (PD) pathology is characterized by the formation of intraneuronal inclusions called Lewy bodies, which are comprised of α-synuclein (α-syn). Duplication or triplication of α-syn gene or mutations in α-syn (A53T, A30P, and E46K) are linked to autosomal dominant PD, implicating the role of α-syn in the pathogenesis of PD. The therapeutic benefit of reducing α-syn levels has been proposed. One mechanism to degrade wild-type α-syn is through chaperone-mediated autophagy (CMA), however, mutant α-synucleins cannot be degraded through CMA and have been shown to inhibit CMA (Cuervo et al. 2004; Xilouri et al. 2009). These observations are further supported by a recent study which showed that enhancing CMA in vivo mitigates wild type α-synuclein-induced neurodegeneration (Xilouri et al. 2013).
Another potential mechanism to degrade both wild type and mutant α-syn is by stimulating macroautophagy, a lysosome-dependent degradation process. The process of macroautophagy (referred to here as autophagy) involves activation of the induction complex, formation of autophagosomes that engulf cytoplasmic materials, subsequent fusion of autophagosomes with lysosomes, and finally, degradation of sequestered material by lysosomal enzymes (Levine and Kroemer 2008). Autophagy was initially identified as a response to starvation. However, it also plays a critical role in the response to various stressors, including hypoxia, high temperature, accumulation of protein aggregates, and infection (Levine and Klionsky 2004; Nakai et al. 2007). In addition, previous studies link autophagy gene mutations to neurodegeneration in both mice and humans (Hara et al. 2006; Haack et al. 2012; Saitsu et al. 2013). The association of autophagy with human neurodegenerative diseases reveals its importance as a cellular surveillance system.
Several studies have shown that mutant α-syn is degraded through autophagy (Webb et al. 2003; Xilouri et al. 2008). The benefit of autophagy modulators in neurodegenerative diseases has also been investigated preclinically (Sarkar et al. 2005, 2007a,b; Berger et al. 2006; Williams et al. 2008; Rodriguez-Navarro et al. 2010; Spilman et al. 2010). These studies show that pharmacological modifiers of autophagy provide neuroprotection in cellular models of neurodegenerative disease by promoting degradation of aggregate-prone proteins. However, these studies rely heavily on in vitro studies using non-neuronal mammalian cell lines, emphasizing the need of further investigation in more physiologically-relevant systems, such as primary neurons.
Glucosylceramide (GlcCer) synthase catalyzes the synthesis of GlcCer, the initial step of glycosphingolipid biosynthesis. Inhibition of GlcCer synthase is commonly referred to as substrate reduction therapy, which was designed to block the synthesis of glucosylceramide, thus reducing glycosphingolipid biosynthesis. Pre-clinical and clinical studies of type I Gaucher disease demonstrate a significant improvement of disease manifestations after treatment with GlcCer synthase inhibitors (Lachmann and Platt 2001; Lukina et al. 2010a,b). Previous studies implicate the glycosphingolipid biogenesis pathway in a neurological context; conditional knock out of GlcCer synthase in mouse brain leads to severe neural defects after birth, highlighting the importance of glycosphingolipid biosynthesis for brain maturation after birth (Jennemann et al. 2005). In addition, the accumulation of GlcCer in lysosomes caused by mutations in glucocerebrosidase (GBA) impairs the degradation of α-syn by lysosomes, and may be involved in the α-syn-associated pathogenesis underlying PD (Mazzulli et al. 2011). Thus, it is intriguing to consider that a decrease in glucosylceramide, by inhibition of GlcCer synthase, could provide therapeutic benefit in PD through modulation of α-syn clearance. Recent studies have shown that pharmacological inhibition of GlcCer synthase provides neuroprotection in several models of lysosomal storage disorders (Ashe et al. 2011; Stein et al. 2012). For example, treatment with GlcCer synthase inhibitors in a Niemann-Pick disease type C feline model (characterized by neurodegeneration and intralysosomal accumulation of cholesterol and glycosphingolipids) decreased GlcCer, reduced production of glycosphingolipids, and improved Purkinje cell survival(Stein et al. 2012).
Here, we investigated whether two previously characterized GlcCer synthase inhibitors, DL-threo-1-Phenyl-2-palmitoylamino-3-morpholino-1-propanol (PPMP) and Genz-123346(Genz) (Lee et al. 1999; Zhao et al. 2007, 2009; Natoli et al. 2010), could enhance autophagy. We found that PPMP and Genz treatment reduced the phosphorylation of AKT and ribosomal protein S6, suggesting that PPMP and Genz act in an mTOR-dependent manner. To confirm the on-target effect, we performed siRNA knock down of GlcCer synthase and found very similar effects as those observed with pharmacological inhibition. Consistent with this, we also found significant increases in lysosomal/late endosomal compartments after GlcCer inhibition. Inhibition of GlcCer synthase in primary neurons led to increased autophagy flux, and reduced the mutant α-syn levels. However, blocking autophagy by 3-MA did not inhibit the PPMP or Genz- induced decrease of mutant α-syn levels, indicating that, despite the robust effects inducing autophagy, the decrease in mutant α-syn was not primarily driven by autophagic degradation. Our results demonstrate a link between glycosphingolipid biosynthesis and autophagy in primary neurons, and suggest therapeutic potential for modulating the glycosphingolipid biosynthesis pathway for PD.
Cell culture experiments
The inducible Flp-In™ T-REx™ HEK293 stable cell line expressing enhanced green fluorescent protein (EGFP)-p62 was generated according to the manual of the Flp-In system (Life Technologies, Grand Island, NY, USA). Cells were cultured in Dulbecco's modified Eagle's medium, GlutaMAX™ (10566; Life Technologies) supplied with 10% fetal bovine serum (FBS) (16140; Life Technologies), 15 μg/mL blasticidin (A1113903; Life Technologies), and 100 μg/mL hygromycin B (10843555001; Roche) in 5% CO2 at 37°C.
Wild-type control and Atg5−/− mouse embryonic fibroblasts (MEFs) were kindly provided by Noboru Mizushima (Universtiy of Tokyo, Japan). MEFs were grown in Dulbecco's modified Eagle's medium, GlutaMAX™ supplied with 10% FBS, 100 μg/mL penicillin, and 100 U/mL streptomycin (15140-122; Gibco, Rockville, MD, USA).
PPMP was purchased from Enzo Life Sciences (Cat # BML-SL215-0005, Farmingdale, NY, USA). Genz-123346 (Genz) was synthesized according to the published structure (Zhao et al. 2007). 3-Methyladenine (3-MA) was purchased from Sigma (St Louis, MO, USA) (Cat # M9281).
Primary neuronal culture
Embryonic primary neuronal cultures were prepared from E16 wild type or A53T transgenic mice of either sex. A53T α-syn mice and their wild-type littermates were ordered from the Jackson Laboratory (Bar Harbor, ME, USA). Homozygous transgenic C57BI/C3H mice were generated expressing human A53T α-syn under the control of the prion promoter. The generation and phenotype of these mice was previously described (Giasson et al. 2002; Paumier et al. 2013). Embryonic brain cortices were microdissected in cold Hank's balanced salt solution (HBSS) with Ca2+ and Mg2+ (14065-056; Gibco) buffer supplemented with 100 μg/mL penicillin and 100 U/mL streptomycin (15140-122; Gibco), 10 mM MgCl2 (M1028; Sigma), 2 mM GlutaMax (35050; Gibco), and 7 mM HEPES (15630-080; Gibco). Cortices were then mechanically dissociated using sterile Pasteur pipette followed by an enzymatic dissociation with 0.05% Trypsin-EDTA (25300-054; Gibco) and 100 units/mL of DNase I (LS2007, Code D; Worthington Biochemical Corp., Lakewood, NJ, USA) for 15–20 min at 37°C. Minimal essential medium (MEM) complete medium was added to stop trypsinization. Minimal essential medium (MEM) complete medium contained MEM Earle's salts (11090; Gibco) with GlutaMax, 100 μg/mL penicillin and 100 U/mL streptomycin, 2 mM GlutaMax, 0.278% (weight/volume) additional glucose (G7021; Sigma), 10% heat-inactivated FBS (SH30070-03; HyClone, Logan, UT, USA). Cells were then counted using cell counter and Trypan blue (15250; Life Technologies). Neurons were plated at the density of 800 000 per well in a 6-well poly-d-lysine coated plate and maintained in Neurobasal medium (21103-049; Gibco) containing B27 supplement (17504; Gibco), 2 mM GlutaMax, 100 μg/mL penicillin, and 100 U/mL streptomycin. All animal experiments or procedures were carried out in accordance with regulations and established guidelines and were reviewed and approved by Pfizer Institutional Animal Care and Use Committee. The ARRIVE guidelines (http://www.nc3rs.org.uk/page.asp?id=1357) was followed.
Cell were lysed in RIPA buffer (R0278; Sigma), containing protease inhibitors (Sigma) and phosphatase inhibitors I and III (P2850, P0044; Sigma), on ice for 30 min and then centrifuged at 10 000 g for 10 min at 4°C. Supernatant was collected and analyzed by western blot. Proteins were resolved using Novex® 4–20% Tris-Glycine gels or NuPAGE® Novex® 4–12% Bis-Tris Gels (Life Technologies). Primary antibodies were incubated at 4°C overnight and LI-COR secondary antibodies were used at 1 : 20 000. The blots were scanned and quantified with the Odyssey infrared scanning system (LI-COR, Lincoln, NE, USA). The following primary antibodies were used: Rabbit anti-α-syn antibody (sc-7011-R; Santa Cruz Biotechnology, Santa Cruz, CA, USA) at 1 : 250, mouse anti-human-α-syn antibody 4B12 (SIG-39730; Covance, Princeton, NJ, USA) at 1 : 1000, polyclonal anti-Actin (A2066; Sigma) at 1 : 2000, monoclonal anti-Actin (A2228; Sigma) at 1 : 25 000, rabbit polyclonal anti-LC3B at 1 : 1000 (L7543; Sigma), mouse anti-p62 at 1 : 1000 (Abcam, ab56416), rabbit anti-pS6 at 1 : 1000 (2211; Cell Signaling Technology, Beverly, MA, USA), rabbit anti-S6 at 1 : 1000 (2217; Cell Signaling Technology), rabbit anti-pAKT at 1 : 1000 (9275; Cell Signaling Technology), rabbit anti-AKT at 1 : 1000 (9272; Cell Signaling Technology), rabbit anti-mTOR at 1 : 1000 (2983; Cell Signaling Technology), rabbit anti-ULK1 (4773; Cell Signaling Technology), at 1 : 11 000, rabbit anti-pERK1/2 at 1 : 1000 (9101; Cell Signaling Technology), and rabbit anti-ERK1/2 at 1 : 1000 (9102; Cell Signaling Technology).
Flp-In™ T-REx™ HEK293 EGFP-p62 cells were cultured in 6-well poly-d-lysine plates (354413; BD Biosciences) as described above. The expression of EGFP-p62 was induced by adding doxycycline (PV6051; Sigma) at final 1 μg/mL to the culture medium for 24 h. The expression was turned off by removing Doxycycline and washing the cells three times with culture medium. The cells were then treated for 12 h with the following compounds: 0.1% dimethylsulfoxide (D2650; Sigma), 1 μM Rapamycin (R8781; Sigma), 1 μM PI-103 (EMD Millipore, Billerica, MA, USA), 400 nM bafilomycin A1 (B1793; Sigma), and 10 μM MG132 (C2211; Sigma). After compound treatment, Flp-In™ T-REx™ HEK293 EGFP-p62 cells in 6-well plates were detached using 0.25% trypsin (25200056; Life Technologies), followed by addition of 5% FBS in HBSS (10-547F; Lonza, Allendale, NJ, USA). Cells were washed once with HBSS, centrifuged and re-suspended in HBSS supplemented with 1% FBS. Fluorescence intensity was then analyzed using a BD FACS Calibur Flow Cytometer with CELLQuest software (BD Biosciences). Geometric means of the green fluorescent protein (GFP) intensity from 10 000 cells were recorded.
Transfection of siRNA was performed with Lipofectamine RNAiMAX (13778075; Life Technologies) according to manufacturer's instructions using 50 nM siRNA. Cells were analyzed 72 h after trans-fection. The following siRNAs were used: GlcCer synthase siRNA-1 (GCSi-1) (s14643; Life Technologies), GlcCer synthase siRNA-2 (GCSi-2) (s14644; Life Technologies), GlcCer synthase siRNA smart pool (GCSi-s) (L-006441; Thermo Scientific Dharmacon, Pittsburgh, PA, USA), mTORi-1 (s603; Life Technologies), mTORi-2 (s604; Life Technologies), and ULK1i (L-005049-00; Thermo Scientific Dharmacon).
Total RNA was isolated from cells using the RNeasy Mini Kit (Qiagen). cDNA was synthesized using iScript™ cDNA Synthesis Kit (Bio-Rad Laboratories, Hercules, CA, USA). The expression levels of GlcCer synthase were assessed using TaqMan® Gene Expression Assay primer/probe for human UGCG (Cat #, Hs00234293_m1, Life Technologies). TaqMan® Gene Expression Assay primer/probe for human Glyceraldehyde 3-phosphate dehydrogenase was used as normalization control (Cat # 402869, Life Technologies). qPCR was performed using StepOne™ System (Applied Biosystems, Wilmington, DE, USA). Data were analyzed using ΔΔCt methods.
Lysotracker red staining and analysis
HEK293T cells were plated onto a poly-d-lysine coated 96-well microplate (Greiner Bio-One, Monroe, NC, USA) and treated with vehicle or the indicated concentrations of drug for 8 h prior to imaging. Cells were then treated with 50 nM Lysotracker Red (Life Technologies) and NucBlue Live ReadyProbes Reagent (1 : 1000, Life Technologies) to label lysosomes and nuclei, respectively, for 30 min and imaged live using the Cellomics Arrayscan (Thermo) at 37°C and 5% CO2. In a single experiment, there were six wells for each treatment condition (dimethylsulfoxide, PPMP or Genz), and 64 fields were imaged per well. Therefore, there were 384 fields imaged for each condition. Experiments were repeated three times, and the quantification graphs were a summary of these three independent experiments. Widefield images were taken using a 40× objective, scanning 64 fields per well. Image Analysis was performed with the Cellomics Arrayscan VTI using the SpotDetector Bioapplication to quantify the number, size, and intensity of lysosomes/late endosomes. Higher magnification confocal imaging was performed using a Yokogawa Spinning Disk Confocal on a Zeiss inverted microscope. Cells were visualized using a 63× NA 1.4 Oil Apochromat objective and illuminated with 405 nm and 561 nm laser light. Images were captured using a Roper Evolve back-thinned EM CCD camera and Zen software (Zeiss, Thornwood, NY, USA) and were subsequently processed in using ImageJ software (National Institutes of Health, Bethesda, MD, USA).
Statistical analysis was performed with anova with Bonferroni correction or Student's t test using GraphPad Prism (La Jolla, CA, USA). Each western blot image shown was from a representative experiment from at least three experiments. Each quantification graph shown was a summary of these experiments if not specified. Data shown were mean ± SEM.
Inhibition of GlcCer synthase increases autophagy flux in HEK293 cells
Many autophagy assays that measured the levels of LC3-II were significantly limited in their ability to distinguish autophagy enhancers from blockers, as the accumulation of autophagosomes could result either from an increase in the formation or a decrease in the clearance of autophagosomes. Therefore, it was critical that a reliable assay was used to measure autophagy flux. We utilized a previously described flow cytometry assay that monitored the degradation of EGFP-p62 (Larsen et al. 2010). SQSTM1/p62 (referred to as p62) was an ubiquitin-binding protein that targeted ubiquitinated proteins for degradation by autophagy, and p62 itself was also degraded in this process. We utilized the known autophagy enhancers, rapamycin, and PI-103, as controls to confirm increased autophagy flux when treating the Flp-In™ T-REx™ HEK293 EGFP-p62 cells (Figure S1a). Conversely, the autophagy inhibitor bafilomycin A1, which blocked the autophagosomes-lysosome fusion, was used as a negative control and led to more than 2-fold increase in EGFP-p62 levels. MG132, a proteasome inhibitor, had almost no effect on the EGFP-p62 levels (Figure S1a), indicating the assay was selective for autophagy degradation.
Using this assay, we identified two GlcCer synthase inhibitors-PPMP and Genz-123346 (Genz) as autophagy enhancers. Both PPMP and Genz increased the degradation of EGFP-p62 in a concentration-dependent manner (Fig. 1a and b). To confirm the effect of GlcCer synthase inhibitors on autophagy flux, we performed the standard autophagy flux assay by measuring LC3-II levels in the presence of bafilomycin A1 at a saturating concentration via western blot analysis. A saturating concentration of bafilomycin A1 led to a complete blockade of LC3-II degradation; therefore, any further increase in LC3-II levels upon compound treatment was because of an increase in the formation of autophagosomes, indicating increased autophagy flux. Using the same cell line as in the EGFP-p62 assay, we observed a dose-dependent response with both PPMP and Genz treatment, which resulted in a further increase in LC3-II levels in presence of bafilomycin A1 when compared with control samples (Fig. 1c and d). This result demonstrated that GlcCer synthase inhibitors PPMP and Genz were autophagy enhancers.
Nishida and co-workers reported Atg5/Atg7-independent alternative autophagy (Nishida et al. 2009). To examine whether GlcCer synthase inhibitors affected the Atg5-dependent canonical autophagy pathway, we treated wild type (WT) or Atg 5−/− mouse embryonic fibroblast epithelium (MEF) cells for 8 h in the absence or presence of 10 μM PPMP or Genz (Fig. 1e). Similar to results in HEK293 cells, GlcCer synthase inhibitor treatment in WT MEF cells (8 h) was associated with a significant increase in LC3-II levels (Fig. 1). However, no accumulation of LC3-II was noted in Atg5−/− MEF cells following 8 h of GlcCer synthase inhibitor treatment. This result suggested that GlcCer synthase inhibitors stimulated autophagy in an Atg5-dependent manner.
Pharmacological approaches could often exhibit off-target effects because of limited selectivity of the molecules. To confirm whether PPMP and Genz elicited on-target effects, we silenced GlcCer synthase using siRNA in the EGFP-p62 assay to investigate its impact on autophagy flux. First, to evaluate the feasibility of adapting the EGFP-p62 assay to the siRNA platform, we performed siRNA knock down of two key autophagy regulatory genes:mTOR and ULK1. Initially, we confirmed the siRNA knock down of mTOR and ULK1 via western blot analysis (Figure S1b), and observed that siRNA knock down of autophagy inhibitor mTOR decreased the accumulation of EGFP-p62 (indicating an increase in autophagy flux), whereas knock down of the essential autophagy gene, ULK1, significantly increased EGFP-p62 levels (indicating a decrease in autophagy flux) (Figure S1b). These results demonstrated the adaptability of the EGFP-p62 assay to a siRNA knockdown platform to evaluate the effects of genes of interest on autophagy flux. To knock down GlcCer synthase in Flp-In™ T-REx™ HEK293 EGFP-p62 cells, we used three different siRNAs (GCSi-1, GCSi-2 and GCSi-s) that targeted different regions of GlcCer synthase transcripts. We confirmed that the siRNA knock down via quantitative RT-PCR, all three siRNAs led to a ~ 90% reduction of GlcCer synthase mRNA levels (Figure S1d). As a result of the unavailability of a reliable antibody against GlcCer synthase, we were not able to confirm the siRNA knock down using western blot analysis. Nonetheless, siRNA knock down of GlcCer synthase resulted in significant decreases in EGFP-p62 levels comparable with siRNA knock down of mTOR (Fig. 2a), indicating an increase in autophagy flux. More importantly, when we knocked down GlcCer synthase using siRNA, PPMP or Genz failed to further decrease EGFP-p62 levels (Fig. 2b). Taken together, these results suggested that PPMP and Genz exerted their effect primarily by inhibiting GlcCer synthase.
Inhibition of GlcCer synthase induced autophagy via the AKT-mTOR signaling pathway
To identify the mechanism by which GlcCer synthase inhibition induced autophagy flux, we examined the effect of PPMP and Genz on the AKT-mTOR signaling pathway, which was a key negative regulator of autophagy (Yang and Klionsky 2010). The activity of the AKT-mTOR signaling pathway could be monitored by the level of phosphorylated forms of AKT and ribosomal S6 protein (S6). We observed a significant decrease in p-AKT and p-S6 levels in HEK293T cells treated with PPMP or Genz for 8 h, whereas the total AKT and S6 levels remained relatively unaltered (Fig. 3a and b). In contrast, extracellular signal–regulated kinase (ERK) signaling was slightly increased by GlcCer synthase inhibitors (Figure S2). More importantly, when we knocked down GlcCer synthase in HEK293T via siRNA, we found a similar reduction in the phosphorylation levels of AKT and S6 (Fig. 3c and d). Our findings were also consistent with a recent in vivo study in which the inhibition of AKT-mTOR was reported in the kidneys of Genz treated mice (Natoli et al. 2010). Furthermore, when we knocked down mTOR in Flp-In™ T-REx™ HEK293 EGFP-p62 cells using siRNA, we did not observe further reductions in EGFP-p62 levels when the cells were also treated with 5 μM PPMP or Genz (Fig. 2b). Based on these results, we concluded that GlcCer synthase inhibitors induced autophagy via inhibition of the AKT-mTOR signaling pathway.
Inhibition of GlcCer synthase led to increases in the size and numbers of lysosomes/late endosomes
Autophagy relied on lysosomes to degrade the sequestered contents of the autophagosome. Alterations in autophagy levels were often accompanied by changes in the lysosomal compartment. To test whether inhibition of GlcCer synthase also affected lysosomes, we used Lysotracker Red, an acid-dependent dye commonly used to label lysosomal/late endosomal compartments in live cells. We treated HEK 293T cells with 5 μM PPMP and Genz for 8 h and found that inhibition of GlcCer synthase resulted in significant increases in the average size and numbers of lysosomal/late endosomal structures (Fig. 4 and Figure S3). The total lysotracker intensity per cell was also significantly increased (Fig. 4), suggesting an enlargement of lysosomal/late endosomal compartments.
GlcCer synthase inhibition increased autophagy flux and reduced mutant α-syn levels in mouse primary neurons – but potentially via indendent mechanisms
Enhancing autophagy flux wasshown to facilitate the clearance of aggregation-prone proteins including mutant α-syn (Ravikumar et al. 2002; Webb et al. 2003). Furthermore, Mazzulli et al. 2011 proposed a pathogenic positive feedback mechanism of α-syn and GBA depletion, which led to the accumulation of GlcCer (Mazzulli et al. 2011). Therefore, we investigated whether GlcCer synthase inhibitor-mediated increases in autophagy flux would facilitate the clearance of mutant α-syn proteins. To examine this, we used primary neurons from A53T transgenic mice, a model for PD in which human A53T mutant α-syn was expressed under the prion promoter (Giasson et al. 2002). We found that treating the primary neurons with 5 μM PPMP or Genz resulted in a significant enhancement of autophagy flux, measured by LC3-II western blot (Fig. 5a). Similar to HEK293T cells, pharmacological inhibition of GlcCer synthase in primary neurons led to a significant decrease in the phosphorylated forms of AKT and S6 proteins, indicating that the AKT-mTOR signaling pathway was inhibited (Fig. 5b). These results demonstrate that inhibition of GlcCer synthase activity also enhanced autophagy in primary neurons.
Next, we examined the levels of human A53T mutant α-syn levels following the GlcCer synthase inhibitor treatments. We observed that 48 h treatment with 5 μM PPMP or Genz resulted in a significant reduction in human A53T mutant α-syn when compared with the control samples (Fig. 5c). To test whether the decrease in mutant α-syn protein was because of enhanced autophagy flux, we used 3-methyladenine (3-MA), which blocked the formation of autophagosomes (Seglen and Gordon 1982; Petiot et al. 2000; Sarkar et al. 2009a). In A53T primary neurons simultaneously treated with 3-MA, we found that PPMP or Genz also reduced A53T mutant α-syn levels to the similar extent in A53T neurons that were not treated with 3-MA (Fig. 5c), which suggested that GlcCer synthase inhibitors mediated the clearance of mutant α-syn protein primarily through mechanisms other than autophagy or via a 3-MA insensitive autophagy pathway (Juenemann and Reits 2012). Furthermore, we analyzed the GlcCer synthase inhibitors in a previously described PC12-α-syn cell line, in which the turnover of hSyn-A53T-HA was dependent on autophagy (Webb et al. 2003; Williams et al. 2008) (Figure S4a). However, we found that PPMP or Genz did not affect the turnover of hSyn-A53T-HA (Figure S4b), which suggested that the mechanism that mediated GlcCer synthase inhibitors induced the reduction of mutant α-syn protein in primary neurons was distinct from the PC12 cells.
Our data demonstrate a link between the glycosphingolipid biosynthesis pathway and autophagy. We discovered that pharmacological inhibition of GlcCer synthase or siRNA knock down of this enzyme leads to enhanced autophagy flux. We also illustrate that GlcCer synthase inhibitors exert their effects on autophagy through inhibition of the AKT-mTOR signaling pathway. Moreover, inhibition of GlcCer synthase resulted in significant increases in the average size and number of lysosomal/late endosomal structures, consistent with an intracellular regulatory role of GlcCer biosynthesis in the autophagy-lysosome pathway. More importantly, we found that GlcCer synthase inhibition promoted the clearance of mutant α-syn protein, suggesting that GlcCer synthase inhibitors possess some therapeutic potential to ameliorate pathogenic α-syn mediated toxicity.
GlcCer biosynthesis and autophagy-lysosome pathway
GlcCer synthase produces GlcCer from ceramide, the initial step of glycosphingolipid biosynthesis. Glycosphingolipids are essential structural components of eukaryotic cell membranes and membrane compartments, and play a role in regulating cell growth, differentiation, intracellular signaling, and trafficking (Futerman and Hannun 2004). Our results indicate that the glycosphingolipid biosynthesis pathway also plays a role in regulating autophagy. A previous study revealed that exogenous ceramide induces autophagy (Scarlatti et al. 2004), suggesting that endogenous ceramide may also mediate the induction of autophagy by other lipids. In addition, a recent study by Mosbech et al. (2013) reported an increase in LGG::GFP (lgg is the C. elegan homolog of Atg8 gene) containing puncta in the seam cells of the hyl-1;lagr-1 double mutant C. elegans that lack two ceramide synthases, indicating an increase in autophagosome numbers. Addition of cell-permeable ceramide also deactivates AKT (Schubert et al. 2000; Scarlatti et al. 2004), which is consistent with our results that inhibiting GlcCer synthase led to a reduction of AKT phosphorylation. The AKT-mTOR signaling pathway is the primary negative regulatory pathway for autophagy, and mTOR inhibits autophagy by controlling the activation of the autophagy induction complex (Nazio et al. 2013). Since PI3K/AKT signaling plays a critical role in lipid metabolism (Taniguchi et al. 2006), our results suggest that AKT-mTOR signaling represents the link between lipid metabolism and autophagy.
GlcCer has also been shown to modulate membrane trafficking along the endocytic pathway (Sillence et al. 2002). Since the lysosomes are considered the terminal point in the endocytic pathway and essential for the autophagy process, we examined the effect of GlcCer synthase inhibitors on lysosome morphology. We found a modest increase in both the abundance and average size of the lysosomes/late endosomes after GlcCer synthase inhibition. Yu et al. 2010 reported a similar increase of lysosomal sizes in normal rat kidney cells under starvation conditions when autophagy is activated (Yu et al. 2010). However, whether inhibiting GlcCer synthase has a direct effect on lysosomes, or the lysosomal changes are secondary to the increase in autophagy flux, still remains to be determined.
Therapeutic value of GlcCer inhibitors for PD
Glycosphingolipids accumulation as a result of defective lysosomal breakdown is the main cause of glycosphingolipid storage disorders (GSDs), although the detailed pathological mechanisms remain unknown. One of the most prevalent GSDs is Gaucher disease, which is caused by mutations in GBA, the lysosomal enzyme that breaks down GlcCer. Substrate reduction therapy with inhibitors of GlcCer synthase that are designed to lower the substrate GlcCer levels, has been investigated in clinical trials of Gaucher disease (Cox et al. 2000; Lachmann and Platt 2001) and in the treatment of mouse models of other GSDs such as Sandhoff and Niemann-Pick C diseases (Jeyakumar et al. 1999; Zervas et al. 2001; Ashe et al. 2011; Stein et al. 2012). Besides this therapeutic value in substrate reduction therapy, our results revealed a novel therapeutic potential for these GlcCer synthase inhibitors – clearance of mutant α-syn, although the detailed mechanism remains to be fully determined.
Pharmacological induction of autophagy can be a useful mechanism to promote the clearance of these aggregate-prone proteins and protect cells against the toxic effects of these proteins in a range of disease models (Ravikumar et al. 2002; Berger et al. 2006; Rose et al. 2010; Hochfeld et al. 2013). Several studies have shown that inducing autophagy by targeting mTOR kinase using rapamycin enhances the clearance of mutant huntingtin fragments and α-syn, reducing aggregate formation and thus protecting against toxicity in models of Huntington's disease and PD respectively (Berger et al. 2006; Sarkar et al. 2009b). Therefore, we initially proposed that enhancing autophagy flux by inhibiting GlcCer synthase may also be able to promote clearance of α-syn. We performed additional experiments to test this hypothesis. However, our results suggested that the GlcCer synthase inhibitor-mediated decrease in A53T α-syn in primary neurons were largely independent on the increase of autophagy flux, because blocking of autophagy by 3-MA has no effect on the decrease of A53T α-syn caused by inhibition of GlcCer synthase. This may be mediated through alternative, 3-MA-insensitive, autophagy pathways (Juenemann and Reits 2012). In addition, Mazzulli et al. (2011) revealed a self-propagating positive feedback process in which elevated levels of toxic α-syn species impaired lysosomal trafficking of GBA that lead to accumulation of GlcCer and further stabilization of α-syn oligomers (Mazzulli et al. 2011). Moreover, GBA mutations or knock down are associated with an impaired lysosome pathway presumably caused by GlcCer accumulation (Sun and Grabowski 2010; Mazzulli et al. 2011). Thus, specific treatments that decrease GlcCer accumulation are expected to break the pathogenic feedback loop of α-syn aggregation and reduce the formation of toxic α-syn oligomers in PD. In our study, we demonstrated that GlcCer synthase inhibition reduced mutant α-syn levels, supporting this hypothesis.
Together, our results suggest that the accumulation of GlcCer may be particularly relevant to PD and the strategies that target GlcCer accumulation may present a novel intervention to reduce α-syn-mediated toxicity in PD and other synucleinopathies. However, the two GlcCer synthase inhibitors, PPMP and Genz, cannot penetrate the blood brain barrier (Larsen et al. 2011 and data not shown), preventing us from pursuing in vivo experiments to investigate the GlcCer synthase inhibition in animal models of PD. Larsen et al. (2011) has reported a potent GlcCer synthase inhibitor that spreads throughout the central nervous system and reduces GlcCer levels in the brain (Larsen et al. 2011). Further investigation using similar GlcCer synthase inhibitors will be of particular interest.
Acknowledgments and conflict of interest disclosure
This work was funded by a Michael J. Fox Foundation Grant (Therapeutic Development Initiative). We are grateful to Noboru Mizushima (Universtiy of Tokyo) for kindly providing wild-type control and Atg5−/− mouse embryonic fibroblasts (MEFs).
All authors were employees and shareholders of Pfizer during the period this study when the data were generated and interpreted. All experiments were conducted in compliance with the ARRIVE guidelines.