Abstract
Glucocerebrosidase gene (GBA) mutations are the most common genetic contributor to Parkinson’s disease (PD) and are associated with decreased glucocerebrosidase (GCase) enzymatic activity in PD. PD patients without GBA mutations also exhibit lower levels of GCase activity in the central nervous system suggesting a potential contribution of the enzyme activity in disease pathogenesis, possibly by alteration of lysosomal function. α-synuclein (ASYN), a protein with a central role in PD pathogenesis, has been shown to be secreted partly in association with exosomes. It is possible that a dysfunction of the endocytic pathway through GCase may result in altered exosome release of ASYN. The aim of this study was to examine whether manipulating GCase activity in vivo and in vitro could affect ASYN accumulation and secretion. GCase overexpression in vitro resulted in a significant decrease of exosome secretion. Chronic inhibition of GCase activity in vivo, by administration of the covalent inhibitor conduritol-B epoxide in A53T-synuclein alpha gene Tg mice significantly elevated intracellular oligomeric ASYN species. Importantly, GCase inhibition, induced a profound increase in the number of brain exosomes released, as well as exosome-associated ASYN oligomers. Finally, virus-mediated expression of mutant GBA in the mouse striatum increased ASYN secretion in the same region. Together, these results provide for the first time evidence that a decrease of GCase or overexpression of mutant GCase in a chronic in vivo setting can affect ASYN secretion. Such effects may mediate enhanced propagation of ASYN, driving pathology in GBA-associated PD.
Introduction
In the past decade, a strong association between Gaucher disease (GD) and Parkinson’s disease (PD) has been demonstrated. Mutations in the GBA1 gene (OMIM 606463), the gene encoding the lysosomal enzyme GCase, have been shown to be an important genetic risk factor for PD (1–4). Pathogenetic mutations in both alleles of GBA1 cause the recessively inherited lysosomal storage disorder Gaucher. While the initial notion of a possible correlation between the GBA1 gene and PD came from observing Parkinsonism in patients with mild GD (5,6), it was through a case–control study that an increased risk of GBA heterozygotes to develop PD was revealed (7).
Understanding the role of GBA mutations in the pathogenesis of synucleinopathies remains a challenge. Experimental data suggest that there may be a direct relationship between levels of GCase activity and α-synuclein (ASYN) proteostasis, while both loss-of-function and toxic gain-of-function mutations in GBA have been proposed as possible contributing factors (8).
Several independent groups have reported increased accumulation of ASYN in the brains of mouse models of GD with different pathogenetic GBA mutations (9–14). Decrease in enzymatic activity of lysosomal hydrolases in GBA mutation carriers may contribute to PD pathogenesis by increasing the level of neurotoxic oligomeric ASYN species (15). On the other hand, we have reported, based on studies in patient erythrocytes and neuronal cell lines, that effects of GBA mutations on ASYN may be lysosome-independent, and relate more to the altered lipid properties of membranes (16–18).
A number of studies have demonstrated that ASYN can be physiologically secreted to the extracellular space while it has also been shown that this release is achieved, at least in part, via exosomes (19,20). Interestingly, oligomeric species can be found in exosomes (19) suggesting that increased exosome associated secretion of ASYN could aid toxic seed formation perhaps via the proposed prion-like mechanisms of pathology progression. Moreover, upregulation of exosome secretion has been correlated with conditions of impaired lysosomal function and increased cytosolic cargo of a misfolded protein (19–23). In agreement with these data, levels of ASYN were recently found to be increased in the media of GBA-N370S Parkinson’s induced pluripotent stem cells (iPSC)-derived dopamine neurons which also exhibited autophagic/lysosomal disturbances (24).
Given the valid hypothesis that deregulation of extracellular ASYN levels could be critical for the initiation or the progression of Parkinson’s disease we sought to investigate the effects of altered GCase activity and expression in the secretion of ASYN in a mouse model of PD. Pharmacologic inhibition of central nervous system (CNS) GCase activity led to ASYN species accumulation in select brain regions. Importantly, this inhibition led to a significant increase in exosome release and exosome-associated oligomeric ASYN. Adenovirus mediated overexpression of the N370S mutant form of GBA in the mouse striatum also resulted in increased ASYN secretion. Together, these results demonstrate that modulation of GCase in the CNS of mouse models of PD could affect ASYN secretion through exosome-associated and free form secretion, suggesting that GBA-linked PD may result from altered homeostasis of ASYN secretory pathways, possibly linked to disease spreading.
Results
Expression of wild-type and mutant GCase does not affect the intracellular and secreted ASYN levels in primary mouse cortical neurons
It has been previously shown that the D409V and N370S GBA mutants can cause ASYN aggregation, both in animal models and iPSC dopamine neurons derived from GBA mutation carriers (10,25,26). Based on these observations, D409V and N370S GBA mutants were chosen for the in vitro experiments. Adenoviral vectors (AV) expressing wild-type (WT) and mutant forms (D409V, N370S) of GBA were used to transduce primary mouse cortical neurons. A green fluorescent protein (GFP)-expressing-AV was used as a control for the viral transduction. GCase levels were measured by Western blotting using a polyclonal GBA antibody. Compared with the GFP transduced cells, WT GBA AV and D409V GBA AV showed a statistically significant increase in GCase levels [WT GBA: (4.80 ± 0.95)-fold increase versus GFP: (1.09± 0.25)-fold increase compared with non-transduced cells, D409V GBA: (4.27 ± 1.14)-fold increase versus GFP: (1.09± 0.25)-fold increase of non-transduced cells]; an increase in GCase levels was observed in the N370S mutant AV infected cells compared with the GFP transduced cells, however, this increase was not statistically significant [N370S GBA: (3.00 ± 0.69)-fold increase versus GFP: (1.09± 0.25)-fold increase of non-transduced cells, Fig. 1A]. GCase activity assay of the cell pellet revealed that the WT GBA infected cells had 64% higher activity compared with that of the non-transduced cells [WT GBA: 11.21 ± 0.54 versus GFP: 5.84± 0.97 relative fluorescent units (RFU)/μg of protein] while the cells infected with the mutant forms of GBA were found to have no difference in the GCase activity assay, when compared with GFP control (D409V GBA: 6.41 ± 1.45 versus GFP: 5.84± 0.97 RFU/μg of protein, N370S GBA: 5.26 ± 0.56 versus GFP: 5.84± 0.97 RFU/μg of protein, Fig. 1B). This suggests that the mutant forms expressed a non-active form of GCase which did not affect the endogenous activity levels. ASYN levels were measured both by western blotting and by sandwich enzyme-linked immunosorbent assay (ELISA). A trend of increase in ASYN levels was observed by western blotting in all transduced cells, none of which was however statistically significant (Fig. 1C). Similar results were obtained for secreted ASYN levels measured by ELISA, as shown in Figure 2A. A trend of increase that was not significant was observed in all mutant GBA-treated cells compared with control (Fig. 2A).
Effects of GBA expression on α-synuclein levels in mouse cortical neurons. Mouse cortical neurons in fifth DIV were infected with control or WT GBA, D409V GBA and N370S GBA expressing AV viruses. Following 2 days of infection samples were analysed for the levels of GCase. (A) All GBA AV-transduced mouse cortical neurons showed an increase in GCase protein levels compared to neurons treated with the GFP AV control virus (n=6–8/group). The increase in GCase levels in WT GBA and D409V GBA infected neurons was statistically significant. (B) Neurons treated with the WT GBA also showed a significant increase in GCase activity (n=3/group). α-Synuclein levels in infected neurons were measured, both by Western blotting and ELISA. No significant differences in α-synuclein levels between the different groups (n=6–8/group) were observed (C). β-Actin was used as a loading control. Data represent mean values±SEM. Differences were estimated using one-way ANOVA followed by Fischer’s LSD test and Student’s t-test. (A) P=0.0312; RFU: relative fluorescent units.
Extracellular levels of α-synuclein in GBA overexpressing mouse cortical neurons. The CM of AV-transduced mouse cortical neurons were used to measure secreted α-synuclein levels by sandwich ELISA (A). No significant changes were observed (n=7–10/group) in secreted total α-synuclein levels. (B) Exosomes from the CM were isolated and their levels were estimated using an AChE assay. Cells treated with WT GBA1 AV, secreted significantly less exosomes compared to the GFP AV control treated cells (n=4–5/group). Exosome α-synuclein levels were measured by ELISA. No changes were observed in total exosome associated α-synuclein levels (C). Similarly, no changes were noted in the ‘per exosome’ α-synuclein cargo, when exosome α-synuclein levels were normalized to the total exosome AChE levels (D) (n=3–6/group). Data represent mean values ±SEM. Differences were estimated using one-way ANOVA followed by Tukey’s post hoc test. (B) P=0.0197.
Overexpression of WT GBA in primary mouse cortical neurons reduces exosome release
Secreted ASYN has been shown to be associated with exosomes (19,20,24). We sought to investigate whether overexpression of GBA pathologic forms could affect the levels of exosomes. To this end, exosomes from the condition medium/media (CM) were isolated and used for an acetylcholinesterase (AChE) activity assay. The activity of AChE, an enzyme specific to exosomes, offers an estimation of the exosome numbers of the fraction collected from the centrifugation. As shown in Figure 2B the cells infected with the WT GBA AV were found to have 37% less exosomes secreted compared with the non-transduced cells (WT GBA: 0.34 ± 0.05 versus non-transduced: 0.54 ± 0.02 ng AChE/μg of protein) as judged by the AChE assay. In order to investigate whether exosome-associated ASYN was different between groups, ASYN levels of the exosome fractions were measured by ELISA. Results were normalized to the total exosome protein measured by Bradford Protein Assay. As depicted in Figure 2C, total exosome associated ASYN levels showed a trend of decrease in the WT GBA overexpressing cells (WT GBA: 0.17 ± 0.05 versus non-transduced: 0.22 ± 0.05 ng ASYN per μg exosome protein). A normalization of exosome fraction ASYN levels with the total exosome fraction AChE levels offered an estimate of the ‘ASYN cargo per exosome’. Again, no differences were found between groups (Fig. 2D).
Pharmacologic inhibition of GCase activity in the presence of AV-induced ASYN overexpression does not significantly affect the intracellular and secreted ASYN levels in primary mouse cortical neurons
We have previously reported that GCase inhibition does not affect endogenous intracellular ASYN levels in cultured rat cortical neurons (18). Effects on secreted ASYN, however, have not been reported. We sought to investigate whether GCase inhibition in the presence of ASYN overexpression could alter intracellular and extracellular ASYN levels. To this end, WT ASYN or GFP were overexpressed using AVs in primary mouse cortical neurons. GCase activity was inhibited in one group by adding conduritol-B epoxide (CBE). The enzyme activity was measured through a fluorescence assay as described in Materials and Methods. ASYN overexpression reduced GCase activity significantly compared with GFP (asyn AV: 15.68 ± 0.70 versus GFP AV: 20.40 ± 3.01 RFU/μg of protein, Supplementary Material, Fig. S1A), in agreement with previous observations in cortical lysates of A53T mice (27). CBE treatment did not influence intracellular Triton X soluble ASYN levels as measured by Western Blotting (CBE + ASYN AV: 120.50 ± 6.88, ASYN AV: 114.50 ± 12.18 levels normalized to β-actin, Supplementary Material, Fig. S1B). The sodium dodecyl sulfate (SDS)-soluble ASYN fraction showed no differences between the two groups (CBE + ASYN AV: 1.93 ± 0.55, ASYN AV: 1.52 ± 0.47 levels normalized to flotillin-1, Supplementary Material, Fig. S1B). Extracellular ASYN levels showed a trend for increase that was not statistically significant, as measured by ELISA (CBE + ASYN AV: 1.71 ± 0.45, ASYNAV: 1.06 ± 0.36 ng ASYN/μg total protein, Supplementary Material, Fig. S1C).
Pharmacologic inhibition of GCase activity does not affect exosome number and exosome-associated ASYN in primary mouse cortical neurons
We next sought to investigate whether inhibiting GCase activity in mouse cortical neurons could affect exosome release. CBE was used to inhibit GCase as described previously. As shown in Supplementary Material, Figure S1D, exosome release in cells treated with CBE was not different compared with non-treated cells (CBE: 0.25 ± 0.01 versus non-treated: 0.25 ± 0.05 ng AChE/μg of protein). Similarly, total exosome associated ASYN levels were not different between groups (CBE: 0.29 ± 0.03 versus non-treated: 0.29 ± 0.03 ng ASYN per μg exosome protein). No differences were found in the ‘ASYN cargo per exosome’ between groups (CBE: 1.15 ± 0.14 versus non-treated: 1.24 ± 0.24 ng ASYN per ng AChE).
Pharmacologic inhibition of GCase activity increases oligomeric ASYN levels in select brain regions in vivo
To further elucidate a role of GBA in ASYN levels in vivo we pharmacologically inhibited GCase in mice overexpressing A53T ASYN with chronic (8 weeks) exposure to CBE. Previous studies indicated that an administration interval of 48–72 h was optimal to achieve a residual brain GCase activity close to 50% (28). The effect of GCase inhibition via intra peritoneal injections was confirmed by measuring the GCase activity in the brain of the treated mice. Cortex or striatal tissue was collected 48–96 h after the last injection and the enzyme activity was measured through a fluorescence assay as described in Materials and Methods. Mice treated with CBE were found to have a 20% reduction in enzyme activity compared with the saline treated mice (Fig. 3A). It has previously been reported that in mice lacking GBA proteasomal activity is compromised (29). To investigate whether the GCase activity inhibition would have an effect in proteasomal activity in our model, proteasomal activity was also measured in a similar manner. As shown in Figure 3A there was no difference found between the CBE and the saline treated groups (CBE: 155.50 ± 40.03 versus saline: 203.30 ± 23.36 RFU/μg total protein).
Increase of α-synuclein levels in CBE injected animals. Cortical and striatal brain tissue of IP injected A53T SNCA overexpressing mice was used to measure GCase and proteasome activity. (A) Tissue from the CBE treated mice showed a small but significant decrease of GCase activity (n=6/group). No differences were observed in proteasomal activity. The levels of a-synuclein in the midbrain, striatal and cortical brain tissue of CBE treated animals were analysed using the C20 antibody. Midbrain Triton X soluble α-synuclein of CBE injected animals showed a significant increase compared to control. Striatal and cortical Triton X soluble samples showed no difference between the groups (B) (n=5/group). Western blotting revealed the presence of Triton X soluble α-synuclein oligomers. Striatal α-synuclein oligomers were found to be significantly increased in the CBE-treated group (C) (n=3/group). No differences were observed in phosphorylated α-synuclein levels between the groups in midbrain, striatum and cortex areas (D) (n=5/group). SDS soluble monomeric and oligomeric α-synuclein levels were found to have no differences in midbrain, striatum and cortex of the saline or CBE treated animals (E) (n=4–6/group). Recombinant α-synuclein (r-asyn) and recombinant phosphorylated α-synuclein (p-asyn) were used as controls for the antibodies. Data represent mean values ±SEM. Differences were estimated using paired Student’s t-test. (A) P=0.019, (C) P=0.0368 (KO: SNCA knockout; CBE: CBE-treated; saline: normal saline 0.9% treated; s: Triton-x soluble fraction; i: Triton-x insoluble fraction). GAPDH and γ-tubulin were used as loading controls.
Brain ASYN levels and phospho ASYN levels were measured by western blotting. In animals treated with CBE a ∼23% increase of total ASYN levels in the midbrain of the Triton X soluble fraction was observed, compared with the control, saline treated animals (4.86 ±0.66 versus 3.96 ± 0.66 levels normalized to GAPDH). The levels of monomeric ASYN in the striatum and the cortex were not different between the two groups (Fig. 3B).
Western blot analysis of midbrain, striatum and cortex brain tissue using C20 and syn1 antibodies revealed the presence of abundant Triton X soluble ASYN oligomers in the CBE treated animals compared with the saline treated. Oligomers were found to be statistically increased in CBE treated mouse striata (2.51 ± 0.31 versus 1.52 ± 0.08 levels normalized to GAPDH, Fig. 3C;Supplementary Material, Fig. S2), whereas no differences were found in cortex and midbrain regions (data not shown). To investigate whether CBE treatment affected the levels of phosphorylated ASYN we analyzed both the Triton X and SDS fractions of the three isolated brain regions using the specific phospho ASYN antibody. No differences were noted between the two groups (Fig. 3D). ASYN levels in the SDS soluble fractions of all three brain regions examined were not different between the two animal groups (Fig. 3E). As expected, no oligomeric ASYN species were detected by C20 in the synuclein alpha gene (SNCA) knockout (KO) mice (Supplementary Material, Fig. S2A).
Inhibition of GCase activity increases brain exosome levels
Exosome associated ASYN has been shown to be released in the medium of neuronal cells in vitro (19,24). In order to investigate the effect of GCase inhibition on the numbers of exosomes released in vivo, brain tissue of the A53T mice was dissected and papain treated. The homogenized brain tissue was sequentially filtered and centrifuged and exosome pellets were collected and assessed as described above. The tissue from CBE treated mice showed a ∼97% increase of exosome number compared with the saline treated mice (CBE: 42.56 ±22.31 versus saline: 21.60 ±11.59 ng AChE/μg of protein, Fig. 4A). The exosome pellets were subsequently sonicated in order to analyze the exosome associated ASYN by sandwich ELISA and Western blotting. ASYN ELISA measurements showed a ∼22.9% increase in levels in the CBE treated brain compared with the saline treated control (CBE: 40.40 ± 1.61 versus saline: 32.18 ±1.62 ng ASYN/μg exosome protein, Fig. 4B). Western Blot analysis of CBE treated mice tissue showed an increase in ASYN oligomers which was statistically significant (CBE: 14.46 ±1.06 versus saline: 9.93± 0.89 levels normalized to flotillin-1). There were no differences in monomeric ASYN levels between the saline and CBE-treated groups (Fig. 4C). Further investigation of phosphorylated ASYN levels by western blotting showed no differences between the groups (data not shown).
Inhibition of GCase activity increases both brain exosome number and exosome associated α-synuclein. Whole brain tissue from the two groups was used to measure exosome number. The CBE-treated mouse brain tissue was found to have a statistically significant higher exosome number compared to the saline-treated group (A) (n=4/group). Exosome fractions were used to measure α-synuclein levels by ELISA. α-Synuclein levels were found to be significantly increased in the CBE treated group (B) (n=4/group). Exosome associated α-synuclein levels of the two different groups were measured by Western blotting. Triton-X soluble α-synuclein oligomers were significantly increased in the CBE treated group (C) (n=3–4/group). Data represent mean values ±SEM. Differences were estimated using Student’s t-test, ratio paired Student’s t-test and Student’s t-test. (A) P=0.001; (B) P=0.03; (C) P=0.0269. Flotillin was used as loading control.
Overexpression of mutant N370S GCase, but not GCase inhibition, increases ASYN secretion in vivo
We next sought to investigate if GCase inhibition affects the release of ASYN in vivo using microdialysis. To determine whether free, non-exosome associated ASYN secretion was affected by GCase inhibition, striatal microdialysates from Tg A53T mice treated for 8 weeks with CBE were collected and analysed with ELISA. No differences between the two groups were detected suggesting that the observed drop of GCase activity in the CBE treated group was not enough to elevate the levels of free ASYN in the interstitial fluid (ISF; Fig. 5A).
N370S GBA overexpression increases secreted α-synuclein levels in vivo. GCase inhibition in the CBE injected animals did not affect striatal α-synuclein secretion, as measured in striatal microdialysates by ELISA (A) (n=5–6/group). A53T mice overexpressing N370S GBA by AAV were found to have a profound increase of α-synuclein levels in the striatum compared to mice overexpressing GFP or WT GBA (B) (n=4–6/group). Data represent mean values ±SEM. Differences were estimated using Student’s t-test and one-way ANOVA followed by Tukey’s post hoc test. (B) P=0.0037.
Then, we examined whether overexpression of either WT or mutant GBA affects ASYN secretion in vivo. Two months old Tg A53T mice received a stereotactic injection of adeno-associated virus (AAVs) overexpressing GFP, WT GBA or N370S GBA in the striatum and protein levels were verified with western blotting and immunohistochemistry (Supplementary Material, Fig. S3A and B). Two months later ISF samples were collected by microdialysis and analyzed by ELISA. As shown in Figure 5B, the ASYN levels of the AAV-N370S GBA injected mice were 1.38-fold higher compared with the AAV-GFP group (N370S GBA: 0.87 ± 0.16 versus GFP: 0.36 ± 0.08 ng/ml ASYN). AAV-WT GBA-injected mice showed unaltered ASYN secretion compared with GFP. None of the expressed GBA forms appeared to saturate the ubiquitin-proteasome system as verified by examining the levels of polyubiquitin proteins (Supplementary Material, Fig. S3D).
Inhibition of GCase activity does not change the microtubule-associated proteins 1A/1B light chain 3AI, LC3II and P62 levels in vivo
We finally investigated whether the GCase inhibition and its effect on ASYN levels and secretion affected autophagy and proteasome pathways. We measured microtubule-associated proteins 1A/1B light chain 3A (LC3)I, LC3II and p62 levels in the Triton X soluble fraction of midbrain and striatal tissue by western blotting. There were no differences between the Triton X soluble fractions between the two groups in the P62 (midbrain: 0.52 ± 0.06 versus 0.52 ± 0.16, striatum: 0.32 ± 0.02 versus 0.52 ± 0.07 levels normalized to β-actin), LC3I (midbrain: 2.09 ± 0.57 versus 2.46 ± 0.73, striatum: 1.60 ± 0.32 versus 1.59 ± 0.34 levels normalized to β-actin) or LC3II (midbrain: 1.38 ± 0.44 versus saline: 1.29 ± 0.39, striatum: 1.16 ± 0.21 versus 0.96 ± 0.18 levels normalized to β-actin, Fig. 6A). Similarly, there were no differences in the SDS soluble fractions of the saline and CBE-injected groups in the p62 (midbrain: 1.20 ± 0.02 versus 1.24 ± 0.28, striatum: 1.06 ± 0.12 versus 1.19 ± 0.04 levels normalized to γ-tubulin), LC3I (midbrain: 1.56 ± 0.35 versus 1.74 ± 0.13, striatum: 0.92 ± 0.31 versus 1.23 ± 0.15 levels normalized to γ-tubulin) or LC3II (midbrain: 0.71 ± 0.09 versus 0.70 ± 0.10, striatum: 0.35 ± 0.18 versus 0.41 ± 0.05 levels normalized to γ-tubulin, Fig. 6B).
Inhibition of GCase activity does not affect LC3I, LC3II and P62 levels. Mibrain and striatal brain tissue of IP (saline or CBE)-injected A53T SNCA overexpressing mice were used to measure LC3I, LC3II and P62 levels by Western blotting. As shown in (A), Triton-x soluble samples showed no differences between the groups (n=3/group). Similarly, SDS-soluble samples showed no differences (n=3/group) (B). β-Actin (A) and γ-tubulin (B) were used as respective loading controls.
N370S mutant GBA significantly increases LC3 levels in vivo
The effect of N370S- or WT- GBA on autophagy indices in striatum was also measured using immunohistochemistry. LC3 fluorescence intensity was significantly increased in the N370S GBA overexpressing group compared with both the WT GBA and the GFP overexpressing group (N370S: 1.27 ± 0.13, WT GBA: 0.80 ± 0.15, GFP: 0.77 ± 0.07 ratio of the ipsilateral side puncta to the contralateral side puncta, Fig. 7A and B). p62 staining showed no differences between the two groups (Fig. 7C and D).
N370S GBA overexpression increases LC3 levels in vivo. Mice received striatal injections of one of the following AAVs; i. GFP, ii. N370S GBA, iii. WT GBA. Eight weeks later, mice were perfused and sections were stained for LC3, TUJ1, and p62. The intensity of fluorescence signal of LC3 and p62 in TUJ1-positive neurons was measured and quantified in ipsilateral (injected) versus contralateral (non-injected) side. LC3 fluorescence intensity was significantly increased in the N370S GBA overexpressing group compared to GFP and WT GBA overexpressing groups (A, B) (n=6/group) (P=0.017, P=0.023, respectively). There were no differences in the P62 staining fluorescence intensity between the groups (C, D). Data represent mean values ±SEM. Differences were estimated using one-way ANOVA followed by Tukey’s post hoc test.
Discussion
The dysregulation of extracellular ASYN as a contributor to PD initiation and progression has been the basis for the hypothesis of a prion-like mechanism of PD pathology (30). Recent studies have also used this hypothesis in order to evaluate ASYN and ASYN aggregates as a potential biomarker for PD (31). Indeed, pathogenic species of ASYN can aid toxic seed formation and potentially act as a prion protein when located extracellularly while pathogenic ASYN has also been found in exosomes (19,32). It has previously been reported that prion proteins can be associated with exosomes (33,34). Taken together, the above data point to the suggestion that ASYN can act as a prion-like protein through exosome-mediated propagation.
To investigate this intriguing hypothesis we set out to examine how GCase modulation could affect free ASYN secretion, exosome secretion and exosome associated ASYN, both in vitro and in vivo.
In line with our previous results (18), GCase inhibition was not sufficient to influence ASYN in neuronal cell cultures, neither intracellular nor secreted. Moreover, exosomes were also not affected by profound GCase inhibition. WT GBA overexpression in neuronal cell cultures, while leading to a reduction of exosome release, did not significantly reduce exosome-associated ASYN. In total, both WT and mutant GBA overexpression or GCase inhibition did not influence intracellular or secreted ASYN significantly. However, our model indicated that GCase activity might have an effect on exosome secretion. These results could be interpreted as demonstrating a lack of interaction between GCase and ASYN at the biochemical level. However, they could also be explained by the major handicap of this primary cellular model. In both experiments, the time frame employed was short; although GBA PD is known to have an earlier age of onset than sporadic PD, it is a neurodegenerative disorder that appears in middle aged or, more commonly, elderly patients. It was thus conceivable that the intrinsic limitations in the time frame of our cell cultures did not allow for a long time course that is needed for pathology to develop. Furthermore, none of our in vitro models had a pathogenic ASYN form that would resemble that found in a PD patient. The GCase inhibition model had a WT ASYN overexpression and the GBA overexpression model only had an endogenous WT ASYN expression. We reasoned that an in vivo setting which could provide a longer time frame for other gene modifiers to interact with a pathogenic ASYN background could be crucial to demonstrate effects on ASYN aggregation and secretion. This prompted us to move on to an in vivo model that would provide such a setting. As suggested by previous studies, lysosome dysfunction (20) or protein aggregation (35) can increase exosome release. In our model, GBA overexpression could have reversed either of the above pathways reducing exosome number.
Based on these observations and on the fact that certain non-significant trends of increase of ASYN were observed in the mutant GBA overexpressing cells as well as upon GCase inhibition, we moved on to evaluate the effects of chronic GCase inhibition on intracellular and extracellular ASYN species in vivo. Our aim was to replicate previously published data that showed an increase of ASYN oligomers in a GD mouse model (36) and investigate whether this GCase inhibition effect altered ASYN secretion. ASYN levels were measured in the midbrain, striatum and cortex of A53T-SNCA transgenic mice following treatment with saline or CBE. A modest, but significant, increase of ASYN monomer was found specifically in the midbrain and, more interestingly, ASYN oligomers were found to be significantly increased in the striatum, suggesting specific effects on the nigrostriatal axis. A number of studies have reported specific nigrostriatal pathway abnormalities in mice treated with CBE including synaptic dysfunction and increased inflammation (11,28). Such alterations may aid the increase of ASYN oligomeric species, as observed in the striatum. These results, combined with the preliminary in vitro results, allowed us to confidently test our hypothesis of an altered ASYN secretion in the presence of GCase activity inhibition. Interestingly, animals overexpressing A53T ASYN in the presence of reduced GCase activity had a two-fold increase of exosome number and exosome associated ASYN. More importantly, this increase was found to be due specifically to an increased presence of ASYN oligomers in the brain of the same group. This confirms previous findings of increased transmission of ASYN aggregates in GCase deficient cells (37) and indicates that this increased transmission could be exosome-mediated. In contrast to these profound effects on exosomal ASYN, chronic CBE treatment did not influence free ISF ASYN. To examine whether GCase inhibition affected ASYN by modifying the autophagy pathway, we measured brain P62 and LC3 levels by Western Blotting. Interestingly, there were no differences observed between the two groups. This suggests that chronic modest GCase loss of function affects ASYN species and exosomal secretion through an autophagy-independent mechanism.
Based on our previous work that examined the effect of a GBA mutant in the deregulation of extracellular ASYN and exosome associated ASYN in human iPSC dopaminergic neurons (24), we redirected our efforts toward examining a possible gain of function effect of the N370S GBA mutant on extracellular ASYN levels of PD model mice. While the CBE experiments simulated a chronic loss of function of GCase, such as is expected in GBA-PD and likely sporadic PD, a gain of function effect of GBA mutations has also been proposed as a mechanism to explain the link of GBA mutations to PD (8). To examine the possible role of a gain of function pathology model in extracellular ASYN, we used stereotactically injected AAVs overexpressing N370S GBA in the striatum of the same mouse model of ASYN overexpression. Through ISF microdialysis, mice overexpressing N370S GBA in the striatum were found to have a profound increase of secreted ASYN compared both with the GFP and the WT GBA overexpressing groups. These results indicate that gain of function effects of GBA mutations, likely through misprocessing of GCase, endoplasmic reticulum (ER) stress upregulation, and autophagic/lysosomal dysfunction, as demonstrated in our previous work (24), may be responsible for excessive release of ASYN. Interestingly, LC3 staining of brain sections from the two groups revealed a significant increase in the N370S GBA-overexpressing group as calculated by fluorescence intensity measurement. N370S GBA overexpression did not alter P62 levels. It has been suggested that misfolded GCase trapped in the ER leads to both an increase in the ubiquitin–proteasome system (UPS) and ER stress (24,38). In our study, viral expression of either WT or N370S GBA did not seem to affect proteasomal activity as judged by the levels of polyubiquitinated proteins detected in the striatum of injected mice. Further experiments are needed to investigate the effect of mutant GBA on ER and the release of ASYN in our model. It remains to be seen whether this profound increase in ASYN secretion caused by mutant GCase can lead to oligomeric ASYN formation, thus effectively contributing to PD pathology. Owing to experimental limitations, it was not possible to test whether mutant GCase overexpresion affects exosomal ASYN.
Thus, in a chronic in vivo setting, either gain of function effects of mutant GCase through autophagic/lysosomal dysfunction or loss of function of endogenous GCase lead to profound alterations in extracellular ASYN. In the former case, effects are observed on free ISF ASYN, whereas in the latter, on exosomal ASYN, and in particular oligomeric species (Fig. 8). Hence, combined effects of loss of function and gain of function may lead to alteration of the homeostasis of ASYN secretion in PD patients with GBA mutations.
Proposed mechanisms of gain-of-function and loss-of-function effects of GCase on α-synuclein. (A) A gain-of-toxic-function mechanism was replicated by intrastriatal injections of N370S mutant GBA in mice overexpressing α-synuclein. An erratic GCase could lead to increased autophagosome formation as well as ER stress. Taken together, these two pathways could alter α-synuclein secretion. (B) A loss-of-function mechanism was achieved by inhibiting GCase through intraperitoneal injections of CBE in mice overexpressing α-synuclein. A decreased GCase activity led to increased exosome release as well as α-synuclein accumulation in the form of soluble oligomers. These changes in the intracellular and exosome-associated α-synuclein levels were mediated through an autophagy independent mechanism.
Taken together, our results provide for the first time in vivo evidence that GCase has a role in extracellular ASYN homeostasis and that exosome associated pathogenic ASYN can be increased in the presence of GCase activity inhibition. Exosome-associated oligomeric ASYN, apart from a therapeutic target, may also prove to be a relevant biomarker for PD, and thus the study of its modifiers, such as GCase shown here, may prove to be clinically relevant.
Materials and methods
Animals
Eight-week-old male and female Tg mice expressing human mutant A53T ASYN under the control of the prion promoter (Jackson Laboratories, Bar Harbor, ME, USA) (39) were used in this study. All efforts were made to minimize animal suffering and to reduce the number of the animals used, according to the European Communities Council Directive (86/609/EEC) guidelines for the care and use of laboratory animals. All animal experiments were approved by the Institutional Animal Care and Use Committee of BRFAA (permit number 1153/12-05-2015).
Cortical neuronal cultures
Cultures of C57/Bl6 (embryonic day 16, E16) cortical neurons were prepared as previously described (40). Dissociated cells were plated onto poly-d-lysine-coated 6-well or 12-well dishes at a density of ∼1.5 × 106 cells/cm2. Cells were maintained in Neurobasal medium (Gibco, Rockville, MD, USA; Invitrogen, Carlsbad, CA, USA), with B27 supplement (Gibco; Invitrogen), l-glutamine (0.5 mM), and penicillin/streptomycin (1%). More than 98% of the cells cultured under these conditions represent postmitotic neurons (40). The time in culture of cells was calculated using days in vitro (DIV), with the day of harvesting designated as 0 DIV.
Adenoviral transduction of primary cortical neurons
Cultures of cortical neurons were transduced with WT, D409V, N370S GBA and WT SNCA AVs on 5 DIV. A multiplicity of infection (MOI) of 10–15 was used. CM were removed and replaced with fresh Neurobasal medium containing the AVs. On 6 DIV CM containing the AVs were removed and CM of 0–5 DIV cultures was reused. Cells were collected on 9 DIV.
Inhibition of GCase activity in primary cortical neurons using CBE
CBE was added in cultures of cortical neurons on 6 DIV in a concentration of 200 μM. Cells were collected 72 h later, on 9 DIV.
Western immunoblotting
Embryonic mouse cortical neurons were washed twice in cold phosphate-buffered saline (PBS) and then harvested in STET lysis buffer (50 mM Tris, pH 7.6, 150 mM NaCl, 1% Triton X-100, 2 mM ethylenediaminetetraacetic acid) with protease inhibitors. Lysates were centrifuged at 10 000g for 10 min at 4°C. Protein concentrations were determined using the Bradford method (Bio-Rad, Hercules, CA, USA). Twenty to fifty micrograms of lysates were mixed with 4× Laemmli buffer prior to running on 10–13% SDS–polyacrylamide gels. Following transfer to a nitrocellulose membrane, blots were probed with the following antibodies: polyclonal α-syn C-20 (1:1000; Santa Cruz Biotechnology), SNCA/syn-1/ASYN monoclonal (1:1000; BD Biosciences), anti-alpha-synuclein phospho S129 antibody monoclonal [EP1536Y] (1:500; Abcam), Anti-glucocerebrosidase (C-terminal) antibody polyclonal (1:500; Sigma-Aldrich), monoclonal b-actin (1:1000; Santa Cruz Biotechnology), monoclonal Anti-γ-Tubulin antibody (1:1000; Sigma-Aldrich), Anti-GAPDH antibody polyclonal (1:1000; Abcam), polyclonal LC3 (1:1000; MBL Life Sciences), polyclonal P62 (1:1000; MBL Life Sciences), polyclonal ubiquitin [Z0458] (1:1000; Dako), monoclonal human specific anti-GBA antibody (1:1000; gift of Pablo Sardi, Sanofi Genzyme; 12). Blots were probed with horseradish peroxidase-conjugated secondary antibodies and visualized with enhanced chemiluminescence substrate following exposure to Super RX film (FUJI FILM, Europe GmbH, Germany). After scanning the images with Adobe Photoshop CS6 (Adobe Systems, USA), Gel analyzer software 1.0 (Biosure, Greece) was used to quantify the intensity of the bands.
Statistical analysis
These data are shown as the mean ± standard error of the mean (SEM). Unless stated otherwise, statistical analysis was carried out with GraphPad Prism 6 (GraphPad Software, San Diego, CA, USA) using Student’s t-test or one-way analysis of variance (ANOVA) followed by Tukey’s or Fisher’s least significant difference (LSD) multiple comparisons post-test. Values of P < 0.05 were considered significant.
GCase activity assay
GCase enzyme activity was assessed using a standard fluorimetric assay (41). Briefly, primary cortical neurons or mouse brain tissue were lysed in 20–50 μl of GCase Activity Assay Buffer (50 mM citric assay, 176 mM K2HPO4, 10 mM sodium taurocholate, 0.01% Tween-20, final pH 5.9). Following incubation on ice for 30 min, samples were centrifuged for 10 min at 4°C. Neuron cell lysate (5 μl) was incubated with either 5 μl of assay buffer or 5 μl of 40 mM CBE (Sigma), for 15 min at room temperature. Subsequently, 25 μl of 5 mM 4-methylumbiferil β-glucopyranoside (4MU-β-Glc) substrates were added to samples, followed by 25 min incubation at 37°C. Samples were then cooled on ice and reactions were stopped with 465 μl Stop Buffer (1 M NaOH, 1 M glycine, final pH 10). RFU were measured at 450 nm in a Perkin-Elmer LS-55 luminescence Spectrometer (Perkin-Elmer, Norwalk, CT, USA). Each sample was measured in duplicate, the average non-specific activity was subtracted from each reading and the final result was normalized to the total protein concentration.
Adenoviral vector construction and virus production
Full-length human WT, D409V and N370S GBA cDNA were cloned into a modified version of the PENTR.GD entry vector and introduced into the pAd/PL-DEST Gateway vector (Invitrogen). Second-generation E1, E3, E2a-deleted recombinant human serotype 5 adenoviruses (rAd) were generated, as described previously (42). Viral vector stocks were amplified from plaque isolates in order to guarantee homogeneity of the production. Final vector stocks were purified and concentrated using double discontinuous and continuous cesium chloride (CsCl) gradients. Viral titers of purified vector stocks were determined by Adeno-X Rapid Titer kit (Clontech). The following titers were obtained, expressed as viral particles (vp)/μl: 2.31 × 108 for rAd-WT GBA, 1, 39 × 108 for rAd-D409V GBA, 5, 9 × 108 for rAd-N370S GBA and 1.51 × 108 for rAd-GFP.
Proteasomal activity assay
Neuronal tissue was lysed as described earlier in lysis buffer. The supernatants were collected for the measurement of the enzymatic chymotrypsin-like activity of the proteasome, essentially as described (43), based on cleavage of proteasomal substrate III (Calbiochem), using a Perkin Elmer fluorimeter LC-55 with excitation at 380 nm and emission at 438 nm. Three independent reactions were performed for each sample, in the presence or absence of the selective proteasomal inhibitor epoxomicin (1 µM), and the mean of the difference between these measurements was recorded.
Partial inhibition of GCase activity in A53T-SNCA mice using CBE
A partial reduction in CNS GCase activity in A53T-SNCA mice (animals expressing WT alleles of GBA1) was achieved by intraperitoneal (IP) administration of a covalent inhibitor of the enzyme, CBE (100 mg/kg, i.p., 3 times per week for 8 weeks). In order to achieve a partial reduction in CNS GCase activity a 3 times per week dosage was adopted (28).
ASYN ELISA assay
An in-house ELISA for the accurate quantification of ASYN concentration was developed by using two commercially available ASYN-specific antibodies: the monoclonal Syn-1 (BD Transductions) as the capture antibody and the polyclonal C-20 (Santa Cruz), as described elsewhere (44). Spark™ 10 M multimode microplate reader (Tecan) was used for reading the plates.
In vivo microdialysis
Guide cannulas were stereotaxically implanted in the striatum under isoflurane anesthesia (4–2.5%) as previously described (45). Animals were kept anesthetized during the whole procedure. Breathing was kept stable using an oxygen/air ratio of 0.5. Bore holes were made above the right striatum according to the mouse atlas of Paxinos and Franklin (coordinates, anterior–posterior = +0.5 mm, medial–lateral = −2.2 mm, dorsal–ventral = −2.4 mm). CMA 12 guide cannulas were inserted and fixed to the skull with stainless steel screws and dental cement. Mice were removed from the stereotaxic device and allowed to recover in individual cages; 72–96 h after surgery, mice were moved to the microdialysis cage. During microdialysis, mice were awake and had free access to food and water (CMA 120 System for Freely Moving Animals). CMA 12 custom made probes were manually inserted and connected to the CMA 402 syringe pump with a constant flow rate of 0.6 µl/min. Prior to sample collection, the probe was allowed to equilibrate for at least 2 h with the same flow rate. Samples were collected bihourly for 12–14 h using a CMA 170 refrigerated fraction collector and stored at −80°C until analyzed by ELISA.
Isolation of externalized membrane vesicles
The isolation of the externalized membrane vesicles was performed as described previously (46). Briefly, the CM was first centrifuged at 4000g for 10 min at 4°C to remove dead cells and debris, and the supernatant was further centrifuged at 100 000g for 2 h at 4°C. The supernatant (S100) was collected, and the pellet (P100) containing the externalized vesicles was reconstituted in 50 μl of PBS.
Preparation of exosome-depleted medium
The depletion of the medium from secreted exosomes released during the first four days of culture was performed as described previously (47). Briefly, Neurobasal medium containing 10% B27, penicillin/streptomycin, and l-glutamine was centrifuged at 100 000g for 2 h at 4°C. The supernatant was carefully removed and sterilized by filtering through a 0.2 μm filter (Whatman) and stored at 4°C until reused in primary neuronal cultures.
Exosome quantification
The amount of exosomes released was assessed based on the activity of AChE, an enzyme specific to these vesicles (48). AChE activity was determined as described by Savina et al. (49). Briefly, 5 μl of the exosome fraction was suspended in 20 μl of PBS. Five microliters of this PBS-diluted exosome fraction were then added to individual wells on a 96-well flat bottomed microplate. A 1.25 mM acetylthiocholine and 0.1 mM 5,5′-dithiobis (2-nitrobenzoic acid) were then added to exosome fractions in a final volume of 250 μl, and the change in absorbance at 405 nm was monitored. These data presented represent AChE enzymatic activity after 30 min of incubation, normalized to the amount of total protein of the exosome fraction measured through a Bradford protein assay.
Whole brain exosome isolation and purification
Exosomes were isolated from mouse brain tissue as previously described (50), with slight modifications. Frozen mouse brain was dissected and treated with 20 units/ml papain (Worthington) in Hibernate E solution (3 ml/hemi-brain; BrainBits, Springfield, IL, USA) for 15 min at 37°C. The brain tissue was gently homogenized in two volumes (6 ml/hemi-brain) of cold Hibernate E solution. The brain homogenate was sequentially filtered through a 40-m mesh filter (BD Biosciences) and a 0.2-m syringe filter (Thermo Scientific). Exosomes were isolated from the filtrate as described previously (47). Briefly, the filtrate was sequentially centrifuged at 300g for 10 min at 4°C, 2000g for 10 min at 4°C, and 10 000g for 30 min at 4°C to discard cells, membranes, and debris. The supernatant was centrifuged at 100 000g for 70 min at 4°C to pellet exosomes. The exosome pellet was resuspended in 60 ml of cold PBS (Invitrogen), and the exosome solution was centrifuged at 100 000g for 70 min at 4°C. The washed exosome pellet was resuspended in 2 ml of 0.95 M sucrose solution and inserted inside a sucrose step gradient column (six 2-ml steps starting from 2.0 M sucrose up to 0.25 M sucrose in 0.35 M increments, with the 0.95 M sucrose step containing the exosomes). The sucrose step gradient was centrifuged at 200 000g for 16 h at 4°C. One-ml fractions were collected from the top of the gradient, and fractions flanking the interphase separating two neighboring sucrose layers were pooled together for a total of seven fractions (a, top 1-ml fraction; b, 2-ml; c, 2-ml; d, 2-ml; e, 2-ml; f, 2-ml; and g, bottom 1-ml fraction). These fractions were diluted in cold PBS and centrifuged at 100 000g at 4°C for 70 min. Sucrose gradient fraction pellets were resuspended in 200 μl of cold PBS.
Stereotaxic injections
Male or female 2 months old transgenic mice (overexpressing A53T SNCA) were anesthetized with isoflurane (Abbott, B506) and stereotaxically injected with the viral vectors into the right substantia nigra (SN). Control animals received GFP AAV. Two microlitres of a working dilution of 6 ×109 vp/μl of each virus were administered. A single needle insertion (co-ordinates: +0.2 mm relative to Bregma, +2.0 mm from midline) into the right forebrain was used to target the inoculum to the dorsal neostriatum (+2.6 mm beneath the dura). Injections were performed using a 10 µl syringe (Hamilton, NV) at a rate of 0.1 µl per min (2.5 µl total per site) with the needle in place for >5 min at each target. Animals were monitored regularly following recovery from surgery.
Tissue staining and fluorescence measurement
Mice were perfused intracardially with PBS followed by paraformaldehyde 4%. Brains were dissected and post-fixed in 4% paraformaldehyde. They were then left in 30% sucrose overnight followed by 15% sucrose and stored in −80°C. Tissue sections (35 μm) were cryostat-cut at the level of the striatum and stored in −20°C. On the day of the experiment, sections were incubated (at room temperature) in PBS (3 ×5 min) and then in blocking solution (10% normal goat serum, 0.1% Triton in PBS for 1 h). Sections were then incubated in the primary antibody [LC3 1:1000, p62 1:1000 (MBL Life Sciences), neuron-specific class III β-tubulin (TUJ1) 1:500 (BioLegend, Inc.) in blocking solution] at 4°C for 48 h. Following, sections were washed in PBS (3 ×15 min) at room temperature and incubated for another 2 h at room temperature in secondary antibody (1:2500) and 4′,6-diamidino-2-phenylindole (1:2000, Sigma Aldrich) in blocking solution. Lastly, sections were washed in PBS (3×10 min) before transferred on glass slides. Images were obtained using the Leica TCS SP5 (Wetzlar, Germany). The intensity of fluorescence signal of LC3 and p62 was measured with Imaris software (v7.7.2, Bitplane AG) as previously described (51). A set of parameters was set in order to identify neuronal cells, using TUJ1 as a filter (masking channel). Then, the quantification of immunofluorescent puncta of p62 or LC3 on TUJ1 positive cells was performed automatically by the spot counter plugin of the software. The individual value of puncta for each subject is the ratio of ipsilateral side puncta to contralateral side puncta. Representative images were created using Fiji-ImageJ software.
Supplementary Material
Supplementary Material is available at HMG online.
Conflict of Interest statement. P.S. is an employee and stockholder of SANOFI company.
Funding
This work was supported by an ARISTEIA I (General Secretariat of Science and Technology, GSRT) grant to K.V.
References
Author notes
Georgia Nikolopoulou and Ivi Antoniadou contributed equally to this work.
