Parkinson disease-linked GBA mutation effects reversed by molecular chaperones in human cell and fly models

GBA gene mutations are the greatest cause of Parkinson disease (PD). GBA encodes the lysosomal enzyme glucocerebrosidase (GCase) but the mechanisms by which loss of GCase contributes to PD remain unclear. Inhibition of autophagy and the generation of endoplasmic reticulum (ER) stress are both implicated. Mutant GCase can unfold in the ER and be degraded via the unfolded protein response, activating ER stress and reducing lysosomal GCase. Small molecule chaperones that cross the blood brain barrier help mutant GCase refold and traffic correctly to lysosomes are putative treatments for PD. We treated fibroblast cells from PD patients with heterozygous GBA mutations and Drosophila expressing human wild-type, N370S and L444P GBA with the molecular chaperones ambroxol and isofagomine. Both chaperones increased GCase levels and activity, but also GBA mRNA, in control and mutant GBA fibroblasts. Expression of mutated GBA in Drosophila resulted in dopaminergic neuronal loss, a progressive locomotor defect, abnormal aggregates in the ER and increased levels of the ER stress reporter Xbp1-EGFP. Treatment with both chaperones lowered ER stress and prevented the loss of motor function, providing proof of principle that small molecule chaperones can reverse mutant GBA-mediated ER stress in vivo and might prove effective for treating PD.


Results
Ambroxol and isofagomine increase GCase mRNA and enzyme activity in fibroblasts. Human dermal fibroblasts express barely detectable levels of α -synuclein compared to neurons ( Supplementary Fig. 1). We therefore chose fibroblasts to investigate small molecule chaperones on mutant GBA, to be confident that any effects observed on ER stress were not due to the contribution of the well characterised relationship between perturbed α -synuclein homeostasis and mutant GCase reported in neurons and brain [14][15][16] . GCase activity was measured in control fibroblasts (wt/wt; CTRL), fibroblasts from PD patients with heterozygous GBA mutations (N370S/wt and L444P/wt) and fibroblasts from idiopathic (sporadic) PD patients (iPD). GCase activity was significantly decreased in N370S/wt and L444P/wt fibroblasts by 32% and 35%, respectively (P < 0.05 vs. CTRL; Fig. 1a). No loss of GCase activity was found in iPD fibroblasts. Similar to previous reports 32, 36 this decrease in GCase activity was concomitant with a significant decrease in both GBA mRNA and GCase protein levels in mutant GBA lines (P < 0.05 vs. CTRL; Fig. 1b-d). Under basal conditions, markers of UPR stress such as increased levels of CHOP mRNA (transcriptionally induced by both PERK and ATF6 arms of the UPR 24 ) and protein levels of the chaperone BiP were significantly increased in L444P/wt by 280% (P < 0.01 vs. CTRL) and 153% (P < 0.05 vs. CTRL), respectively (Fig. 1e,f). No markers of the UPR were significantly increased in the N370S/wt or iPD lines.
BiP levels were decreased in L444P/wt cell lines when treated with ambroxol (86 ± 12% of vehicle treated cells) or isofagomine (81 ± 15% of vehicle treated cells), although not significantly (Fig. 2e). Neither chaperone treatment had any effect on CHOP levels in L444P/wt cells (Fig. 2f). Surprisingly, 60 μ M ambroxol treatment considerably increased CHOP levels in CTRL cells. Titration of ambroxol indicated that 5 μ M ambroxol treatments of control, N370S/wt and iPD cells was enough to significantly increase GCase activity ( Supplementary Fig. 2), while not increasing CHOP levels in CTRL cells. However, this dose was not sufficient to increase GCase activity in L444P/wt cells. Generation of Drosophila expressing human WT, N370S or L44P GBA. To investigate further the impact of GBA mutations on ER stress and the potential protective effects of ambroxol and isofagomine in vivo we generated transgenic Drosophila to express human wt, N370S or L444P GCase. GBA mRNA levels and protein expression of human GCase were similar in the wt and mutant lines (Fig. 3a,b). Despite equivalent expression levels, the enzyme activity of GCase was significantly decreased (P < 0.01) by 82% in N370S flies and 75% in L444P flies compared to the wt line (Fig. 3c). Immunofluorescence revealed an abundant co-localisation of GCase with ER, however, unlike the wt, N370S and L444P caused abnormal aggregates and swellings within the ER (Fig. 3d).
Pan-neuronal expression of the GCase variants did not significantly perturb life-span ( Supplementary Fig. 3) or cause overt degeneration of the eye even after extensive aging ( Supplementary Fig. 4). However, while expression of wt GBA had no effect on motor (climbing) ability, the two mutant GBA lines showed a progressive climbing defect (Fig. 4a). A significant decrease in climbing was detected after 10 days in N370S flies which deteriorated further by 20 days. The L444P line exhibited a significant decrease in climbing after 20 days. Furthermore, the number of dopaminergic neurons in both N370S and L444P flies after 30 days was significantly decreased compared to wt GBA flies (Fig. 4b).
To investigate the possible induction of ER stress we used a ER stress reporter transgene, Xbp1-EGFP, where GFP is only expressed following an ER stress-induced splice event 37 . Using the developing eye tissue as a tractable system, as previously reported 34 expression of wt GBA resulted in increased Xbp1-EGFP levels compared to control (Fig. 5). However, expression of both N370S and L444P GBA induced significantly higher Xbp1-EGFP levels, relative to wt GBA flies (Fig. 5), indicating these forms caused an increased level of ER stress.
Ambroxol and isofagomine reduce ER stress and reverse locomotor deficits in Drosophila with mutant GBA. GBA expressing flies were next raised on food containing ambroxol (500 μ M) or isofagomine (50 μ M). Isofagomine has been reported to bind more potently to GCase than ambroxol 29,38,39 . Therefore to reduce the chance of toxic side effects we used a lower dose of isofagomine. Exposure to either chaperone during development significantly reduced the levels of Xbp1-EGFP compared to untreated flies (Fig. 6a,b). Notably, treatment with isofagomine was able to reduce the high Xbp1-EGFP levels induced by all GBA variants back to levels comparable to control flies. Thus, these treatments were able to reverse the elevated ER stress in vivo.
To determine whether the reduction of ER stress by these chaperones provided a functional benefit to the flies, we assessed the locomotor ability of GBA expressing flies fed with ambroxol or isofagomine for 10 days. Strikingly, both chaperones were able to completely prevent the age-dependent decline in climbing ability caused by the expression of mutant GBA (Fig. 6c). Isofagomine but not ambroxol significantly increased GCase activity by 170 ± 11% in wt flies (P < 0.05 versus vehicle treated wt flies) and 194 ± 7% in L444P flies (P < 0.05 versus vehicle treated L444P flies; Fig. 7a). The increased GCase activity following isofagomine treatment in wt and L444P flies was concomitant with an increase in protein levels ( Fig. 7b). Treatment of human cell lysates with endoglycosidase H has been used to identify GCase species trapped in the ER 22,23 . Treatment of fly lysates with endoglycosidase H yielded the same pattern in wt and mutant GBA lines and was unaffected by chaperone treatment (Fig. 7b).

Discussion
In this study we have used human cell lines and Drosophila models to investigate the biochemical consequences of the N370S and L444P GBA mutations that cause GD and significantly increase the risk for PD, and to evaluate the effects of two chaperones to reverse the abnormalities induced by these mutations. Our results confirm that the mutations cause reductions in GCase activity and elevated ER stress in the L444P fibroblasts and both the N370S and L444P flies. The flies exhibited reduced climbing activity and a loss of tyrosine hydroxylase positive neurons. Effects were partially reversed in flies by ambroxol and isofagomine.
Similar to other studies, we find that treatment of human dermal fibroblasts for several days with ambroxol or isofagomine increases GCase activity, protein expression and co-localisation with lysosomes 29,30,32,36,40 . Furthermore, this increase in GCase activity and protein expression occurs in control fibroblasts and fibroblasts from idiopathic PD to a similar extent as fibroblasts from PD patients with heterozygous GBA mutations or GD fibroblasts 32,36 . Both ambroxol and isofagomine treatment of fibroblasts increased GBA mRNA levels suggesting that these compounds elicit cellular responses to promote GCase expression in addition to their chaperoning ability. We opted to investigate treatment of cells with ambroxol or isofagomine over several days as we anticipate that this is the period of time required to re-fold GCase and for increased GCase trafficking to lysosomes to elicit measurable improvements in lysosomal function. Macrophages derived from inducible pluripotent stem cells (iPSC) from GD patients required five days of ambroxol or isofagomine treatment to reduce production of proinflammatory cytokines and improve the clearance of phagocytosed red blood cells 41 . Furthermore, ALP inhibition and perturbed α -synuclein Ambroxol treatment has been shown in fibroblasts to induce the coordinated lysosomal expression and regulation (CLEAR) network mediated via transcription factor EB 32 . Activation of the CLEAR network increases lysosomal biogenesis, clearance of damaged proteins and organelles by macroautophagy and maintains mitochondrial homeostasis [43][44][45][46] . In addition to increased GCase activity and expression levels, ambroxol treatment has been reported to increase the expression/activity of lysosomal cathepsins, the GCase transporter LIMP2, and saposin C, the endogenous activator of lysosomal GCase activity 32,36 . Recently, ambroxol has been reported to affect calcium homeostasis in lung type II pneumocytes 47 , which if applicable to other cell types, would modulate many cellular pathways.
Further studies on different ambroxol concentrations and duration of treatment are needed to dissect the physiological actions of this chaperone. The beneficial effects of ambroxol on GCase activity, lysosomal localisation, and inflammatory mediators have typically used concentrations of 50-100 μ M 30,32,36,41 . The K m and K i of GCase for ambroxol at pH 7.0 has been reported to be about 5-10 μ M 30,38 , while the current therapeutic doses of ambroxol (90-120 mg/day) yield a plasma drug concentration of 0.6-1.2 μ M 47 . We found that 5 μ M ambroxol significantly increased GCase activity in control, N370S/wt and iPD fibroblasts, and was not accompanied by increased CHOP mRNA levels seen in control fibroblasts at the higher dose. Unlike 100 μ M ambroxol, 1 μ M does not mobilise calcium 47 . However, higher doses may be required for the L444P GCase mutation as 5 μ M was ineffective at increasing GCase activity, and has previously been shown to require higher doses (50 μ M) to increase GCase activity by 30% and reduce ER trapping in L444P/L444P GD fibroblasts 31 . It is apparent from this and other studies that further characterization of the effects of ambroxol and isofagomine on ER stress makers in human models for the common GBA mutations is required.
Transgenic expression of human GBA in Drosophila generated substantial GCase activity, compared to background endogenous levels. However, the expression of N370S and L444P GCase resulted in significantly less GCase activity, increased ER stress, progressive loss of locomotor ability and the loss of dopaminergic neurons.   Importantly, since Drosophila do not express endogenous α -synuclein, these findings show that mutant GCase mediated-ER stress is capable of causing loss of dopaminergic neurons in the absence of α -synuclein toxicity. Human and mouse models of pathogenic GCase have previously shown phenotypes relevant to PD, but because of the concomitant inhibition of the ALP and subsequent accumulation of α -synuclein, it was not possible to differentiate between the deleterious contribution of ER stress and autophagy impairment [14][15][16]42 . ER stress has been observed in other Drosophila models expressing human pathogenic mutant GCase, such as N370S, L444P and R120W, RecNcil 34,35 . During preparation of this manuscript ambroxol has also been shown to partially reverse the ER stress and loss of dopaminergic neurons in Drosophila expressing N370S and L444P GBA 48 . We have extended these observations by showing that isofagomine, as well as ambroxol 34 , can reverse the ER stress, and importantly, that both these chaperones can rescue locomotor deficits. Recently, deletion of the GBA homolog dGBA1b in Drosophila has been shown to affect locomotor ability and decrease lifespan 49 . Ubiquitin deposits and accumulation of Ref(2)P, the Drosophila homolog of p62/SQSTM1, also occurred, suggestive of impaired autophagy/quality control mechanisms. However, no change in dopaminergic neuronal number was observed, although evidence of other neurodegeneration was detected 49 .
The mechanism by which ambroxol and isofagomine exert these effects is not clear. We did observe a significant increase in GCase activity in the wt and L444P fly models following isofagomine treatment, while ambroxol had no significant effect on either wt or mutant GCase. It appears increasingly likely that these compounds may have multiple effects that may include the chaperoning of mutant GCase trapped in the ER, reversal of ER stress and transcriptional activation. The latter property may be most clearly reflected in our human-derived cell model where increases in GBA mRNA, protein levels and activity are seen even in the presence of only wt alleles. The N370S GBA mutant fly model response to treatment without a significant increase in GCase activity may in turn reflect effects on the reversal of ER stress and as yet unidentified additional mechanisms. In this respect it is notable that isofagomine was found to be protective in a neuronopathic GD mouse model by preventing the dysregulation of several mRNA and miRNA species in the brain involved in several pathways including neuroinflammation, mitochondrial function and axon guidance 50 . This protection occurred despite isofagomine not reducing the accumulation of substrate.
Regardless of the mechanism, reversal of ER stress and full restoration of locomotor function in our fly models by small molecule chaperones is an important proof of principal. Data suggest that the wild-type GCase becomes trapped in the ER and trafficking to lysosomes becomes impaired when cellular α -synuclein levels are elevated 8,9,14 . It is envisaged that if wild-type/mutant GCase can be refolded in the ER of humans, reducing ER stress and improving the trafficking of GCase to lysosomes, this will have important benefits not only for neuronopathic forms of GD (Type 2 and Type 3), for which no therapy is currently licensed, and PD patients with GBA mutations, but also sporadic forms of PD. Increased lysosomal GCase will help to reduce the accumulation of substrate in type 2 and type 3 GD, while in PD this should improve autophagy, thus reducing the formation of pathogenic α -synuclein species with a beneficial effect on disease pathogenesis 51 .

Materials and Methods
Fibroblast tissue culture lines. Human dermal fibroblasts were generated from skin biopsies taken from five controls (wt/wt), five PD patients with N370S/wt GBA mutations, five PD patients with L444P/wt GBA mutations and five patients with idiopathic PD. Controls were age-matched to PD patients and mutations were confirmed by Sanger sequencing. Fibroblasts were cultured in Dulbecco's modified Eagle media (4.5 g/L glucose) supplemented with 10% (v/v) fetal calf serum, 1 mM sodium pyruvate and penicillin-streptomycin. Fibroblasts were not used after passage 12. All biopsies were obtained following informed patient consent following Hampstead Research Ethical committee approval (10/H0720/21/21). The methods were carried out in accordance with the relevant guidelines in the above-mentioned ethics approval.  10 cm plates. Fibroblasts were treated with vehicle (dimethyl sulfoxide) or respective chaperone on days 0, 2, and 4. Fibroblasts were harvested on day 6 by trypsinisation and washed once in phosphate buffered saline (PBS) prior to freezing.
Endoglycosidase H treatment of fly lysates. Fly lysates were prepared as for immunoblotting and 20 μ g protein treated with 1000 units of endoglycosidase H (New England Biolabs) according to manufacturer's instructions for 2 hours at 37 °C. Reaction was stopped by addition of gel loading dye and lysates immunoblotted for human GCase as above.
Quantitative real time PCR. For cell samples, RNA was extracted from fibroblasts using RNeasy mini kit (Qiagen) and converted to cDNA (QuantiTect, Qiagen or nanoscript, Primer Design) and relative mRNA levels for GBA and β-actin were measured using TaqMan assays (GBA, Hs00986836_g1; β-actin, Hs99999903_m1; Applied Biosystems) using a STEP One PCR machine (Applied Biosystems). CHOP mRNA levels were measured using QuantiTect SYBRgreen (Qiagen) using the following primers: CHOP, ACC AAG GGA GAA CCA GGA AAC G, TCA CCA TTC GGT CAA TCA GAG C; β-actin, TCT ACA ATG AGC TGC GTG TG, GGT GAG GAT CTT CAT GAG GT. Relative GBA and CHOP mRNA levels were calculated using the 2 −ΔΔCT method and normalised against β-actin.
For Drosophila samples, RNA was extracted using a RNeasy RNA purification kit (Qiagen), and cDNA was synthesized using a Protoscript II first-strand cDNA synthesis kit (New England BioLabs) according to the manufacturers' instructions. Relative GBA mRNA levels were calculated using the 2 −ΔΔCT method and normalized to the ribosomal reference gene, RNA18S. The primers used for Drosophila qRT-PCR were as follows: GBA: TGGGCAGTGACAGCTGAA, CTGGAAGGGGTATCCACTCA; RNA18S: TCTAGCAATATGAGATTGAGCAATAAG, AATACACGTTGATACTTTCATTGTAGC.
Salivary glands and eye imaginal discs were dissected from third instar larvae and fixed with 4% PFA/ PBS solution for 20 min at room temperature, then permeabilised using 0.1% Triton X-100/PBS solution for 3 × 10 min. Samples were blocked for 1 h in 1% bovine serum albumin/PBS solution before addition of the primary antibody for 2 h. Antibodies used were: mouse anti-hGBA (clone 2E2, 1:100, Merck Millipore). AlexaFluor 488 and 568 secondary antibodies from Invitrogen (1:200) were added in 1% PBS for 2 h, then cell nuclei were stained with 0.1 μ g/ml Hoechst for 5 min. UAS-KDEL-EGFP and UAS-Xbp1-EGFP were imaged using the native GFP. Fluorescence quantification was performed using ImageJ using the analyze particles function and measuring the area.
Drosophila brains were dissected from 30-day old flies and immunostained with anti-tyrosine hydroxylase (Immunostar Inc. 22491, 1:100) as described previously 53 . Brains were imaged with an Olympus FV1000 confocal with SIM-scanner on a BX61 upright microscope using a 63X oil 0.95 NA objective. Tyrosine hydroxylase-positive neurons were counted under blinded conditions. independent lines were initially isolated and assessed for consistent effects before selecting single lines for further analysis. Climbing assays were performed as previously described 54 . For all the experiments only males were used.
Chaperone treatment of flies. For chaperone feeding, flies of the specified genotypes were maintained on standard media containing ambroxol hydrochloride (Sigma-Aldrich A9797) or isofagomine (D-tartrate) (Cayman Chemical 16137), dissolved in water, to final concentrations of 500 μ M and 50 μ M. For eye imaginal discs analysis, third instar larvae were raised in food containing drug (4 days feeding during larval stages), and imaginal discs were analyzed by immunohistochemistry and quantified using Image J. At least 10 discs were analyzed per genotype. For climbing assays at least 50 adult males collected from crosses raised on standard media were aged on food containing drug for 10 days before analyzing behavior.
Life span. At least 100 recently emerged adult males were collected under light anesthesia and housed at a density of 25 males per vial. Flies were passaged every other day, and the number of dead flies was recorded.
Light microscopy imaging. Light microscopy imaging was assessed using a Nikon motorized SMZ stereo zoom microscope fitted with 1x Apo lens. Extended focus images were then generated using Nikon Elements software, using the same settings for all the genotypes. Flies were anaesthetised during the process. All animals of a given genotype displayed essentially identical phenotypes and randomly selected representative images are shown.
Statistical analysis. Calculations were performed using GraphPad Prism 6.0. Adult climbing analysis is not normally distributed so the data were analyzed using Kruskal-Wallis non-parametric test with Dunn's correction for multiple comparisons. Protein and mRNA expression levels in fibroblasts and Drosophila, and dopaminergic neurons count significance was determined by one-way analysis of variance (ANOVA) with the Bonferroni's multiple comparison test. For life span experiments Log-rank (Mantel-Cox) test was used for the analysis. Significance levels are indicated in figure legends. Unless specifically indicated, no significant difference was found between a sample and any of the other samples.