GBA-associated PD: chances and obstacles for targeted treatment strategies

Given the clear role of GBA in the pathogenesis of Parkinson’s disease (PD) and its impact on phenotypical characteristics, this review provides an overview of the current knowledge of GBA-associated PD with a special focus on clinical trajectories and the underlying pathological mechanisms. Importantly, differences and characteristics based on mutation severity are recognized, and current as well as potential future treatment options are discussed. These findings will inform future strategies for patient stratification and cohort enrichment as well as suitable outcome measures when designing clinical trials.


Introduction
Over the last decades, research in genetically defined forms of Parkinson's disease (PD) led to the identification of specific pathways underlying the pathophysiology of the disease. Next to defects in vesicular trafficking, mitochondrial and importantly lysosomal dysfunction represent the most relevant pathways (Jankovic and Tan 2020). Studying these early events provide entry points to develop novel therapeutic targets for stratified patient groups as an important step towards precision neurology. The present article exemplifies such strategies focusing on PD patients with different variants in the glucocerebrosidase (GBA) gene (PD GBA ). Also, obstacles of translational research into patient cohorts and study designs for clinical trials are discussed.

GBA variants are the most important genetic risk factor for PD
Biallelic variants in the GBA gene cause Gaucher's disease (GD), the most common lysosomal storage disorder with tissue accumulation of glucosylceramides due to deficiency of the lysosomal enzyme glucocerebrosidase (GCase). Interestingly, about 25% of GD patients report a first-or second-degree relative to present with PD (Goker-Alpan et al. 2004;Halperin et al. 2006). This important clinical observation was the hint to the fact that heterozygous variants in the GBA gene are associated with PD. Subsequently, a large multi-centre study across four continents analysed 5691 PD patients of different ethnic origin compared to 4898 controls and confirmed that with an overall odds ratio (OR) of 5.43, heterozygous variants in the GBA gene represent the most important genetic risk factor for PD (Sidransky et al. 2009). This has now been confirmed across different ethnic populations with Caucasian, Asian (Japanese, Chinese, Taiwanese), Hispanic, and African ancestry (den Heijer et al. 2020;Neumann et al. 2009;Lesage et al. 2011;Chen et al. 2014;Mahungu et al. 2020).
To date, more than 100 different variants have been associated with the risk of PD. However, the pathogenicity of different variants varies largely (Table 1). Whereas variants classified as severe variants (e.g. p.L444P) show an odds ratio of 10-15 for developing PD and mild variants (e.g. p.N370S) have an odd ratio of ≤ 5 for PD, some variants that are non-pathogenic for GD have been proven to increase the risk for PD e.g. p.E326K and p.T369M Zhang et al. 2018;Straniero et al. 2020). These variants show the lowest odds ratios and are thus classified as risk variants. Consequently, GBA-subgroup classification for PD patients is often based on variant severity according to established genotype risks reported for PD (PD GBA_severe , PD GBA_mild , PD GBA_risk ). Interestingly, we see a huge variability of variant distribution among different ethnicities. About 20% of PD patients with Ashkenazi Jewish ancestry carry a GBA variant, with the large majority harbouring the mild p.N370S (> 70%), whereas the severe p.L444P variant is identified in about 5%. Together, the two variants account for about 80% of variants in Ashkenazi Jewish PD patients. In non-Ashkenazi Jewish PD patients, p.L444P is detected in about 30-40% of patients and p.N370S in about 20%, together accounting for 50-60% of variants (Sidransky et al. 2009), indicating that about 40% of variants could be missed if focusing solely on p.N370S and p.L444P. These findings highlight the need for full-gene sequencing and stratification according to variant severity. Moreover, penetrance and disease risk in PD GBA are age-dependent (Anheim et al. 2012;Straniero et al. 2020) and further modified by the composite PD-associated polygenetic risk score (PRS) and single-nucleotide polymorphisms in SNCA, CSTB and TMEM175, the two latter genes encoding proteins associated with lysosomal homeostasis and protein clearance (Blauwendraat et al. 2020b).

PD GBA : severe clinical trajectories with early cognitive decline
Detailed investigation of the phenotypical spectrum, longitudinal trajectories, and rate of progression of motor and non-motor symptoms is of utmost importance to estimate effect sizes and design clinical trials for disease-modifying therapies (duration, sample sizes, progression rates, expected spectrum of symptoms, etc.).
In general, PD GBA show an earlier age at onset compared to PD patients without GBA variants (PD GBA_wildtype ) with a median onset in the early 50s (Sidransky et al. 2009;Blauwendraat et al. 2019). Of note, this effect is not only attributable to GBA variants per se, but is driven by GBA variant severity and variant burden with severe variants as well as homozygous and compound heterozygous variants predisposing to the youngest age at onset (Malek et al. 2018;Thaler et al. 2017). Moreover, age at onset is further reduced in PD GBA by non-coding variants in SNCA and TMEM175 (Blauwendraat et al. 2020b). Although younger, PD GBA present with a higher prevalence of cognitive impairment and more frequently suffer from additional non-motor symptoms including neuropsychiatric disturbances (depression, anxiety, and hallucination), autonomic dysfunction and sleep disturbances such as REM-sleep-behaviour disorder (RBD) when compared to PD GBA_wildtype , (Brockmann et al. 2011;Barrett et al. 2014). These findings have been replicated consistently over the following years in other PD cohorts worldwide, the latest large clinical genome-wide association study in 4093 PD patients . Importantly, GBA variants that are classified as severe (PD GBA_severe ) have been associated with a more aggressive clinical phenotype suggesting a relevant effect depending on GBA variant severity (Cilia et al. 2016;Thaler et al. 2018;Petrucci et al. 2020;Lerche et al. 2021a).
Data from longitudinally investigated cohorts of PD GBA confirm findings from cross-sectional evaluations and revealed that PD GBA , although younger in age and age at onset, present with an accelerated disease progression in terms of motor impairment and cognitive decline as compared to PD GBA_wildtype . Moreover, survival rates are shorter when compared to PD GBA_wildtype (Brockmann et al. 2015b;Cilia et al. 2016). In a British cohort, after 10 years of disease duration, 46% of PD GBA remained dementia-free in    (Stoker et al. 2020). Similar results were reported in a large longitudinal cohort of Italian patients with a clearly more aggressive pattern depending on GBA variant severity (Cilia et al. 2016). Interestingly, a recent study reports that PD GBA who are treated with deep brain stimulation (DBS) in the subthalamic nucleus (STN) showed an even more rapid cognitive decline compared to PD GBA without DBS as well as PD GBA_wildtype with and without DBS. This finding suggests that the additive effect of GBA variants and STN-DBS negatively impact cognition and that presurgical genetic screening should be considered (Pal et al. 2022). Further studies are needed for replication and to evaluate the underlying pathophysiological mechanisms. The typical motor manifestation of PD is preceded by a prodromal phase that is characterized by a variety non-motor and early motor signs (Berg et al. 2015). Non-motor symptoms include among others hyposmia, autonomic dysfunction, and neuropsychiatric symptoms, whereas reduced arm swing and bradykinesia indicate early motor signs. However, type, prevalence, time of occurrence, and rate of progression of these prodromal symptoms are variable between patients. Given the findings from the manifest disease phase in PD GBA with the pronounced non-motor profile and more rapid disease progression, we retrospectively assessed patient's perception of their individual prodromal phase before PD diagnosis. Comparing PD GBA and PD GBA_wildtype , we could show that: (i) PD GBA demonstrate a higher prevalence of prodromal symptoms and a shorter prodromal phase with almost parallel beginning of non-motor and early motor signs before PD diagnosis. Contrary, PD GBA_wildtype show a long prodromal interval starting with non-motor symptoms long before early motor signs manifested. (ii) PD GBA with severe variants reported the highest total amount of prodromal signs. These findings suggest that complexity of symptoms known from the manifest disease might be present already in the prodromal phase (Zimmermann et al. 2018). Similarly, prospective studies found that prodromal GBA variant carriers present with more pronounced deterioration of motor and non-motor symptoms, specifically cognitive decline and hyposmia when compared to healthy controls without GBA variant (Avenali et al. 2019;Beavan et al. 2015;Mullin et al. 2019). Another study in patients with REM-sleep behaviour disorder (RBD) reports that GBA variants are associated with accelerated phenoconversion to PD and/or dementia in this specific cohort (Honeycutt et al. 2019).

GBA variants are an important genetic risk factor for Dementia with Lewy Bodies (DLB)
The important finding that PD GBA shows pronounced and early development of dementia prompted the community to perform a large multicenter analysis across 11 centres evaluating GBA variants in 721 cases with DLB, which represents a clinico-pathological continuum to PD. With an even higher OR than seen in PD, GBA variants are also strongly associated with DLB (8.28). Similar to PD, GBA variants predispose to an earlier age at onset, more pronounced disease Variant position based on NM_001005742. Suggested PD severity mainly based on reported GD severity. Additionally, frameshift and nonsense variants were categorized as "severe". Variants described as pathogenic in GD, but with unknown GD severity were categorized as "mild/ severe". Variants described as not pathogenic in GD, but have been detected as risk factors for PD were categorized as "risk". Missense and splice site variants not described in GD and of unknown significance for PD were categorized as "VUS" = variants of unknown significance severity/progression and rather "pure" form of DLB without concomitant Alzheimer's profile as defined by CSF p-tau/ Aβ1-42 ratio (Nalls et al. 2013;van der Lee et al. 2021). This study further supports GBA variants as a significant genetic risk factor for synucleinopathies and confirmed the overall impression that GBA-associated Parkinsonism predisposes to an increased incidence of dementia (Fig. 1).

Pathomechanisms in PD GBA
Experimental evidence from cell models suggests that GBA variants result in disrupted protein folding of GCase in the endoplasmic reticulum (ER), impaired trafficking of GCase from the ER to Golgi and ultimately in lower lysosomal GCase enzyme activity. This in turn causes a build-up of glucosylceramides (GlcCer) and glucosylsphingosines (GlcSph) (Beutler 1992) and impairs lysosomal function and thereby the degradation of α-synuclein (Mazzulli et al. 2011).

GBA variants predispose to accelerated α-synuclein aggregation and Lewy-body pathology
Post-mortem studies show enhanced aggregation and propagation of α-synuclein not only in the substantia nigra and putamen but also wide-spread neocortical Lewy-body pathology in brain tissue of PD GBA and DLB GBA (Neumann et al. 2009;Gundner et al. 2019).

Fig. 1
Pathogenic mechanisms underlying PD GBA . Loss of lysosomal GCase activity results in impaired autophagy affecting the degradation of both physiological (red dot) and misfolded α-synuclein (red dot complex) resulting in the aggregation of α-synuclein (red strains). GBA variants also cause the GCase protein to misfold in the ER (brown enzyme) with impaired trafficking to the lysosome which also affects α-synuclein degradation. Accumulation of GCase substrates (GlcCer and GlcSph, yellow) also causes α-synuclein misfolding and aggregation, as may changes in the lipid homeostasis (both sphingolipids and phospholipids) of cellular membranes (yellow) due to decreased lysosomal function. In PD GBA_wildtype , the trafficking of wild-type GCase (green enzyme) can be inhibited by increased levels of α-synuclein (red dot complex) and α-synuclein fibrils (red strains), and contribute GCase deficiency (brown enzyme) irrespective of a GBA mutation. This figure was adapted from "Brainstem with Callout" and "Structural Overview of an Animal Cell", by BioRender. com (2022). Retrieved from https:// app. biore nder. com/ biore ndertempl ates The field of PET imaging markers to assess the cerebral load of α-synuclein in-vivo is difficult. However, this month [03(2022)] first positive results were reported at the AD/PD Conference for a new PET tracer developed by AC Immune to distinguish multiple system atrophy (MSA) from healthy controls and patients with other forms of α-synuclein (PD, DLB). Therefore, research in PD has focused on CSF. Yet, it is unclear whether CSF profiles of α-synuclein species reflect brain pathology. Cross-sectional and longitudinal analyses in PD GBA_wildtype and PD GBA demonstrated decreased CSF levels of total α-synuclein compared to healthy controls with the highest decrease in PD GBA patients carrying severe variants (Malek et al. 2014;Mollenhauer et al. 2019;Lerche et al. 2020Lerche et al. , 2021a. Correspondingly, the same pattern was also reported in patients with DLB GBA (Lerche et al. 2019a). However, a substantial inter-individual variability and overlap with healthy controls is seen, so that CSF levels of total α-synuclein are not ideal. Recently, the ultrasensitive assays real-time quaking-induced conversion (RT-QuIC) and protein misfolding cyclic amplification (PMCA) have been successfully implemented. These assays exploit the seeding capacities of prion or prion-like proteins as an amplification strategy to reveal minute amounts of disease-specific protein aggregates in CSF (Fairfoul et al. 2016;Shahnawaz et al. 2017). Both methods are highly sensitive (88-96%) and specific (83-98%) for α-synuclein aggregates and Lewy-body pathology in PD and DLB as assessed in matched CSF/brain samples compared to healthy controls and other forms of dementia and parkinsonism (Rossi et al. 2020;Kang et al. 2019). However, histopathological findings in some genetic forms of PD are remarkably variable. While PD GBA show extensive Lewy-body pathology, most PD patients with bi-allelic mutations in the recessive gene PRKN (PD recessive_bi-allelic ) show nigral degeneration without Lewybodies (Schneider and Alcalay 2017). Also, histopathology in PD patients with LRRK2 mutations (PD LRRK2 ) is variable, including typical Lewy-body pathology, misfolded tau deposition, or nigral degeneration without Lewy-body (Zimprich et al. 2004;Heckman et al. 2016;Kalia et al. 2015). This prompted us to evaluate CSF α-synuclein seeding capacities with RT-QuIC in two large cohorts of PD and DLB patients enriched for genetic forms. Remarkably, PD GBA (93%) and DLB GBA (100%), especially those carrying severe variants, showed the highest percentage of positive α-synuclein seeding and the most pronounced α-synuclein seeding kinetics. In contrast, PD recessive_bi-allelic did not show CSF α-synuclein seeding at all, whereas those carrying heterozygous mutations in these recessive genes showed less α-synuclein seeding than PD wildtype (91%) with a reduced positivity rate of 59%. Also, PD LRRK2 showed a reduced rate of α-synuclein seeding (78%) compared to PD wildtype (Brockmann et al. 2021). The heterogeneity in α-synuclein seeding activity among the different genetic forms mirrors histopathological findings in these cases and highlight the value of α-synuclein seeding activity as an in-vivo marker of Lewy-body pathology.
The accelerated cognitive decline PD GBA makes this subgroup of PD a good model to study CSF profiles that are associated with cognitive impairment. In general, limbic and/or cortical Lewy-body pathology is hypothesized to be the main substrate forcing driving cognitive decline in PD (Aarsland et al. 2005). In more recent years, it became clear that a considerable proportion of PD patients who developed dementia in their disease course show concomitant amyloidbeta and tau pathology at autopsy in addition to the typical Lewy-body pathology (Halliday et al. 2008;Compta et al. 2011). Correspondingly, reduced CSF levels of Amyloidbeta 1-42 (Aβ 1-42 ) and/or elevated CSF levels of total-Tau (t-Tau) and phospho-Tau (p-Tau) have been reported to be associated with cognitive impairment in PD (Brockmann et al. 2015a(Brockmann et al. , 2017Lerche et al. 2019b;Kang et al. 2016). However, this seems not to be the case in PD GBA as CSF levels of Aβ 1-42 , t-Tau, and p-Tau are similar to those seen in healthy control individuals. In light of the CSF profiles of reduced total levels of α-synuclein and the prominent α-synuclein seeding activity, the pronounced cognitive decline in PD GBA is driven by α-synuclein aggregation and cortical Lewy-body pathology.
Taken together, these histopathological and CSF characteristics of predominant and accelerated α-synuclein-driven Lewy-body pathology make PD GBA and DLB GBA a role model to study pathways leading to α-synuclein aggregation and highlight these patient cohorts as prime candidates for clinical trials targeting α-synuclein.

GCase deficiency and α-synuclein aggregation
Heterozygous variants in the GBA gene are associated with a reduction of GCase protein levels and GCase enzyme activity in cell and animal models as well as in a variety of patient-derived biomaterials (Lerche et al. 2021a;Alcalay et al. 2015Alcalay et al. , 2020Schondorf et al. 2014;Paciotti et al. 2019). Again, the degree of reduction is dependent from variant severity. Interestingly, GCase activity is also reduced in PD GBA_wildtype , albeit to a lesser degree (Parnetti et al. 2017).
There is reasonable evidence from different cell models including induced pluripotent stem (IPS) cell-derived human dopaminergic midbrain neurons and human midbrain organoids that deficiency of the GCase enzyme is paralleled by increased levels of intracellular α-synuclein, specifically α-synuclein species susceptible to aggregation such as high molecular weight and decreased tetramer/monomer ratio (Schondorf et al. 2014;Kim et al. 2018;Magalhaes et al. 2016;Aflaki et al. 2016;Mazzulli et al. 2016;Jo et al. 2021). Correspondingly, post-mortem studies in PD GBA and DLB GBA and to a lesser degree also in PD GBA_wildtype and DLB GBA_wildtype revealed that reduced GCase protein levels and reduced GCase enzyme activity are accompanied by increased levels of intracellular α-synuclein. Notably, these findings are not restricted to the substantia nigra and putamen but also identified in cortical regions (Murphy et al. 2014;Gegg et al. 2012;Moors et al. 2019;Gundner et al. 2019).
More specifically, it is suggested that lysosomal GCase and α-synuclein are linked in a bidirectional pathogenic loop as shown in cell cultures and IPS cell-derived dopaminergic midbrain neurons: (I) functional loss of GCase activity compromises lysosomal degradation of α-synuclein and promotes its aggregation. (II) α-Synuclein itself inhibits the activity of GCase (Mazzulli et al. 2011;Schondorf et al. 2014). Consequently, PD GBA fulfill both conditions of this bidirectional loop in parallel leading to a self-reinforcing mechanism. Thereby, α-synuclein aggregation and propagation might be accelerated which possibly explains the widespread neocortical Lewy-body pathology and rapid clinical progression.
However, this bidirectional loop between GCase and α-synuclein might be oversimplified, since we have clear evidence for a more complex impairment of the autophagy-lysosomal pathway including disrupted macroautophagy with reduced fusion of autophagosomes with lysosomes and decreased expression/activity of other proteolytic lysosomal enzymes such as cathepsin B and D (Aflaki et al. 2020;Blauwendraat et al. 2020a;Lerche et al. 2021b).

Disturbance in sphingolipid homeostasis and α-synuclein aggregation
Adding to the complexity of the underlying pathophysiology are additional alterations of intracellular and membraneassociated sphingolipid homeostasis. In GD patients, GD post-mortem brain studies, and IPS cell-derived human dopaminergic midbrain neurons with bi-allelic and heterozygous GBA variants, GCase deficiency results in accumulation of the GCase substrates GlcCer and GlcSph. PD GBA patients, specifically PD GBA with severe variants, show not only elevated levels of the GCase substrates GlcCer and GlcSph in CSF and plasma but also increased CSF levels of downstream-products (Cer) and by-products (SPA, S1P) when compared to healthy controls and PD GBA_wildtype (Lerche et al. 2021a;Surface et al. 2022). Assessments in plasma from PD GBA_wildtype as well as in aging mouse models further support findings that with decreasing GCase, activity levels of downstream/by-products are also elevated in addition to upstream substrates (Hallett et al. 2018;Mielke et al. 2013). In GD with pronounced GCase deficiency, GlcCer are alternatively processed to GlcSph and exit the lysosome into the cytosol (Hein et al. 2007;Elleder 2006;Ferraz et al. 2016). Cytosolic GlcSph is further hydrolyzed to ceramides, sphingosine, and sphingosine-1-phosphate. Recent studies highlight the role of ceramides and sphingosine-1-phosphate as key players in the regulation of cell death and survival with involvement in ER stress, autophagy, protein and lipid transport, exosome secretion with neurotoxic protein spreading, neuroinflammation, and mitochondrial dysfunction (Wang and Bieberich 2018). Data from α-synuclein/ GBA transgenic mice and HEK cell cultures show that Glc-Cer, GlcSph, sphingosine, and sphingosine-1-phosphate promote the formation of oligomeric α-synuclein (Taguchi et al. 2017). Expanding these findings, recent data in human dopaminergic midbrain neurons suggest that conformational changes of α-synuclein towards an aggregation-prone pattern can be even induced by the presence of glycosphingolipids alone irrespective of GCase deficiency due to variants in the GBA gene (Zunke et al. 2018). Post-mortem studies show increased levels of GlcCer, GlcSph and ceramides in the substantia nigra and frontal cortex of PD GBA_wildtype (Rocha et al. 2015a;Huebecker et al. 2019;Kurzawa-Akanbi et al. 2021). However, no differences were seen in the putamen of PD GBA and PD GBA_wildtype compared to controls (Gegg et al. 2012). More post-mortem studies with uniformly assessed brain regions and cell types as well as stratification according to GBA variant severity are needed to shed light on these seemingly discrepancies.
Enhanced activation of phosphocholine cytidyltransferase resulting in increased synthesis of phosphatidylcholine as major component of phospholipid cell membranes was reported in GD (Bodennec et al. 2002). Interestingly, alterations in the lipid bilayer composition of membranes cause impaired α-synuclein membrane binding and enhance aggregation-prone fibril formation (Piccinini et al. 2010). Combined 1 H and 31 P magnetic resonance spectroscopic imaging revealed that PD GBA patients display a disturbed membrane phospholipid metabolism in the putamen and midbrain with reduced levels of the precursor choline and increased levels of the membrane-related phospholipid degradation product glycerophosphoethanolamine. These changes were accompanied by neuronal loss in these brain regions as measured by reduced levels of the neuronal marker N-acetyl-aspartate (Brockmann et al. 2012).

Therapeutic targets in GBA-associated PD
Based on the knowledge of the molecular mechanisms underlying PD GBA , pathway-specific treatment options are beginning to emerge.

GCase
The significant reduction of GCase protein levels and GCase enzyme activity offer a plausible therapeutic rational to either increase GCase protein levels or enhance enzyme activity. Unfortunately, intravenous enzyme replacement therapy is not possible due to insufficient central nervous system penetration.
Gene therapy with adeno-associated virus (AAV)-based vectors promoting GBA overexpression. This approach reduced α-synuclein accumulation, improved lysosomal function and lipid turnover, and attenuated deficits in working memory and fine motor performance in α-synuclein mutant/overexpressing wild-type and GD rodent models (Rocha et al. 2015b;Glajch et al. 2021;Sardi et al. 2011). The AAV9-based vector PR001 increased GCase activity, reduced glycolipid substrate accumulation, and improved motor deficits in two mouse models of GCase deficiency (Abeliovich et al. 2021). Based on these results, a phase 1/2a non-randomized clinical trial with a single administration of PR001 into the cisterna magna is currently under investigation in PD patients with at least one pathogenic GBA variant. The study duration is 5 years. During the first year, patients will be evaluated for safety, tolerability, immunogenicity, biomarkers, and clinical efficacy measures. Patients will continue to be followed for an additional 4 years to monitor safety and selected biomarker and efficacy measures (NCT04127578).
GCase-enhancing small-molecule chaperones refold misfolded GCase in the ER and promote proper trafficking, thereby increasing lysosomal GCase protein levels. Interestingly, experimental data in cell and animal models with GBA variants suggest that the expectorant Ambroxol increases GCase availability via such mechanism (Kopytova et al. 2021;Ambrosi et al. 2015;Maegawa et al. 2009;Magalhaes et al. 2018;McNeill et al. 2014;Yang et al. 2022;Migdalska-Richards et al. 2016). These findings led to a proofof-principle phase 2 open-label study with Ambroxol in 17 PD patients with and without GBA variants. Ambroxol was well tolerated and CSF GCase protein levels as well as CSF levels of α-synuclein increased by 35% and 13%, respectively. However, CSF GCase enzyme activity decreased by 19% which might be explained by an inhibitory effect of Ambroxol on GCase activity within acellular human CSF with a neutral pH (Mullin et al. 2020).
More strikingly, a recent publication could show that the small-molecule S-181 increases wild-type GCase activity in iPSC-derived dopaminergic neurons not only from PD GBA but also from PD wildtype as well as from patients with other PD-related gene mutations in LRRK2, DJ-1, and PARKN who also had decreased levels of GCase activity. S-181 treatment of these PD iPSC-derived dopaminergic neurons partially restored lysosomal function and lowered accumulation of oxidized dopamine, GlcCer, and α-synuclein (Burbulla et al. 2019). These recent findings highlight not only the importance of lysosomal dysfunction in the pathophysiology of the prototype PD GBA but also the significance of this pathway, possibly in concert with additional pathways such as mitochondrial dysfunction, for PD in general.

Substrate reduction therapy
Substrate reduction therapy to reduce GlcCer production with penetration into the central nervous system is available for oral application in GD. Venglustat has been evaluated in a phase 2 randomized trial (MOVES-PD, NCT02906020) in PD GBA . The compound clearly reduced CSF levels of GlcCer in a dose-dependent manner in plasma and CSF. However, the study was stopped prematurely, since patients in the verum group showed enhanced clinical deterioration suggesting an off-target effect with possible anti-dopaminergic activity.

Alpha-synuclein-targeting compounds
Targeting alpha-synuclein also seems a reasonable treatment option given the predominant α-synuclein aggregation and wide-spread Lewy-body pathology in PD GBA .

Conclusion and outlook
GBA-associated PD is remarkable for several reasons. The phenotypical trajectories show a faster disease progression with pronounced early cognitive decline and a clear dependency based on mutation severity. Importantly, the development of dementia is not associated with Amyloid-β pathology as shown instead in a relevant proportion of PD without GBA variants but rather due to predominant α-synuclein aggregation. The identified pathophysiological mechanisms highlight GCase deficiency and lysosomal dysfunction resulting in disrupted glycosphingolipid homeostasis and ultimately impaired α-synuclein degradation with enhanced aggregation. Again, these are dependent on mutation severity and offer different targets for individualized treatment options. However, the failure of the MOVES-PD trial (NCT02906020) demonstrates the challenges we are facing in translational research. Findings from GD as typical and clearly defined young-onset lysosomal storage lipid disorder due to bi-allelic mutations in GBA are not simply transferable into PD, a multifactorial disease of the elderly with possibly additional contributing factors (e.g., mitochondrial dysfunction and lifetime environmental exposure). Specifically, the pathophysiological mechanisms of impaired glycosphingolipid homeostasis leading to impaired α-synuclein degradation need more investigation. In this context, longitudinal patient cohorts with repeated collections of biomaterials, ideally starting in the prodromal stage followed up until death with brain donation, might inform us on biomarkers that reflect the underlying pathological processes and possible read-outs for target engagement.
Future clinical trials in PD GBA might incorporate the knowledge learned over the last years: (i) Patients should be stratified according to GBA variant severity with those carrying severe mutations to be preferentially included in proof-of-concept trials. (ii) The early cognitive decline based on predominant α-synuclein-driven pathology offers the opportunity to address PD-associated dementia with diseasemodifying agents in a clearly defined prodromal phase preceding dementia and based on clear biological stratification.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.