Cross-talks among GBA mutations, glucocerebrosidase, and α-synuclein in GBA-associated Parkinson’s disease and their targeted therapeutic approaches: a comprehensive review

The GBA gene encoding the lysosomal enzyme GCase comprises 10 introns and 11 exons and is located on chromosome 1q21, which is a gene-rich site composed of 2 pseudogenes (GBAP and MTX1) and 9 genes within a long sequence of about 100 kb. The highly untranslated pseudogene GBAP is homologous to GBA, which share a 96% exonic sequence identity in the coding region, complicating the detection of GBA mutations. The second convergently transcribed pseudogene MTX1 (Metaxin-1) encodes a protein found in the mitochondrial external (outer) membrane and is located downstream to the GBAP sequence [5, 12, 13, 20‐22]. GCase (D-glucosyl-N-acylspingosine glucohydrolase) is a 497-amino-acid (AA) active protein with 5 glycosylation regions [17, 23‐27]. The protein is synthesized in the ER, transported by the lysosomal integral membrane protein-2 (LIMP2), which is a GCase transporter encoded by SCARB2 gene, and becomes active upon reaching the acidic lumen of the lysosome by interacting with its activator protein saposin C. In the lysosomal compartment, the enzyme hydrolyses glucose moieties from glucosylceramide and glucosylsphingosine [21]. GCase consists of 3 domains: domain-1, an antiparallel β-sheet; domain-2, an immunoglobulin-like domain comprised of two closely associated β sheets; and domain-3, a (β/α)8 triosephosphate isomerase (TIM) barrel [17, 23].

GCase goes through several structural alterations during the transportation to the lysosome. It has 2 functional ATG-initiation sites, transcribed as a 516- or 536-AA protein, which is further processed into the 497-AA functional enzyme [23, 28]. It has been proposed that the transport of GCase into the ER is differentially affected by the different leaders and the 19-AA leader peptide cleavage takes place when it enters the ER. Oligosaccharide modulations also occur but they do not affect the intracellular stabilization or the catalytic activity of GCase; hence, their significance remains unclear [29, 30]. The transportation of GCase to the lysosome is independent of mannose-6-phosphate as suggested by the following lines of evidence. In the lysosomal disease state, the transportation of lysosomal enzymes into the lysosome could not be targeted via the mannose-6-phosphate-dependent pathway [31, 32]. Studies in different animal and cellular models have suggested that LIMP2, which is responsible for GCase transportation to the lysosome, is independent of mannose-6-phosphate [26].

An important function of GCase is to degrade glucocerebroside (also known as glucosylceramide) into glucose and ceramide. It also potentially cleaves β-glucosides such as glucosylsphingosine. GCase deficiency renders glucocerebroside accumulation within cells of the reticuloendothelial system [21].

The correlation between GBA mutations and a higher risk for developing PD was initially observed in GD clinics about 14 years ago [33, 34], when GD patients and their relatives, who were supposed to be GBA mutation carriers, were found to have a higher incidence of PD than the general population [35]. This finding promoted investigation of patients diagnosed at the GD clinic, which revealed a PD incidence of about 25% [36]. A survey conducted at the GD clinic (Jerusalem) elucidated a similar finding to the earlier report [6, 37, 38]. In 2004, it was reported that the first-degree relatives of GD patients had a much higher incidence of PD relative to the normal population [6, 38]. This was when various Parkinsonism-associated GD populations were screened for this gene mutation to assess the vital role of GBA in the PD pathogenesis worldwide [38]. Animal models for the analysis of GBA mutations in GD had been established before the recognition of its role as a risk factor for PD and now new models have been designed especially for the evaluation of the correlation between GBA mutations and PD. In addition, there is a wide range of cell-based designs using e.g., the pluripotential induced stem cells (iPSCs) [39, 40]. Broad Gaucher Registry statistics suggest that the risks for developing PD in GD patients are 9%–12% and 5%–7% before the age of 80 years and 70 years, respectively, although the GBA mutations occur in 5%–10% of PD patients [41]. This made GBA mutations the most significant genetic risk factor for PD [42]. Macrophages of GD patients are associated with glycolipid stress and load and are thus regarded as “Gaucher cells”, which are characteristic of the disease pathology. The single heterozygous mutations were originally believed to be harmless, but after analysis of early cases of Parkinsonism in GD patients, the heterozygous mutations were considered to present a substantial risk of developing PD [6, 43, 44].

The associations of GBA mutations with dementia with Lewy bodies (DLB) and Parkinson’s disease dementia (PDD) have also been well established. GBA1 mutations are a risk factor for the development of DLB, suggesting some correlations among GBA1, dementia and parkinsonism. Patients with DLB are 8 times more likely to be the GBA mutation-carriers than controls. Such a risk is greater than that reported for PD, and seems to correlate with disease severity, earlier age onset and progression of disease. GBA is involved in the DLB pathology, but the precise cause of this predisposition is not known. Recently, the association of DLB with the PD-linked SCARB2 has highlighted the significance of the lysosomal pathway in DLB. Genetically, DLB seems to be heterogeneous, with common risk factors and relatively rare contribution from pathogenic causative mutations [45, 46]. There is no significant difference between DLB patients with and without a GBA mutation in the neuropathological data, though the GBA - mutation-carriers show a reduced activity of GCase and more prominent lipid profile alteration in the brain. GBA expression is reduced in PDD and DLB cases in the caudate nucleus and the temporal cortex [36, 47]. PD mild cognitive impairment (MCI) and PDD are two types of cognitive impairment and the most prevalent and disabling non-motor symptoms of PD. MCI is the early warning signal of late dementia in PD. Research evidence has shown that about 30% of newly diagnosed PD patients have MCI, > 40% of PD patients with normal cognitive function will develop MCI within 6 years, and about 80% of patients may develop PDD at the late phase of PD. Various studies have suggested that the GBA mutations are more likely to be associated with PD-MCI [47, 48].

About 495 mutations and rearrangements across the exons of GBA are known to be associated with GD, including splicing mutations, frameshift mutations, point mutations, deletions, insertions and null alleles, and most of them are “missense mutations”. The most common mutations are point mutations including N370S (seen exclusively in type 1 GD patients in the Europe, USA and Israel) and L444P (observed worldwide) [21, 35, 38]. There is a 10%–30% probability of developing PD among the heterozygous mutation carriers at the age of 80, a 20-fold rise as compared to the non-carriers [49‐53]. The idiopathic PD is undifferentiated from GBA + PD, with the only distinctive characteristic being the earlier age of onset and a greater incidence of cognitive features [49, 54, 55]. Moreover, the variant E326K is a notable example that is not considered a GD-causing mutation but a risk factor for PD [54, 56]. This implies a precise, and perhaps a different process by which mutations make their carriers prone to PD. The enzymatic activity of GCase is differentially influenced by different GBA mutations, as many of the mutations lead to no residual activity whereas others result in decreased activity [39, 45].

In comparison to GD, a smaller number of GBA mutations has been observed in PD patients (~ 130 GBA mutations) [33, 34]. Only certain mutations that are most generally linked with PD have been screened in several studies. Hence, the PD-related mutants that are less persistent could go unidentified. Generally, c.1448 T > C (L444P) and C.1226A > G (N370S) are the two most persistent mutations amongst others. They even contribute to 70%–80% of GBA mutants linked with PD in certain populations [57]. In a population of Colombian PD patients, an increased frequency of the variant K198E was observed in comparison with controls, consistent with the recent finding in GD patients [33]. Severe GBA mutations such as c.115 + 1 G > A (IVS2 + 1), c.84dupG (84GG), c.1297 G > T (V394L), c.1263del + RecTL, c.1342 G > C (D409H), and c.1448 T > C (L444P), appear to be correlated with a higher risk of PD development compared to the mild mutations such as c.84dupG (84GGG) and N370S [58]. p.T369M is one of the GBA variants that do not cause GD in homozygous carriers and may modify the GCase activity. In some studies, the substitution of p.T369M was associated with PD, while in other studies it had similar or increased frequency in controls. It is of interest that the GBA p.T369M substitution was demonstrated to be correlated with declined GCase activity in PD patients and controls compared to that in non-carriers. A recent meta-analysis showed that the p.T369M allele was associated with an increased risk for PD [59, 60]. Besides, the severe mutations mentioned above are also linked with an earlier onset age, and also have greater incidence and elevated cognitive ability involvement [58‐62]. A previous study showed that PD patients with extreme GBA mutations had significantly worse motor symptoms together with some non-motor symptoms such as insomnia and rapid-eye-movement sleep disturbances, compared to subjects with moderate mutations or idiopathic PD [63]. Interestingly, GBA is nothing more than a risk element for PD. The latter implies that the disease will not develop in every carrier. The implications of all these mutations remain to be clarified. Information from a wide array of human studies shows that around 9.1% of GBA carriers will develop PD. Several reports have indicated that the prevalence of PD in GD patients at age of 80 is about 30%, but more research is needed to validate this data [50]. Homozygous GBA variant-linked patients that are affected by GD have a greater risk of developing PD and generally have an earlier age of symptom onset [63]. It is important to note that even in the case of severe mutations, most GD subjects do not develop PD. The heterozygous and homozygous GBA variant-carriers exhibit a 5- and 10-times greater risk of having PD, respectively [50, 64, 65]. GBA mutations are present in ~ 2%–30% of PD patients [66] (Table 1).

Mitochondrial dysfunction, neuroinflammation, and excessive ROS formation are key processes that lead to dopaminergic neuronal death, wherein the mitochondrial dysfunction is involved in the pathophysiology of both familial and idiopathic PD [112, 113]. Evidence has suggested that α-synuclein interacts with the mitochondrial import proteins through a cryptic mitochondrial import signal [114]. Mutations of PTEN-induced putative kinase (PINK1) and PARK2 contribute to the monogenic PD, and are assumed to affect mitochondrial function by elevating the susceptibility to toxins [115]. In a neuronopathic GD mouse model (K14-lnl/lnl), Ossellame et al. found that the proteasomal and autophagic pathways in astrocytes and neurons were impaired and that insoluble α-synuclein deposits occurred in neurons [107, 116]. Depletion in GCase function in cell studies led to a gradual deterioration of the potential of mitochondrial membrane needed for ATP output, fragmented mitochondria, respiratory complex function loss, and oxidative stress.

Ultimately, dysregulation of calcium occurred in the mitochondria, resulting in a distorted potential of the membrane [117]. In addition, ROS are generated after mitochondrial dysfunction, inducing persistent oxidative damage that may trigger α-synuclein misfolding and activate other deteriorating channels in the neuron [118, 119]. Accordingly, secondary mitochondrial dysfunction may arise from the primary lysosomal insult (i.e., GCase activity loss). Cellular disturbances including ROS, impairment of mitophagy, and ER stress can further deteriorate the cellular homeostasis and promote α-synuclein accumulation [120]. The transcript level of the antioxidant NQ01 has been found to be augmented in GD patients and GBA heterozygotes with or without PD, which may serve as a potential compensatory process. Thus, the GCase insufficiency can increase the oxidative stress of neurons. Elevating the levels of glutathione as an antioxidant may be a therapeutic strategy, which can be achieved by administration of N-acetylcysteine [121, 122].

Chemotactic factors are a central part of the process of neuroinflammation in that they can allocate immunologic facilitators to the inflammatory site [123, 124]. Iron is increased in PD patients in the substantia nigra. Excessive iron can normally be chelated by neuromelanin or ferritin. However, with natural ageing, the residual iron is stored in the brain. Excessive levels of iron induced by reactive nitrogen species or ROS can contribute to neuronal death [125]. Disruption of zinc homeostasis can aggravate or cause neurodegeneration via induction of protein misfolding and oxidative stress. Zinc augmentation also occurs in the substantia nigra of PD patients [126]. Mutations of PARK9/ATP13A2, which encode a zinc pump that brings zinc into the membraned components, are correlated with juvenile-onset PD. ATP13A2 overexpression causes extracellular zinc tolerance and facilitates the exosomal transfer of α-synuclein [127, 128]. Silenced ATP13A2 expression in primary neurons causes a decline in lysosomal chelation of Zn²⁺ and increases the expression of Zn²⁺ transporters. The increased Zn²⁺ leads to lysosome dysfunction that leads to the GCase–α-synuclein pathological cascade. Either chelation of zinc or improvement of ATP13A2 expression can improve the phenotype [129, 130].

Current promising therapies for GD include enzyme replacement therapy (ERT) and substrate replacement therapy (SRT), both have been approved by the FDA and work through the production and maintenance of a more standard ratio of GCase substrate in patients. These therapies have markedly enhanced the visceral symptoms of GD but are unable to gain access to the blood-brain barrier, so the neuronopathic symptoms of GD cannot be alleviated or reversed [2]. Considering the significant involvement of GCase in the PD pathogenesis, innovative therapies that can restore the levels of neural GCase, may not only enhance the life quality of neuronopathic patients, but also delay the development of PD in populations vulnerable to the GD-associated PD or idiopathic PD. Currently, brain-penetrating variants of SRT are under clinical trials with PD patients who are carriers of heterozygous GBA mutation [36, 129]. Numerous companies have been trying to tackle this problem, achieving very encouraging outcomes in animal and cell models. The interventions available in clinical trials so far mainly resolve pathways that are considered to be counterproductive in connecting GBA mutations to PD. The hypothetical pathways that lead to impairment of GCase and the associated therapies targeting these pathways are shown in Fig. 3.

One hypothesis demonstrates that the mutated GBA proteins are not adequately foldable in the ER, triggering protein accumulation in the cellular compartment, which induces dopaminergic neurons to respond to the stress, contributing to their damage and death. In addition, the β-GCase entanglement in the ER causes a declined enzyme level within the cell, inducing the aggregation of α-synuclein [102]. To target this pathogenic mechanism, distinctive chaperones, which are capable of facilitating the substrate replenishment, have been tested [103, 104, 131, 132]. A clinical trial on ambroxol, a chaperone that demonstrated extremely interesting preliminary findings, was established in 2016 (NCT02914366). This clinical trial at Phase 2 was aimed to evaluate the efficacy and safety of ambroxol to enhance the cognitive and motor characteristics of GBA + PD patients [125, 133]. Research has shown that administering chaperones to raise the native protein levels of chaperone could be the gateway to GCase refolding and restoration of normal enzymatic function in the brain. Arimoclomol is one of those chemical substances, which activates the heat shock reaction, thus magnifying Hsp70 and other proteins from heat shock. The administration of arimoclomol to fibroblasts from L444P genotype patients increased GCase activity at a frequency comparable to about 1 unit of the regular ERT drug, alglucerase [132]. Another pharmacological chaperone isofagomine, has been studied in vivo and in vitro to determine the capacity to modify the GBA mutation-induced phenotype and to stabilize the GCase [134].

Another pathway to be explored for the treatment of GBA + PD is the accumulation of glucosylceramide (substrate of GCase) in the dopaminergic neurons due to GBA mutation [131, 135, 136, 137]. Recently, a phase 2 multicenter, placebo-controlled, randomized, double-blind study was initiated to analyze the pharmacokinetics, pharmacodynamics, and safety of an oral molecule, ibiglustat (SAR402671), which is capable of reducing beta-GCase levels in early-stage GBA + PD. Compared with the wild-type form, mutated GCases are much more unstable. Modification of GCase degradation may be another effective technique to increase the enzymatic function and thereby combate aggregation of α-synuclein and neurodegeneration. Hsp90 is responsible for the breakdown of misfolded GCase, together with another heat shock protein PARK-2. In general, specific HSP inhibitors and histone deacetylase inhibitors (HDACis) can enhance the levels of GCase, limiting its degradation. HDACis, in effect, prevent contact between GCase and Hdp90 by hyperactivating one of its domains [36, 138, 139, 140]. It has been proposed that the inactivation of protein phosphatase 2A may reflect the possible process via which the deficiency of GCase blocks autophagy and facilitates the aggregation of α-synuclein. Autophagy upregulation by the mTOR inhibitors polyphenols and rapamycin showed favorable effects on synucleinopathy in animal and cell models by decreasing the intracytoplasmic aggregates of protein and ultimately the death of cells. Surprisingly, these pathways tend to underlie the cross-talk between α-synuclein and GCase and can influence the propagation of disease [136, 141, 142].