Introduction—the Discovery of GBA Mutations in Parkinson’s Disease

Unlike most genetic findings in recent decades, which are generally based on linkage analysis, genome-wide association studies (GWAS), and whole exome sequencing studies, the discovery of the association between GBA mutations and Parkinson’s disease (PD) came from clinical observations [1, 2]. Initial reports on the frequency of GBA mutations in PD were inconsistent, until later studies unequivocally demonstrated a strong association between GBA mutations and PD [3,4,5]. The first case-control study was on Ashkenazi-Jews (AJ), in which GBA mutations are particularly common, and suggested that around 30% of AJ PD patients carry a GBA mutation [3]. Subsequently, follow-up studies from the USA and Israel reported that 15–20% of AJ PD carries a GBA mutation [6••, 7]. A large meta-analysis from 16 centers worldwide has since clearly demonstrated that GBA mutations are associated with PD [8], which was replicated in many other populations [4, 9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30]. In large-scale GWASs, the GBA locus was associated with the strongest risk for PD, driven by the p.E326K variant [31,32,33].

Gaucher’s disease (GD, OMIM #230800, #230900, and #231000) is caused by a recessively inherited deficiency of the lysosomal enzyme glucocerebrosidase (GCase), encoded by the GBA gene (OMIM #606463). Traditionally, GD is divided into three types according to increasing severity of the disease and the degree of neuronal involvement, where type 1 is non-neuronopathic, type 2 is acute neuronopathic, and type 3 is chronic neuronopathic [34].

We hereby review the current knowledge on the structure and function of the GBA gene, its protein product GCase, and their phenotypic correlations. We also discuss implications for diagnosis, counseling, and treatment and suggest future directions for research.

The Normal Structure and Function of the GBA Gene and Its Protein Product, Glucocerebrosidase

The GBA gene contains 11 exons spanning 7.6 kb on chromosome 1q21, in a gene-rich region that includes 9 genes and 2 pseudo-genes within a 100-kb-long sequence. The GBA promoter region includes putative TATA and CAAT-like boxes approximately 250 bp upstream to the ATG start site and lacks the GGCGGG motif [35]. The two pseudo-genes, GBAP and MTX1P, are located between the GBA and MTX1 genes and appear to have resulted from a duplication event which took place approximately 27–40 million years ago. These pseudo-genes are found in humans and primates, but not in other species [36, 37]. GBAP shares 96% sequence identity with GBA but spans only 5 kb due to intronic Alu sequences, a form of transposon that can relocate throughout the DNA during evolution, which are present in the GBA but not in the GBAP gene sequence. Additionally, a 55 bp deletion in exon 9 flanked by a short inverted repeat also distinguishes the pseudo-gene from the GBA gene [36].

The active protein transcribed by GBA, GCase, is a 497-amino acid (AA) lysosomal hydrolase. The main function of GCase is to degrade glucocerebroside into ceramide and glucose, but it also cleaves glucosylsphingosine and potentially other β-glucosides [38]. During its transport to the lysosome, GCase undergoes several modifications. Having two functional ATG initiation sites, GCase is transcribed as a 536 or 516 AA protein [39], which is further processed into the functional 497 AA enzyme. This cleavage of the 19 or 39 AA long leader peptide occurs while entering the endoplasmic reticulum (ER) [40], and it was suggested that the different leaders differentially affect transport into the ER [41]. Oligosaccharide modifications also occur, but they have no effect on the catalytic activity or the intracellular stabilization of GCase; therefore, their importance is not clear [42]. Several lines of evidence suggest that the transport of GCase from the ER to the lysosome is mannose-6-phosphate independent [37,38,39]. First, GCase is located within the lysosome in I-cell disease (ICD), which is caused by deficiency of the enzyme N-acetylglucosamine 1-phosphodiester N-acetylglucosaminidase [43]. In this disorder, lysosomal enzymes that are usually transported into the lysosome via the mannose-6-phosphate-dependent pathway cannot be targeted into the lysosome [44]. Second, by using cellular and animal models, it was shown that the lysosomal integral membrane protein type 2 (LIMP2/SCARB2) is a mannose-6-phosphate independent receptor which transfers GCase to the lysosome [45].

Within the lysosome, GCase is peripherally associated to the inner membrane [46, 47], where it exerts its activity together with Saposin C and negatively charged lipids that are essential for its proper function [48, 49]. X-ray studies showed that GCase has three tertiary structure domains: domain I (residues 1–27 and 383–414), which contains two disulfide bridges (residues 4–16 and 18–23), domain II (residues 30–75 and 431–497), an immunoglobulin-like domain, and domain III (residues 76–381 and 416–430), a TIM (triophosphate isomerase) barrel. The TIM barrel contains three free cysteines and the catalytic site (with glutamate residues in positions 235 and 340), which is covered by three loops (residues 312–319, 341–350, and 393–396) [50].

GBA Mutations and Their Effect on Glucocerebrosidase Structure and Function

Approximately 300 mutations and gene re-arrangements in the GBA gene have been described [40], which are classified according to the type of GD (I, II, or III) that might develop [51]. Severe (including null) GBA mutations are those that when inherited from both parents result in the severe types of GD (types II or III), while mild mutations are those that if inherited in a homozygous or compound heterozygous manner, cause the mild type of GD (type I) [51]. In addition, there are variants which have an unclear role in GD but are clearly risk factors for PD, the most notable example being E326K. The following is a review of the possible effects of different GBA mutations on the structure and function of GCase.

Effects of GBA Mutations on GCase Enzymatic Activity

Different GBA mutations differentially affect the enzymatic activity of GCase, as some mutations result in almost no residual activity whereas others show only reduced activity. In several cases, the level of enzymatic activity does not correlate with the severity of GD [34], and the enzymatic activity range of severe and mild GBA mutations can overlap. For instance, the measured enzymatic activity of GCase with the p.N370S mutation, which is always associated with type I GD, may be lower than the measured enzymatic activity of GCase with severe mutations such as p.L444P, p.G390R, p.N382K, and others [38, 52, 53•]. One clinical example of such lack of correlation between residual enzymatic activity and severity of disease was shown in an infant with very severe neuropathic type II GD, who was homozygous for the p.G202R mutation, but had “only slightly reduced activity” of GCase [54]. Ideally, GCase activity should be measured within the lysosome only, which may better reflect the true function of GCase. Additionally, factors other than the residual GCase activity, whether genetic/biologic or environmental, may determine the severity of the disease.

Effects of GBA Mutations on GCase Traffic, Binding, and Interaction Properties

In the infant homozygous to the severe p.G202R mutation described above [54], the degree of ER retention of GCase was examined, and it was shown that the mutated GCase was not transported to the lysosome and was retained in the ER. It was therefore possible that the severity of the disease was not only due to the enzymatic level of activity, but also due to the fact that GCase could not exert its function where it belongs, within the lysosome. Subsequent studies demonstrated that the transport of GCase with the severe p.D409H and p.L444P mutations to the lysosome is also restricted and that the transport of GCase with the mild p.N370S mutation was partial [55]. Similarly, the degree of ER retention of GCase was determined in seven patients with type I GD and four with types II or III. Generally, ER retention was higher in GD patients with the severe type of the disease, carrying mutations such as p.P415R, p.L444P, and p.D409H, with the exception of one patient with type III GD that had a comparable degree of ER retention to that of type I GD patients [56]. These studies suggest that although mutated GCase may have residual enzymatic activities in cellular assays, they may not reach the lysosome in vivo, and therefore cannot exert their function. This will result in a de facto severe deficiency of GCase. Indeed, it is possible that impaired transportation of GCase to the lysosome may also lead to PD. GWASs and other genetic studies have identified variants around SCARB2, the transporter of GCase from the ER to the lysosome, as a genetic risk factor for PD [31, 57,58,59]. An association with SCARB2 was also suggested in dementia with Lewy bodies (DLB) [60]. These data may indicate that even when GCase is functioning based on enzymatic activity assays, perturbed transport to the lysosome may increase the risk for synucleinopathies.

GBA mutations may also cause structural effects that can influence the function of GCase. Experiments in human fibroblasts containing the mutated p.N370S GCase demonstrated a reduced capacity of the enzyme to interact with its activator, saposin C, and with anionic phospholipids that are necessary for its proper function [61]. Supporting this observation, in a structural model of the interaction between GCase and saposin C, the p.N370S mutation was mapped to the interacting surface of the two proteins, and so was the severe mutation p.L444P [62]. In addition, it was suggested that the GBA p.N370S mutation may affect the stability of the helical turn conformation of loop 1 [63].

Potential Mechanisms of GBA Parkinson's Disease

Several pathways were suggested to be involved in GBA-associated neurodegeneration. Mazzulli and colleagues demonstrated that the substrate of GCase, glucosylceramide, may lead to α-synuclein accumulation, and inversely, α-synuclein accumulation may lead to reduced GCase activity [64]. This relationship has been supported by several lines of research: overexpression of α-synuclein led to decreased GCase activity in mice and cell models [65], and reduced GCase levels have been shown in brain tissue [66,67,68], CSF [69], and peripheral blood of PD patients compared to controls [53•], independent of GBA mutation carrier status. Further, a recent study found that in human-induced pluripotent stem-derived neuronal models (iPSn), treatment with GCase inhibitors led to increased α-synuclein aggregation [70]. This aggregation was reversible by treating GD and PD patient iPSns with a glucosylceramide synthase inhibitor. These studies suggest a vicious, neurotoxic cycle between α-synuclein and GCase that may partially explain the mechanism underlying GBA PD, but the vulnerability of specific neuron types is still not understood. Further, no study to date has shown elevated concentration of glucosylceramide in GBA heterozygotes.

Another mechanism by which GBA mutations may result in PD is ER-associated degradation (ERAD) impairment and ER stress-related cell death. α-synuclein accumulation may cause ER stress, impair degradation of ERAD substrates, and inhibit ER to Golgi traffic [71]. Supporting this observation is the role of some PD-associated genes, such as PARK2, in ERAD [72, 73]. Taken together, it is possible that ERAD and ER stress could be important in PD pathogenesis. Along this line, the ER retention detected in experiments with some of the mutated forms of GCase [54,55,56] may suggest that ER stress is also involved in the pathogenesis of PD in carriers of some GBA mutations. It was also shown that mutated GCase interacts with parkin, which promotes the accumulation of GCase in aggresome-like structures [74].

However, the mechanisms suggested above are challenged by the fact that null GBA mutations which do not result in a protein product [51, 52, 54], such as 84GG, IVS2+1, R359X [51], and others, also increase the risk of developing PD [4, 75]. If the protein does not exist, it cannot accumulate or assist in fibrillization. It is possible that the ER stress observed in models with GBA mutations may be due to α-synuclein accumulation rather than accumulation of GCase itself. It is also likely that GBA mutations increase susceptibility to PD in more ways than one and that both suggested mechanisms contribute to disease development.

The ceramide metabolism pathway is also a mechanism that may be involved in PD pathogenesis. Since LBs are the pathologic hallmark of PD, but are also found in other diseases, it was noted that some of the genes underlying these diseases (such as GBA, SMPD1, ASAH1, GALC, PANK2, and PLA2G6) have a significant role in the ceramide metabolism pathway [76,77,78], and recent studies suggested the involvement of SMPD1, ASAH1, and GALC in PD [31, 77,78,79]. Ceramide may play a role in some PD-related mechanisms such as stress-induced cell death [80] or inflammation [81]. Moreover, ceramide binds cathepsin D and triggers its cleavage to the catalytic form [82], which is one of the main lysosomal enzymes responsible for α-synuclein degradation [83]. Therefore, the ceramide metabolism pathway may be related to LB formation in GBA PD [76].

Genotype-Phenotype Correlations in GBA Parkinson's Disease

The initial reports on GBA-PD reported atypical parkinsonism with seizures and dementia among Gaucher patients who are homozygous GBA mutation carriers and a phenotype similar to idiopathic PD (iPD) in GBA heterozygotes [34]. Subsequently, a wealth of information was collected on GBA PD. Similar to idiopathic PD, phenotype is heterogeneous and symptomatology can vary (Table 1). On an individual level, GBA heterozygotes or homozygotes with PD are indistinguishable from idiopathic PD patients. However, rate of motor progression is faster in GBA PD compared to idiopathic PD [84••]. In addition, GBA mutation carriers are more likely to manifest non-motor symptoms which are prevalent in iPD including cognitive impairment [89], REM sleep behavior disorder (RBD) [93], hyposmia [85], and autonomic dysfunction [89, 90]. The rate of cognitive change is faster in GBA PD than in idiopathic PD, and therefore, patients tend to have faster motor and cognitive progression than iPD (Table 2). In addition, multiple studies suggest a “dose effect”, where Gaucher PD (PD with homozygous or compound heterozygous GBA mutations) is associated with earlier age at onset and more cognitive changes than GBA PD (PD with heterozygous mutations) [85, 97]. Among GBA PD cases, those who carry severe mutations (e.g., L444P) have faster motor progression as measured by the UPDRS, faster rate of dementia and younger age-at-death than carriers of mild mutations (e.g., N370S) [6••, 84••, 86•, 91•, 98]. Referral to deep brain stimulation may happen earlier in GBA PD [95], but has also been associated with early cognitive impairment after DBS.

Table 1 Clinical characteristics of GBA PD compared to idiopathic PD
Table 2 Longitudinal characteristics of GBA PD compared to idiopathic PD

GBA Mutations and Other Synucleinopathies

GBA in REM Sleep Behavior Disorder

Most individuals with RBD are likely to develop one of the synucleinopathies: PD, DLB, or MSA [99]. Interestingly, when comparing the clinical presentation of GBA PD and RBD-associated PD, there are many similar characteristics. PD patients with GBA mutations and PD patients with RBD are both likely to have non-motor symptoms: autonomic dysfunction [89, 100], cognitive decline, and faster progression to dementia [89, 101, 102]. Furthermore, both are also associated with similar motor characteristics: rapid motor progression [89, 103] and the postural-instability-gait-dysfunction phenotype [104, 105]. Therefore, it is no surprise to find that GBA mutations were associated with RBD in a cohort of idiopathic RBD patients [93] and in subsequent studies [106, 107]. This association was even stronger than the association with PD in a similar population [108], suggesting that GBA mutations are more specifically associated with the RBD subtype of PD. Furthermore, in clinical PD patients screened with an RBD questionnaire, GBA mutations were associated with probable RBD [93]. It was also demonstrated that among biallelic GBA mutation carriers, as well as among heterozygous carriers who did not have PD, RBD scores were significantly worse than non-carriers [109]. From a pathological point of view, both RBD-associated PD and GBA PD probably have a more diffused spread of α-synuclein accumulation compared to idiopathic PD [110, 111]. Overall, these studies strongly support the association between GBA and RBD.

GBA in Dementia with Lewy Bodies

Several case control studies demonstrated an increased proportion of GBA mutation carriers among DLB patients compared to controls, ranging in frequency from 3.5 to 31% depending on the cohort [17, 112]. The largest multicenter study to date compared 721 clinically diagnosed DLB cases and 1962 controls and found that 7.49% of DLB patients carried a GBA mutation (odds ratio of 8.28 compared to controls) [113]. This relationship has been further supported by studies on neuropathologically confirmed DLB [10, 110, 114], which show increased frequency of GBA mutation carriers among DLB patients compared to controls, and a reverse association between GBA mutations and severity of Alzheimer’s pathology [10, 115]. This suggests that GBA mutations specifically facilitate the development of synucleinopathies, although the mechanism for this relationship remains unclear. However, lysosomal dysfunction is likely a contributing factor [60, 116], as several lysosomal genes have been associated with DLB, including SCARB2, SMPD1, and MCOLN1 [60, 116].

GBA Mutations and Multiple System Atrophy

The association between GBA mutations and MSA is less consistent, possibly because MSA is a rarer disease. In total, eight studies examined the association between GBA variants and MSA [110, 117,118,119,120,121,122,123]. Six studies have found no association between GBA variants and MSA. However, two studies reported an association between GBA and MSA, including the largest international MSA clinical study to date [122], and a study at Columbia University which was enriched for Ashkenazi Jews in both MSA cases and controls [123].

Clinical Implications—Diagnosis, Counseling, and Treatment

The vast research data obtained in the past decades regarding GBA PD has yet to significantly impact patient care. Genetic testing is not currently standard practice in PD treatment, even though genetic information could potentially inform diagnosis, prognosis, and treatment [124, 125]. In addition, patient knowledge regarding genetics in PD is limited, with two reports suggesting that the majority of patients think genetic testing would be useful in PD while only a minority are familiar with GBA or LRRK2 [126, 127]. Though these attitudes can differ across cultures [128], initiatives to improve genetic counseling access for GBA PD patients are warranted.

It is very likely that the launch of precision medicine clinical trials targeting GBA mutation carriers will change patient interest in receiving genotype information. Patients can now be screened for a multicenter randomized clinical trial assessing a glucosylceramide synthase inhibitor for treatment of GBA PD (clinicaltrials.gov identifier: NCT02906020). Another potential therapy is ambroxol hydrochloride, a small molecular chaperone that has been shown to increase brain GCase activity in mouse and primate models [129]. Originally identified through a library screen of FDA-approved compounds [130], ambroxol is now being tested in clinical trials for PD (clinicaltrials.gov identifier: NCT02941822) and PD with dementia (clinicaltrials.gov identifier: NCT02914366). These trials suggest a promising future of precision medicine approaches to genetic PD, though the efficacy of these compounds for improving symptoms remains to be seen. Nonetheless, ascertaining eligibility for these trials is a compelling reason to want to know one’s genotype, and the growing development of such targeted initiatives will likely lead to increased demand for genetic counseling regarding GBA and other genetic risk factors for PD.

The counseling of GBA mutation carriers is complicated by the incomplete penetrance of GBA, where the majority of GBA mutation carriers will not go on to develop synucleinopathy. Despite the clear relationship between GBA and PD, no formal genetic counseling guidelines address how to approach this topic with GBA mutation carriers [131], particularly with family members of patients with GBA PD. Some GBA mutation carriers may also find out their genetic status when undergoing prenatal testing or when their newborns are tested for Gaucher disease, a practice that is becoming increasingly common. One study has suggested that most prospective parents do want to be informed of their PD risk prior to genetic testing for Gaucher [132]. Navigating this relationship will continue to be an issue given the rise of direct-to-consumer genetic testing, which may or may not offer genetic counseling for results.

Future Challenges and Directions for Research

Much has been learned about the genetics of PD over the last 20 years, but there is still more to be uncovered. The mechanistic link between glucocerebrosidase and α-synuclein remains unsolved, and understanding this relationship will be instrumental in designing more targeted therapies for GBA PD. This has not stopped development of the first clinical trial for GBA PD (clinicaltrials.gov identifier: NCT02906020), but more research is needed to develop truly targeted therapies, especially given the phenotypic heterogeneity of disease. It also remains unclear how GBA is related to RBD, and other synucleinopathies such as DLB and MSA. Future studies will need to further investigate these relationships and explore whether alternate mechanisms are at play.

GBA PD patients have a more rapid disease progression than non-carrier PD patients, and there is a dose effect on the rate of disease progression [6••, 84••, 85]. However, the clinical utility of genetic testing in PD is still debated. Some clinicians argue that genotype information is useful for prognostication, and therefore symptom prevention [124]. With the advent of genotype-specific clinical trials, genetic testing is also useful for determining eligibility for experimental therapies. Yet, there are several drawbacks, especially with regards to pre-symptomatic at-risk individuals, including the lack of preventative treatment, the ambiguity of negative results, and the limited ethnic populations in whom genetic testing can identify risk mutations [125]. Tackling these issues will be increasingly important as direct-to-consumer genetic testing becomes cheaper and more widely used. More research is needed to determine how knowing genetic status influences clinical outcomes and to better elucidate barriers to genetic testing in clinical settings. To further improve upon prognostication, better phenotyping of specific GBA mutations may also be warranted.

Conclusion

There has been a breadth of research on GBA since it was identified as a genetic risk factor for synucleinopathies. The mechanism of this relationship remains unclear, though several lines of study suggest interactions with α-synuclein, lysosomal dysfunction, ER stress, and ceramide metabolism. Phenotypically, GBA mutation carriers can present with Parkinson’s disease or dementia with Lewy bodies, though the relationship with multiple system atrophy is still uncertain. In PD patients, GBA mutation carriers are more likely than non-carriers to have younger age of onset, cognitive impairment, and RBD and tend to have quicker motor progression and cognitive decline. Further, there appears to be a dose effect where carriers of severe mutations have quicker progression than carriers of mild mutations, and homozygotes or compound heterozygotes have quicker progression than heterozygotes. Yet, the vast majority of GBA mutation carriers will never develop a synucleinopathy, which complicates genetic counseling for carriers without PD. Now that clinical trials are underway specifically targeting GBA PD, further research is warranted to identify barriers to genetic testing and counseling and to improve upon mutation-based prognostication. Further research is also still needed to elucidate how GBA mutations contribute to disease pathology.