Abstract

It has become increasingly clear that “simple” recessive disorders provide unique insight into the complexities of common diseases. For years, research on Gaucher disease, a rare inherited disorder resulting from a deficiency of the lysosomal enzyme glucocerebrosidase, focused on its cell pathology and genetic basis. Clinical research showed that Gaucher disease manifests with broad phenotypic variation typical of many metabolic disorders, ranging from neonatal lethality to asymptomatic octogenarians. This clinical spectrum now overlaps with different disorders including Parkinson’s disease and other Lewy body disorders, myoclonic epilepsy, and infantile neuro-degenerative disorders. In fact, unraveling the factors contributing to heterogeneity in a single gene disorder may have a direct impact on studies of the pathophysiology and therapeutic options available for these more common and complex neurologic diseases.

Gaucher Disease

Gaucher disease is an autosomal recessively inherited disease found in all ethnicities, but with increased frequency among Ashkenazi Jews. Gaucher disease was first recognized by Philippe Gaucher who described it in his medical school thesis in 1882 (Gaucher, 1882). The disease manifests with vast phenotypic variation. It is a disorder primarily of the reticulo-endothelial system, and lysosomes within macrophages become enlarged because of the accumulation of the lipid glucocerebroside, giving rise to the characteristically appearing “Gaucher cells.” Glucocerebrosides are abundant in cell membranes and undergo degradation as a result of the turn-over of blood cells. Commonly encountered manifestations of Gaucher disease include anemia, thrombocytopenia, hepatosplenomegaly, and bone involvement, with osteoporosis, pain crises, or pathologic fractures. Some patients may also develop pulmonary and/or a diverse assortment of neurologic manifestations (Figure 1).

Many aspects of Gaucher disease serve as a prototype for all of the lysosomal storage disorders. Over the past decades this was the first lysosomal disorder to have the enzymatic deficiency elucidated (Bradyet al., 1965), the gene identified (Ginnset al., 1984;Sorgeet al., 1985), an animal model generated (Tybulewiczet al., 1992), and enzyme replacement therapy developed (Bartonet al., 1991). Moreover, Gaucher disease exemplifies many of the lysosomal storage disorders because of the wide range of phenotypes encountered. Classically, the disorder was classified into three distinct types (type 1: OMIM#230800; type 2: OMIM#230900; type 3: OMIM#2301000) (Knudson and Kaplan, 1962). Type 1 Gaucher disease, by definition, was considered to be non-neuronopathic, and is by far the most common of the types. Type 2 Gaucher disease is the acute neuronopathic form, and was viewed as a rapidly progressive neurodegenerative disorder of late infancy, resulting in death within the first year or two of life. Type 3, or chronic neuronopathic, Gaucher disease was a “catch all” encompassing patients who survived infancy but had some form of neurologic involvement. Often the neurologic manifestation was solely the slowing of the horizontal saccadic eye movements, but other patients developed neurodegeneration, myoclonic epilepsy, or psychiatric manifestations.

Thus, defects in glucocerebrosidase are frequently associated with some degree of neurologic abnormalities. However, detailed clinical evaluations of diverse patients with Gaucher disease have illuminated difficulties in classifying patients into distinct disease categories based upon the rate and severity of neurologic involvement (Figure 1). For example, neurological manifestations can occur with type 1 Gaucher disease when spinal compression fractures cause nerve compression syndromes (Grewalet al., 1991) and, at times, children with type 2 Gaucher disease can live for many years when treated (Goker-Alpanet al., 2003). This spectrum greatly complicates genetic counseling for this disorder and provides a challenge for physicians.

It has long been known that neither the amount of lipid stored, nor the residual enzymatic activity detected, correlates well with symptom severity. With the advent of DNA diagnostics, there was hope that genotypic information could be used for prognosis and therapeutic decisions. However, this did not always turn out to be the case. The gene for glucocerebrosidase(GBA) is on chromosome 1q21, and encompasses 11 exons. There is a highly homologous pseudogene sequence located nearby, sharing approximately 96% of the sequence in coding regions (Horowitzet al., 1989). Almost 300 mutations and polymorphisms inGBA have been identified. Most are point mutations, but others, including insertions and deletions, splice site alterations, and recombinant alleles, have been described (Hruskaet al., 2008). The recombinant alleles can be due to rearrangements, duplications, or gene fusions ofGBA with the nearby pseudogene, complicating their detection.

Gaucher disease and mutations inGBA are found at a surprisingly high frequency amongst the Ashkenazi Jewish population, although the disorder is panethnic. Among Ashkenazi Jews, the combined frequency of the N370S and c.84dupG mutations is 0.0343, providing for a combined incidence of roughly 1:855 (Beutler, 1992). Estimates of the true prevalence of Gaucher disease diverge, due both to the vast clinical variability leading to misdiagnosis, and to undiagnosed patients lacking overt symptoms. Nevertheless, the incidence has been reported to range from 1:640 to 1:3,969 (Beutler, 1992). Despite the low overall prevalence, Gaucher disease is the most common of the sphingolipidoses and has been the source of considerable attention, due in part to its relation to other more common disorders (Sidransky, 2004).

Patients sharing the sameGBA genotypes can exhibit considerable clinical heterogeneity, while clinically similar patients can harbor many different mutant alleles. Thus, as has been the case with many Mendelian disorders, genotype/phenotype correlations in Gaucher disease are not straightforward, and the vast clinical variability encountered is still not well explained (Goker-Alpanet al., 2005;Koprivicaet al., 2000). However, some important generalizations can be made. For example, a common mutant allele, N370S, is exclusively associated with type 1 Gaucher disease, and has been termed the “neuroprotective allele.” Patients homozygous for mutation L444P generally have some neurologic involvement. Yet there are other examples, such as among L444P homozygotes, where patients sharing the same DNA mutations manifest with vastly different symptoms (Goker-Alpanet al., 2005). Siblings with the same genotypes can have different disease manifestations, organ involvement, and responses to therapy. There are even reports of identical twins with differing disease severity (Lachmannet al., 2004). Conversely, patients with similar or unique phenotypes can have multiple different genotypes. Thus it is clear that modifier genes and environmental factors must play a role (Sidransky, 2004).

Our attempts to better understand the phenotypes associated with glucocerebrosidase deficiency have led to an expanded awareness of the clinical diversity in this single gene disorder. Novel disease manifestations continue to be noted and described. This phenomenon is not unique to Gaucher disease. As our ability to establish rare diagnoses has improved, so too has our appreciation of the ends of the phenotypic spectra in these disorders.

Patients with Gaucher Disease and Parkinsonism

During the past decade, clinicians treating patients with Gaucher disease observed that some adult patients develop parkinsonian symptoms. Previously, while a number of case reports and case studies describing such patients were published, this association had largely been viewed as coincidental. And, although it is true that patients with rare diseases are not immune from common disorders, the rigorous pursuit of the extent and nature of this association has ultimately proven quite worthwhile.

The first publications establishing this association focused on individuals or small groups of patients with Gaucher disease manifesting parkinsonian syndromes (Bembiet al., 2003;Machaczkaet al., 1999;Neudorferet al., 1996;Tayebiet al., 2001;Varkonyiet al., 2003). A series of 17 patients with Gaucher disease who developed parkinsonism was then assembled (Tayebiet al., 2003). While among the 17 cases, 12 differentGBA genotypes were represented, the common N370S allele was the most frequentGBA mutation identified, demonstrating that referring to this allele as “neuroprotective” could result in confusion. Generally, patients in this series were noted to have a relatively early onset of Parkinson’s disease (PD), mild Gaucher manifestations, and relatively classic parkinsonian manifestations, and demonstrated a mixed response to dopamine. Subsequently, other cases have been studied and published, exhibiting a wide range of parkinsonian manifestations with variations in both the age of onset and treatment response.

Another observation made was that probands with Gaucher disease appeared to have a greater number of relatives with parkinsonism (Goker-Alpanet al., 2004;Halperinet al., 2006). Often this was a parent or grandparent who was a known or obligate carrier.

Next, a serendipitous finding launched this study in a new direction. While evaluating a pathologic sample from a subject with Gaucher disease and PD, two “control” samples from subjects with PD alone were also assayed. All three brain samples were noted to have deficient glucocerebrosidase when compared to disease-free controls. Next, DNA was isolated from the three specimens, and surprisingly, all carried mutations inGBA. One subject was an N370S homozygote, while the other two were heterozygotes. This immediately prompted a further study, conducted using brain bank samples preserved from 57 subjects with pathologically confirmed PD (Lwinet al., 2004). In this study, DNA extracted from the brain samples was sequenced, demonstrating the presence of aGBA mutation in 12 samples (14%), and two other alterations, T369M and E326K, were identified in four of the cases. This, together with the observations made in Gaucher families, led to the suggestion that even heterozygosity forGBA mutations might be associated with parkinsonism. In a subsequent publication from Northern Israel, investigators screened 99 Ashkenazi patients with idiopathic PD and 1,543 healthy Ashkenazi Jews for six specificGBA mutations commonly found in this population. 31.3% of the patients with PD carried one of these mutations, versus 6.2% of healthy controls (p<0.001) (Aharon-Peretzet al., 2004).

Following these initial studies, different cohorts with PD from around the world have been screened forGBA mutations. Some investigators focused on the presence or absence of specific common Gaucher alleles (most often N370S and L444P), but others sequenced the entireGBA gene (Braset al., 2009;Clarket al., 2007;Eblanet al., 2006;Gan-Oret al., 2008;Kalinderiet al., 2009;Lesageet al., 2011;Mataet al., 2008;Mitsuiet al., 2009;Neumannet al., 2009;Nicholset al., 2009;Satoet al., 2005;Tanet al., 2007;Toftet al., 2006;Wuet al., 2007;Ziegleret al., 2007). While the actual numbers varied depending upon screening method and ethnicity, each of the studies reported a higher frequency ofGBA mutations among both Ashkenazi Jewish and non-Jewish subjects with PD when compared to matched controls, although in some cases the results were not statistically significant. Overall, the frequency of heterozygousGBA mutations ranged from 10.7% to 31.3% among Ashkenazi Jewish cases with PD, and 2.3% to 9.4% in non-Ashkenazi Jewish patients. Generally, the carrier frequency forGBA mutations among Ashkenazi Jews is cited as between 1 in 14 and 1 in 18, and mutation N370S accounts for roughly 70% of the mutant alleles (Beutler, 1992). In other ethnic groups the mutation rate can be quite variable, but generally mutations are found in less than 1% of the population and a vast range ofGBA mutations are encountered. The frequency ofGBA mutations was also explored in familial PD, and here too mutations were more frequent than in controls (Mitsuiet al., 2009;Nicholset al., 2009).

However, these initial studies were met with considerable skepticism. Many were not as large as desired, and others had deficiencies as a result of factors including screening for only a limited number ofGBA mutations, the lack of appropriate controls, the uncertain or mixed ethnicity of samples, and inaccurate definitions ofGBA mutant alleles. To address these concerns, a large multicenter collaborative study was assembled. The final study included patients from 16 different research centers in four continents, totaling 5,691 patients with PD and 4,898 controls (Sidranskyet al., 2009). All patients were screened for two mutations, N370S and L444P, which were identified in 15% of 780 patients and 3% of 387 controls of Ashkenazi Jewish ancestry, while in patients with PD from other ethnicities, the combined frequency of one of these two mutations was 3%. Among approximately 1,700 non-Ashkenazi Jewish patients with PD whereGBA was fully sequenced, it was determined that 7% were mutation carriers. Overall, the odds ratio for carrying aGBA mutation in subjects with PD was 5.43 (95% CI, 3.89 to 7.57), conclusively rendering mutations in this gene a common risk factor for Parkinson’s disease.

GBA Mutations in Other Lewy Body Disorders

These investigations were also extended to other Lewy body disorders including Lewy body dementia [encompassing dementia with Lewy bodies (DLB) and Lewy body variant Alzheimer disease] and multiple system atrophy. Each of these disorders is characterized by the deposition of fibrillated α-synuclein in inclusion bodies found primarily in the brainstem or cortical regions (McKeithet al., 2005;Wenning and Jellinger, 2005). Several studies evaluatingGBA in cases with pathologically confirmed DLB reported the presence ofGBA mutations in 6–28% of cases screened (Clarket al., 2009;Farreret al., 2009;Goker-Alpanet al., 2006). AlthoughGBA mutations were not found exclusively in cases with Lewy bodies,GBA carriers are significantly more likely to have Lewy bodies as a pathological finding. However, no association was shown in three different studies of multiple system atrophy (Goker-Alpanet al., 2006;Jamroziket al., 2010;Segaraneet al., 2009).

Clinical Features ofGBA Mutation Carriers

The first patients reported with Gaucher disease and Parkinson’s disease developed parkinsonian features in their forties and fifties, at an age earlier than most sporadic Parkinson’s disease cases (Tayebiet al., 2003). Other studies report that the age of onset of motor impairment was 1.7 to 6 years earlier inGBA mutation carriers than in patients without mutations (Aharon-Peretzet al., 2004;Clarket al., 2007;Neumannet al., 2009;Nicholset al., 2009;Tanet al., 2007;Wuet al., 2007). Even when focusing on subjects where PD developed before the age of fifty,GBA mutation carriers had an earlier age of onset of clinical symptoms (Clarket al., 2007;Gan-Oret al., 2008).

Several reports using clinical evaluation scales including the unified Parkinson’s disease rating scale (UPDRS), mini-mental state examination (MMSE), and Hoehn and Yahr, indicated that there was no significant difference in clinical manifestations and disease progression betweenGBA carriers and controls (Sidranskyet al., 2009). However, there have been recent studies suggesting a higher frequency of cognitive decline (Alcalayet al., 2010;Goker-Alpanet al., 2008a;Neumannet al., 2009;Seto-Salviaet al., 2012), bradykinesia (Gan-Oret al., 2009), olfactory dysfunction (Brockmannet al., 2011;Goker-Alpanet al., 2008a;McNeillet al., 2012), and a lower frequency of rigidity (Clarket al., 2007) associated withGBA mutations.

Several groups have now pursued brain imaging studies to better understand the disease mechanisms underlyingGBA-associated parkinsonism (Goker-Alpanet al., 2012;Konoet al., 2010). Initially, several case reports using positron emission tomography (PET) scanning investigated dopaminergic function inGBA-associated parkinsonism and demonstrated presynaptic dopaminergic dysfunction characteristic of patients with PD. In a recently published study from the National Institutes of Health, 107 study participants were divided into four groups: subjects with both Gaucher disease and Parkinson’s disease; those with Parkinson’s disease withoutGBA mutations; those with Gaucher disease and a family history of parkinsonism; and healthyGBA-mutation carriers with a family history of Parkinson’s disease (Goker-Alpanet al., 2012). Each group was matched with a comparison group of healthy, age-matched controls. PET imaging showed that subjects with Parkinson’s disease, with and withoutGBA mutations, had a similar reduction in dopamine levels. However, cerebral blood flow was reduced in patients with Gaucher disease with parkinsonism, particularly in specific brain regions affected in subjects with different forms of dementia, including Alzheimer’s disease. This finding provides biologic confirmation of the clinical impression of impaired cognition in some patients withGBA-associated Parkinson’s disease. Furthermore, among 7 mutation carriers and 14 patients with Gaucher disease without Parkinson’s disease, only two had reduced dopamine level. The observation that most of the patients withGBA mutations did not demonstrate early signs of Parkinson’s disease may provide some reassurance to those considered at-risk.

Possible Mechanisms for This Association

Elucidating the nature of the relationship betweenGBA mutations and the pathogenesis of Parkinson’s disease and other Lewy body disorders remains a challenge. Both gain- and loss-of-function theories have been postulated (Westbroeket al., 2011;Hardyet al., 2009). Since the clinical features and neuropathology seen resemble other synucleinopathies, glucocerebrosidase could contribute by enhancing α-synuclein aggregation. Most mutant alleles identified in patients withGBA associated Parkinson’s disease result in a misfolded protein, supporting a gain-of-function role for these mutations. This abnormal conformation could contribute to parkinsonism by compounding α-synuclein aggregation. However, the misfolded protein could contribute in other ways. Among those considered are impairment of lysosomal function and overburdening of the ubiquitin-proteasomal pathways. Alterations in autophagy, reported in several neurodegenerative lysosomal storage disorders, could also result (Sun and Grabowski, 2010). It has also been proposed that parkin, an E3 ubiquitin ligase, may play a role in the degradation and accumulation of mutant glucocerebrosidase (Ronet al., 2010).

Another different theory is that parkinsonism could be a consequence of glucocerebrosidase deficiency, or a loss of enzymatic function, altering lipid homeostasis and impacting lysosomal function. Glucocerebrosidase is not the only ceramide-related gene implicated in PD (Braset al., 2008). There are reports that α-synuclein can bind to other lipid raft-associated gangliosides (Fortinet al., 2004). Lipid rafts, composed of lipids in the same family as glucocerebroside, contribute to the proper trafficking of α-synuclein to presynaptic membranes.

However, none of these different theories is fully convincing. Some of theGBA mutations identified in patients with parkinsonism are null mutations, which would not have a gain-of-function. The fact that most patients with Gaucher disease never develop Parkinson’s disease, despite having a significant deficiency of glucocerebrosidase, does not support a gain-of-function role. One study reports a physical association between native glucocerebrosidase and α-synuclein that occurs only at lysosomal pH (Yapet al., 2011). Recent work performed in human tissues, cellular models and mice have attempted to further our understanding of this association and have generated exciting new theories (Choiet al., 2011;Cullenet al., 2011;Manning-Boget al., 2009;Mazzulliet al., 2011;Sardiet al., 2011;Yapet al., 2011). Great progress in this field is anticipated in the coming years.

Type 3 Gaucher Disease

Type 3 Gaucher disease is also a continuum of different clinical presentations. Some patients with type 3 Gaucher disease exhibit slowed horizontal saccadic eye movements as their sole neurologic manifestation (Sidranskyet al., 1992b). The possible role of mutations inGBA in the pathways controlling horizontal eye movements remains to be explored.

Furthermore, some patients with type 3 Gaucher disease are found to have associated developmental delays, language difficulties, and learning disabilities, which are also not well understood (Goker-Alpanet al., 2008b). A recent study of patients with type 3 Gaucher disease, all sharing the same Gaucher genotype L444P/L444P, demonstrated a range of associated phenotypes, including successful college students and children with autism-spectrum disorder, implicating a role for genetic modifiers in establishing phenotype (Goker-Alpanet al., 2005).

One interesting Gaucher phenotype is a rare variant of type 3 Gaucher disease, which appears to be associated with a specificGBA mutation, D409H. Affected patients have atypical manifestations including cardiac and/or aortic calcifications or fibrosis, impaired saccadic eye movements, and, at times, hydrocephalus and skeletal anomalies. The pathogenesis of this unusual glucocerebrosidase-associated syndrome remains an enigma (Uyamaet al., 1992).

Gaucher Disease and Myoclonic Epilepsy

Another subset of patients with type 3 Gaucher disease develops a treatment-refractory form of progressive myoclonic epilepsy (Parket al., 2003). It has been shown that this phenotype is not associated with any specific genotype, although someGBA mutations, particularly N188S, V394L, F213I, G202R, and G377S, are seen more frequently in such patients (Kowarzet al., 2005;Parket al., 2003). A recent discovery into the intracellular pathways involved in the transport of glucocerebrosidase may provide new insights into the pathogenesis of progressive myoclonic epilepsies.

Unlike many lysosomal enzymes, glucocerebrosidase does not use the mannose 6-phosphate targeting pathway to traffic to the lysosome. Instead, it was recently reported that glucocerebrosidase is transported to this organelle by LIMP-2, a lysosomal integral membrane protein (Blanzet al., 2010). Genomic studies have implicated LIMP-2 in various forms of inherited myoclonic epilepsy, including action myoclonus-renal failure (AMRF) (Balreiraet al., 2008;Dardiset al., 2009). Mutations inSCARB2, the gene encoding LIMP-2, have been identified in both the homozygous and heterozygous form in such patients (Dardiset al., 2009). In at least one patient with Gaucher disease manifesting myoclonic epilepsy, a mutation was detected in the gene encoding LIMP-2 (DePaoloet al., 2009). Moreover, several of the mutations identified in patients with Gaucher disease and myoclonic epilepsy appear to cluster around exon 4, the binding site for LIMP-2 (Blanzet al., 2010). While the association of glucocerebrosidase and myoclonus are still not clear, this link does open new areas for exploration. Thus, once again, the evaluation of rare patients with neuronopathic Gaucher disease may be instrumental in reinforcing the role of LIMP-2 in the myoclonic epilepsies and could lead to basic science studies to elucidate the consequences of mutations in this protein.

Type 2 Gaucher Disease

Type 2 Gaucher disease is a progressive neurodegenerative disorder resulting in death by age 1–3 years. However, unlike other lipid storage disorders, the degree of pathologic findings in the brain can be quite minimal. Evaluations of the first mouse model for Gaucher disease, a knock-out model, led to the recognition of a prenatal/neonatal form of type 2 Gaucher disease, presenting as hydrops fetalis or with a collodian baby phenotype (Eblanet al., 2005;Sidranskyet al., 1992a). While scattered case reports had described such infants, it was only after the description of the mouse model that it became clear that these extreme manifestations resulted from glucocerebrosidase deficiency. Subsequently, many newborns have prospectively been diagnosed with this severe and lethal form of Gaucher disease.

Moreover, ultrastructural evaluations of skin from both these mice and babies revealed unique and similar changes, consisting of disruption to the bilayer structure of the outer layers of the stratum corneum (Holleranet al., 1994). Functional studies demonstrated that the structural alterations were accompanied by an altered ratio of ceramides to glucosylceramides in the outer epidermis, resulting in abnormalities in skin barrier function. Since the structural abnormality is seen in all of the babies with type 2 Gaucher disease tested, this observation may be diagnostically useful as a specific biomarker for type 2 Gaucher disease (Chanet al., 2011). Ultimately, these studies have led to a better appreciation of the role of glucocerebrosidase in maintaining the skin barrier.

A puzzling observation was that the null-allele Gaucher mice had minimal brain pathology. Further biochemical evaluations demonstrated that a different toxic substrate, glucosylsphingosine, accumulated in brains of Gaucher mice (Orviskyet al., 2000). Levels of glucosylsphingosine were found to also be elevated in brain tissue from fetuses and infants with type 2 Gaucher disease (Nilsson and Svennerholm, 1982;Orviskyet al., 2002). In Krabbe disease, a sphingosine derivative was also found to be the toxin responsible for the brain damage that ensues. Here again, careful basic science evaluations of this rare disease may lead to a better understanding of the pathophysiology of this and related neurological disorders.

Conclusion

Thus, detailed analysis of aspects of Gaucher disease has provided insights not only applicable to this Mendelian disorder, but also very relevant to other rare and common diseases. A greater appreciation of the complexities of other “simple” Mendelian disorders is similarly likely to expand our understanding of unrelated common diseases. Moreover, such pursuits will yield new techniques and strategies that can be applied to the challenge of the genetic dissection of complex disorders, and to the design of new therapeutic approaches for common diseases.

Acknowledgments

This work was supported by the Intramural Research Programs of the National Human Genome Research Institute and National Institutes of Health. I thank Julia Fekecs for her assistance in drafting the figure and Rafi Tamargo for his help in preparing the manuscript.

Footnotes

References