The Genetics of Parkinson Disease

Abstract

Parkinson disease (PD) is the second most common neurodegenerative disorder. In most instances, PD is thought to result from a complex interaction between multiple genetic and environmental factors, though rare monogenic forms of the disease do exist. Mutations in 6 genes (SNCA, LRRK2, PRKN, DJ1, PINK1, and ATP13A2) have conclusively been shown to cause familial parkinsonism. In addition, common variation in 3 genes (MAPT, LRRK2, and SNCA) and loss-of-function mutations in GBA have been well-validated as susceptibility factors for PD. The function of these genes and their contribution to PD pathogenesis remain to be fully elucidated. The prevalence, incidence, clinical manifestations, and genetic components of PD are discussed in this review.

Parkinson Disease

Introduction

Prevalence and Incidence

Parkinson disease ([PD] OMIM #168600) is the second most common neurodegenerative disorder. The incidence is similar worldwide, with the prevalence increasing in proportion to regional increases in population longevity with more than 1% affected over the age of 65 years and more than 4% of the population affected by the age of 85 years.¹ In most instances, PD is multifactorial, likely arising from a combination of polygenic inheritance, environmental exposures, and gene–environment interactions. Approximately 20% of patients with PD report a family history of the disease and monogenic forms of PD are relatively rare.^2,3

Clinical Manifestations

Clinically, PD has traditionally been defined by the presence of cardinal motor signs: tremor, rigidity, bradykinesia, and postural instability. However, a large body of evidence now indicates that nonmotor features such as autonomic insufficiency, cognitive impairment, depression, olfactory deficits, psychosis, and sleep disturbance are very common during the course of the disease. Motor dysfunction is thought to arise from progressive loss of dopaminergic cells within the substantia nigra pars compacta and becomes evident when approximately 80% of striatal dopamine and 50% of nigral neurons are lost.⁴ Levodopa is still the most effective treatment for PD, but its use is complicated by the emergence of motor fluctuations and dyskinesias. Other treatment options include dopamine agonists, anticholingergics, amantadine, monoamine oxidase (MAO) inhibitors, and deep brain stimulation.⁵ However, neuroprotective treatment that delays or arrests neurodegeneration in PD remains an unrealized goal.⁶

Clinical Diagnosis

The clinical diagnosis of PD is typically based on the presence of cardinal motor features, absence of atypical findings suggestive of an alternate diagnosis, and response to levodopa.⁵ The diagnostic criteria most frequently used in clinical research are those of the UK Parkinson's Disease Society Brain Bank.⁷ Differentiating PD from other parkinsonian disorders such as progressive supranuclear palsy (PSP) and multiple system atrophy can be challenging early in the course of the disease. Neuroimaging techniques have been used to facilitate early diagnosis of PD but currently are not widely used in clinical practice.⁸

Neuropathological Diagnosis

The diagnosis of PD can be confirmed at autopsy. The pathological hallmarks of the disease are dopaminergic cell loss within the substantia nigra pars compacta and the presence of Lewy bodies (LBs) and Lewy neurites (collectively known as Lewy-related pathology) in vulnerable populations of neurons. Lewy bodies are intracytoplasmic inclusions and within the brain stem have a dense eosinophilic core and a clearer surrounding halo on hematoxylin and eosin staining. The principal component of LBs is α-synuclein, a small (140 amino acids) protein that is predominantly expressed in the neocortex, hippocampus, substantia nigra, thalamus, and cerebellum.⁹ Lewy neurites are nerve cell processes that contain aggregates of α-synuclein and are most numerous in the CA2/3 region of the hippocampus and in the substantia nigra.¹⁰ Lewy bodies and Lewy neurites are best visualized immunohistochemically, using an antibody to α-synuclein.¹⁰

The neuropathological changes characteristic of PD are thought to evolve sequentially across the brain, and a classification scheme comprising 6 stages has been proposed.¹¹ In the early presymptomatic stages (1 and 2), pathology is limited to the medulla oblongata and pontine tegmentum (including the dorsal motor nucleus of vagus and the locus coeruleus) and olfactory bulb. As patients become symptomatic in stages 3 to 4, the substantia nigra and other nuclei within the midbrain and forebrain, such as the nucleus basalis of Meynert, become involved. Finally, in patients who survive into the late stages of the disease (5-6), pathology advances from mesocortex into neocortex.

Genetics of PD

Introduction

Historically, PD was considered largely sporadic in nature without genetic origin. However, in the past decade, genetic studies of PD families from different geographical regions worldwide have strengthened the hypothesis that PD has a substantial genetic component. The first gene unequivocally tied to PD (SNCA, PARK1 locus) was discovered through analysis of a large multigenerational Italian family (the Contursi Kindred) in which parkinsonism segregated in an autosomal dominant pattern.^12,13 Since then a total of 18 PD loci have been nominated through linkage analysis (PARK1-15) or genomewide association studies ([GWASs]; PARK16-18; Tables 1 and 2).^3,14-26 Mutations within the genes at 6 of these loci (SNCA, LRRK2, PRKN, DJ1, PINK1, and ATP13A2) have conclusively been demonstrated to cause familial parkinsonism.²⁹ In addition, common polymorphisms within 2 of these same genes (SNCA and LRRK2) and variation in 2 other genes not assigned to a PARK locus (MAPT and GBA) are now well-validated risk factors for PD.^27,30-33 Here we discuss the most important genes relevant to PD and their associated clinical features.

Table 1. Genes/Loci Underlying Monogenic Parkinsonism.

PARK Locus	Gene	Map Position	Inheritance	Type of Parkinsonism
Well-validated loci/genes
PARK1/PARK4	SNCA	4q21	AD	EOPD
PARK2	Parkin	6q25.2–q27	AR	Juvenile and EOPD
PARK6	PINK1	1p35-p36	AR	EOPD
PARK7	DJ-1	1p36	AR	EOPD
PARK8	LRRK2	12q12	AD (incomplete penetrance)	LOPD
PARK9	ATP13A2	1p36	AR	Kufor-Rakeb syndrome
Putative loci/genes
PARK3	Unknown	2p13	AD	LOPD
PARK5	UCHL1	4p14	AD	LOPD
PARK10	Unknown	1p32	Not clear	LOPD
PARK11	GIGYF2	2q36–q37	AD (incomplete penetrance)	LOPD
PARK12	Unknown	Xq21-25	Not clear	Not clear
PARK13	Omi/HTRA2	2p12	Not clear	Not clear
PARK14	PLA2G6	22q13.1	AR	Adult onset dystonia-parkinsonism
PARK15	FBXO7	22q12–q13	AR	Early onset parkinsonian-pyramidal syndrome

Abbreviations: AD, autosomal dominant; AR, autosomal recessive; EOPD, early onset Parkinson disease; LOPD, late onset Parkinson disease.

Table 2. Susceptibility Genes/Loci for Parkinson Disease.

PARK Locus	Gene	Map Position	Risk variants	Approximate Odds Ratio for PD
Well-validated loci/genes
PARK1/PARK4	SNCA	4q21	REP1 repeat polymorphism, multiple SNPs in 3′ half of gene	1.2-1.4
PARK8	LRRK2	12q12	G2385R, R1628P	2.0-2.2
Not assigned	MAPT	17q21.1	H1 haplotype	1.4
Not assigned	GBA	1q21	>300 mutations including; N370S and L444P	5.4
Putative loci/genes
PARK16	Unknown	1q32	Multiple SNPs from GWASs	1.3-1.4
PARK17	GAK	4p16	Multiple SNPs from GWASs	1.5
PARK18	HLA-DRA	6p21.3	Multiple SNPs from GWASs	1.3

Abbreviations: GWASs, genomewide association studies; PD, Parkinson disease; SNP, single nucleotide polymorphism.

Adapted from PDGene Web site (http://www.pdgene.org) and references 18, ²⁷, and ²⁸.

Genes Associated With Autosomal Dominant PD

PARK1/PARK4: SNCA

Inheritance and clinical features

PARK1/PARK4-linked PD displays autosomal dominant inheritance, is often of early onset, and usually progresses rapidly. Affected family members sometimes have atypical clinical features including myoclonus and hypoventilation. Three missense mutations, A53T,¹² A30P,³⁴ and E46K,³⁵ duplications^36-39 and triplications^40,41 of SNCA are known (Figure 1). The A53T substitution was the first mutation identified in a large family with autosomal dominant disease.¹³ Later, A30P and E46K substitutions were identified in a German and Spanish family, respectively, with clinical features described as dementia with LBs.^34,35 SNCA missense mutations and multiplications are both extremely rare causes of familial parkinsonism.^36-41

Figure 1.

SNCA (PARK1/PARK4) gene and protein structure. All exons and functional domains are shown. All 3 point mutations identified to date are also shown. ATG indicates the beginning of the coding region (gray); Ex, exon; aa, amino acid.

Gene location and structure. SNCA

Is located on chromosome 4q22.1, has 6 exons, and encodes a 140-amino acid protein. The N-terminus consists of an amphipathic α-helical domain that associates with membrane microdomains, known as lipid rafts.⁴² The central region contains a fibrillization region and the C-terminus contains an aggregation inhibition region (Figure 1).⁴³

Gene function and expression. SNCA

Is expressed throughout the mammalian brain and enriched in presynaptic nerve terminals.⁹ The protein can adopt partially folded structures but in its native form is unfolded and can assume both monomeric and oligomeric alpha helix and β-sheet conformations, as well as morphologically diverse aggregates, ranging from those that are amorphous to amyloid-like fibrils.⁴⁴ These fibrillar moieties are a component of LBs in both familial and idiopathic PD,⁴⁵ but it is unclear whether the fibrils themselves, or the oligomeric fibrilization intermediates (protofibrils), are toxic to the cell. Interestingly, SNCA genomic multiplications in familial PD are associated with an increase in protein expression,⁴¹ and brain samples of triplication mutant carriers show protofibril formation is enhanced with an increase in SNCA expression.⁴⁶ In vitro, A30P, A53T, and E46K mutant proteins show an increased propensity for self-aggregation and oligomerization into protofibrils, compared with wild-type protein^47,48 that may be related to the membrane permeabilization activity of these protofibrils that form pore-like and tubular structures.⁴⁹ It appears that only A53T and E46K promote formation of the fibrils,^50,51 whereas A30P has been reported to disrupt the interaction between α-synuclein and the lipid raft and possibly redistributing the protein away from the synapse.⁴²

A mouse spontaneous deletion strain is viable, fertile, and phenotypically normal,⁵² whereas overexpression of wild-type SNCA in a mouse model has many features of PD, such as loss of dopaminergic terminals in the striatum, mislocalization and accumulation of insoluble α-synuclein, and motor abnormalities.^53-55 Both A30P and A53T mutant mouse models display neuronal cell loss and motor changes.⁵⁶

Genetic variation

As described in previous sections, multiplications and 3 missense mutations in SNCA have been reported to cause PD, and there is some evidence of genotype–phenotype correlation. For example, American and European families with SNCA triplication show different clinical features than families with a duplication where the phenotype closely resembles idiopathic PD, with late age-of-onset, slow progression, and no atypical features, suggesting that SNCA gene dosage might play a role in disease progression.^36,37 SNCA duplications are rarely associated with dementia.^38,39 SNCA triplications and the E46K mutation are more commonly associated with dementia than the A30P mutation and gene duplications. The A53T mutation has been associated with dementia and the presence of cortical LBs.^57,58

Several common polymorphisms in SNCA have consistently been shown to associate with PD. These variants include a complex repeat polymorphism (REP1) located approximately 10 kb upstream from the translation start site, and several single nucleotide polymorphisms (SNPs) at the 3′ end of the gene.^30,59,60 SNCA has been among the genes most significantly associated with PD in all 5 of the PD GWASs conducted in the past 2 years.^15-18,61 The mechanism by which common SNCA variants modify susceptibility for PD is not yet known. However, there is evidence to suggest that SNCA alleles associated with increased PD risk are also correlated with higher α-synuclein expression in vitro and with elevated peripheral levels of α-synuclein in vivo.^30,62

PARK8: LRRK2

Inheritance and clinical features

Autosomal dominant PARK8-linked PD was first identified in a Japanese family known as the Sagamihara kindred.⁶³ Affected individuals have clinically typical late onset PD and dementia is not a common feature.⁶⁴ Pathologically, the disease appears to be heterogeneous with the reports of LB pathology, tau pathology, neuronal loss without intracellular inclusions,^65,66 and motor neuron disease.⁶⁴

Gene location and structure

The gene for PARK8 was identified as Leucine Repeat Rich Kinase 2 (LRRK2) in 2004 (also called dardarin, from the Basque word for tremor), in families from the Basque region of Spain, Britain, Western Nebraska, and in an American kindred of German descent.^64,67 It is located on chromosome 12p12 and encompasses 144 kb, consisting of 51 exons (7449 bp complementary DNA [cDNA]), and encodes a protein consisting of 2517 amino acids. The LRRK2 gene contains several functional domains including ARM (armadillo domain), ANK (ankyrin repeat domain), LRR (leucine-repeat-rich), ROC (Ras of complex proteins), COR (carboxy terminal of ROC), MAPKKK (mitogen-activated protein kinase kinase kinase), and a WD40 domain that is rich in tryptophan and aspartate repeats.

Gene function and expression

The function of LRRK2 is not well known although it has been identified as a tyrosine kinase–like protein.⁶⁸ The ROC domain is able to bind guanosine triphosphate (GTP) and is essential for the MAPKKK domain to exert kinase activity but does not have GTPase activity.⁶⁹ Some of the LRRK2 mutations appear to exert increased kinase activity.^70,71 Other functional domains are believed to be important in protein–protein interactions.⁶⁴ LRRK2 also interacts with other familial PD proteins. For example, LRRK2 appears to interact with parkin through the ROC domain; however, the interaction with parkin does not seem to enhance polyubiquitylation of LRRK2.⁷² LRRK2 expression has been described in the central nervous system (cerebral cortex, medulla, cerebellum, spinal cord, putamen, and substantia nigra), heart, kidney, lung, liver, and peripheral leukocytes.^64,67 LRRK2 protein is found in the cytosol and mitochondrial outer membrane,⁷¹ plasma membrane, lysosomes, endosomes, transport vesicles, Golgi apparatus, a cytoskeleton protein microtubule, synaptic vesicles, and lipid rafts.^73,74 Interestingly, α-synuclein is also expressed in the presynaptic membranes and lipid rafts.⁴²

There is currently very limited postmortem data on pathogenic LRRK2 mutations, but it appears that typical LB pathology is seen in most LRRK2-related patients. One clinicopathological study reported substantia nigra cell loss, LB formation, and small numbers of cortical LBs.⁷⁵ In the same study, the 18F-dopa positron emission tomography (PET) in a proband, but not unaffected family members, showed a pattern of nigrostriatal dysfunction typical of idiopathic PD.⁷⁵ The mechanism that links LRRK2 protein to SNCA protein accumulation remains unknown, but evidence suggests that there may be a direct interaction between LRRK2 and the SNCA protein.⁷⁶

Genetic variation

Over 40 missense or nonsense mutations have been reported in LRRK2, but the pathogenicity of most of these rare variants has not yet been determined (Figure 2).^63,64,67,68 There is currently convincing evidence to suggest that 6 mutations are disease-causing (R1441C, R1441G, R1441H, Y1699C, G2019S, and I2020T). By far the most prevalent mutation is G2019S, which occurs in 1% to 2% of PD patients of European origin, 15% to 20% of Ashkenazi Jewish patients, and approximately 40% of North African Arabs with PD.^77-80 G2019S exhibits reduced penetrance, though penetrance estimates at 80 years of age vary widely from 24% to nearly 100%.^78,81-84 In contrast, the next most frequent mutations, which occur in codon 1441, are highly penetrant.⁸⁵

Figure 2.

LRRK2 (PARK8) gene and protein structure. Established pathogenic mutations (boldface) and risk variants (gray) are shown. Variants specific to Asian populations are indicated by an asterisk. ARM indciates Armadillo region; ANK, Ankyrin repeat region; LRR, leucine-rich repeat domain; ROC, Ras of complex; COR, C terminal of Ras (GTPase);. MAPKKK, mitogen-activated protein kinase kinase kinase; Ex, exon; aa, amino acid.

Two common LRRK2 polymorphisms (G2385R and R1628P) that occur only in Asian populations are now well-validated risk factors for PD.^86,87 A recent meta-analysis of 9 studies of G2385R reported a combined odds ratio in favor of PD of 2.55 (95% CI, 2.10-3.10).⁸⁸ Whether LRRK2 polymorphisms that convey risk for PD exist in other populations remains to be determined.

Genes Associated With Autosomal Recessive PD

PARK2

Inheritance and clinical features

PARK2-related parkinsonism is autosomal recessive and usually of early onset. Age at onset is typically between childhood and age 40 years.⁸⁹ Dystonia is frequently present and patients are levodopa responsive. Pathologically, the substantia nigra undergoes severe neuronal loss and gliosis, whereas the locus coeruleus is much less severely involved and usually no LBs are seen^90,91 although rare LB-positive cases have been reported.^92-94

Gene location and structure

Linkage analysis of families with autosomal recessive juvenile parkinsonism mapped the PARK2 locus to chromosome 6q26, near the sod2 locus.^95-97 By screening a BAC library using the D6S305 marker at this region, a cDNA was cloned containing the 1395 bp open reading frame of the novel PARK2 gene.⁹⁸ The PARK2 gene contains 12 exons and spans approximately 1.38 Mb.^98,99

Gene function and expression. PARK2

Encodes parkin, a 465-amino acid protein that belongs to the “ring between ring fingers” (RBR) family of E3 ubiquitin ligases. The RBR domain of these proteins is composed of 2 RING fingers linked by a cysteine-rich “in-between-RING” (IBR) motif. The RBR domain interacts with ubiquitin-conjugating enzymes (E2s) to catalyze attachment of ubiquitin to protein targets, thus tagging these proteins for destruction by the proteosome.^100,101 Many ubiquitination substrates have been proposed for parkin including the aminoacyl-tRNA synthetase cofactor, p38, and a rare, 22-kDa glycosylated form of α-synuclein.^102-104 Parkin is predominantly a cytosolic protein but also co-localizes to synaptic vesicles, the Golgi complex, endoplasmic reticulum, and the mitochondrial outer membrane.^100,104-107

Many PD-linked point mutations alter wild-type parkin cellular localization, solubility, or propensity to aggregate.^108-110 Other mutations, including insertions and deletions, result in parkin loss-of-function. Several lines of PARK2 knockout mice have been generated, but surprisingly none display nigral neuronal degeneration or signs of motor dysfunction.^111-113 Subtle behavioral abnormalities and reduced numbers of noradrenergic neurons in the locus ceruleus have been reported in some strains but not others.^112-114 In 1 knockout mouse line, reduced numbers of mitochondrial oxidative phosphorylation proteins, a decrease in mitochondrial respiratory capacity, and age-dependent increases in oxidative damage were reported.¹¹⁵ Mitochondrial defects have also been reported in parkin knockout Drosophila, suggesting that defects in parkin ubiquination function might be secondary in the course of pathogenic events.^116,117 In vitro studies of a PARK2-knockdown SH-SY5Y cell line showed apoptotic cell death and an increase in the auto-oxidized forms of levodopa and dopamine, suggesting that parkin might have important antioxidative properties.¹¹⁸

Although patients with PARK2-related parkinsonism exhibit loss of pigmented nigral dopamine neurons, LBs are usually not observed.⁹² However, there are exceptions, as nigral LBs have been reported in 2 patients with compound heterozygous mutations (a 52-year-old with R275W and an exon 3 deletion and a 73-year-old with an exon 7 deletion and del1072T) and in a patient homozygous for an exon 3 deletion.^92-94

Genetic variation

Reported mutations in PARK2 now exceed 100 including missense and nonsense mutations, as well as exonic deletions, rearrangements, and duplications.^119-124 These mutations span the entire length of the gene and include the N-terminal UBL domain and the RING-IBR-RING domain (Figure 3). The most common mutations are (1) deletions of exon 4 (n = 28), (2) deletions of exon 3 (n = 27), (3) deletions of exons 3 to 4 (n = 23), (4) a point mutation in exon 7 (924C>T; n = 38), and (5) a single base pair deletion in exon 2 (255/256delA; n = 17). These 5 variants account for 35% of all PARK2 mutations. Hot spots for mutations are concentrated in exons 2 and 7, whereas exon rearrangements are more likely to occur in introns 2 through 4¹²⁵ (Figure 3).

Figure 3.

PARK2 gene and protein structure. More than 100 mutations have been identified. Number of mutations in each exon is shown as reported in the PD mutation database http://grenada.lumc.nl/LOVD2/TPI/home.php?select_db=PARK2. Hot spots for parkin mutations are concentrated in exons 2 and 7, whereas hot spots for exon rearrangements occur in introns 2 through 4. ATG indicates the beginning of the coding region (gray); UBL, ubiquitin-like domain; Ex, exon; aa, amino acid.

PARK2 mutations account for over 50% of patients with juvenile onset parkinsonism (age at onset ≤20 years), but mutation frequency diminishes substantially with increasing age at onset.^89,126,127 In late onset PD (>50 years), the proportion of patients homozygous or compound heterozygous for PARK2 mutations is <1%.^123,128,129

Whether simple heterozygotes for PARK2 mutations are at increased risk for PD remains controversial.¹³⁰ Some studies have reported a significantly higher frequency of PARK2 mutations in patients with PD versus controls.^131,132 Also, an analysis of 183 PD families found that the mean age at onset in affected individuals with 1 PARK2 mutation was 11.7 years earlier than affected individuals with no mutations.¹³³ Consistent with these data, Khan and colleagues reported that asymptomatic heterozygous PARK2 carriers show significant striatal dopaminergic dysfunction by 18F-dopa PET, suggesting that such individuals might have haploinsufficiency.¹³⁴ However, the largest case–control study conducted to date, which included 2091 unselected patients with PD and 1686 controls found no significant difference in mutation frequency between the 2 groups.¹²⁹ Further larger scale studies will be needed to determine the role of heterozygous PARK2 mutations in the modifying risk of PD.

PARK7: DJ1

Inheritance and clinical features. DJ1

Recessively inherited missense and exonic deletion mutations were first identified in 2 European families with an age of onset of 20 to 40 years.¹³⁵ PARK7-linked PD appears to be very rare.^135-137 Very few DJ1 patients have been reported in the literature and thus there is limited knowledge on the clinical features, neuropathology, and genotype–phenotype correlation for DJ1-related PD. In addition to parkinsonism, clinical characteristics that have been reported in some patients include psychiatric symptoms,¹³⁸ short statue, and brachydactyly.¹³⁹

Gene location and structure. DJ1

Has been cloned and is located on chromosome 1p36.23. It has a transcript length of 949 bps with 7 exons.¹⁴⁰ It encodes a protein consisting of 189 amino acids.¹⁴¹

Gene function and expression. DJ1

Is a homodimer that belongs to the peptidase C56 family of proteins.¹⁴² It is a cytoplasmic protein but can also translocate into the mitochondria¹⁴³ and appears to act as an antioxidant.^140,144-146 Its antioxidant properties may depend on a cysteine residue at position106, which on oxidation forms a disulphide bond.¹⁴⁵ DJ1 might act as either a redox-sensor protein that can prevent the aggregation of α-synuclein or an antioxidant.^145,147-151 It might also act as a reactive oxygen species scavenger through auto-oxidation.¹⁵² These proposed functions for DJ1 could be particularly important in nigral dopamine neurons that are exposed to particularly high levels of oxidative stress.

Expression of DJ1 is ubiquitous and abundant in most mammalian tissues including in the brain where it is found in both neuronal and glial cells.¹⁵³ Downregulation of endogenous DJ1 protein in neuronal cell lines by small interfering RNA (siRNA) enhances oxidative stress-induced cell death, ER stress, and proteasome inhibition but not by proapoptotic stimulus.^152,154 The L166P mutant protein has a reduced antioxidative activity.¹⁵⁵ Mutant DJ1 appears to interact with parkin,¹⁴⁶ whereby parkin acts as an E3 ligase to remove mutated DJ1. DJ1 null mice are hypersensitive to oxidative stress and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyrindine (MPTP).¹⁵⁶ DJ1 protein expression is increased on oxidative stress induced by paraquat.¹⁴⁹

DJ1 does not appear to be an essential component of LBs as only a small proportion of LBs from PD brains display DJ1 immunoreactivity.¹⁵³ Although DJ1 mutations are rare even in early onset PD, recent studies suggest that DJ1 protein might play an important role in sporadic late-onset PD. Sporadic PD brain has greater oxidative damage to DJ1 and a significant increase in total protein levels, compared with normal controls.¹⁵⁷

Genetic variation

In DJ1-related parkinsonism, mutations are found in the homozygous or compound heterozygous state, putatively resulting in a loss of protein function. The L166P mutation causes destabilization through unfolding of the C-terminus, inhibiting dimerization and enhancing degradation by the proteasome.^142,158,159 In addition, probably consequent to instability, L166P reduces the neuroprotective function of DJ1.¹⁵² Reduced nuclear localization, in favor of the mitochondria, is also seen for L166P, as well as for the M26I and D149A mutations.^141,160 The mutations L166P, E64D, M26I, A104T, and D149A have been shown to create structural perturbations of DJ1 protein that lead to global destabilization, unfolding of the protein structure, heterodimer formation, or reduced antioxidant activity.^155,161,162

PARK6: PINK1

Inheritance and clinical features

PARK6 was first mapped to chromosome 1p35-p36 in a large consanguineous Italian family with autosomal recessive, early onset PD. Subsequently, phosphatase and tensin homolog (PTEN)-induced putative kinase 1 (PINK1) was determined to be the disease-causing gene. PINK1 mutations have been reported to account for approximately 1% to 3% of early onset PD in populations of European ancestry,^163-167 8.9% of autosomal recessive PD in a sample of Japanese families,¹⁶⁸ and 2.5% of early onset PD in a sample of ethnic Chinese, Malays, and Indians.^169,170 Although age at onset for PINK1-related PD is usually in the fourth to fifth decade, clinical features are similar to late onset PD, with slow progression, excellent response to levodopa, and in some instances dementia.^{164,167,170,171} There is some evidence to suggest that heterozygous carriers of PINK1 mutations might be at increased risk for PD, but definitive large-scale studies have not yet been performed.^166,172 In support of this idea, heterozygous PARK6 carriers have been reported to have a 20% to 30% reduction in caudate and putamen 18F-dopa uptake in comparison with controls.¹⁷³

Gene location and structure. PINK1

Is located on chromosome 1p36.12, has 8 exons, and a cDNA that spans 1.8 kb. It encodes a protein with 581 amino acids. It has a serine/threonine protein kinase domain. However, its function is not known.¹⁶⁴ It is a mitochondrial protein located in the matrix and the intermembrane space that is ubiquitously expressed in the brain and systemic organs and contains a mitochondrial-targeting motif and a conserved serine/threonine kinase domain.¹⁷⁴

Gene function and expression

Functional studies have shown that PINK1 is localized to mitochondria both in vitro and in vivo.¹⁷⁵ Wild-type PINK1 appears to be important in neuroprotection against mitochondrial dysfunction and proteasome-induced apoptosis, whereas the G309D mutation impairs this protective effect, possibly by interfering with adenosine diphosphate (ADP) binding and thus inhibiting kinase activity.^164,176 E240K and L489P mutants disrupt PINK1's protectivity by either enhancing the instability of the protein or disrupting the kinase activity of the protein.¹⁷⁷ In vitro studies indicate that cells transfected with PINK1 mutants have disrupted mitochondrial membrane potential under stressful conditions.¹⁷⁸ Knockout models of the Drosophila PINK1 ortholog have defects in mitochondrial morphology and increased sensitivity to oxidative stress and appear to be rescued by human parkin.¹⁷⁹

Genetic variation

The first mutations discovered were the G309D missense and a W437X truncating mutation found in the families of Italian and Spanish descent.^164,176 Since then, several point mutations, frameshifts, and truncating mutants have been identified.^167,170,180 Interestingly, in contrast to PARK2, the majority of PINK1 mutations reported are either missense or nonsense mutations.^{165,168,171,176,181} One family with a large homozygous deletion has been reported involving exons 6 to 8.¹⁶⁸ Most of the reported point mutations in PINK1 are located in a highly conserved amino acid position in the protein kinase domain.¹⁷⁸ However, in a large Sudanese kindred with early onset parkinsonism and dopa-responsive dystonia, the pathogenic mutation (A217D) is located in the highly conserved adenosine triphosphate (ATP) orientation site of the PINK1 kinase domain.^182,183

PARK9: ATP13A2

Inheritance and clinical features

Homozygous and compound heterozygous mutations in the P-type ATPase gene (ATP13A2) have been demonstrated in a Jordanian family^184,185 and a Chilean family¹⁸⁶ with Kufor-Rakeb syndrome, a form of recessively inherited atypical parkinsonism which is clinically characterized by very early age of onset (11-16 years), levodopa-responsive parkinsonism, pyramidal signs, dementia, and supranuclear gaze palsy^184,187 Magnetic resonance imaging (MRI) shows significant atrophy of the globus pallidus and the pyramids, and generalized brain atrophy in later stages. Some develop facial, faucial, and finger mini-myoclonus, visual hallucinations, and oculogyric dystonic spasm.¹⁸⁷

Gene location and structure

The disease locus designated as PARK9 was mapped to 1p36 with a maximum logarithm of the odds (LOD) score of 3.6, a hot spot for autosomal recessive familial PD. The disease gene was subsequently identified as ATP13A2. The transcript has 29 exons and is 3854 bps in length. The ATP13A2 protein contains 1180 amino acids and has 10 transmembrane domains.¹⁸⁶

Gene function and expression

ATP13A2 is a lysosomal membrane protein with an ATPase domain.¹⁸⁶ It is a member of the P5 subfamily of ATPases, which transports inorganic cations and other substrates. The exact function of the ATP13A2 protein is still unknown. ATP13A2 is predominantly expressed in brain tissue with the highest levels reported in ventral midbrain (which includes the substantia nigra).¹⁸⁶ ATP13A2 mRNA levels are approximately 10-fold higher in the nigral dopamine neurons of sporadic patients than control participant brains

Genetic variation

All known ATP13A2 mutations appear to directly or indirectly affect transmembrane domains.¹⁸⁶ In vitro evidence indicates that wild-type ATP13A2 is localized to the lysosome membrane of transiently transfected cells, whereas unstable truncated mutants are retained in the endoplasmic reticulum and degraded by the proteasome.¹⁸⁶ A homozygous missense mutation (G504R) has been identified in 1 sporadic case from Brazil with juvenile parkinsonism.¹⁸⁸ This patient had symptom onset at age 12, levodopa-responsive severe akinetic-rigid parkinsonism, levodopa-induced motor fluctuations and dyskinesias, severe visual hallucinations, and supranuclear vertical gaze paresis, moderate diffuse atrophy but did not have dementia or pyramidal deficits. In this same study, 2 Italian patients with early onset PD without atypical features each carried a single novel missense mutation (T12M or G533R),¹⁸⁸ raising the question of whether heterozygous ATP13A2 mutation carriers are at increased risk of PD.

Genetic Risk Factors for PD

Common variants in 3 genes (LRRK2, MAPT, and SNCA) and loss-of-function mutations in another (GBA) are now well-established risk factors for PD. In addition, 3 new putative susceptibility loci for PD (PARK16-18) have been identified from recent GWASs. Of these 7 genes/loci, LRRK2 and SNCA have been described in previous sections. In this section, we briefly describe the remainder.

GBA

Loss of function mutations in the glucocerebrosidase (GBA) gene, which encodes the enzyme glucocerebrosidase, result in Gaucher disease (GD), an autosomal recessive glycolipid storage disorder with multisystemic manifestations, including involvement of the liver, spleen, bone marrow, lungs, and nervous system.¹⁸⁹ Nearly 300 pathogenic mutations have been described, but in populations of European origin, 2 mutations (N370 and L444P) account for approximately two thirds of the disease alleles seen in GD.^189,190 A small subset of patients with GD develop parkinsonism with brain stem or diffuse Lewy-related pathology.¹⁹¹ An increased incidence of parkinsonism has also been reported in relatives of patients with GD.¹⁹² These observations prompted a number of case–control association analyses seeking to determine whether GBA mutations are a risk factor for PD. In 2009, Sidransky and colleagues published the definitive study on this topic, a pooled analysis of 5691 patients with PD and 4898 controls from 16 centers in 12 countries.²⁷ In the subset of participants in which the entire GBA coding region was screened, loss-of-function mutations were observed in 6.9% of cases and 1.3% of controls (odds ratio, 5.4; 95% CI, 3.9-7.6). Among the subset of individuals of Ashkenazi Jewish ancestry, higher mutation frequencies were seen: 19.3% in cases and 4.1% in controls. The clinical characteristics of patients having PD with and without GBA mutations were very similar. There is also evidence to suggest that GBA mutations are a risk factor for dementia with LBs.^193,194

The mechanism by which GBA mutations modify risk for PD is not yet known, but both gain-of- and loss-of-function hypotheses have been proposed. One putative gain-of-function mechanism is based on data suggesting that some GBA mutations result in misfolded protein.¹⁹⁵ Misfolded glucocerebrosidase might then contribute to neurodegeneration by inducing lysosomal insufficiency, by impairing autophagic pathways necessary for degrading α-synuclein, or by overwhelming the ubiquitin–proteasome pathway.¹⁹⁶ A loss-of-function hypothesis recently proposed centers on observations that α-synuclein binds to lipids in the plasma membrane,¹⁹⁷ and that binding of lipid might reduce the formation of fibrillar forms of α-synuclein. Glucocerebrosidase degrades glucocerebroside to ceramide, and thus GBA haploinsuffiency might cause this and other polyunsaturated lipids to accumulate and alter the sphingolipid composition of cell membranes. This in turn could disrupt membrane binding of α-synuclein, thus enhancing its aggregation in the cytoplasm.¹⁹⁶

Genetic Variation

Microtubule-Associated Protein Tau

The microtubule-associated protein tau, encoded by the MAPT gene, is primarily expressed in neurons and plays a key role in the organization and integrity of the cytoskeleton.¹⁹⁸ Filamentous neuronal tau inclusions define a set of neurodegenerative diseases referred to as the “tauopathies,” which include Alzheimer disease, corticobasal degeneration (CBD), PSP, and frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17).^199,200 MAPT was first linked to the pathogenesis of tauopathies by the discovery of mutations resulting in FTDP-17 and subsequently an extended common haplotype (H1) across the gene was shown to associate with disease risk in PSP and CBD.^201-203 Because PD shares some clinical features with the tauopathies, a number of studies have been conducted over the past decade on the relationship between common MAPT variants and PD risk. Two large case–control studies published in 2007 provided strong evidence that the MAPT H1 haplotype is associated with PD risk, and a metaanalysis of 18 studies yield an odds ratio for PD (under a dominant model) of 1.40 (95% CI, 1.30-1.50; P = 2 × 10⁻¹⁹).^28,204 Subsequently, the MAPT H1 haplotype has been confirmed as a PD risk factor in all 4 of the GWASs conducted in populations of European origin in the past 2 years.^16-18,61

The MAPT H1 and H2 haplotypes actually represent 2 distinct families of subhaplotypes, which arose from an inversion of 900 kb on chromosome 17q21 approximately 3 million years ago.²⁰⁵ One such subhaplotype within the H1 family, designated “H1c,” is clearly associated with PSP and CBD.²⁰⁶ However, this same subhaplotype is not associated with PD, so it appears that different MAPT variants convey risk of PD versus the tauopathies.²⁰³

In FTDP-17, MAPT mutations are thought to alter tau function in a manner that promotes tau fibrillization and impairs its ability to regulate microtubule dynamics, thus causing neurotoxicity.^199,207 However, the mechanism by which common MAPT variation modifies risk of PD, a disease in which tau aggregates are rarely seen, remains to be determined.

PARK16-18

Recent GWASs have nominated 3 new susceptibility loci. Satake and colleagues reported an association between multiple SNPs on chromosome 1q32 and PD in a large, 2-tiered GWASs of participants of Japanese ancestry.¹⁵ This region, designated PARK16, contains several candidate genes, including RAB7L1, a small GTP-binding protein that regulates exo- and endocytotic pathways, and NUCKS1, a nuclear DNA-binding protein expressed in brain and other tissues that might regulate chromatin structure and its activity.^208-210

Pankratz and colleagues performed a GWASs in 857 familial PD cases and 867 controls.¹⁷ Although none of their results met genomewide significance, the region most significantly associated with PD under an additive model occurred on chromosome 4 (rs11248060, odds ratio 1.69, P = 3.4 × 10⁻⁶). This region was later replicated in a second, larger GWASs and designated PARK17.¹⁸ The PARK17 locus contains several genes, include GAK (cyclin G-associated kinase), a promising candidate gene that functions as a cell cycle regulator and is differentially expressed within the substantia nigra of PD brains in comparison to controls.^17,211

Finally, the most recently conducted GWASs by Hamza and colleagues nominated the HLA region, designated PARK18, as a putative susceptibility locus for PD.¹⁸ This fits well with a growing body of literature, indicating that chronic inflammation and humoral immunity play a role in the pathogenesis PD.^212,213

Because of concerns for type I error that are inherent in all GWASs, PARK16-18 must be replicated in multiple independent samples before being considered bona fide PD susceptibility loci.

Summary and Conclusions

A total of 6 genes (SNCA, LRRK2, PRKN, DJ1, PINK1, and ATP13A2,) have conclusively been linked to monogenic forms of parkinsonism, and all but ATP13A2 result in disease that closely resembles the clinical features of idiopathic PD. Common variation in 3 genes (MAPT, LRRK2, and SNCA) and loss-of-function mutations in GBA are now well-established risk factors for PD. Recent PD GWASs have nominated 3 new putative susceptibility loci (PARK16-18), though these findings require rigorous replication before firm conclusions can be drawn.

Additional susceptibility loci will likely be uncovered in the near future, as the wealth of recent data from GWASs is further analyzed. Such efforts will include meta-analysis, consideration of gene × gene and gene × environment interaction, and analysis of copy number variation. Traditional linkage analysis will also continue to be an invaluable tool for gene discovery, and efforts to examine understudied populations across the world could prove particularly fruitful.

Although important progress has been made, the mechanisms by which variation in PD-linked genes leads to neurodegeneration remains poorly understood. However, data accumulated thus far has implicated mitochondrial dysfunction, oxidative damage, aberrant protein aggregation, and deficits in ubiquitin-mediated protein degradation as playing key roles in the etiopathogenesis of PD.