Main

Parkinsonism is a broad term referring to all clinical states characterized by tremor, slowed movement (bradykinesia), rigidity, and postural instability. Parkinson disease (PD) is the primary and most common form of parkinsonism, and is the second most common neurodegenerative disorder after Alzheimer disease (AD). It affects more than 1% of 55-year-old individuals and more than 3% of those over 75 years of age.1 The age of disease onset is widely variable, ranging from juvenile to very late in life with an average age of onset of 60 years. Generally, individuals with onset before age 20 are considered to have juvenile-onset; those with onset between 20 and 50 years of age are classified as having early-onset; and those with onset after age 50 are referred to as late-onset.

The overall age- and gender-adjusted incidence rate is 13.4 per 100,000. There is a higher prevalence among men (19.0 per 100,000) than among women (9.9 per 100,000).2 PD seems to have a similar incidence across most ethnicities; however, it may be less common among African Americans.2

Psychiatric manifestations can be a prominent feature of disease and may include depression and visual hallucinations. Depression occurs in 25–50% of PD patients.3–5 Later in disease progression, dementia eventually occurs in 20–40% of cases.6

A number of risk factors have been evaluated for their role in disease susceptibility. Smoking has consistently been reported to result in a 50% decrease in the risk of disease.7–11 The effect of smoking is dose-dependent and temporal, with the protective effect of smoking greater for those who quit later in life.8,10,11 Caffeine has a similar well-documented, dose-dependent protective effect on PD.8,12–14 The protective effect of caffeine in women seems to be modulated by the use of postmenopausal hormones, with protection only being conferred in women that did not undergo estrogen replacement therapy.15,16 Serious head trauma has also been found in multiple studies to increase the risk of PD.17,18 Other factors that have been inconsistently reported to modulate disease risk include well water,19,20 pesticide use,21–23 and rural living.19,24 The results of 29 studies are summarized in Lai et al.25

In the mid 1980s, a contaminate found in a synthetic form of heroin was found to cause rapid-onset, levodopa-responsive parkinsonism.26 The responsible toxin, 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), is metabolized and then preferentially enters dopaminergic neurons at terminal reuptake sites. Toxicity results from the inhibition of complex I of the electron transport chain, thus implicating mitochondrial dysfunction as a notable cause of PD. Subsequently, it was shown that complex I activity is selectively diminished in the substantia nigra of patients with idiopathic PD as well.27 The identification of an environmental agent that can induce parkinsonism also provided a means of creating animal models of PD. Various doses and protocols have been able to mimic acute and chronic forms of neurodegeneration by eliciting neurochemical changes such as decreased levels of dopamine, increased oxidative damage, decreased concentrations of antioxidants and aggregates containing alpha-synuclein, a principal component of Lewy bodies.28 Another important contribution from the study of MPTP is the reminder that PD has a very complex etiology. Although over 400 individuals were thought to have been exposed to MPTP, only a small fraction developed this severe form of parkinsonism,29 thus illustrating the multifactorial nature of disease risk.

Studies from around the world have provided evidence that genetic risk factors are involved in the pathogenesis of the idiopathic form of PD. Estimates of the relative risk to first-degree relatives of an affected individual range from 2.7 to 3.5 in the United States,30,31 2.9 in Finland,32 6.7 in Iceland,33 7.7 in France,34 3.2 in three centers within Europe,35 5.0 in Canada,36 13.4 in Italy,37 and 7.1 in Germany.38

DIAGNOSTIC FEATURES OF PD

The cardinal pathologic feature of PD is the loss of dopaminergic neurons in the substantia nigra with intracytoplasmic inclusions (Lewy bodies) in the remaining, intact nigral neurons.39,40 Symptoms of PD typically onset when 50–80% of the dopaminergic neurons in the substantia nigra are no longer functioning. Traditionally, the presence of Lewy bodies was required for pathologic confirmation of PD; however, it has been recognized recently that nigral pathology may occur in the absence of Lewy bodies.

Because a diagnosis of PD can only be made through documentation of salient clinical features and postmortem verification of Lewy bodies, some diagnostic uncertainty is unavoidable. The careful application of diagnostic criteria derived from existing clinicopathologic studies can increase the positive predictive value of diagnosis to over 95%.41 Presence of resting tremor, response to dopamine agents, asymmetric onset of symptoms, and the absence of atypical features that suggest other diagnoses are all criteria that can be used to increase the certainty of diagnosis. Functional imaging techniques such as positron emission tomography or single photon computed emission tomography using radioactively labeled ligands of the presynaptic dopaminergic neurons can support the diagnosis but are usually limited to a research setting. By maximizing the specificity of the criteria, the sensitivity of the criteria falls dramatically, thereby excluding as many as one third of true cases.42 Although these diagnostic criteria are ideal for a genetic research study, they may not be useful for making a clinical diagnosis.

Several other neurologic entities must be considered as part of the differential diagnosis of PD. They include parkinsonism-predominant multiple system atrophy (formerly called striatonigral degeneration), progressive supranuclear palsy, corticobasal degeneration, essential tremor, drug-induced parkinsonism, postencephalitic conditions, Lewy body dementia, and AD. Parkinsonism can also be a prominent feature of some Mendelian disorders, including spinocerebellar ataxias (i.e., MJD/SCA3 and SCA2), Huntington disease, dopa-responsive dystonia, familial prion disease, frontotemporal dementia with parkinsonism-17, Wilson disease, and X-linked dystonia-parkinsonism syndrome (Lubag). Laboratory or radiologic studies cannot be used to confirm PD; however, they are useful in excluding alternative diagnoses, such as stroke, tumor, and thyroid disease.

GENETICS OF PD

For many years PD was thought to be strictly influenced by environmental factors without a substantial genetic contribution to disease etiology. However, research has since demonstrated the importance of genetic factors in at least a subset of PD patients (Table 1). Linkage analyses in both autosomal dominant and autosomal recessive kindreds have identified the genes resulting in Mendelian (single gene) forms of parkinsonism. Mutations in SNCA (PARK1) and LRRK2 (PARK8) result in autosomal dominant PD. Linkage to a third locus (PARK3) has been reported, but the gene has not yet been identified. A fourth locus, UCHL1 (PARK5) has been implicated, but not confirmed. Mutations in three known genes, PRKN (PARK2), DJ-1 (PARK7), and PINK1 (PARK6), result in autosomal recessive PD. The role of these genes and the clinical characteristics of patients with mutations are described in more detail below. Importantly, the mutations identified to date in these genes likely cause PD in fewer than 5% of patients.

Table 1 Causative genes for Parkinson disease
Full size table

Our understanding of the genetic contribution to the risk of PD is likely still incomplete, despite the identification of several PD causative genes. Therefore, analyses are ongoing that seek to identify genetic variants that increase or decrease the risk of PD. Typically, these analyses have focused on genes that are not causative but rather are susceptibility loci. Several susceptibility genes have also been implicated; although their role in most cases is not yet conclusively replicated.

Autosomal dominant PD

PARK1: SNCA

The first mutation identified in PD subjects was found in an autosomal dominant pedigree of Italian descent, called the Contursi kindred.43 Members of this family have a mutation in the gene (SNCA; A53T in exon 4) that codes for the protein alpha-synuclein. The affected members of the family have similar clinical and pathologic findings to those with idiopathic PD, including a response to levodopa and the presence of Lewy bodies. However, the mean age of onset of affected individuals is 46 years. The same mutation (A53T in exon 4) was also found in nine Greek families44; given the close historical ties between Greece and Southern Italy, it has been suggested that the mutation may be the result of a common founder effect.45

Additional point mutations in SNCA have been reported. A German family was found to have an A30P in exon 3, which resulted in disease.46 A Spanish kindred with autosomal dominant parkinsonism and Lewy body dementia was found to have an E46K mutation.47 Subsequent sequence analyses in thousands of patients have shown that point mutations in SNCA are an uncommon cause of familial or simplex PD.48–54

Further analyses have shown that other types of alterations in SNCA can result in PD. A triplication of a large chromosomal region containing SNCA has been shown to cause autosomal dominant PD.55 Duplication of SNCA has also been found to cause disease.56–58 Analyses of these families suggest that gene triplication results in an earlier onset of disease as compared with gene duplication. Extensive molecular analyses initially indicated that gene multiplication is a rare mechanism for PD causation.59–61 However, a recent study found three SNCA duplications in a sample of 906 screened subjects (<1%).62 Only one of the three positive patients had a positive family history of PD. Importantly, in each of these three families, asymptomatic members were found who carried the duplication, suggesting that penetrance of a SNCA duplication may be incomplete.

Although mutations in SNCA have been known to cause PD for nearly a decade, the mechanism by which these mutations lead to disease is poorly understood. It is thought that aberrant aggregation of the protein results in cell damage and ultimately neuronal death. However, much more research is required to understand how mutations in SNCA or multiplication of SNCA result in parkinsonism.

Variation in the promoter region of SNCA has been reported to increase susceptibility for PD,63 and a meta-analysis of 2692 cases and 2652 controls has further bolstered evidence that this marker, termed Rep1, is associated with a slight, but significant, increase in the risk of PD.64 Rep1 is a dinucleotide repeat sequence that has three prominent allele sizes. Analysis of the surrounding DNA suggests that two domains flanking the Rep1 repeat interact to enhance expression of SNCA whereas the repeat acts as a negative modulator.65 In addition, different alleles can vary the expression levels of SNCA in SH-SY5Y cells by upto threefold.65 It is possible that even a subtle increase in expression could, over the course of many decades, predispose an individual to develop PD.

PARK8: LRRK2

Mutations in the most recently identified gene, LRRK2, have been found among patients with a later age of onset and seem to result in typical, idiopathic PD.66,67 Nearly a dozen different mutations have been reported; the most common, G2019S, has been found in approximately 5–7% of familial, autosomal dominant PD68–70 and 1–2% of sporadic cases.71 These estimates have been derived from mostly North American and Northern European populations; the G2019S mutation seems to be extremely rare in East Asia.72,73 Nevertheless, the G2019S mutation is the most common single cause of PD identified to date.

The age of onset for individuals with the G2019S mutation is highly variable (from age 35 to 78 years). Studies in autosomal dominant pedigrees segregating the G2019S mutation estimated the penetrance of the mutation to be relatively high. In one study, the penetrance at age 50 was 17% but rose to 85% by age 70 years.70 A second study found the penetrance to be 33% at age 55 and 100% by age 75.74 However, it is quite likely that the estimated penetrance of the G2019S mutation is greater in pedigrees with autosomal dominant inheritance as compared with estimates from pedigrees without such a strong family history of disease. Therefore, to avoid an ascertainment bias, a recent study ascertained LRRK2 mutation carriers through the molecular screening of a large series of consecutive PD patients who were tested regardless of their family history of PD.75 Testing of the relatives of the PD patients found to have the G2019S mutation yielded much lower estimates of disease penetrance. In this sample, the penetrance by age 50 was 17%, and at age 70 was 54%. In this study, five carriers of the G2019S mutation were over the age of 75 (up to age 89) and did not have signs of disease. The relatively low penetrance rates raise substantial concerns with regard to molecular testing, as discussed later.

The frequency of the G2019S mutation has been reported to be substantially higher among Ashkenazi Jews76 and North African Arabs,77,78 and to a lesser extent in Portugal.79,80 Haplotype studies suggest a founder effect, which may explain the lower to null frequencies of this mutation in populations further from these regions. Homozygotes and heterozygotes for the G2019S mutation have similar clinical features and both genotypes demonstrate reduced penetrance.81

Imaging studies of patients with G2019S, R1441C, and Y1699C LRRK2 mutations have been indistinguishable from those obtained from patients with idiopathic PD not known to carry LRRK2 mutations.82 The vast majority of LRRK2 cases that have come to autopsy have been found to have brainstem or transition Lewy body disease, typical of idiopathic PD.83 However, some LRRK2 mutation-positive cases have been reported with disparate neuropathology, which has included Lewy bodies restricted to the brainstem, diffuse Lewy bodies, neurofibrillary tangles and abnormal tau deposits, and frontotemporal lobar degeneration with ubiquitinated neuronal intranuclear inclusions.66,82,84,85

LRRK2 codes for a protein kinase that contains five functional domains in the C-terminal half of the protein.86 Allelic heterogeneity has been observed in this gene with disease-producing mutations identified in all five domains. Despite its large size, exhaustive screening of all 51 exons has been performed by several studies and additional mutations have been reported.87–91 However, each of these novel mutations has proven to be quite rare and was found in only one or a few families.92,93 Several LRRK2 substitutions have been reported (i.e., R1514Q) that have been found at similar frequency in PD cases and controls and therefore are unlikely to be disease-producing.94

Two studies have corroborated that LRRK2 is primarily found in the cytosol, but a small proportion of the protein associates with the outer membrane of the mitochondria.95–97 It also seems that mutations in LRRK2 do not affect the protein's steady-state levels, localization, or turnover. Instead, the mutations studied thus far at the molecular level all seem to up-regulate kinase activity and increase autophosphorylation.95,96

Other autosomal dominant loci

The PARK3 locus was originally identified in several families of German descent who were segregating an autosomal dominant form of disease.98 Clinical symptoms in these families are similar to those in typical PD, with a mean age of onset of 59 years and Lewy body pathology. In 1998, a linkage study mapped a putative gene to Chromosome 2p13 and the locus was termed PARK3.98 The causative mutation has not been definitively identified; however, there is some evidence that it is influenced by variation in or around SPR, which codes for an enzyme called sepiapterin reductase that is implicated in dopamine biosynthesis.99,100 No pathogenic mutations in sepiapterin reductase (SPR) gene have been identified in families with PD101; however, a mutation was identified in an individual with dopa-responsive dystonia.102

Analyses in a single sibling pair of German heritage reported the cosegregation of disease with an I93M mutation in UCHL1 (PARK5). The clinical features in these siblings were similar to those seen in idiopathic PD and included a response to levodopa and age of onset at 49 and 50 years.103 Molecular testing of hundreds of individuals with PD has not identified any others carrying the I93M or any other mutations in UCHL1; thus, the initial report103 may be the result of a coincidental polymorphism.104,105

Autosomal recessive PD

PARK2: PRKN

Mutations in the PRKN gene (PARK2), located on Chromosome 6, were initially reported in a sample of Japanese families with autosomal recessive, juvenile parkinsonism.106 Subsequent molecular screening has identified the majority of the mutations in subjects with onset up to age 40. Patients with PRKN mutations have typical PD features, often with lower-limb dystonia. Disease progression is slow. Sustained response to levodopa is observed as well as early, often severe, dopa-induced complications (fluctuations and dyskinesias).107 Interestingly, there have been reports that patients with PRKN mutations in some instances lack the characteristic Lewy bodies found in most cases of idiopathic PD.108,109

As a result of extensive molecular testing, over 100 mutations have been reported throughout PRKN.110 Mutations have included point mutations as well as exon rearrangements, including both deletions and duplications.106,111–118 Several studies have sought to determine the origins of certain mutations by analyzing their surrounding haplotypes110,119 In general, whole exon rearrangements are thought to represent independent events whereas certain missense mutations may be the result of a founder effect. Mutations have been found in each of the 12 exons of PRKN.

Parkin is an E3-type ubiquitin protein ligase that is involved in the degradation of specific proteins, including alpha-synuclein, a primary component of Lewy bodies. Mutations in PRKN are thought to disrupt this important E3 activity and result in a loss of normal protein function. Although all parkin substrates have yet to be definitively identified, it is hypothesized that the accumulation of these proteins, which were to have been ubiquitinated, results in selective cell death of neurons in the substantia nigra and locus coeruleus.120

Several studies have reported that a PRKN mutation in only one of the two copies of the PRKN gene may increase susceptibility for PD or may even result in an autosomal dominant pattern of inheritance.109,118,121,122 However, a recent study123 has found similar rates of heterozygous missense PRKN mutations in controls and PD subjects. Although this study did not examine the rates of dosage mutations in cases and controls, it was the first to fully sequence a large number of controls, suggesting that the higher rates of heterozygous PRKN mutations previously observed in cases may have been due to bias. These results would seem to reduce the likelihood that a single mutation in PRKN may increase the risk for PD. Further data are clearly warranted to address these disparate results and also to provide more accurate estimates of the penetrance of PRKN mutations as well as genotype/phenotype correlations.

PARK6: PINK1

Mutations in PINK1 were initially identified in early onset, autosomal recessive kindreds. Point mutations, frameshift mutations, and truncating mutations have been reported throughout the gene. PINK1 mutations account for 1–7% of early onset or autosomal recessive PD in white patients.124–126 The frequency seems to be higher in Japanese patients with estimates that nearly 9% of autosomal recessive PD patients have a mutation in this gene.127 Patients with PINK1 mutations seem to have clinical features that resemble late onset PD; however, they may have atypical features such as dystonia at onset, sleep benefit, and psychiatric disturbances.124,128–130 Penetrance of disease mutations seems to be high.127,128,131 Wild-type PINK1 encodes a protein that localizes to the mitochondria124 and has been hypothesized to have a neuroprotective role against mitochondrial dysfunction and proteasomally induced apoptosis. It is hypothesized that mutations in PINK1 may result in increased susceptibility to reactive oxygen species and other cellular stressors and thereby may result in PD.132 Similar to other forms of autosomal recessive forms of PD, there is debate as to whether a PINK1 mutation in the heterozygous state may increase the risk for PD.133–136

PARK7: DJ-1

Only seven families have been identified worldwide with autosomal recessive, early onset PD due to mutations in DJ-1. The reported mutations consist of missense mutations, whole exon deletions, a frameshift mutation, and a splice site mutation found in either a homozygous or compound heterozygous state.137–142 Through extensive molecular screening, it is estimated that mutations in DJ-1 account for <1% of all cases of early onset PD.136,141,143–146

DJ-1 encodes a ubiquitous, highly conserved protein that plays a role in oxidative stress.147,148 In particular, DJ-1 seems to act as an intracellular sensor for such damage when the cysteine residue at position 106 becomes oxidized.149,150 Once oxidized, it is thought that DJ-1 acts as a chaperone for proteins such as alpha-synuclein, thereby preventing alpha-synuclein fibrillation, protein aggregation, and misfolding.151,152DJ-1 protein levels are elevated in the cerebrospinal fluid of individuals with idiopathic PD, particularly for those in the earlier stages of disease, suggesting that DJ-1 might be useful as a biomarker for neurodegenerative disease.153DJ-1 knockout mice have nigrostriatal dopaminergic deficits and are hypersensitive to the effects of the neurotoxin MPTP.154

Susceptibility genes for PD

Although the study of large families segregating Mendelian forms of PD have provided substantial insight regarding the etiology and pathogenesis of PD, mutations in these genes have been found in fewer than 5% of all PD patients. Previously published data suggest that the risk of PD among the first-degree relatives of an affected individual is 2 to 14 times higher than the risk in the general population. These data would suggest that additional loci contribute to the risk of PD.

Several different approaches have been used to identify these additional loci. Several research groups have identified multiplex PD families, typically those with at least a sibling pair with disease, and have performed a whole genome linkage screening to identify chromosomal regions linked to the risk of PD or the age of PD onset.155–161 Chromosomes 5 and X have appeared in multiple linkage studies. Analyses combining data from two studies was not able to replicate linkage to Chromosome 5.162 Additional analyses on the X chromosome have identified several candidate genes; however, none have been verified.

Another approach that has been used to identify genes contributing to PD susceptibility has been to compare allele and genotype frequencies in a sample of PD cases and neurologically normal controls. This approach has been used in analyses of the entire genome as well as others that were limited to the study of a single candidate gene. Two genome-wide association studies have been performed.163,164 Unfortunately, there has been little overlap between the two studies and a few independent studies have been published that have not confirmed the initially associated regions or single-nucleotide polymorphisms.165–168 Although discouraging, the design of the two initial genome-wide association studies were quite different in the types of PD cases and controls selected, and these factors may account for the discrepant results. In addition, there is evidence that the sample size necessary to detect and replicate susceptibility alleles with small effect sizes may require a sample that is 10 times that of those used in these first two studies.

The evaluation of particular candidate genes has led to the identification of several susceptibility genes; however, most have failed to consistently replicate in other populations.169 An exhaustive review of all genes analyzed as a candidate gene for PD is beyond the scope of this review. Rather, we have summarized the genes that have been the subject of the most intense focus in recent years.

NR4A2

NR4A2, also known as Nurr1, encodes a member of a nuclear receptor superfamily that is essential for the differentiation of the nigral dopaminergic neurons.170–172 Mice in which both alleles of the NR4A2 gene have been inactivated lack mesencephalic dopaminergic neurons.173,174 Mice in which only one copy of the NR4A2 gene is inactivated demonstrate greater susceptibility to nigral injury and have features consistent with PD.175 A polymorphism in exon 6 of NR4A2 was found to be more frequent in familial cases of PD. A year later, two different mutations in exon 1 were found to segregate with PD in 10 families.176 The age at onset of disease and clinical features of these 10 probands did not differ from those of individuals with typical PD. Le et al.176 also presented data suggesting that dopaminergic dysfunction can result from mutations in NR4A2. However, neither the association with the polymorphism nor either mutation has been found in other large studies of familial and sporadic PD.177–183

SNCAIP

Similar to alpha-synuclein (the protein encoded by SNCA), synphilin-1, the protein encoded by SNCAIP, is a substrate of parkin (PRKN). It has been shown to interact directly with alpha-synuclein and is found, along with parkin and alpha-synuclein, in Lewy bodies. The same mutation (R621C) was reported in two individuals with late-onset idiopathic PD (age of onset: 63 and 69 years) who had no apparent family history of disease.184 In a group of 328 German individuals with familial or sporadic PD, the R621C mutation was the only genetic variant found and was not seen in 351 control individuals. The two mutation carriers share the same rare alleles for five of the six microsatellite markers genotyped in the chromosomal region around the SNCAIP gene, suggesting that this variant was inherited from a common ancestor.184

Functional studies of synphilin-1 indicate that abnormal protein can form cytoplasmic inclusions in transfected cells and that cells transfected with the R621C polymorphism were more susceptible to apoptosis than cells expressing wild-type synphilin-1. The role of synphilin-1 in PD susceptibility has not been replicated185–187; therefore, although it may be important in disease pathogenesis, it is likely to be a rare cause of disease.

APOE

Variation in apolipoprotein E (APOE) has been strongly associated with the risk of AD. Variation in two amino acid residues results in three observed alleles (ε2, ε3, and ε4). The ε4 allele is found in roughly a third of whites and is associated with a 2-fold risk of disease when found in the heterozygous state and a 5–10-fold increase in risk when homozygous.188 Dementia is estimated to occur in 30% of patients with PD,6 and a large longitudinal study of AD patients found that at least one motor sign of PD was detected in 44% of patients during at least one of their study visits.189 Because of this overlap, several studies have evaluated whether APOE genotype is a risk factor for PD.

Evidence for the role of APOE in PD has been inconsistent.190,191 Several recent studies have found the ε4 allele to be associated with PD192–194; however, an earlier meta-analysis found a marginal association only with the ε2 allele.195 A more consistent association of the ε4 allele has been with an increased risk of dementia within a PD sample.190,196,197 Several larger genetic studies have also found the ε4 allele to be significantly associated with an earlier age of PD onset.191,197–199

GBA

Individuals with mutations in both copies of GBA, the gene encoding glucocerebrosidase, have Gaucher disease. This disease is found at a high rate among Ashkenazi Jews. A single mutation in GBA has been reported to confer an increased risk of PD (odds ratio [OR]: 7.0; G-test P < 1 × 10−10), particularly in populations of Ashkenazi Jewish descent.200

All studies that have sought to confirm this association have found an increased frequency of GBA mutations in cases versus controls; however, a few studies failed to find a difference that was statistically significant, possibly due to lack of power (β < 0.20).201 Studies of a Canadian cohort (OR: 7.3; P = 0.03),202 two US cohorts (OR: 5.6; P = 0.01),203,204 a Venezuelan cohort (OR: 4.1; P = 0.17),205 a Norwegian cohort (OR: 1.3; P = 0.56),206 and a Chinese cohort (OR: 4.1; P = 0.16)201 have all found heterozygous GBA mutations at higher rates among cases.

MAPT

The encoded protein, more frequently referred to as tau, is abundantly expressed in neurons and plays an important role in organizing and maintaining cell structure.207 Aggregation of tau is a pathologic hallmark of several neurodegenerative disorders, collectively known as tauopathies. These include Pick disease and AD, as well as several disorders with parkinsonian features: progressive supranuclear palsy, corticobasal degeneration, and frontotemporal dementia with parkinsonism. Even some individuals that present as typical PD can have tau pathology, as was discussed earlier regarding individuals with LRRK2 mutations. A variety of mutations have been identified in the MAPT gene, primarily in patients with frontotemporal dementia; however, clinical and phenotypic heterogeneity, even within the same family, argues that the tau protein's role in the brain is complex. Together with additional genetic or environmental factors, tau dysfunction can have an effect on many neuronal processes.207 Recent evidence demonstrates that one of those factors that tau interacts with is alpha-synuclein208 and that one of tau's neuronal perturbations may lead to PD.

Two meta-analyses performed using several thousand cases and controls suggest that a large haplotype block containing the MAPT gene is associated with a small but significant increase in risk for PD.209,210 This was recently confirmed by another large study of 1762 PD patients and 2010 control subjects and was found to be consistent across age, gender, and family history.211 The deleterious haplotype (H1) and the protective haplotype (H2) actually represent groups of subhaplotypes that have formed independently; however, associations with any of these subhaplotypes have so far been inconsistent.211 The single-nucleotide polymorphisms that define the parent haplotypes of H1 and H2 are in complete linkage disequilibrium with each other (r2 = 1), indicating that the functional variation could be anywhere within this large 900-kb region and not necessarily within any segment of MAPT. Complex permutations of alternative splicing lead to many different isoforms of tau; so if the association with H1 is due to variation that were to upset this delicate balance of isoforms, it may help to explain the variety of different neurodegenerative phenotypes that exhibit tau pathology.

Mitochondrial DNA

Mitochondrial dysfunction, particularly with regard to complex I of the electron transport chain, has been implicated in the pathogenesis of PD. Variations within mitochondrial DNA have been thought to influence susceptibility to neurodegeneration for many years212,213; however, exhaustive sequencing of the mitochondrial genome has not yet revealed mutations that consistently associate with PD.214,215 Population-based studies have also been undertaken to identify mitochondrial haplotype groups that increase the risk of PD216–218; however, these too have provided conflicting results. More recently, two groups have shown that somatic deletions of mitochondrial DNA are found at a higher rate within substantia nigra neurons from individuals with PD as well as the elderly in general.219,220

MOLECULAR DIAGNOSTIC TESTING IN PD

Molecular diagnostic testing is currently available for PRKN (PARK2), PINK1 (PARK6), DJ-1 (PARK7), and LRRK2 (PARK8). However, molecular testing and genetic counseling is challenging in PD patients and their families because of the varied patterns of inheritance that has been observed. Clear review of the family history and clinical symptoms is essential in prioritizing the molecular tests. Importantly, clinical recommendations are not altered based on the presence or absence of a particular molecular mutation. Therefore, molecular diagnostic testing is not essential in the current management of PD patients.

For PD patients with an early onset of disease, it has been estimated that 50% of those with onset before age 40 have a mutation in PRKN, PINK1, or DJ-1. However, given the rarity of PINK1 or DJ-1 mutations, initial testing of PRKN would seem to be the most cost effective strategy. Based on the distribution of age of onset of patients with PRKN mutations, it has been recommended that those patients with onset prior to age 40 be considered for screening for PRKN mutations. Although mutations have been found in those who onset above age 40, the rate is quite low and therefore testing is not likely to be cost effective.

The significance of a mutation in only one of the two PRKN alleles is not yet known. Therefore, it is strongly recommended that PRKN testing be undertaken only in those who have already onset with disease; molecular screening of PRKN is not recommended as a test for presymptomatic individuals. Another important caveat in PRKN testing is the presence of sequence variants that seem not to be disease-producing. A number of such variants have been reported in PRKN and several are at relatively high frequency. Therefore, careful attention to the type of sequence variation reported in molecular testing is essential.

In patients with early onset PD who have been tested for PRKN and found to be negative, it may be appropriate to consider screening PINK1 and DJ-1. However, mutations in both of these genes are much less frequent than those in PRKN.

For patients with typical, later onset idiopathic PD, screening for the three most common mutations in LRRK2 may be considered. Based on previous studies, it is clear that this mutation is much more likely in those PD patients with a family history consistent with autosomal dominant PD. Typically these patients will report affected parents or siblings. Of note, given the size of the LRRK2 gene and rarity of the other reported mutations, diagnostic molecular testing is only available for mutations found at codons 1441, 2019, and 2020. As a result, a false-negative result is possible.

Similar to the caveats with the testing of autosomal recessive genes, it is strongly recommended that LRRK2 testing be performed only for those patients with a diagnosis of PD. Studies have clearly demonstrated that the G2019S mutation is not fully penetrant, even when homozygous. Therefore, to avoid providing a molecular test result without accurate predictive data, it is strongly recommended that the G2019S test not be used as a presymptomatic test for PD.

SUMMARY

Over the past decade, substantial progress has been made in our understanding of the genetic contribution to PD susceptibility. Five genes have been identified that cause either autosomal dominant or autosomal recessive PD. Molecular diagnostic testing is available for the genes contributing to the most common type of autosomal recessive (PRKN) and autosomal dominant (LRRK2) PD. In addition, molecular screening is also available on a clinical basis for two genes that cause relatively rare forms of autosomal recessive PD (PINK1, DJ-1). In all cases, molecular testing is only recommended for individuals with diagnosed PD. For both PRKN and LRRK2, all mutations are not equally penetrant, and accurate penetrance estimates are not yet available. Therefore, it is strongly recommended that molecular diagnostic testing not be performed for presymptomatic individuals.