|PD Phenotypes associated with specific mutations|
|Flowchart for familial parkinsonism phenotypes. Adapted from Crosiers et al. (2011).|
Parkinson’s disease (PD) is the second most common neurodegenerative brain disorder affecting about 1% of individuals over the age of sixty and causing a dramatic decrease in life expectancy . It is classically categorized as a movement disorder since affected individuals normally exhibit postural instability, tremor at rest, and akinesia. These symptoms are generally attributed to the loss of dopamine producing neurons in the substantia nigra, a region in the midbrain. Despite developments in symptom management, the cause of PD is not currently known. Research is currently directed towards finding the genetic etiology of PD in hopes of establishing its molecular pathology, implementing procedures for early detection and developing more effective treatments. Classic linkage studies and positional cloning strategies have led to the discovery of multiple genes that cause monogenetic autosomal-dominant or autosomal-recessive forms of PD. These genetic analyses have also yielded rare genetic variants and environmental risk factors for PD. However, these known Mendelian forms of PD and other identified risk factors only explain 20-30% of cases in the general population. The advent of new genetic techniques such as genome wide association studies and whole genome sequencing promises to uncover many common and rare genetic variants that may cause or predispose one to PD.
Table of Contents
1.1 Genetic Techniques
1.1a Classical Gene Linkage Studies
|PD associated genes|
|Genes associated with PD and respective chromosomal positions
found using classical linkage analysis and positional cloning. AD autosomal
dominant, AR autosomal recessive. Adapted from Gasser et al. (2011).
Classical gene linkage studies and positional cloning strategies have been critical in identifying a substantial number of genes that cause monogenic forms of PD. Mutations in these single pairs of genes have traditional Mendelian autosomal dominant or autosomal inheritance patterns. Classical gene linkage relies on the principle that crossing over between linked genes during meiosis is directly related to the distance separating genes on a chromosome; the smaller the distance the genes the lower the chance that non-sister chromatids would cross over in that area. Relative distances between genes are calculated using recombination frequencies. Distances based on recombination frequencies are often quite crude and inaccurate compared to current gene mapping methods. These genetic maps were originally constructed by examining pedigrees of extended families that had a high prevalence of PD. The first study to this effect was carried out by Henry Mjönes in a Swedish family in 1949 who claimed to find an autosomal-dominant gene with 60% penetrance. However, his conclusions were long discounted as many claimed he had no justification in including relatives with atypical clinical presentations or those that were oligosymtomatic as secondary cases. Erroneous classification of individuals as affected leads to incorrect linkage results when examining pedigrees. The original family is now known to carry an α-synuclein multiplication. Technological advances such as the identification of microsatellite repeat elements and the development of DNA amplification techniques such as polymerase reaction (PCR) have since then facilitated the fine mapping of a large number of genes associated with PD through positional cloning. The most common positional cloning technique called “chromosome walking” involves using regions of known genetic polymorphisms that flank the target gene. DNA fragments from nearby genetic markers are progressively cloned and sequenced until the locus of interest is identified. Developments in biotechnology and statistics applied to genetic mapping have led to the discovery of many genes with Mendelian transmission patterns related to PD in rapid succession.
1.1b Genome Wide Association Studies (GWAS)
|Manhattan plot of genes associated with PD from GWAS studies|
|Meta-analysis combining results from 7,123,986 random-effects
association loci from the PDGene database Adapted from Lill et al. (2012).
Genome-wide association studies (GWAS) operate based on the ‘common disease-common variant hypothesis’ which states that heritability for complex diseases such as PD can be partly attributed to allelic variants inherently present in the population. On a technical level, GWAS involves comparing the entire genomes of affected individuals to controls to look for differences in single nucleotide polymorphisms (SNPs). If the frequency of a particular SNP is higher in affected individuals, this variant is deemed to be “associated” with the disease. Dozens of potential risk loci for PD have been identified in over 800 genetic association studies conducted to date. The high volume of information has prompted the scientific community to create a PDGene database that keeps track of all SNPs found to be associated with Parkinson’s disease.
Studies using large cohorts of affected individuals are often run to increase the power of GWAS. A recent study in a Scandinavian population analyzed samples from 1345 unrelated PD patients and 1225 control subjects and replicated associated signals in 11 genes at p<0.05 (SNCA, STK39, MAPT, GPNMB, CCDC62/HIP1R, SYT11, GAK, STX1B, MCCC1/<AMP3, ACMSD and FGF20)9. The same study failed to replicate associations in some very well established loci which speak to the many shortcomings of GWAS. Results from GWAS are usually very difficult to replicate due to the use of small samples sizes of low power and genetic heterogeneity across populations causing false positive results. Even genes that have shown reproducible results usually have a very low odds ratios and account for a small portion of estimated heritability. The identification of a gene through GWAS does not indicate causality and there are often many hypotheses regarding the biological mechanisms through which a particular trait is affected. New statistical techniques that attest to the fact that genes do not work in isolation are currently being developed to analyse GWAS data using pathway based analysis and functional SNP annotation.
1.1c Next Generation Sequencing (NGS)
|Next generation sequencing platform|
|The Illumina sequencing by synthesis method is the most used next generation
sequencing platform that supports massive parallel sequencing
An alternative explanation for the cause of sporadic neurodegenerative diseases such as PD is the ‘common disease – rare variants’ hypothesis. This posits that rare variants with intermediate strengths in a number of genes are responsible for complex neurological disorders. Identifying these mutations requires massive parallel sequencing of the entire genome which is now feasible with the advent of next generation sequencing (NGS). In contrast to the traditional capillary based Sanger sequencing method, NGS is capable of generating a readout of an individual’s entire genome in a matter of days compared to weeks. After the three billion base pairs that constitute a genome are obtained, they are mapped to a reference genome. A minimum depth of coverage is necessary to accurately detect variations compared to the reference genome. Typically 8-10 reads per base pair are needed to confidently detect a heterozygous SNP. High accuracy of reads is important to distinguish true variants from sequencing artefacts and data analysis is mindful of the intrinsic error rate associated with each NGS platform.
Stringent guidelines need to be established with regards to using sufficient sample sizes and analyzing sequencing results before concluding the pathogenicity of individual rare variants.
A key consideration is that a fair proportion of rare variants are specific to certain populations. An example in the PD phenotype is the Gly2019Ser mutation in the LRRK2 gene that is highly prevalent in affected Ashkenazi Jews and North African Arabs and virtually absent in Asian, South African and some European populations. As with GWAS and classical gene linkage studies, the functional consequence of a particular variant is difficult to assess. Genes may interact in unpredictable manners influenced by other mutations and environmental factors. However, despite current limitations regarding data analysis and interpretation, rapid developments make whole genome sequencing the most promising technique for establishing the genetic etiology of PD and other complex diseases. Whether PD is caused primarily by a large number of rare risk mutations or a small number of common risk mutations will have profound implications for genetic counselling and clinical treatment of PD.
2.1 Autosomal Dominant forms of Parkinson’s disease
2.1a PARK1 (SNCA)
|Positive SNCA staining in a Lewy body of a PD patient.
Adapted from Muri et al. (1998).
This particular autosomal dominant form of PD was first identified by examining pedigrees of an Italian family and three unrelated Greek kindred. Affected individuals in these families were found to have a mutation which caused an alanine to threonine substitution at position 53 (A53T) in PARK1. They presented with a form of Parkinson’s that had an early age of onset and rapid disease progression. Further studies found other nonsynonymous mutation in PARK1 that resulted in different PD phenotypes. An alanine to proline substitution at position 30 (A30P) caused an age of onset in the third to fifth decade while a glutamate to lysine substitution at position 46 (E46K) had a late onset age. However, PARK1 point mutations are rare and not found in cases of sporadic PD. The more common cause of PD associated with PARK1 is due to duplications or triplications of the gene. Duplications cause late onset PD which is responsive to levodopa treatment while triplications cause a more severe early onset phenotype with rapid progression and dementia, . The dosage dependent pathogenic effect of this gene led to the logical conclusion that there is a toxic gain of function in the protein it encodes, α-synuclein. α-synuclein is a small high acidic protein that is the main component of Lewy bodies and is found in neuronal and glial cells of the hippocampus, neocortex, thalamus and substantia nigra. The current hypothesis is that amino acid changes in this protein result in an increased tendency to form fibrillar aggregates which eventually lead to neuronal dysfunction and cell death.
2.1b PARK8 (LRRK2)
Most forms of autosomal dominant PD are due to mutations in the LRRK2 gene. Missense mutations in LRRK2 were first discovered to co segregate with PD in four Basque families, one German-Canadian family and one American family. There have been more than 100 described variants of LRRK2 though only seven have been found to be pathogenic: R1141G, R1441C, R1141H, Y1699C, G2019S, I2020T and N1437H27. Of these seven, the G2019S mutation is the most common across populations. In contrast to PARK1, mutations in PARK8 have variable penetrance where only a proportion of individuals with a mutation will develop PD. G2019S, for example, has been shown to have an age-dependent penetrance which ranges from 28% at age 59 to 74% at age 79. There have even been reports of healthy individuals that carry the G2019S mutation. However, here is typically little phenotypic variation in individuals that do develop PD. Most patients have a late onset form of the disease that is clinically indistinguishable from idiopathic PD and often has Lewy body pathology. To date, there are numerous possible functions for the normal function of the protein LRRK2 encoded by PARK8. This protein contains an Ankyrin repeat region, a DFG-like motif, a RAS domain, an MLK-like domain, a WD40 domain, a kinase domain and a la leucine-rich repeat domain. Studies suggest that this protein, associated with the mitochondrial outer membrane but largely absent in the cytoplasm, interacts with α-synuclein but not DJ-1 or tau. Mouse models are currently under development to study the molecular pathways compromised by mutations in LRRK2.
2.2 Autosomal Recessive forms of Parkinson’s disease
2.2a PARK2 (Parkin)
|PINK1 and Parkin interaction|
|Mechanism by which PINK1 and Parkin may interact.
PINK1 associates with inner mitochondrial membrane. Parkin
localizes mainly in the cytosol and on the mitochondrial outer
membrane. Adapted from Dodson et al. (1997).
The PARK2 mutation was first found to co-segregate with a juvenile onset form of PD in five Japanese patients. Classical linkage analysis mapped the gene to the telomeric region in the long arm of chromosome 6. To date, a number of different mutations have been documented including small insertions and deletions, single or multiple exon copy number variations and point mutations. Mutations in PARK2, both heterozygous and homozygous, are found in 10-20% of sporadic early-onset patients and 40-50% of familial early-onset cases. Heterozygotes with PARK2 mutations are frequently reported though the phenotypic profile of these individuals is still under investigation. There has been evidence to suggest that they display a small yet significant decrease in striatal 18F-DOPA uptake as shown by PET imaging suggesting subclinical dysfunction in the nigrostriatal pathway. Clinically, PARK2 mutations carriers are indistinguishable from typical early-onset PD patients excepting for a lower levodopa equivalent dose and a delay in developing levodopa-related motor complications. This form of PD is generally not accompanied by Lewy body formation though some reports show the presence of Lewy bodies in the substantia nigra and locus coeruleus. Parkin, the protein encoded by PARK2, is an E3 ubiquitin ligase found to be associated with the endoplasmic reticulum, synaptic vesicles, the Golgi apparatus and mitochondria. Parkin has been shown to be recruited to damaged mitochondria, having low membrane potential, where it is a key player in mediating the engulfment of mitochondria by autophagosomes. Studies in Drosophila and mammalian systems have shown evidence to suggest that the parkin and PINK1 mediated mitophagy pathway is involved in PD pathogenesis,.
2.2b PARK6 (PINK1)
Mutations in the PARK6 gene were first associated with an early onset form of PD in three consanguineous families. A number of mutations have been reported including missense and nonsense point mutations, single or multiple copy number variants(CNVs) and insertions and deletions all of which lead to a loss of function of PINK1, the protein encoded by PARK6. Though there are some common mutations in the PARK6 gene (Q129fs, A168P, R246X, Y258X, T313M, W437X, Q456X, D525fs and deletion of exons 6–8) the majority are private. The phenotype associated with this type of parkinsonism is one of slow disease progression and more favorable response to levodopa treatment compared to idiopathic PD. Functionally, PINK1 has been found to interact with Parkin in the mitochondrial autophagy network. Animal models have shown that overexpression of Parkin can rescue PINK1 phenotypes with the converse not being true. This suggests that Parkin may function downstream of PINK1.
2.2c PARK7 (DJ-1)
Positional cloning in an Italian and Dutch family found mutations in the PARK7 gene to cause an autosomal recessive early onset form of PD. The majority of PARK7 mutations are private and generally very rare across populations. Though there is considerable phenotypic variability between homozygotes and compound heterozygotes, this form of parkinsonism seems to show a slow disease progression, dystonic features and brisk tendon reflexes. The protein encoded by PARK7 has been shown to have a number of functions ranging from a peroxidise, a protein chaperone, and RNA binding protein and a regulator of mitochondrial integrity. Recent evidence postulates that DJ-1 may act in parallel with the PINK1 mitochondrial autophagy pathway.
3.1 Rare genetic risk factors
|GBA mutations cause cell toxicity|
|Mutations in GBA may cause cells to be unable to degrade
α-synuclein leading to eventual cell death. Adapted from Sidransky et al. (2012).
Autosomal dominant or autosomal recessive phenotypes represent a one-to-one relationship between gene and disease where a mutant copy (or two in the case of homozygous recessive) will certainly lead to development of PD. Rare genetic risk factors have less appreciative effects. The most well investigated genetic risk factor is the GBA gene which also causes Gaucher’s disease. Gaucher's disease is an autosomal recessive disorder that affects the mononuclear phagocyte system. Affected individuals have an engorgement of lysosomes due to an excess of stored lipid. Close clinical observation found that patients with Gaucher’s were at a significantly increased risk for developing PD. To date, there have been more than 300 different mutations identified in GBA; though a nearby pesudogene with 96% exonic sequence homology complicates sequencing and mutation detection. Patients exhibited a form of PD with an earlier age of onset and favorable response to levodopa treatment.
There have been a number of genes implicated as risk factors. Mutations in the F-box only protein 7 (PARK15, FBX07) have been shown to lead to an early onset form of PD with pyramidal symptoms. ATP13A1, which usually is found highly expressed in the cortex and dopamine neurons in the substantia nigra, was found to be linked to an early onset form of parkinsonism with dementia and pyramidal symptoms. To date, differences in ethnic populations and irreproducible results in GWAS studies have made it difficult to provide concrete evidence for many reported risk factors. The advent of high throughput sequencing promises to uncover many more risk factors in the near future and to test the validity of current candidate genes.