Article Text

Download PDFPDF
Molecular pathogenesis of Parkinson's disease: update
  1. Shinji Saiki,
  2. Shigeto Sato,
  3. Nobutaka Hattori
  1. Department of Neurology, Juntendo University School of Medicine, Bunkyo, Tokyo, Japan
  1. Correspondence to Professor N Hattori, Department of Neurology, Juntendo University School of Medicine, 2-1-1 Hongo, Bunkyo-ku, Tokyo 113-8421, Japan; nhattori{at}


Parkinson disease (PD) is a neurodegenerative disease characterised by progressive disturbances in motor, autonomic and psychiatric functions. Much has been learnt since the disease entity was established in 1817. Although there are well established treatments that can alleviate the symptoms of PD, a pressing need exists to improve our understanding of the pathogenesis to enable development of disease modifying treatments. Ten responsible genes for PD have been identified and recent progress in molecular research on the protein functions of the genes provides new insights into the pathogenesis of hereditary as well as sporadic PD. Also, genome wide association studies, a powerful approach to identify weak effects of common genetic variants in common diseases, have identified a number of new possible PD associated genes, including PD genes previously detected. However, there is still much to learn about the interactions of the gene products, and important insights may come from chemical and genetic screens. In this review, an overview is provided of the molecular pathogenesis and genetics of PD, focusing particularly on the functions of the PD related gene products with marked research progress.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Parkinson's disease (PD) is the second most common progressive neurodegenerative disease, named after James Parkinson's who provided a classic account of the condition in 1817. Affecting 1–2% of the population over the age of 65 years, the prevalence of PD increases by approximately 4% in those older than 85 years. Ten genes that contribute to the genetic aetiology of hereditary PD (hPD) were identified, mainly through positional cloning strategies in inherited PD patients and families (table 1).1 2 Several responsible genes for hPD have been identified, and based on functional studies in vitro and in vivo of gene products, some have been found to interact with each other in various cellular systems for homeostasis, such as synaptic homeostasis (α-synuclein), mitochondrial maintenance (PINK1, parkin, DJ-1, Omi/HtrA2), autophagy–lysosome pathway (α-synuclein, parkin, PINK1, Omi/HtrA2), axonal transport (LRRK2) and ubiquitin proteasome systems (α-synuclein, parkin, DJ-1, UCH-L1). Impairments in a number of cellular systems have been suggested to underlie hPD (figure 1). Also, more recent studies revealed that mutations in the same genes can be involved in familial PD and be risk factors for sporadic PD (sPD), suggesting that inherited and sPD could have common pathological mechanisms.3 Therefore, understanding the function of the proteins encoded by hPD genes will hopefully further our understanding of the mechanisms leading to inherited and sPD.

Table 1

Genetic and clinical characteristics of hereditary Parkinson's disease

Figure 1

Schematic representation of the possible pathogenesis in hereditary Parkinson's disease. ALP, autophagy–lysosome pathway; ERAD, endoplasmic reticulum associated degradation; Ub, ubiquitin; UPS, ubiquitin proteasome system.

In this review, we will summarise the latest research progress in the molecular mechanisms of hPD and genetic association studies of sPD.

Hereditary PD

α-Synuclein (PARK1 and PARK4)


SNCA was the first causal PD gene identified in a large Italian family.4 Mutations (A30P, E46K and A53T), duplications and triplications of the SNCA gene have been reported.2 Clinical features of patients with the E46K mutation are similar to those of dementia with Lewy bodies, while A30P is not associated with severe dementia. Individuals with SNCA triplication developed an early onset form of PD with rapid progression and more extended neurodegeneration.5

Recent genome wide association studies (GWAS) have demonstrated a strong association between common single nucleotide polymorphism within the SNCA locus and PD in European and Japanese population, consistent with the finding that variation at the SNCA locus increases PD susceptibility.6–9 Although the SNCA single nucleotide polymorphism associated with sPD show a low OR (1.2–1.4), these findings are consistent with α-synuclein aggregation pathology.

Molecular biology

α-Synuclein is mainly expressed in the presynaptic terminal of the CNS. The protein binds with lipids and unfolds in the steady state. Although the exact function remains unclear, it regulates dopamine homeostasis in presynaptic vesicle cycling.5 The phenotype of α-synuclein knockout mice is unremarkable and only shows a mild decrease in dopamine levels in the striatum and a mild decrease in synaptic vesicles in the hippocampus. Compared with the wild-type α-synuclein, mutant forms easily aggregate in neuronal cells in vitro and in vivo.10 11 Transgenic mice with wild or mutant α-synuclein under various promoters have shown neuronal inclusions, mitochondrial abnormalities and neurodegeneration.12–14 Which type of α-synuclein species is the most toxic to cells remains unclear but some studies assert that mature aggregates are not themselves the toxic moiety but rather an attempt by the cell to clear small toxic oligomers.15 Hsp90 modulates the assembly of α-synuclein in an ATP dependent manner by restricting conformational fluctuations of α-synuclein.16 Recent advances in research on the protein degradation system associated with PD revealed the importance of ubiquitin proteasome and the autophagy–lysosome pathway in disease pathogenesis.17 Wild-type α-synuclein is degraded by both chaperone mediated autophagy and macroautophagy, while A30P and A53T are degraded mainly by the latter.17–19 Furthermore, macroautophagy itself is blocked by α-synuclein via Rap1a dysregulation.20

Several lines of evidence have shown that permeabilised α-synuclein from a neuron may be toxic to neurons and/or glias they are next to. Actually, grafted healthy neurons can gradually develop the same pathology as host neurons in PD brains.21 These findings have suggested that non-cell autonomous cell death as well as cell autonomous cell death may have an important role in disease pathogenesis.

Parkin (PARK2)


The first genetic locus for autosomal recessive juvenile parkinsonism was mapped to chromosome 6, and the disease gene named parkin (PRKN) was identified in consanguineous families.22–24 Mutations in the PRKN gene are most common in autosomal recessive juvenile parkinsonism and many mutations have been reported.3 The clinical picture is similar to that of sPD except for earlier onset, dystonic features, brisk reflexes and sleep benefit. Pathologically, no Lewy bodies were seen in most cases.25–27 Whether or not heterozygous PRKN mutations may cause or increase the susceptibility to late onset typical PD remains controversial. [18F]Fluorodopa uptake by positron emission tomography was reduced in heterozygous carriers without symptoms.28 29 In addition, heterozygous carriers of PRKN mutations have been reported to have either minor motor signs or present with late onset parkinsonism, suggesting a link between heterozygous mutations and disease pathogenesis.27 30 31 On the other hand, screening for PRKN mutations in late onset PD and healthy controls revealed similar frequencies of genetic variants.32 33

Molecular biology

Parkin is associated with the ubiquitin proteasome system as an E3 ubiquitin ligase.34 The C terminal binds with ubiquitin E2 enzymes and recognises a substrate whereas the N terminal interacts with the 19S subunit of proteasome. A nonsense mutation lacking the rear RING finger motif had no E3 activity and sole IBR-RING2 retained E3 activity, and thus most parkin mutations do not lead to loss of kinase activity.35 α-Synuclein and synphilin-1 were identified as parkin substrates and consist of Lewy bodies.36 37 Parkin mainly localises in the cytoplasm as well as in plasma membranes and partly in mitochondria. Under physiological or pathological conditions, parkin is involved in mitochondrial maintenance and recent evidence revealed that parkin with PINK1 physically associate and functionally cooperate to identify and label damaged mitochondria for selective degradation via autophagy (mitophagy).38–42 Protein–protein interactions between parkin and other PD related genes are detailed in each gene section.



PARK6 was first identified on chromosome 1p36.43 The disease gene was identified as PINK1 (PTEN induced kinase 1) containing eight exons.44 The clinical characteristics are autosomal recessive, early onset, slow disease progression and L-dopa responsive parkinsonism. Most mutations were missense mutations, but whole gene deletions were also reported.45 46 Many putative pathogenic mutations were also observed in a heterozygous state in familial and sPD patients as well as in healthy controls. However, most of the studies have not checked the copy number variants, causing the mutation pathogenicity to remain controversial.2 Lewy bodies, neuronal loss and astrocytic gliosis in the substantia nigra were detected in a patient with PINK1 compound heterozygous mutations.47

Molecular biology

PINK1 has eight exons encoding 581 amino acids, including a mitochondrial targeting sequence, transmembrane domain and kinase domain.48 The gene product is ubiquitously expressed in the brain and systemic organs. The protein mainly localises in mitochondria, especially in the outer membrane. PINK1 is a serine–threonine kinase and several pathological mutations in PINK1 have been reported to change their kinase activities.49–52 In addition, Rictor (a component of mTORC2),53 tumour necrosis factor receptor associated protein 1 (TRAP1; a mitochondrial chaperone),50 Omi (PARK13 gene product) and parkin (PARK2 gene product) were identified as substrates for PINK1.54 55

PINK1 regulates mitochondrial dynamics and respiratory functions.38 53 56–58 Mitochondrial fission is accelerated by PINK1 overexpression accompanied by parkin.59 60 PINK1 ablation with siRNA in neurons reduces resistance against oxidative stress while its overexpression provides resistance.61 Using genetically modified Drosophila models, we see that PINK1 deficiency causes the same phenotype as parkin deficiency and the PINK1 deficiency phenotype is rescued by parkin complementation, suggesting that parkin is downstream of PINK1.62–64 Several lines of evidence have provided new aspects of the PINK1/parkin pathway associated with mitochondrial elimination via macroautophagy (mitophagy). When mitochondrial membrane potentials are lost, endogenous PINK1 is accumulated followed by parkin recruitment, and subsequently the depolarised mitochondria were eliminated by mitophagy.40 41 65 66 Mitochondrial targeting sequence, kinase activity of PINK1 and the linker domain of parkin are indispensable for the PINK1/parkin mediated mitophagy.

DJ-1 (PARK7)


Clinical features of PARK7 are characterised by early onset parkinsonism with scoliosis, blephalospasm and psychiatric symptoms, similar to those of PARK2 and PARK6. The disease gene was identified as DJ-1, which has eight exons encoding 189 amino acids. Three missense mutations (L166P, M26I, E64D) in exons 1–5 of the gene have been identified in Italian, Dutch and Uruguayan families. DJ-1 protein was detected around Lewy bodies, suggesting DJ-1 is not in the main structure of Lewy bodies. However, the protein was detected in astrocytes and in a part of the cytoplasmic inclusions positive to tau in brains with corticobasal degeneration, progressive supranuclear palsy and multiple system atrophy.67–69

Molecular biology

DJ-1 is almost ubiquitously expressed in organs, including the brain. Endogenous DJ-1 is present in synaptic terminals, mitochondria and membranous organelles.70 71 DJ-1 with the L166P mutation lost more stability compared with the wild-type and mutant DJ-1 (M26I, E64D).72 In DJ-1 knockout mice, no significant loss of dopaminergic neurons and decreased susceptibility to oxidative stress were noted.73 DJ-1 is a multifunctional redox sensitive protein regulating mitochondrial oxidative stress and increases expression levels of SOD1 in an Erk1/2-Elk1 pathway dependent manner,74 and facilitates prosurvival factor Akt, leading to suppression of apoptosis.75 Also, the protein inhibits TRAIL induced apoptosis by blocking Fas associated protein death domain mediated pro-caspase-8 activation.76 Along with parkin and PINK1, DJ-1 has various cellular functions such as regulation of mitochondrial morphology as well as misfolded protein degradation by forming an E3 ligase complex with those proteins.77



Clinical features of PARK8 are essentially similar to those of sPD except for earlier onset age. The disease gene was identified as the leucine rich repeat kinase 2 gene (LRRK2) linked to autosomal dominant inherited PD encoding 2517 amino acids.78–80 PARK8 is the most common form of hPD in the world. Until now, 20 missense or nonsense mutations have been reported.81 LRRK2 mutations were also found in some sPD cases; neuropathological findings were heterogeneous.82 83 Most of the cases with LRRK2 mutations showed various degrees of Lewy bodies but intraneuronal aggregations positive to tau were rarely detected.79 84 85 The G2019S mutation in LRRK2 is the most common genetic cause of PD, accounting for a significant proportion of both autosomal dominant and sPD cases.

Molecular biology

LRRK2 protein, containing a GTPase domain, a Ras of complex domain, a C terminal of Ras complex domain and a mitogen activated kinase domain, is highly expressed in the brain, and mRNA levels are rich in the striatum and hippocampus compared with other regions.86 Intracellular LRRK2 is mainly distributed in the plasma membrane and vesicular structures.87 88 Immunoprecipitation techniques have revealed that LRRK2 interacts with parkin.89 In transgenic flies, neurodegeneration by LRRK2 with or without a mutation is modified by overexpression or siRNA knockdown of parkin, PINK1 or DJ-1, suggesting genetic interaction between them.90 91 Activity changes of LRRK2 kinase and GTPase have been suspected as a key factor in LRRK pathogenesis. Changes in LRRK2 activity cause alterations in mitogen activated protein kinase, translational control, tumour necrosis factor α/Fas ligand and Wnt signalling pathways with the cell biological functions of LRRK2 such as vesicle trafficking.80 The most common pathological mutation in LRRK2, G2019S LRRK2, causes neurite retraction by activation of Rac1 small GTPase.92 LRRK2 mutations inhibit an endogenous peroxidase by phosphorylation promoting dysregulation of mitochondrial function and oxidative damage.93 G2019S human LRRK2 transgenic rat models specifically expressed in the nigrostriatal system have shown progressive degeneration of nigral dopaminergic neurons.94 In terms of LRRK2 control, PKA has been identified as a potential upstream kinase of LRRK2 at S935, on which binding of 14-3-3 with LRRK2 depends.95 However, the exact biological function of LRRK2 remains largely unclear because no physiological substrates have been identified to date.



PARK9, also known as Kufor–Rakeb syndrome, is an autosomal recessive parkinsonian disorder characterised by early onset (14–16 years old), good response to L-dopa treatment, pyramidal feature, supranuclear gaze palsy and dementia.96 The gene locus was mapped to 1p36 and the disease gene was identified as ATP13A2, which localises in lysosomal membranes.97 Various types of mutations in the ATP13A2 have been reported.

Molecular biology

ATP13A2 is predicted to be a lysosomal P5-type ATPase that plays important roles in regulating cation homeostasis. Although ATP13A2 function remains unclear, it might be involved in protecting cells against manganese and mutant α-synuclein toxicity.98 Wild-type ATP13A2 localises mainly in lysosomes whereas three separate mutants with a mutation involved in PD cause retention of the protein in the endoplasmic reticulum, and are eliminated by the endoplasmic reticulum associated degradation pathway.99 Wild-type ATP13A2, but not pathogenic mutants, reduced intracellular manganese concentration and prevented cytochrome C release from the mitochondria.100

Omi/HtrA2 (PARK13)


Missense mutations in the gene coding for Omi/HtrA2 were reported to be associated with four patients with sPD, presenting with typical parkinsonism.55 G399S and A141S mutations were detected and resulted in defective activation of the protease activity of Omi/HtrA2. Pathologically, accumulation of Omi was found in neuronal and glial inclusions in brains with α-synucleinopathies as well as in Lewy bodies.101 The largest association study revealed no overall strong association of Omi/HtrA2 variants with sPD in populations worldwide.102

Molecular biology

Omi/HtrA2 is a nuclearly encoded mitochondrial protein consisting of 458 amino acids, originally identified as a proapoptotic protein binding with an apoptosis inhibiting protein.103 104 Omi knockout mice presented with neuronal loss in the striatum and died within 30 days of birth.105 Cells overexpressing Omi mutant with G399S have shown mitochondrial morphological changes followed by dysfunction and increased susceptibility against oxidative stress.55 Interestingly, wild-type Omi/HtrA2, not protease defective mutant, activates autophagy through digestion of Hax-1, a Bcl-2 family related protein that represses autophagy via Beclin-1 inhibition, suggesting an insufficient protein degradation system may play a key role.106



PARK14 is an autosomal recessive parkinsonian syndrome characterised by early onset rapidly progressive parkinsonism, dystonia, cognitive decline, and cerebral and cerebellar atrophy. Through homozygosity mapping and direct sequencing, two different homozygous mutations in PLA2G6, which also causes infantile neuroaxonal dystrophy and neurodegeneration with brain iron accumulation, were identified.107 108 Cranial MRI did not detect iron accumulation in the basal ganglia in most cases with this disorder.108 109

Molecular biology

The PLA2G6 gene encodes a group VIA calcium independent phospholipase A2, also known as calcium independent phospholipase A2 β, which hydrolyses the sn-2 acyl chain of phospholipids, generating free fatty acids and lysophospholipids. In an in vitro assay, wild-type PLA2G6 associated with infantile neuroaxonal dystrophy/neurodegeneration with brain iron accumulation failed to catalyse fatty acid release from phospholipids, while PARK14 associated mutations ((R741Q, R747W and R632W) did not, implying that other functions of PLA2G6 include interactions with calmodulin and that PLA2G6 might also be associated with calcium/calmodulin dependent protein kinase II-β.110 111



Only three families with mutations in FBXO7 have been reported.112 113 Affected individuals had juvenile onset (10–19 years old) of progressive parkinsonism associated with spasticity, and variable response to L-dopa. No pathological studies have been reported.

Molecular biology

Fbox7 is a member of the F box containing protein (FBP) family with an F box domain. F box containing proteins are expected to function as molecular scaffolds in the formation of the protein complex; however, the exact function of FBXO7 remains unclear.

Other genes associated with Parkinson's disease

GWAS have uncovered a number of candidate genes involved in PD in European and Japanese populations, indicating a substantial contribution of genetics underlying susceptibility to both early onset and late onset PD.6 7 114–119 These studies have shown repeatedly a common variation in SNCA and an inversion of the region containing the MAPT. Recent genetic studies revealed mutations in the GBA gene, the most widespread genetic risk factor for parkinsonism identified to date.120–124 In this section, we summarise the molecular mechanisms of the two genes, MAPT and GBA.


Mutations in MAPT, encoding microtubule associated tau, result in tauopathies, including progressive supuranuclear palsy, corticobasal degeneration and frontotemporal lobar degeneration.125 Tau is a soluble protein, but insoluble aggregates are produced during the formation of neurofibrillary tangles which disrupts microtubule associated dynamics and neuronal functions. Considering the interplay between α-synuclein and tau reported previously,126 it is interesting that there would be a common pathogenesis associated with aggregation formations.


Early observed patients with Gaucher disease and their heterozygous relatives present with parkinsonism.127 In addition, autopsy studies have shown the presence of mutant glucocerebrosidase (GCase) in α-synuclein positive Lewy bodies in Gaucher disease patients and carriers with α-synucleinopathies.128 GCase is a lysosomal hydrolase with 497 amino acids that catalyses the metabolism of the glycolipid glucosylceramide to ceramide and glucose. Cells overexpressing mutant GCase promoted α-synuclein accumulation in a dose and time dependent manner.129 α-Synuclein GCase interacts selectively under lysosomal solution conditions (pH 5.5) and the interaction site was mapped to the α-synuclein C terminal residues 118–137.130 Insufficient functions of the lysosomes may have an effect on chaperone mediated autophagy or macroautophagy.

Concluding remarks

In the 14 years since the first causative gene (α-synuclein) in PD was discovered, great advances have been made in understanding the biology of the disease. Recent evidence shows that the environment plays no role in the aetiology of PD.131 In addition, GWAS suggest that a number of genes influence susceptibility.3 The PD associated genes provide valuable clues regarding the molecular pathogenesis of PD because the pathomechanism for sPD would have certain pathways in common with those of hPD. Importantly, basic biological studies in PD have led to numerous potential therapeutic strategies. For example, a specific inhibitor for LRRK2 phosphorylations at Ser910 and Ser935 was recently developed.132 In the future, it becomes more important to translate laboratory data, including molecular pathogenesis as well as genetic associations, into clinical treatments, leading to disease modifying therapies to conquer the disease onset and/or progression.



  • Funding The authors are very grateful for the CREST Grant from the Japan Science and Technology Agency (NH), grants from the Ministry of Health, Labour and Welfare (NH) and the Ministry of Education, Culture, Sports, Science and Technology (NH), Grant-in-Aid for Young Scientists (A) (S Saiki), a promoted grant from Juntendo University (S Saiki) and grants from the Takeda Scientific Foundation (S Sato, S Saiki) and the Life Science Foundation (S Saiki).

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.