For the past 40 years, research into Parkinson's disease (PD) has been predominantly the province of epidemiologists interested in pursuing the connection between the disease and environmental factors such as viral infection or neurotoxins. Hereditary influences were actually discounted because of a high monozygotic twin discordance rate found in studies that were later shown to be inadequate and inconclusive. There has recently been a resurgence of interest in investigating hereditary factors in PD when it became more and more apparent that a positive family history was a major risk factor for the disease. Meanwhile, it also became increasingly apparent from neuropathological studies that the common, idiopathic form of Parkinson's disease had, in fact, a pathological correlate, i.e., the existence of Lewy bodies, an eosinophilic cytoplasmic inclusion body, distributed diffusely throughout the substnatia nigra, hypothalamus, hippocampus, autonomic ganglia and olfactory tracts. Although candidate gene approaches to linkage in PD families have not been rewarding, a genome wide scan mapped PD to 4q21-23 in one large family with PD with diffuse Lewy bodies, where a candidate gene, [alpha]-synuclein, resides. This gene encodes a presynaptic protein of which a peptide fragment is known to be a constituent of Alzheimer's disease plaques. The identification of a missense mutation in the [alpha]-synuclein gene in four independent PD families suggests that at least some fraction of familial PD with diffuse Lewy bodies is the result of an abnormal protein that interferes with normal protein degradation leading to the development of inclusions and ultimately neuronal cell death. There may be common pathogenetic mechanisms involved in [alpha]-synuclein mutations in PD and [beta]-amyloid and presenilin gene mutations in Alzheimer's disease.
Parkinson's disease (PD) was first described by James Parkinson in 1817 (1 ) and for 180 years the diagnosis has remained a primarily clinical one. The disease is common with a prevalence of between 500 000-1 000 000 in the United States (2 ). Life-time risk is 1 in 40, making PD the second most common neurodegenerative disease after Alzheimer's disease (AD) (2 ). For well over 100 years, neurologists have been aware of a significant role for heredity in the development of PD but a number of factors have, until quite recently, hampered genetic research in PD.
The first factor was a lack of agreement on precisely what constitutes `typical' PD and whether familial forms of the disease are the `real' PD or not. This attitude is typified by one medical school professor who, during hospital rounds with a colleague of ours, asserted that a lack of family history was a diagnostic criterion for PD. As a result, research into PD that emphasized familial forms of the disease was seen to be off the mark and not directed at real PD. The idea that there was little or no genetic contribution to the etiology of PD was strengthened by a series of twin studies (3 -6 ) during the 1980s that failed to demonstrate increased concordance rates for PD in monozygotic versus dizygotic twins.
A second reason that research efforts may have not emphasized the role of heredity in PD was the striking results of epidemiological studies linking PD to environmental agents such as viruses or neurotoxins. Much of the research community's energies and efforts became focused on identifying the nature and effects of environmental agents thought to be responsible for causing the disease. The 1918 influenza pandemic associated with encephalitis lethargica and, as a late sequela, post-encephalitic PD, pointed to viral infection as a major cause for PD (7 ). The significance of this association seemed so strong at one point that one investigator suggested that PD would disappear by 1980 as a clinical entity as the survivors of the pandemic grew old and died (7 ). Other strong support for an environmental cause for PD was the occurrence of irreversible symptoms and signs of PD occurring after the injection of illicit drug preparations contaminated with N-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) (8 ,9 ). MPTP metabolites interfere with complex I of the electron transport chain, thus depleting cells of the products of oxidative phosphorylation and exposing the dopaminergic neurons of the substantia nigra to irreversible damage from energy depletion and oxidative stress from free radicals (2 ,10 ).
Although viral encephalitis or an environmental toxin such as MPTP is clearly capable of damaging the substantia nigra, the 1990s has seen a re-emergence of interest in the genetic causes for idiopathic PD, primarily through the efforts of neurologists such as R. Duvoisin and Z. Wszolek. As a result, there is currently much more research investigating the role of heredity and gene-environment interactions in causing PD. Rather than choosing between the false dichotomy of an exclusive environmental or genetic etiology, geneticists are attempting a synthesis of the epidemiological and genetic data by focusing on three areas of inquiry: (i) inherited and somatic mitochondrial genome abnormalities causing defects in energy metabolism; (ii) candidate gene linkage and association studies of genes encoding drug metabolizing enzymes; (iii) whole genome scans to locate and identify the genes involved in families with highly penetrant, autosomal dominant PD. Two excellent, relatively recent, reviews of the genetic aspects of Parkinson's disease have been published (11 ,12 ).
As in many areas of genetics, definition of the phenotype is one of the most difficult and contentious obstacles facing the geneticist. An accurate clinical diagnosis is crucial when trying to categorize patients and families for genetic studies in what is certainly a heterogeneous disorder with multiple etiologies. It is generally accepted that the motor symptoms of PD, or parkinsonism, result from a deficiency or dysfunction of dopamine, or of the neurons producing dopamine, in the pars compacta of the substantia nigra, regardless of etiology. However, idiopathic PD, the most common form of the disease, has been difficult to approach for genetic study because the name can mean different things to different researchers.
For the clinician, the diagnosis of Parkinson's disease is based on finding at least two of the four primary motor signs of the disease: tremor, bradykinesia, muscular rigidity and postural instability. Also frequently seen, but not considered primary diagnostic criteria, are dementia, defects in gastrointestinal motility, dysautonomia and responsiveness to l-DOPA therapy. The disease is progressive and usually has an insidious onset in mid to late adulthood. In one study (13 ) of 100 patients, only three-quarters of patients carrying a diagnosis of PD as determined by qualified neurologists actually had the disease confirmed by autopsy. In this study, if clinical diagnostic criteria were made more stringent retrospectively, specificity rose to 93% but at the expense of a drop in sensitivity to 68%.
For the neuropathologist, autopsy demonstration of diffuse Lewy body disease has come to be the hallmark of PD (14 ). In diffuse Lewy body disease, the Lewy body, a characteristic intracellular cytoplasmic inclusion named at the beginning of the 20th century after its discoverer, is found in many regions of the brain, particularly in the neurons of the substantia nigra, in other nuclei in the brain stem, in the hypothalamus, the hippocampus, in the autonomic neurons of the esophagus and gastrointestinal tract and occasionally in the cortex (12 ). Lewy bodies are very frequently found in the brains of patients dying with PD even though there can be extreme variation in the age of onset, severity and clinical manifestations of the PD. In 100 patients carrying a diagnosis of PD who came to autopsy, the 24 patients who did not have Lewy bodies were found not to have PD and instead to have other causes for their clinical signs (15 ). However, the presence of Lewy bodies per se is not uniquely pathognomic of PD. A form of Alzheimer's disease, known as Lewy body variant AD, has been described in which a few Lewy bodies are found in brain cortex along with the pathognomic amyloid plaques and neurofibrillary tangles of AD (16 ). Conversely, it is possible to have the motor dysfunction characteristic of PD, even on a familial basis, without the pathological process whose characteristic marker is the Lewy body (reviewed in 12 ). Examples of hereditary parkinsonism syndromes without diffuse Lewy body disease include juvenile autosomal recessive parkinsonism mapping to chromosome 6q25-27 (17 ), the rapidly progressive autosomal dominant parkinsonism associated with dementia, dystonia and disinhibition, located on chromosome 17q21, and X-linked dystonia-parkinsonism (18 -20 ). These are well-defined but rare familial disorders that have the motor signs of parkinsonism in common with diffuse Lewy body PD but are certainly different diseases. In the case of acquired parkinsonism due to MPTP toxicity, the typical Lewy body pathology of PD is not seen after acute MPTP exposure in humans nor in monkeys undergoing chronic MPTP administration although some atypical inclusions have been reported in the brains of these animals (21 -23 ). Thus, over the past 10 years, neurologists have come to accept that the highly prevalent disorder, idiopathic PD, is a distinctive disorder usually occurring in middle to late adulthood with characteristic pathology of diffuse Lewy body disease and the classical motor and central nervous system manifestations of parkinsonism (14 ,24 ,25 ). For the purposes of this review, we will concentrate on the genetics of this common form of the disease.
Familial clustering of PD was observed a century ago (11 ,25 ) when 10-15% of PD patients were reported to have a positive family history. Accurate measurement of the frequency of PD among relatives of affected patients is made very difficult by the inaccuracy of making a diagnosis of PD from history, the relative insensitivity and lack of specificity of the clinical exam, the variability in age of onset, and the rarity of autopsy data required for making the diagnosis of diffuse Lewy body disease in relatives reported to have PD. During the 1980s, a genetic contribution to PD was thought highly unlikely when a series of twin studies failed to demonstrate an increased concordance for PD in monozygotic versus dizygotic twins (3 -6 ). As has been pointed out, however, (26 ) these studies are not interpretable because the number of twins studied and the duration of observation to either support or reject a genetic contribution to PD are inadequate. As nearly 80% of the function of dopaminergic neurons of the substantia nigra must be lost before symptoms occur (27 ), reliance on clinical diagnosis alone and a lack of follow-up makes a determination of concordance extremely insensitive. Further evidence that clinical evaluation to determine concordance in twin studies is far too insensitive comes from the use of positron emission tomography to measure 18fluorodopa uptake into basal ganglia (28 ). Concordance rate among monozygotic twins increased when 18fluorodopa uptake into basal ganglia was used to diagnosis preclinical dysfunction of the substantia nigra (29 ). As a result, the lack of increased concordance in twin studies can be essentially discounted as the result of inconclusive and insensitive methodology.
More recently, a number of epidemiological studies (30 -36 ) with differing methodological approaches and study populations have been published and found to support a familial contribution to PD. In case control studies, positive family history was found to be the single greatest risk factor for PD (25 ,30 -36 ). In family studies, a family history positive for PD was found in 10-24% of patients, and the relative risk for PD in first degree relatives of PD patients ranged from 4 to 10. In the largest of such studies, the frequency of PD was 2% in 1458 first degree relatives of 233 PD patients, a significantly higher frequency than the 1% seen in the 7834 first degree relatives of 1172 age-matched controls (30 ).
In an attempt to integrate our current knowledge of the familial nature of PD, the possible role of defects in energy metabolism in causing the disease, and the toxicity of MPTP, research has proceeded on two parallel tracks: looking for mitochondrial genome abnormalities in PD and applying linkage and association methods to affected individuals using polymorphisms in drug metabolizing enzymes. Modest deficiencies of complex I of the electron transport chain (reviewed in 2 ) have been reported in the substantia nigra (21 ), muscle (37 -39 ) and peripheral blood cells (39 ,40 ) of PD patients but whether these abnormalities are primary or secondary to the disease process is unknown. The mitochondrial genome in aging brain can demonstrate striking departures from normal (41 ,42 ), but an analysis of the mitochondrial genome in the blood and brain of PD patients has failed to reveal a consistent hereditary or somatic deletion in PD patients as compared with age-matched control blood and brain samples (reviewed in 2 ). A few small studies have suggested a significant association of certain mitochondrial DNA polymorphisms with PD but the sample sizes were small and replication is ongoing (42 ,43 ). Thus, there appears to be some consistent abnormalities in oxidative phosphorylation, particularly of complex I, in PD but a direct causative role for inherited or acquired mutation of the mitochondrial genome remains unproved.
In view of the possible involvement of neurotoxins in PD, research has also been directed towards identifying variation in certain drug metabolizing enzymes that might predispose certain individuals to greater toxicity from environmental agents. Association studies in PD have focused on polymorphisms in drug metabolizing enzymes such as cytochrome P450IA1 (encoded by the CYPIA1 gene) and the debrisoquine 4-hydroxylase cytochrome P450 (encoded by the CYP2D6 gene), as well as the monoamine oxidase A and B enzymes (encoded by MAOA and MAOB). Such association studies are always fraught with difficulties arising from unrecognized ascertainment bias and stratification. Unfortunately, the nature of a late onset disease means that parental genotypes required to permit the use of the more dependable transmission disequilibrium test for association are usually not available (44 ). Nonetheless, a number of studies have suggested significant association between PD and loci encoding cytochrome P450 enzymes and monoamine oxidase. In Japan, an MspI restriction fragment length polymorphism in CYPI1A was found to be associated with a 2-fold higher risk for PD in heterozygotes and 6.5-fold higher in homozygotes (45 ). Association studies with the CYP2D6 gene revealed a highly significant 4-fold increased risk among patients with disease onset before age 50 in a Spanish population although no such association was seen in United States and British samples (46 ,47 ). A particular haplotype at the MAOA locus was found to be 3-fold more common among PD patients as compared to controls while another haplotype was 3-fold less common among the affected patients (48 ). No such associations were found at the MAOB locus (48 ,49 ). Thus, preliminary data support a modest association for certain alleles at these loci with PD but, as has been the history with many association studies, replication of these findings is difficult and requires larger, carefully controlled studies before they can be accepted as definite.
In view of the association of the apolipoprotein E allele 4 with late-onset AD (50 ), it was clearly of interest to investigate whether a similar association exists between PD, either with or without dementia, and the apolipoprotein E4 allele. At least two such studies have failed to show any such association (51 ,52 ). Of note, one study did find a significant association between apoE4 and the Lewy body variant of AD, but not PD, further supporting the idea that the Lewy body variant of AD and PD are distinct disorders (16 ).
A genetic contribution to PD has received substantial support from the identification of a number of families in which PD with diffuse Lewy body disease was inherited as an autosomal dominant trait (32 ,34 ,53 -58 ). Penetrance appeared to be age-dependent, rising from 43% early in adulthood to 85% after age 70. In one such family, from the town of Contursi near Salerno in Italy (56 ), linkage analysis has allowed the gene for PD with autopsy-proven diffuse Lewy body disease to be mapped to the long arm of chromosome 4 at 4q21-q23 (59 ). No other positive mapping studies have been published to date, although essentially negative linkage studies have been reported with a variety of candidate genes in some pedigrees with apparent autosomal dominant inheritance of Lewy body positive PD (60 ). Of note, anticipation was seen in some families with multiple generations affected with PD (57 ,58 ) but was ascribed primarily to ascertainment bias in other families (55 ). The repeat expansion detection method failed to show evidence for an expanded CAG trinucleotide repeat in the Contursi PD kindred (61 ).
Over the last few years, a major shift in thinking and focus for PD research has resulted in a rebirth of interest in investigating a genetic contribution to the disease. This change in focus has already begun to bear fruit in the form of a developing consensus on the relationship between diffuse Lewy body disease and PD, leading to much improved genetic epidemiological investigations of the disorder, the identification of families with clear autosomal dominant inheritance patterns, and the identification of at least one gene in which a particular missense mutation predisposes with high probability to the development of the disease. PD research is now poised to begin to ask fundamental questions of pathogenesis in human patients with classical Parkinson's disease.
Human Molecular Genetics
Pages
Introduction
Clinical Phenotype
Genetic Epidemiology Of PD
The Future
References
REFERENCES
This page is maintained by OUP admin. Last updated Fri Sep 12 18:09:23 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1997