Preclinical Biomarkers of Parkinson DiseasePreclinical Biomarkers of Parkinson Disease FREE

Parkinson disease (PD), the second most common neurodegenerative disorder, is characterized by a large number of motor and nonmotor features that can have a serious effect on the function and quality of life of the affected individual.^{1 - 2} Currently, the clinical diagnosis of PD is based on the presence of a combination of cardinal signs, including rest tremor, bradykinesia, rigidity, and loss of postural reflexes.^{3 - 4} The pathological process leading to PD begins decades before the typical motor symptoms and, by the time the diagnosis is made, about 70% to 80% of striatal dopamine (DA)⁵ and at least one-third of substantia nigra (SN) neurons⁶ and striatal dopaminergic fibers^{7 - 8} are already lost.

Preclinical PD can be defined as a state that precedes the onset of the characteristic motor features of PD, although some nonmotor features may be present or subclinical abnormalities may exist that can be detected only by neuroimaging or by biological or physiological test results (Figure).^{3 ,9} Clinicopathological studies have suggested that the preclinical period of PD has a relatively short duration of only about 4 to 7 years.^{7 ,10 - 11} This is also supported by various imaging studies indicating a 4% to 13% annualized rate of reduction in striatal uptake of fluorodopa F 18 activity or in DA transporter (DAT) binding in patients with PD compared with the 0% to 2.5% annual reduction in healthy control subjects.^{7 - 8} Other studies, however, including data based on the onset of PD in concordant monozygotic twins,^{12 - 13} have suggested a much longer latency. We have proposed that, at least in some cases, the preclinical period might date even to the prenatal or perinatal period.¹⁴

Biomarkers are tools that assist in detection and diagnosis of the disease, track the progression, and help identify therapeutic targets. Development of highly reliable and sensitive markers of preclinical PD or early PD is a high research priority for the following reasons: (1) It will allow the detection of at-risk individuals before classic symptoms develop, a phase in which neuroprotective therapies are expected to have their greatest impact; (2) it will help discriminate PD from other causes of parkinsonism; and (3) it will accelerate research into etiopathogenesis and pathogenesis-targeted therapeutics of PD.

Growing evidence suggests that a variety of symptoms can precede the classic motor features of PD.^{1 ,4 ,9 ,15 - 16} Symptoms and signs, suggested as possible premonitory features of PD, include essential tremor (ET), olfactory dysfunction, sleep disturbances, depression, anxiety, apathy, low impulsiveness, constipation, and other dysautonomic features.

Although a link between ET and PD has been debated for a long time,^{17 - 18} recent neuropathological studies provide additional support for the association of the 2 disorders.¹⁹ Furthermore, clinical epidemiological studies have indicated that a higher-than-expected proportion of patients with ET have coexistent PD; patients with PD often have preexisting ET, even since childhood; and there is a high prevalence of ET among their first-degree relatives.^{18 ,20 - 21} More recently, a large cohort study of 3813 older people in central Spain found that patients with ET are 4 times more likely than controls to develop PD.²²

Olfactory loss (hyposmia or anosmia) occurs in up to 90% of patients with PD and involves several impairments of odor detection, identification, and discrimination.^{23 - 25} An association between impaired olfaction and the future development of PD has also been found in population-based prospective studies.²⁴ Furthermore, results of odor discrimination measures have been found to be related to disease severity, possibly indicating that at least some aspects of olfactory dysfunction in PD may be secondary to ongoing degenerative processes in PD. These and other studies support that olfactory dysfunction is a very early sign of idiopathic PD.²³ Olfactory testing, combined with neuroimaging and neurochemical markers of central DA measured by the ratio of 6-fluorodopa F 18 uptake between the putamen and SN and between the putamen and caudate, and peripheral DA levels, indicated by cardiac 6-[¹⁸F]fluorodopamine–derived radioactivity, distinguished PD from multiple-system atrophy.²⁶ These results suggest that olfactory loss can predate the development of clinical PD by 2 to 7 years. The mechanism of olfactory loss in PD is not well understood, but it appears to be of central origin rather than due to damage to the olfactory epithelium.²⁷

Many studies have found that patients with PD have a variety of sleep disturbances, including rapid eye movement behavioral disorder (RBD), that precede the onset of motor symptoms by many years or even decades.^{28 - 29} In a long-term clinical follow-up of 93 patients with RBD, 14 developed parkinsonism, 7 met clinical criteria for Lewy body (LB) dementia, 4 met clinical criteria for Alzheimer disease, and 1 developed multiple-system atrophy.²⁹ Patients with RBD are significantly less likely to have a tremor-dominant form of PD, have higher frequency of cognitive impairment³⁰ and falls, and are less responsive to medications.³¹ Furthermore, in addition to motor impairment, patients with RBD have been found to have a high incidence of impaired color discrimination, olfactory dysfunction, and dysautonomia.³¹

Loss of central dopaminergic pathways involved in oculomotor control probably accounts for the impairment of saccades, observed even in early PD,³² but saccadic latency and amplitude have not been found to be particularly sensitive measures of disease progression, chiefly because of marked intrasubject and intersubject variability and low reproducibility over time.³³

Severe dysautonomia is generally present in the advanced stages of PD, when progression of the disease and adverse effects of treatment play a role in the development of these symptoms. When prominent dysautonomia occurs early in the course of the disease, it is usually indicative of atypical parkinsonism, such as multiple-system atrophy.^{4 ,34} Of the various symptoms of autonomic dysfunction associated with PD, constipation tends to appear earliest, and epidemiological studies have shown that constipation can precede motor symptoms by many years.³⁵ Other gastrointestinal tract motility problems such as regurgitation, nausea, and epigastric discomfort, likely related to gastroparesis, have also been reported to predate motor symptoms in some cases of PD.³⁶ Sexual disturbances, orthostatic hypotension, and urinary disturbances are common in advanced stages of PD, but these autonomic features occasionally precede motor symptoms of PD.³⁵ Because these various dysautonomic symptoms are very common in the general population and relatively nonspecific for neurodegenerative disorders, it is unlikely that they will emerge as reliable biomarkers of PD.

In untreated patients with early PD, depression (37%), apathy (27%), sleep disturbances (18%), and anxiety (17%) are the most common neuropsychiatric symptoms.³⁷ Several lines of evidence support the concept that depression in PD probably has a biological rather than a pure psychological reactive basis³⁸ because it is detected in up to 27.6% of patients with PD in the early stages of the disease³⁹ and it may precede the development of motor manifestations by several years.³⁷ Several investigators have also commented on premorbid parkinsonian personality. Twin and other studies have suggested that patients with PD, even since childhood, are more introverted, serious, cautious, tense, nervous, obsessive-compulsive, inflexible, industrious, and honest; that they avoid risk-seeking behavior such as smoking; and that they are less likely to exhibit impulsive or novelty-seeking behavior compared with controls.^{40 - 43} Despite these characteristic psychological and personality changes occurring even in early stages of PD, they are relatively nonspecific, and it is unlikely that these traits can be used to reliably predict increased risk of PD in the general population.

The pathological hallmark of PD is degeneration of dopaminergic neurons in the SN pars compacta, coupled with intracytoplasmic proteinaceous inclusions known as LBs. Braak et al,⁴⁴ however, have suggested that neurodegeneration of nondopaminergic neurons, particularly those in the caudal brainstem such as the dorsal motor nucleus of the vagus and the olfactory regions, precedes the onset of dopaminergic pathological changes in the SN pars compacta. This Braak hypothesis, which proposes that synuclein pathological changes in PD start in the lower brainstem and progress following a predictable caudal-rostral pattern, reaching the SN in the mesencephalon only after extensive involvement of the brainstem has occurred, lends support to the notion that nonmotor features reflecting this prenigral involvement can antedate the classic motor features of PD. Although the Braak hypothesis is supported by early olfactory, sleep, and autonomic involvement in patients with PD, the staging proposal has been challenged for many reasons and inconsistencies, such as absence of cell counts to correlate with the described synuclein pathological features, absence of immunohistochemistry to identify neuronal cell types, absence of observed asymmetry in the pathological findings that would correlate with the well-recognized asymmetry of clinical findings, absence of bulbar symptoms as early features of PD, and the observation that brain synucleinopathy consistent with Braak stages 4 and 6 has been found in individuals without any neurological signs.⁴⁵ In addition, there is controversy as to the classification of dementia with LBs, viewed by Braak et al⁴⁴ as part of stage 6 of the progressive pathological changes but regarded by others as a separate entity because these patients often have behavioral and psychiatric problems before the onset of motor or other signs of PD. Furthermore, the Braak hypothesis is not consistent with the natural history of PD.⁴⁶ There are several other concerns about the Braak hypothesis,^{2 ,47} such as the possibility of a multicentric origin of PD based on the presence of α-synuclein pathological features of Braak stage 3 in cases with incidental LBs and the involvement of the rostral brainstem and spinal cord without caudal medullary inclusions.⁴⁸ Furthermore, many patients with Braak stages 5 and 6 pathological features at autopsy have no overt cognitive or motor impairment⁴⁹ ; 24% of PD cases that come to autopsy satisfy the criteria for dementia with LBs and evidence of diffuse cortical involvement at disease onset; there is lack of evidence of caudorostral spread in 47% of 71 cases of PD; and 7% to 8% of PD cases had no synuclein inclusions in the dorsal motor nucleus of the vagus.^{47 ,50} Some have also criticized the study by Braak et al⁴⁴ because of potential selection bias in favor of cases with involvement of dorsal motor nucleus of the vagus, exclusion of cases of dementia with LBs, and no mention of spinal cord or peripheral system involvement. Despite these concerns, there is compelling evidence that pathological changes of PD extend well beyond the nigrostriatal system, involving cholinergic neurons of the nucleus basalis of Meynert, norepinephrine neurons of the locus ceruleus, serotonin neurons in the midline raphe, and neurons in the cerebral cortex, brainstem, spinal cord, and peripheral autonomic nervous system (eg, epicardium and myenteric plexus).⁴⁸ These changes may occur even before the involvement of the dopaminergic system.² Indeed, based on the interesting observation that human embryonic dopaminergic cells implanted in the striatum of patients with PD may develop PD pathological features, including classic LBs, has led to the hypothesis that α-synuclein may act like a prion and spread, possibly even from the periphery, and propagate throughout the nervous system in PD.⁵¹

Lewy bodies are detected in the brains of about 10% of clinically normal people older than 60 years, and some have suggested that cases with such incidental LBs (iLBs) represent preclinical PD.^{2 ,44 ,48 ,52} This hypothesis is supported by the finding of reduced immunoreactivity for tyrosine hydroxylase and vesicular monoamine transporter 2 in the putamen of these brains compared with those of healthy controls, and the decline in dopaminergic immunoreactivity correlates inversely with SN neuron loss and PD stages.⁵² Decreased tyrosine hydroxylase immunoreactivity in the striatum and in the epicardial nerve fibers of the same individuals was also found in another study that compared the brains and hearts of patients with iLBs with those of healthy controls.⁴⁸ The distribution of LBs in iLB disease is similar to the distribution in PD, but neuronal populations vulnerable to LB pathological changes do not show significant neuronal loss in brains with iLBs. These findings indicate that brains with iLBs have nigrostriatal pathological features that are intermediate between those found in healthy controls and those with PD and provide evidence that the presence of iLBs might represent preclinical PD.^{48 ,52}

The anterior olfactory region in particular has been suggested to be one of the earliest sites of neurodegeneration in PD.^{2 ,44} Hawkes et al⁵³ have revised the original LB staging scheme and proposed that synucleinopathy in PD begins in the olfactory bulb and within enteric cell plexuses. This has led to the so-called dual-hit hypothesis, which proposes that an unknown neurotropic pathogen initiates the pathological process underlying sporadic PD in an anterograde direction via olfactory pathways and in a retrograde direction, via enteric plexuses and preganglionic vagal fibers. The pathogen will then presumably enter the pons and then midbrain until the SN is reached and typical changes associated with PD emerge.⁵³

Several studies have provided evidence that the paravertebral sympathetic ganglia, the enteric system, and the epicardium are affected by synuclein pathological changes at early stages of PD.^{54 - 55} In an autopsy cohort of 98 neurologically unimpaired subjects 64 years or older, examination of the brain, the sacral and thoracic autonomic nuclei of the spinal cord, and several components of the peripheral autonomic nervous system with α-synuclein immunostaining showed that the autonomic nuclei of the spinal cord and the peripheral autonomic nervous system are consistently the earliest affected regions, after medullary structures and the olfactory nerves in neurologically unimpaired older individuals.⁵⁴ This provides a pathological basis for early premotor autonomic dysfunction at a prodromal stage of PD. The frequent presence of α-synuclein aggregates in peripheral autonomic neurons (eg, abdominopelvic autonomic plexuses) may represent an early presymptomatic phase in the development of LB disorders.^{35 ,55} The early involvement of the peripheral autonomic nervous system in PD is also supported by decreased cardiac uptake of iodine 123–labeled metaiodobenzylguanidine in single-photon emission computed tomography (SPECT) studies.^{26 ,56} The loss of sympathetic innervation of the heart may occur even before changes in the dorsal vagal nucleus,⁵⁷ previously thought to be the earliest pathological features of PD.⁴⁴ Although these studies suggest that meta-iodobenzyl guanidine I 123 may be a useful technique in detecting early stages of PD, further studies are needed to validate these observations.

With the use of punch skin biopsy with immunostaining and a panneuronal marker for protein gene product 9.5, PD-related changes are found in the small-caliber cutaneous fibers, indicating involvement of the peripheral (skin) autonomic nervous system in early PD.⁵⁷ There is no difference in skin biopsy scores between treated and untreated patients; therefore, the morphological findings are not related to anti-PD medications and probably reflect peripheral manifestation of PD. The presence of α-synuclein deposits in the dermis of a patient with pure autonomic failure provides evidence that this disorder and others associated with autonomic failure, including PD, dementia with LBs, and multiple-system atrophy, should be viewed as variant synucleinopathies.⁵⁸

In addition to pathological studies, various imaging techniques have been used to study early and presymptomatic stages of PD.^{8 ,59 - 61} Functional imaging can be used to detect preclinical evidence of DA deficiency in people deemed to be at increased risk of PD because of genetic or environmental risk or other premotor features (Table 1).

SPECT used to image DAT has consistently found reduced binding in patients with PD, which correlates with the severity and asymmetry of motor clinical scores.^{8 ,64} Measures of striatal N-ω-fluoropropyl-2β-carbomethoxy-3β-4-[¹²³I]iodophenyl-nortropane ([¹²³I]-FP-CIT) binding in patients with de novo PD correlated well with clinical scales of disease severity, even in the initial phases of the disease, suggesting that this imaging abnormality is a useful biomarker of early PD.⁶² DAT SPECT is also helpful for differentiating patients with PD from those who exhibit parkinsonian symptoms but have scans without evidence of dopaminergic deficit (SWEDDs).⁶³

Positron emission tomography (PET) using carbon 11–tagged dihydrotetrabenazine to label the vesicular monoamine transporter 2, a variety of ¹¹C- or ¹⁸F-labeled ligands for DAT, or 6-fluorodopa F 18, which assesses uptake and decarboxylation of levodopa and vesicular storage of radiolabeled dopamine, can all be used to detect premotor PD or the preclinical period of PD.^{7 ,61 ,64} Reduced DAT binding was the earliest indication of dopaminergic dysfunction in these individuals, indicating that DAT imaging provides a very sensitive index of subclinical deficits related to abnormalities in the nigrostriatal pathway.⁶⁸ Imaging presynaptic dopaminergic terminals with ¹¹C-dihydrotetrabenazine PET can also show nigrostriatal deficits in preclinical PD and might serve as an objective biomarker of PD severity.⁵ Although using PET for the preclinical diagnosis of PD is expensive and not widely available, continued longitudinal follow-up of the asymptomatic at-risk subjects is essential in studying the phenoconversion from preclinical dopaminergic dysfunction to clinical disease, in improving the estimate of the duration of the preclinical period, and in better characterizing the progression of PD.

Several studies have indicated that SN hyperechogenicity, typically found in clinically manifest PD, might also be seen in subjects with a preclinical impairment of the nigrostriatal system. A significant positive correlation was found between the echogenic area of the SN and the concentration of iron and of H- or L-chain ferritins in postmortem brains compared with healthy subjects at different ages.⁶⁵ The use of a combined approach, such as identification of olfactory and autonomic dysfunction, depression, RBD, and other premotor symptoms and detection of hyperechogenicity of the SN, may be an effective strategy for early diagnosis of PD.⁶⁶

Similar to sonography, high-field-strength magnetic resonance imaging (MRI) demonstrated abnormalities in the lateral SN pars compacta of early PD consistent with increased iron content, corresponding to the known distribution of neuronal loss.⁶⁷ Furthermore, using high-resolution diffusion tensor imaging, fractional anisotropy was found to be reduced in the SN of patients with early untreated PD compared with controls.⁶⁰ Post hoc analysis showed, as expected, that reduced fractional anisotropy in patients with PD was greater in the caudal compared with the rostral region of interest. A receiver operator characteristic analysis in the caudal SN showed that sensitivity and specificity were 100% for distinguishing patients with PD from healthy individuals. These findings provide evidence that high-resolution diffusion tensor MRI focusing on the SN distinguishes patients with mild PD from healthy individuals and has the potential to serve as a noninvasive biomarker of early PD.

Although the presymptomatic phase of PD has not been well characterized, there is a general consensus that neuronal damage, as a result of complex genetic, environmental, and other factors, starts years or even decades before the onset of symptoms.^{14 ,44} Understanding these interactions may enable identification of early biomarkers of degeneration and aid in preclinical screening and, eventually, identification of molecules that can halt or delay the neurodegenerative process.^{69 - 70} This research will also provide insights into the variable penetrance of the various genetic forms of PD.⁷¹

Using a variety of assays to measure α-synuclein protein concentrations, several studies have shown that patients with PD have significantly lower α-synuclein levels in their cerebrospinal fluid (CSF) than the control groups, even after adjusting for sex and age.^{72 - 73} Levels of α-synuclein in the CSF have been found to be decreased in patients with parkinsonism linked to synucleinopathy, such as PD and dementia with LBs,⁷⁴ and in patients with PD (n = 177) compared with patients with Alzheimer disease (n = 50) and healthy controls (n = 132).⁷³ One study suggested that α-synuclein levels in lymphomonocytes are affected by age, sex, and 20S proteasome activity in controls.⁷⁵ However, further studies are needed before quantification of α-synuclein levels in the CSF or blood can be adopted as a potential laboratory marker in the clinical diagnosis of PD⁷⁶ (Table 2).

Although much attention has focused on the role of α-synuclein in the pathogenesis of PD, a more common autosomal dominant PD detected in about 10% to 15% of familial cases is caused by mutations in LRRK2 (*609007).⁷⁷ Although responsible for only 1% to 2% of sporadic cases of PD, in some genetically isolated populations such as the Ashkenazi Jews⁸⁶ and North African Arabs,⁸⁷ the prevalence of some LRRK2 mutations, such as G2019S, is much higher and reaches 30% to 40% in familial cases. Testing for such common gene mutations may evolve into a valuable screening tool in these enriched populations.

Levels of DJ-1 (OMIM *602533), another multifunctional protein with loss of its function in gene mutations associated with familial, autosomal recessive PD, have been found to be decreased in the CSF of patients with PD as determined by highly sensitive and quantitative assays (Luminex Corporation, Austin, Texas).⁷³ When combined with other CSF markers, including β-amyloid 1-42, total tau, and interleukin (IL) 8, these assays seem to be able to differentiate patients with PD from those with other neurodegenerative disorders, such as Alzheimer disease and multiple-system atrophy, and from healthy controls.⁸⁸ Mutations in the parkin gene (OMIM *602544), identified as the major cause of autosomal recessive early-onset PD and several types of tumors,⁸⁹ may result in an early-onset PD, and detection of such mutations may serve as a potential biomarker of PD.⁹⁰

NURR1 (NR4A2) (OMIM *601828), encoding a member of nuclear receptor superfamily, is a transcription factor essential for the development, survival, and functional maintenance of midbrain dopaminergic neurons.¹⁰ We studied 278 patients with PD, 166 healthy controls, and 256 controls with neurological disease and found that NURR1 gene expression in human peripheral blood lymphocytes, as measured by quantitative real-time polymerase chain reaction analysis, was significantly decreased in patients with PD.⁷⁹ Although it is unlikely that this will evolve into a reliable test for preclinical PD because of moderate overlap between the various study populations, measuring gene expression signals in blood is a potentially fruitful area of research into biomarkers for the prediction of PD. A transcriptome-wide scan in 105 individuals to interrogate the molecular processes perturbed in the cellular blood of patients with early-stage PD identified a pattern of 22 differentially regulated genes in peripheral blood leukocytes of patients with PD.⁸⁰ These include the cochaperone ST13 (OMIM *606796), which stabilizes heat shock protein 70, a modifier of α-synuclein misfolding and toxicity. The investigators found that the number of ST13 messenger RNA copies is lower in patients with PD than in controls (mean [SE], 0.59 [0.05] vs 0.96 [0.09] copies; P = .002) in 2 independent populations. Thus, measuring gene expression signals in the blood can facilitate the search for biomarkers of PD. Furthermore, the simultaneous detection of multiple misregulated genes and their messenger RNA or protein expressions may have advantages compared with the determination of single proteins and may allow the identification of networks of interacting proteins that reflect the underlying disease process.

There are many other clinical or laboratory abnormalities that can potentially serve as biomarkers for the prediction of PD. For example, several studies have found that urate, a potent antioxidant, could be a protective factor or a biomarker in the diagnosis of PD (Table 2).^{78 ,81 ,91 - 93} Using metabolomic profiling with high-performance liquid chromatography coupled with electrochemical coulometric array detection, Bogdanov et al⁸¹ found that uric acid levels were significantly reduced in patients with PD. Other studies have shown that higher urate levels in the blood and CSF may decrease the risk of PD and predict slow progression.^{78 ,91} These findings have led to a prospective, longitudinal study of inosine—a nutritional supplement known to raise urate levels—designed to determine whether pharmacologically elevating serum urate levels could be an effective disease-modifying therapy in PD.⁹² Despite its many limitations, this study is an example of how discovery of a potential biomarker can lead to an early therapeutic intervention.⁹³

One study found that men with high plasma concentrations of IL-6, an index of inflammation thought to be important in the pathogenesis of PD, have an increased risk of developing PD.⁹⁴ However, this finding should be interpreted with caution because of the small sample size and the lack of associations with other biomarkers of inflammation. More recently, another study also indicated that the elevation of IL-10 levels and the significant correlation between IL-10 and IL-12, a proinflammatory cytokine, suggests that immunological disturbances are involved in patients with PD,⁹⁵ but its role as a biomarker for the early diagnosis of PD needs to be validated.

Other studies provide a signal for a potentially increased risk of PD. In a study of 236 patients, low levels of low-density lipoprotein were associated with a 3.5 times higher-than-expected risk for development of PD, and the follow-up study confirmed the association between low cholesterol levels and the risk of PD.⁹⁶ Increased PD risk has been associated with decreasing darkness of hair color, with highest risk in red-haired and blond individuals, suggesting a role of pigmentation in the pathogenesis of PD.⁹⁷ These reports, although interesting and worth pursuing, require validation before the findings can be considered as potential biomarkers of PD.

Progress in our knowledge of the system biology tools, including genomics, proteomics, lipidomics, and metabolomics, and the use of microarray, spectroscopy, and other techniques to measure multiple-gene messenger RNA expression signals in various tissues, including blood, should facilitate the development of more sensitive and specific biomarkers for the prediction of PD.^{80 ,84} These advances should provide the basis for the development of new and reliable biomarkers of PD that eventually may be instrumental in the development of new treatment strategies.^{71 ,81 ,83} More recently, glycoproteomics, a branch of proteomics that catalogs and quantifies glycoproteins, provides a powerful means to systematically profile the glycopeptides or glycoproteins of a complex mixture that are highly enriched in body fluids, with the great potential to be diagnostic and even prognostic markers of neurodegenerative disorders such as PD.⁸² Metabolomic profiling that uses electrochemical coulometric array detection to search for biomarkers in plasma may also uncover potentially useful biomarkers. One study, for example, showed that 8-hydroxy-2-deoxyguanosine levels (a marker of oxidative damage to DNA) were significantly increased in patients with PD but overlapped with those of controls.⁸¹ Novel methods involving the validation of blood transcriptomic markers for PD could also be promising avenues of future research.⁸⁵

Development and validation of disease-specific biomarkers for the diagnosis of early PD represents one of the most urgent unmet needs in neurology. Most of the current markers of the disease, such as neuroimaging abnormalities detected by SPECT, meta-iodobenzyl guanidine I 123, sonography, PET, or MRI, have been tested in patients with moderate or advanced disease, are not disease specific, and have not been validated in early preclinical stages of the disease.⁹⁸ The sensitivity and specificity of the various imaging techniques are currently too low and the cost is too high to use these techniques as screening tools in the general population. The search for biomarkers is hampered by a number of challenges, including the difficulties in collecting postmortem specimens for validation studies. The future goal is to find a reliable biomarker of neurodegenerative diseases in peripheral blood, the CSF, or other readily accessible tissue. Such preclinical biomarkers are essential in identifying at-risk populations that might be potential targets for neuroprotective or disease-modifying therapeutic strategies. Although the concept of preclinical PD as a natural target for neuroprotective therapies and primary prevention is not new, neuroprotective trials have yet to be performed in at-risk individuals, not only because of the lack of reliable predictive risk markers for PD but primarily because of the lack of proven disease-modifying therapies.^{99 - 100} In addition to sensitive and specific diagnostic biomarkers, there is also an urgent need for reliable surrogate biomarkers that can be used to accurately track progression of the disease. The progress in developing sensitive, specific, accurate, and reliable biomarkers of PD has been rather slow. However, with further refinements of various clinical, neurophysiological, and neuroimaging tests and when coupled with other laboratory biomarkers using blood, CSF, or other tissues and applying novel techniques in genomics, proteomics, lipidomics, and metabolomics, it is likely that clinically useful biomarkers for the prediction of PD will eventually be developed. Currently, a multicenter biomarker study (Parkinson's Progression Markers Initiative [PPMI]) is under way in North America and Europe, a landmark observational clinical study designed to comprehensively evaluate a cohort of recently diagnosed patients with PD and healthy subjects using advanced imaging, biological sampling, and clinical and behavioral assessments to identify biomarkers of Parkinson disease progression (http://www.ppmi-info.org). These biomarkers can then be used effectively to aid in detecting asymptomatic individuals with early disease who may benefit from early therapeutic intervention.

Correspondence: Joseph Jankovic, MD, Parkinson's Disease Center and Movement Disorders Clinic, Department of Neurology, Baylor College of Medicine, 6550 Fannin, Ste 1801, Houston, TX 77030 (josephj@bcm.tmc.edu).

Accepted for Publication: March 15, 2010.

Author Contributions:Study concept and design: Wu, Le, and Jankovic. Acquisition of data: Wu. Analysis and interpretation of data: Wu and Jankovic. Drafting of the manuscript: Wu and Le. Critical revision of the manuscript for important intellectual content: Wu, Le, and Jankovic. Statistical analysis: Wu. Administrative, technical, and material support: Wu and Jankovic. Study supervision: Le.

Financial Disclosure: Dr Jankovic reports having received grants and compensation from the Michael J. Fox Foundation for Parkinson Research, including a collaborative grant with Dr Le for a study of biomarkers in Parkinson's disease, which does not constitute a relevant conflict of interest related to the article.

Funding/Support: This study was supported by grant 5U10NS044441-07 from the National Institute of Neurological Disorders and Stroke, the Clinical Center for the Study of Neuroprotection from PD, and the National Parkinson Foundation through their support of the National Parkinson Foundation Center of Excellence at Baylor College of Medicine. The authors acknowledge the joint participation by the Diana Helis Henry Medical Research Foundation through its direct engagement in the continuous active conduct of medical research in conjunction with Baylor College of Medicine and this program. This work was partially supported by the Research Program of Shanghai Municipal Health Bureau (2010106) (Dr Wu) and a grant from the National Basic Research Program of China (the 973 Program) (2011CB707500) (Dr Wu).