|Home | About | Journals | Submit | Contact Us | Français|
Cognitive impairment, including dementia, is commonly seen in those afflicted with Parkinson disease (PD), particularly at advanced disease stages. Pathologically, PD with dementia (PD-D) is most often associated with the presence of cortical Lewy bodies, as is the closely related dementia with Lewy bodies (DLB). Both PD-D and DLB are also frequently complicated by the presence of neurofibrillary tangles and amyloid plaques, features most often attributed to Alzheimer disease. Biomarkers are urgently needed to differentiate among these disease processes and predict dementia in PD as well as monitor responses of patients to new therapies. A few clinical assessments, along with structural and functional neuroimaging, have been utilized in the last few years with some success in this area. Additionally, a number of other strategies have been employed to identify biochemical/molecular biomarkers associated with cognitive impairment and dementia in PD, e.g., targeted analysis of candidate proteins known to be important to PD pathogenesis and progression in cerebrospinal fluid or blood. Finally, interesting results are emerging from preliminary studies with unbiased and high throughput genomic, proteomic and metabolomic techniques. The current findings and perspectives of applying these strategies and techniques are reviewed in this article, together with potential areas of advancement.
Parkinson disease (PD), the second most common neurodegenerative disease after Alzheimer disease (AD), has been recently recognized as a disease that encompasses more than just nigrostriatal degeneration with motor dysfunction (43, 74). More specifically, a significant portion of PD patients suffer from non-motor symptoms, including anosmia, constipation, depression, autonomic failure, and cognitive dysfunction. There is growing evidence that cognitive symptoms are frequently present at the time of diagnosis (40), contribute heavily to disability (144), and progress to dementia at a rapid rate. PD with dementia (PD-D) shares a common neuropathological pattern, cognitive profile, and clinical course with a closely related condition, dementia with Lewy-bodies (DLB). The pathological hallmark of both PD-D and DLB is the presence of Lewy bodies — ubiquitin-positive and α-synuclein (α-Syn)-enriched intracytoplasmic neuronal inclusions — in neocortical and paralimbic regions, which seems to be one of the main pathological correlates of dementia in PD (3, 9). The cognitive profile of PD-D (and DLB for that matter) also overlaps partially with that of AD, the most common form of geriatric dementia, and AD-type pathology (such as neurofibrillary tangles and amyloid plaques) frequently co-exists with Lewy body pathology, making correct diagnosis of PD-D difficult even in the best hands. The sensitivity and specificity of the recently published PD-D diagnostic criteria (35, 38) have not yet been explored, but, as expected, the diagnosis will be challenging when patients fall in between the two classic profiles of DLB and PD-D (77). Even for clinical DLB, most studies based on the consensus criteria for clinical diagnosis of DLB (89, 90) have reported a high specificity (80–100%) but a low sensitivity (20–60%) (87), meaning that the consensus criteria are appropriate for confirmation of diagnosis but are of limited value in screening for DLB. In addition, distinguishing the Lewy body disorders from non-Lewy body dementias, including AD, remains difficult when atypical presentations occur during or at early stages of the diseases (78, 79, 88).
As such, identifying individuals who are predisposed to developing cognitive impairment and dementia in PD and other neurodegenerative disorders at early stages of disease progression will allow physicians to better assess the individual's prognosis and make informed decisions on the best course of management (141). In addition, potential neuroprotective or neurorestorative therapies and secondary prevention efforts might be expected to have their greatest impact during the earliest phases of the disease process. Therefore, during the development and implementation of neuroprotective therapies, it will be important to develop robust biomarkers - whether based on clinical assessment, structural/functional imaging, or biochemical measurement in body fluids - that can aid in specific and accurate diagnosis, particularly at early disease stages, and can objectively monitor progression and response to treatment. Biochemical/molecular markers, particularly, could also potentially reveal mechanisms involved in disease pathogenesis and thereby provide new targets for therapy.
Significant effort has been spent in several areas to define biomarkers for cognitive impairment in PD, including physiological/sensory testing, structural/functional neuroimaging, and biochemical determination of various molecules. Table 1 lists advantages, disadvantages, and the current status of the major techniques employed for PD-D and DLB biomarker discovery. Of note, the applications of structural and functional neuroimaging and electrophysiological techniques on the diagnosis of PD-D and DLB have already been extensively discussed previously (1, 6, 43, 112, 129).
Differences exist in the electrophysiological profiles of different types of dementia, both in the amount of slowing as well as the distribution of electrophysiological changes. Several electrophysiological studies have shown more slowing of electroencephalographic (EEG) rhythm in DLB than in AD (2). A recent study has shown that patients with DLB and PD-D have more posterior slow-frequency variation than patients with AD and healthy controls (18). Similarly, another study, which assessed mismatch negativity (MMN) of the auditory event-related potential (ERP), concluded that reduced latency and areas were found in PD-D, compared with AD, PD without dementia (PD-ND) patients, and healthy controls; furthermore, more pronounced changes were found in PD-D than those observed in DLB (22). Studies also reported that PD-D patients had distinctly slower baseline EEG activity than PD-ND patients (43), and that a positive correlation occurred between EEG frequency and cognitive scores (126). Although these results indicate that EEG and ERP may provide a potential diagnostic marker to distinguish between Lewy body disease (including PD-D and DLB) and other dementia diseases, most of these studies were based on small numbers of patients and provided no evidence indicating sensitivity and specificity values.
Additionally, olfactory dysfunction has been identified in PD-D and DLB (55, 117); but this observation is nonspecific because patients with AD have similar olfactory deficits. It could be that the main utility of olfactory testing may be to differentiate neurodegenerative diseases from non-neurodegenerative dementia, such as multi-infarct state, that have no impairment of olfaction (45).
Neuroimaging studies using structural magnetic resonance imaging (MRI) have demonstrated that PD-D is associated with cortical atrophy and that the pattern of atrophy differs from that of AD (24), with significant differences between PD-D and PD-ND reported bilaterally in the occipital lobes. Greater whole-brain atrophy rates were observed in PD-D patients than in PD-ND patients and healthy control subjects (23). Regional changes in key areas such as caudate (7), hippocampus (25), and amygdalae (69) were also reported to be associated with PD-D. Notably, atrophy of the amygdalae and the head of hippocampus were reported recently in PD patients without dementia, indicating early memory impairment (20, 68). Similarly, two recent studies (12, 92) have reported areas of brain atrophy in PD with mild cognitive impairment (MCI) for early changes. Still, there is considerable overlap of regional patterns of atrophy that occur between PD-D and PD-ND and between PD-D and DLB (25, 69). Moreover, sample sizes have not been large enough in most studies to allow for statistical modeling to control for the increased motor impairment and longer disease duration often seen in PD-D compared with PD-ND patients (38, 43).
Functional neuroimaging modalities, including positron emission tomography (PET), single photon emission computer tomography (SPECT), and functional MRI (fMRI), provide valuable information on brain function in the neuropsychiatric aspects of PD. A characteristic posterior perfusion deficit or decreased cerebral blood flow has been shown by SPECT in PD-D as well as DLB (8), but whether this observation is sufficiently distinctive for PD-D and clinically useful for diagnostic purposes remains to be investigated (38, 43). fMRI has also been used to identify brain changes associated with visual hallucinations in PD (132), and the use of 18F-fluorodeoxyglucose (FDG) PET to study non-demented PD patients while they were performing executive and memory tasks, has identified a metabolic pattern (60). It has appeared that changes in this network could distinguish PD with and without MCI (59), and more importantly, changes in this network with time (i.e. progression) were identified (61).
Both PET studies using 18F-fluorodopa and SPECT studies using a range of ligands that bind to the dopamine transporter (DAT) have been shown to identify dopaminergic deficits in PD (45). A growing body of literature shows that these same dopaminergic deficits can be found in PD-D or DLB and that these imaging modalities may be used to differentiate Lewy body disorders from AD. 18F-dopa uptake in PD-D has been reported to be significantly decreased in several areas compared with PD (64); however, in other studies, no differences in dopaminergic loss and rate of dopamine decline were found in PD-D and PD-ND patients (32, 115). Also, functional imaging of the cholinergic system has shown substantial cholinergic denervation in PD-D patients as compared with PD-ND patients (17, 57), although differences were not always significant (17).
In PET studies, 11C-Pittsburgh Compound B (PIB) binds to amyloid with a pattern similar to the histopathologically known distribution in AD. Elevated PIB-binding in the brain stem of PD was recently reported, and in a subgroup, cortical PIB binding was indeed found (81, 82), indicating that PIB-PET could be used to further differentiate PD-D with respect to cortical amyloid.
Finally, though currently preliminary, a few novel functional neuroimaging based biomarkers for PD-D have been investigated. For instance, lactate/N-acetyl aspartate ratio was found to be increased in the occipital lobes in PD, particularly in PD-D patients as compared to healthy controls, suggesting that the impairment of oxidative energy metabolism is greater in PD-D (21, 38). Additionally, using proton magnetic resonance spectroscopy, N-acetyl aspartate was found to be decreased in the occipital lobes in PD-D patients as compared to healthy control or PD-ND subjects (134). Also, the uptake of cardiac metaiodobenzylguanidine (MIBG), a specific marker for noradrenergic transporters, is significantly reduced in PD (135, 147) as well as in DLB (105). Although specific studies in PD-D are lacking, MIBG has been found to be more accurate than occipital hypoperfusion using SPECT (51) or CSF markers (140) as a means of discriminating between DLB and AD. Cardiac MIBG may differentiate Lewy body-related dementias from non-Lewy body related types, but cannot differentiate PD-D and DLB.
Taken together, group differences between PD-D/DLB and PD-ND patients have been documented in electrophysiological/sensory studies as well as structural and functional imaging investigations; however, none of these techniques can be recommended for current routine diagnostic purposes because of inadequate sensitivity or specificity, or relatively high expenses (43). Consequently, additional biomarkers with higher sensitivity and specificity, which can also be applied widely and at a reasonable cost, are needed for greater accuracy and diagnostic utility, ideally at preclinical or early stages of disease when therapy is likely to have the greatest impact.
Familial aggregation of PD-D and an increased frequency of dementia in the relatives of PD or PD-D patients have been reported in some studies (73, 85), probably reflecting an influence of genetic factors in the development of dementia in PD. The precise role of genetic factors in development of MCI and dementia in PD is unclear currently, though a few candidate genes have been implicated.
Based on these findings, it is clear that genetic testing, at least in its current form, is not a convenient or reliable source of biomarkers for diagnosis of cognitive impairment in PD.
Human CSF, being most proximal to the CNS, is one of the ideal sources for identifying biomarkers for neurodegenerative diseases. An important consideration is relative availability of CSF obtained by lumbar puncture, making it possible to conduct longitudinal molecular analyses of changes in CSF during the course of diseases. In addition to CSF, the use of plasma, urine, and saliva for biomarker discovery have also been explored as more practical/acceptable approaches; nevertheless, with only few exceptions these attempts have met with limited success, primarily because of the complexity of peripheral samples and confounding factors introduced by other organ systems (123). It was proposed that the best approach might be to establish an ensemble of effective biomarkers in well-characterized human CSF of a relatively small number of individuals involved in research settings, then subsequently pursue their quantification in plasma or urine/saliva for more wide spread application (148). The biochemical/molecular candidate markers will be reviewed in two categories – targeted analysis and unbiased profiling.
Other candidate molecular/biochemical markers include hypocretin, a protein related to the regulation of sleep and wakefulness (94, 146), tissue transglutaminase (138) and serpins (102) that might be associated with aggregation of α-syn, homocysteine (63), neurotransmitter metabolites such as homovanillic acid (HVA) and methoxy-hydroxy-phenyl-ethylene glycol (MHPG) (5, 27, 108), growth hormone in serum (31, 70, 110, 111, 118), osteopontin (80), amino acids relevant for synthesis of nitrogen monoxide (72, 84, 95), and metals such as iron and zinc (41, 42, 66, 114). Unfortunately, studies involving these candidate markers are either contradictory or very preliminary, and thus still difficult to interpret.
In summation, though the emerging body of work on the CSF and blood biomarkers of PD is encouraging (a list of recent findings, along with the associated references, is summarized in Table 2), one of the issues that has plagued the field is a lack of reproducibility between researchers. The disparity can be attributed to several factors, including variation in antibodies that might detect different species of the target protein, limited numbers of patients in some investigations, a less than vigorous quantitative method, or inadequate control for important confounding factors, particularly blood contamination in CSF (148). Further studies are needed to clarify these issues, ideally in CSF and blood samples with pathological verification of clinical diagnosis of each case.
While targeted analysis of biomarkers as discussed above is largely an extension of physiological studies that investigate known pathways involved in PD pathogenesis, emerging technologies are beginning to allow a systematic, unbiased characterization of variation in genes, messenger RNA (mRNA), proteins and metabolites associated with disease conditions and identification of novel biomarkers (49, 120, 123).
Over the last few years, proteomic techniques have been utilized to discover novel markers in different neurodegenerative disease settings, with AD being studied most extensively (33, 121, 123, 130). On the other hand, it has also become clear that the proteomics results reported by different groups vary substantially. Some of these variations may be biological, as human beings are extremely heterogeneous, and most diseases studied are very complex. To resolve this issue, multiple markers may be needed, as a combination of independent markers likely enhances the performance of the putative markers required (4, 29, 125). In addition, standardization of data reported is also critical to reducing inconsistency among different studies. It is almost certain that the differences in databases has contributed to the current inconsistent (and often contradicting) results in CSF biomarker discovery (148). As one of the initiatives to address this issue, we have recently summarized and standardized all of our CSF proteomics results, as well as made the original MS data available to investigators in the field (107) so that future discoveries can be compared to the published results meaningfully.
In proteomic studies for PD biomarker discovery in CSF, one recent study employed an unbiased quantitative proteomic approach called isobaric tagging for relative and absolute protein quantification (iTRAQ) to label human CSF from AD, PD, and DLB patients and healthy controls, and identified more than 1500 proteins; of those, 136, 72, and 101 proteins appeared to be uniquely associated with AD, PD, and DLB, respectively (4). Some of the unique proteins, including apolipoprotein H/beta-2-glycoprotein 1 (Apo-H), ceruloplasmin and chromogranin B (secretogranin I) that are unique to PD as well as apoC1 and T-cadherin that are unique to DLB, were further confirmed in the initial study using Western blotting. Additionally, the ability to distinguish DLB from PD and AD increased when several markers were combined. Besides the proteins confirmed in this initial investigation, several other markers have been tested in a different and larger set of subjects, in which a panel of eight CSF proteins (BDNF, IL-8, VDBP, β2-microglobulin, apoAII, apoE, plus tau and Aβ1-42) appears to be able to classify PD patients with 95% sensitivity and 95% specificity using bead-based Luminex multi-analyte profiling technology (149). Whether these markers are useful in detecting PD-D/DLB remain to be investigated.
It cannot be stressed enough that biomarker discovery does not have to start with body fluids. This is because protein biomarkers identified in human brain may also ultimately enter CSF and/or plasma that can be collected clinically in a living patient. These brain specific markers are difficult to identify in unbiased direct profiling of CSF because they are typically low in abundance, and all current proteomics technology is biased towards abundant proteins. Thus, defining proteins in well-characterized brain tissue, followed by targeted validation of candidate proteins in the CSF and blood is indeed an approach currently taken for PD and DLB biomarker discovery. In this regard, several tissue-based proteomics studies have identified many differentially expressed proteins in PD tissues with respect to controls (11, 67, 71, 106, 122, 124, 145). Some of the proteins, including mortalin (124) and glutathione S-transferase pi (122), were also identified to display significant differences between different PD stages. A subset of these proteins is under intensive investigation as to their contribution to PD development and progression. The deregulated proteins in PD tissue are also being interrogated as potential biomarkers in body fluids, including CSF and plasma.
The advantages of early diagnosis, combined with suboptimal clinical diagnostic accuracy, highlight the need for a valid biological marker(s) for PD with MCI and PD-D. Promising results have been reported from several neuroimaging and hypothesis-driven targeted analysis of proteins in CSF and blood. Unbiased discovery based on genomics, metabolomics, and proteomics has also yielded hopeful results. It is expected that future studies using these strategies will provide novel, more accurate, and less expensive biomarkers and explore whether a combination of different biomarkers can improve diagnostic accuracy. Another potential fruitful area in discovery of biomarkers for cognitive impairment might be the research on post-translational modifications (PTMs) of proteins, e.g. phosphorylation, ubiquitination, and glycosylation, as aberrant PTMs of proteins have now been recognized as an attribute of many mammalian diseases, including neurodegenerative diseases (10, 44, 91). To this end, phosphorylated and/or glycosylated α-syn have been clearly linked to PD pathogenesis, and it is possible that it is the modified α-syn, rather than the total level of α-syn, that reflects disease activity and prognosis. More detailed discussions of the current technologies being utilized as well as potential shortcomings and limitations of the analysis of PTMs as biomarkers of neurodegenerative disease can be found in recently published reviews (26, 148).
Finally, for future investigations, it is critical to address several important issues, including standardization of sample collection protocols, the specificity of the antibodies and the technologies used, and the need to recruit a larger number of subjects and controls, allowing for stratification against important variables typically associated with clinical investigations. These variables have likely contributed to inconsistent results, whether via targeted or unbiased profiling, obtained thus far by various investigators. Independent studies on different cohorts, including cases with pathological verification, are also needed for validation.
This work was supported by grants from the NIH (NS057567, NS060252, ES004696, AG033398, AG025327, and NS062684) and the C-M Shaw Endowment.