Platelets: Peripheral Biomarkers of Dementia?

Body fluids and blood cellular components for diagnostic use

In AD, the clinical diagnosis is supported by amyloid-β (Aβ₄₂), total tau and phosphorylated-tau₁₈₁ measures in the CSF [6]. CSF extraction (obtained via lumbar puncture) is invasive, and it is not widely used for diagnostic purposes. Furthermore, some of the identified CSF biomarkers (i.e., cytokines, adhesion molecules and growth factors) may be of limited diagnostic use as some CSF biomarkers are involved in secondary inflammatory processes [7] rather than being causative agents of dementia. Since clinically-beneficial biomarkers should be easily accessible and directly modified by dementia, their presence in peripheral body fluids (including blood and urine) and skin has been further investigated. In urine, a smaller concentration of brain-derived proteins, such as the AD-associated neuronal thread protein (AD7c-NTP), can also be detected, with higher levels found in subjects with both AD [8] and mild cognitive impairment (MCI) [9], though their clinical relevance need to be further researched. On the other hand, plasma biomarkers have been widely investigated in proteomic studies [10] revealing a great extent of heterogeneity with their clinical reliability and reproducibility questioned [11]. Nevertheless, blood remains the focus in the development of novel dementia diagnostic tools, and its cellular components including platelets, red blood cells, lymphocytes, leucocytes, and mononuclear cells may also prove to have diagnostic value [12–16].

Platelets as peripheral neuronal-like systems

Platelets are the closest and most accessible peripheral neuronal-like cellular system likely to provide a wealth of information about neuronal functioning in dementia. Since discovered by Giulio Bizzozero in 1882 [17], platelets have been exploited for their clinical value. Platelets house mitochondria, dense granules [18], surface proteins, and adhesive proteins making them ample protein reservoirs. Platelets are also a source of growth factors, coagulants and inflammatory mediators involved in their main function of wound healing. Interestingly, platelet proteomes are very similar to neurons but with a relatively lower biological variation when compared to other “housekeeping” proteins such as glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and tubulin [19]; platelets also liken to serotonergic neurons in their uptake of serotonin [20]. Platelets are anucleate which makes them more susceptible to disease states shown through their involvement in the pathology of a range of diseases (i.e., infection, anemia, pre-eclampsia, renal failure) [21–23]. Platelet AβPP is also processed via identical amyloidogenic and non-amyloidogenic pathways as brain AβPP [24]. The downregulation of non-amyloidogenic pathway enzymes (i.e., ADAM10) and upregulation of amyloidogenic pathway enzymes (i.e., BACE1) characteristic of AD pathology is also reported in platelets in AD subjects [25]. Upregulation of the amyloidogenic pathway leads to increased Aβ production and the formation of Aβ aggregates (Fig. 1).

OPEN IN VIEWER

Fig.1

Pictorial presentation of platelet AβPP processing via amyloidogenice and non-amyloidogenic pathways. An increase in the amyloidogenic pathway (combined with a decreased in the non-amyloidogenic pathway) leads to Aβ aggregation followed by a cascade of events resulting in activation of mitochondria and increased intra-platelet phosphorylation of proteins, including tau protein. Aβ, amyloid-β; AβPP, amyloid-β protein precursor; ADAM10, a disintegrin and metalloproteinase 10; BACE1, amyloid-β cleaving enzyme 1; sAβPP, soluble amyloid-β protein precursor; P, phosphorylated tau protein sites; MAO-B, monoamino-oxidase; Ca^2 +, calcium; ROS, reactive oxygen species; PLA₂, phospholipase 2. Modified from Catricala et al. [26].

The sensitivity and reactiveness of platelets to their biochemical environment, thus, makes them largely suitable for use in clinical diagnosis. Here we review changes in platelets that occur during the dementia process, and discuss the viability of unique platelet proteins, enzymes, and receptors as peripheral dementia biomarkers alongside the practicality of their use in routine clinical setting.

Mean platelet volume

A number of studies have confirmed an association between MPV and dementia. MPV was higher in control subjects, lower in patients with MCI, and lower still in AD [28, 29] (Table 1) suggesting that higher MPV is an indication of higher cognitive function; Mini-Mental State Examination (MMSE) scores were positively correlated with MPV [29]. The reason for this change in MPV seen in AD is unclear. It is intriguing to speculate that if platelets are theclosest peripheral cells to neurons, MPV may be used as a by-proxy measure of the central brain neuronal cell death that occurs with the progression of AD. In this respect, MPV levels may well be a marker of either pro-inflammatory disease mechanisms, or loss of synapses and development of neurofibrillary and amyloid pathology associated with AD.

Nevertheless, this may not be the case in other dementias such as VaD. Thus, MPV increases in silent stroke in comparison to healthy subjects but decreases in VaD, similarly to what happens in AD [30]. However, higher levels of MPV in idiopathic thrombocytopenia (ITP) patients are thought to increase the progression from ITP to VaD [31], thus suggesting that higher MPV in diseases such as ITP may be used as an indicator of diseases progression. Further work in relation to elucidating the relationship between MPV and distinct dementia mechanisms is needed to understand the pathophysiology of these changes and how they are related to distinct dementia syndromes.

There is also a need for further research on the potential of any clinical application of MPV as a biomarker of AD and other forms of dementia. Firstly, the standardization of MPV measures across the population needs to be established. Many other factors would similarly need to be considered. For instance, does MPV vary between the sexes and according to ethnicity? Does MPV change during the course of healthy aging? How is MPV influenced by comorbidity and polypharmacy? These all will need to be investigated in greater detail before MPV finds its clinical use as an indicator of dementia, distinct dementia syndromes and dementiaprogression.

Coated platelets

Coated platelets are produced when there is a dual stimulation of platelets with collagen and thrombin. These platelets are able to retain an augmented concentration of pro-coagulant, full-length AβPP molecules on their membrane [26]. Using flow cytometry, coated platelets have been quantified and shown to be significantly increased in the early stages of AD (also known as amnestic MCI), but reduced in later stages of the disease [26, 32] (Table 1). This inverse correlation between the presence of coated platelets and the severity of disease was most evident in AD, whereas FTLD patients had similar number of coated platelets in both early and advanced stages of FTLD [33]. In the light of these findings, coated platelets have a real potential to be used specifically as a biomarker of the progression of AD: amnestic MCI patients with elevated levels of coated platelets are at a higher risk of progressing to AD. Therefore, coated platelets may have further utilization in preclinical stages of cognitive impairment, namely to identify those at risk of developing AD.

Tau protein

Tau protein stabilizes axonal microtubules. Tauopathies cover a wide range of neurological diseases, i.e., AD, FTLD, and several subcortical dementia subtypes [34], including progressive supranuclear palsy, corticobasal degeneration, dementia pugilistica, post-encephalitic Parkinsonism, and chronic brain encephalopathy. Hyperphosphorylation of tau protein causes its dissociation from microtubules which decreases neuronal stability and promotes cell death [35–37].Tau aggregates in the form of neurofibrillary pathology are closely related to the symptoms of dementia, in particular, cognitive impairment and decreased language skills [34].

There are only a handful of dementia studies that have explored tau protein in platelets. Although a proteomic study failed to identify tau protein in platelets [38], several research groups have identified and quantified semi quantitatively and quantitatively platelet tau. Thus, the increase in platelet glycogen synthase kinase 3 beta (GSK3β) activity, known to hyperphosphorylate tau protein (Fig. 2), was significantly elevated even in the early stages of AD (during amnesic MCI) [39] (Table 1). However, there were no significant changes in total GSK3β levels during the course of AD [39]. A later quantitative study reported similar levels of GSK3β between AD and control subjects, with the elevation of platelet phosphorylated GSK3β occurring only following 6 months anti-dementia treatment with donepezil; this was accompanied by the reduction of another enzyme that is involved in the amyloidogenic pathway, phospholipase A₂ (PLA₂) [40]. This would suggest that cholinesterase inhibitors have a dual impact on the platelet dementia biomarkers: they act on both the reduction of amyloid deposits (via increasing in PLA₂ activity) and tau pathogenesis, due to the inhibition of the tau protein phosphorylation. This bears similarities with the neuropathological studies of decreased brain amyloid [41] and neuroprotective and/or neurorestorative effects involving both synapses and neuronal cytoskeleton [42] following anti-dementia treatment. This is another argument in support of platelets reflecting the central brain dementia mechanisms at the periphery.

OPEN IN VIEWER

Fig.2

Overview of some of the pathological changes that occur in platelets during the course of dementia as based on current published peer-reviewed research. A number of molecular changes that occur in platelets mirror the brain-specific dementia processes, in particular the altered AβPP processing, tau posttranslational modifications etc. However, there are still some gaps in research, i.e., MAO-B relationship to mitochondrial fluidity Aβ, amyloid-β: AβPP, amyloid-β protein precursor; ADAM10, a disintegrin and metalloproteinase 10; BACE1, AβPP cleaving enzyme 1; APOE, apolipoprotein E; GSK3β, glycogen synthase kinase 3 beta; MAO-B, monoamine oxidase B; PLA₂, phospholipase A₂; ROS, reactive oxygen species; sAβPP-β_, secreted AβPPβ.

The direct evidence for platelet tau protein involvement in AD largely comes from two research groups. Thus, the higher ratio of high and low molecular weight tau protein, as measured in semi-quantitative studies [36], appears to correlate with both the extent of cognitive impairment [37] (Table 1) and brain atrophy in cognitively strategic regions, as measured with neuroimaging tools [43]. In another quantitative study, however, the levels of both total and phosphorylated platelet tau protein were similar between AD and control subjects, and in AD, the platelet tau protein levels were not correlated with any of the cognitive and non-cognitive (i.e., depression and behavior) measures [44] (Table 1). However, the older AD patients (above the age of 80 years) with higher MMSE scores had higher levels of platelet tau protein [45]. While these studies suggest that changes in platelet tau can be utilized as a dementia biomarker, there is a need for wider research in platelet changes of tau protein in different types of dementia and devising more sensitive analytical tools for its measurements.

Amyloid-β protein precursor

AβPP is a transgenic membrane protein that exists in many isoforms, and human platelets are perhaps the biggest reservoir of AβPP in the body [45]. The pathogenesis of AD is largely attributed to the altered processing of AβPP which is due to the changed activity of secretase enzymes (α-, β-, and γ–secretases) (Fig. 1). These enzymes determine the rate and direction of AβPP splicing to Aβ peptide products and/or AβPP cleaved commodities. The main species of Aβ released from platelets include Aβ_1 - 40 with Aβ_1 - 42 that are also the most predominant within neuronal plaques in the brain tissue in older people with and without dementia, especially AD [46–48]. A reduction in platelet α-secretase levels and an increase in β-secretase activity have been reported in AD (Fig. 2) [7]. These both were attributed to decreased ratios of 120 kDa–130 kDa to 110 kDa platelets AβPP fragments [49] (Table 1). It is important to note that this increase in β-secretase did not correlate with the cognitive deficits as measured by the MMSE suggesting that the increase in β-secretase may be a pre-symptomatic physiological change occurring in dementia and not secondary to the onset of the disease. However, two independent studies, one measuring coated platelets [26], and another measuring the N-terminal end of AβPP [12] reported a negative correlation of these measurements with cognitive scores as measured by the MMSE. The differences in reported findings may be due to the small number of cases enrolled in the above studies, as well as the different methodologies used to determine platelet AβPP expression (Table 1). However, another study reported that platelets Aβ may not be a specific enough biomarker for AD because the changes in its platelet expression may be due to the aging process alone [50]. Further independent studies conducted on a larger number of clinically diagnosed AD patients with different stages in clinical progression are needed to determine the real nature of the noted platelets AβPP and Aβ changes in dementia.

α-Synuclein

α-Synuclein is found in abundance in erythrocytes and platelets [15, 65]. α-Synuclein is closely associated with the pathogenesis of Parkinson’s disease with and without dementia, as well as with dementia with Lewy bodies [66, 67]. The exact function of α-synuclein in normal brain is not yet known; however, the apparent upregulation of α-synuclein in the presynaptic terminal during acquisition-related synaptic rearrangement [68], and especially in preclinical stages of AD [69], suggests that α-synuclein is closely related to cognitive function and neural connections in the brain. In platelets of healthy individuals, α-synuclein levels remained stable with age [67], and do not discriminate between older people with AD and those who are cognitively intact [12].

The gene that codes for α-synuclein (SCNA) appears to have an impact on platelet volume. Thus, studies using SCNA^–/– knockout (KO) mice described a reduction in MPV in KO mice compared to the wild type [70]. As discussed above, these findings provide further support for MPV to be considered as an early marker of platelet activation. The link between SCNA, MPV, platelet activation, and dementia progression indicates that the SCNA has an impact upon platelets and their activation in dementia patients, but its product, peripheral, platelet α-synuclein, may not be necessarily altered as a function of aging and/or cognitive impairment.

Clusterin

Clusterin is a complement inhibitor protein involved in fundamental cell functions such as cell adhesion, apoptosis and lipid transport. It is a Golgi chaperone that facilitates the folding of secreted proteins in an ATP-independent pathway. Clusterin is found in high levels on platelet surfaces [71], mediates Aβ transport across the blood brain barrier and promotes Aβ fibrillization [72], A change in clusterin levels may induce oxidative stress which can lead to clinical symptoms of dementia. Although platelet clusterin levels were similar in control and AD participants [12], plasma/platelet clusterin ratios were increased in AD (but decreased in VaD) and were correlated with apathy and indifference as well as aberrant motor behavior (characteristic behaviors of AD) [12] (Table 1).These findings suggest that clusterin blood levels have the potential to be used to differentiate between distinct dementia subtypes, as well as be closely linked to a number of behavioral problems seen in dementia. Whether this protein is the cause agent or a result product of the dementia process, remains unknown.

Platelet immunoglobulin G

α-Granule platelet IgG levels are proportional to plasma IgG concentrations [73]. The total platelet IgG was consistently increased in patients with ITP and those with thrombocytopenia. This is likely due to increased platelet destruction of non-immune mechanisms [73]. Natural, self-reactive antibodies (such as IgG) in peripheral blood may control the accumulation of toxic proteins in the human brain [67]. There was an identified increase in platelet IgG in AD (16.5%) while plasma levels of IgG remained unaltered [12] (Table 1). Furthermore, the findings of the latter pilot study suggest that the anti-dementia drugs can normalize the platelet expression of IgG in subjects with AD [12]. There is yet to be identified whether this upregulation is being caused by age, duration of dementia and dementia treatments, or the degree of the disease.

Serotonin

A number of studies have addressed the divergences of both uptake and regulation of platelet serotonin in dementia subjects. Serotonin is not produced in platelets but is taken up via similar mechanisms seen in serotonergic neurons [20]. The platelets’ dense granules are responsible for the uptake of serotonin from the blood; once in platelets, serotonin is complexed with ATP or calcium. This stable serotonin-ATP/calcium complex is then released and acts as a platelet agonist on 5HT₂ receptors [74]. The number of platelet dense granules decrease in AD patients in comparison to controls [20], while plasma concentrations increase in AD. The distinction between the upregulation and downregulation of dense granules in different diseases supports the notion of dense granules (and ultimately serotonin levels) as dementia biomarkers. For this research to be applied clinically, standardization and baseline of dense granule levels in a healthy population is also required. It is doubtful that platelet serotonin levels alone could be used as a dementia biomarker as serotonin is implicated in a vast number of clinical syndromes, especially anxiety and depression, and may be altered via the additional use of the wider use of antidepressants, in particular the selective serotonin re-uptake inhibitors.

Depression is a concomitant clinical symptom of dementia. It is found in approximately one third of AD patients [75, 76] and is linked to low levels of serotonin. There is a severely lower rate of serotonin uptake in platelets in AD which leads to increased plasma serotonin levels [7, 77], whereas the platelet serotonin levels progressively become lower as the disease progresses [77–80] (Table 1). How this is related to the underlying depression seen in dementia sufferers, and in AD, in particular, still remains unknown. The most obvious explanation would be linking the decrease in the severity of depression with the progression of AD with the decrease in plasma serotonin levels; a similar decrease was seen in platelet MAO-B activity over the course of AD. The fact that depression is more prevalent in early dementia than in late dementia raises an important research question as to whether reduced serotonin levels, similarly to the higher prevalence of affective disorders [81] may also increase the risk to developing dementia.

Cholesterol levels

Platelet membrane cholesterol has an impact on the pathogenesis of AD. Thus, cholesterol to phospholipid ratio is significantly lower in AD platelets than in platelets of control subjects [82]. According to Cohen et al., there was an increase in internal platelet membranes, which parallels the increase in platelet membrane cholesterol, but no change in the overall membrane lipid composition in platelets from AD subjects [82]. The increase in platelet membrane cholesterol leads to other biological changes that occur in AD, such as an increase in AβPP β-secretase activity [83] which reduces the de novo cholesterol biosynthesis via a negative feedback. The physiological disruption in the homeostatic control of cholesterol biosynthesis appears to be responsible for the abnormal accumulation of cholesterol in people at risk of developing dementia, particularly AD. Liu et al. reported that cholesterol alongside monosialotetrahexosylganglioside (GM1) levels in platelet lipid rafts were higher in AD in comparison to controls whilst other platelet proteins remained unchanged [84]. However, this change in cholesterol levels was not seen in MCI patients [84]. Interestingly, IgM antibodies against GM1 appear to be increased in dementia and some other autoimmune diseases [85]. In the light of lack of reproducible data, it is somewhat difficult to comment about their diagnosticutility.

Surface receptors

Platelet surface receptors have also been quantified as markers of the activation state of platelets, and platelet activation is known to be increased in dementia. Both platelet CD62p (P-selectin) expression and CD41 (GPIIb-IIIa) complex activation were significantly elevated in AD patients. CD41 is a platelet activation marker in dementia and the increase in CD41 complex expression in platelets was associated with faster cognitive decline in AD [86]. However, there is an overwhelming lack of further similar clinical studies.

Another platelet receptor, CD62p, is present on activated platelet membranes [87] and promotes platelet adhesion and thrombin formation [88] (Fig. 2). Increased levels of CD62p were found in circulation in AD patients, but there was no significant change in the membrane-bound CD62p [89, 90] (Table 1). Kniewallner et al. investigated AD postmortem brain tissue and reported that, although both CD41⁺ and CD62p⁺ platelets were directly associated with Aβ plaques, the majority of CD62p⁺ platelets were present in vessels not associated with Aβ plaques or Aβ vascular deposits [91] (Table 1). These findings suggest that prior to the onset of overt AD clinical symptoms, increased levels of CD62p are expressed on platelet surfaces. Furthermore, the higher CD62p platelet expression in AD appears to be closely linked to sex, the severity and duration of AD, depression, agitation, and/or family history of dementia, but not age [77]. Increased platelet membrane-bound CD62p may, thus, have a potential to act as a clinical indicator of dementiarisk.

Monoamine oxidase type B (MAO-B)

MAO-B is an enzyme located in the outer mitochondrial membrane of platelets and is largely responsible for the degradation of dopamine and vasoactive amines to reactive oxygen species. MAO-B (spot B645 in particular) is upregulated with age in healthy participants, and it is this excessive upregulation with age that gives rise to symptoms of Parkinson’s disease dementia (PDD) and influences cognitive function [32] (Fig. 2). In addition, there were notably higher levels of platelet MAO-B in late-onset (non-familial) AD and PDD in comparison to early-onset (familial) AD and control subjects [92]. Furthermore, the levels of MAO-B activity appeared to increase with the progression ofdementia [97].

However, there is some inconsistency in reported findings of platelet MAO-B activity in dementia. Thus, a number of studies reported a decrease of MAO-B activity in later stages of dementia [7, 93], whereas others did not find a significant correlation between MAO-B activity and neurodegeneration [94]. Similarly, AD subjects treated with anti-dementia drugs, appear to have very similar platelet MAO-B levels to those detected in age-matched controls (Mukaetova-Ladinska et al, personal communication). The differences in findings may, thus, closely reflect the differences in distinct clinical dementia phenotypes, since AD subjects with and without psychosis have very different MAO-B platelet profiles [16] (Table 1). Namely the non-psychotic AD individuals have higher MAO-B platelet levels compared to those who have psychotic symptomatology [95]. Interestingly, the effectiveness of MAO-B inhibitors in improving cognitive deficits in dementia though explored [96], needs to be further investigated. In order to pursue this venture, we will need to gather further clinical evidence for the role of MAO-B in the dementia pathogenesis and, in particular, how MAO-B changes at the periphery as a function of dementia progression and clinical phenotypes.